**Part 2: Policies to End Extreme Poverty**

## **How Can the International Community Eradicate Poverty and Hunger by 2030?**

**Stephan Klasen**

#### **1. Introduction**

The last 25 years have seen an extraordinary reduction in absolute poverty. Above all, the strong economic growth of China, India, and other populous Asian countries has led to unprecedented poverty reduction across the planet. In East Asia, absolute poverty, measured as the proportion of the population living on less than USD 1.90 a day, fell from over 60 percent in 1990 to under 3 percent in 2015 (World Bank 2016). In South Asia, poverty receded from 48 percent to around 12 percent. These tremendous successes in Asia led to the global Millennium Development Goal (MDG) of halving absolute poverty between 1990 and 2015 being more than met (from 43 percent in 1990 to 11 percent in 2015, see Figure 1). Of particular importance is the fact that this poverty reduction was accomplished even as inequality in some of these countries, especially China and India, rose drastically. As a result, poverty reduction was significantly slower than it would have been if income had risen equally for all income groups (Ravallion and Datt 2002). However, other regions—Latin America and sub-Saharan Africa—have also made notable contributions to poverty reduction since the mid-1990s, even if the goal of halving the poverty rate was not reached in Africa. Progress in reducing hunger and malnutrition has been slower, but still noteworthy.

Encouraged by these successes, the global community has committed itself, in the Sustainable Development Goals (SDGs), to the complete elimination of extreme poverty and hunger.

Is this a realistic goal? Is it possible? Which measures are necessary, and from whom, to achieve this ambitious goal?

I would like to address these questions in this short article. After briefly discussing some measurement challenges, I will argue that an entirely different approach than in the past will be necessary to reach this goal. It will not be possible to depend on economic growth in low-income countries, nor will development cooperation in its present form be able to play a major role. Instead, a much broader approach will be necessary, in which political and economic stability, the promotion of structural change, the development of social security systems in middle-income

countries, new trade policies, the reduction of inequality, a more active population policy in Africa, and investments to overcome country-specific challenges will have to play a central role.

**Poverty Rate**

**Figure 1.** Percentage of population in the developing world living on less than International \$1.90 a day. Source: Povcalnet, accessed 3 February 2020 and World World Bank (2016). Note: Based on the 1.90 poverty line and 2011 Purchasing Power Parity (PPP). Developing world excludes high-income countries in Europe, North America, and Australasia.

#### **2. Measuring Poverty and Hunger**

The SDGs, with their 17 goals, 169 targets, and 250 indicators, suggest a high degree of precision when it comes to monitoring the global development agenda. A closer look reveals, however, that many targets are impossible to measure, indicators are missing, and that the necessary data are often simply unavailable. This problem already arose with the MDGs, but is even more problematic for the SDGs, especially since far too little has been done in recent years to develop the necessary database for the SDGs.

These problems concern not only the new targets or particularly challenging indicators, but also the main indicators for extreme poverty and hunger. The absolute poverty numbers depend on the international poverty line, which is updated every five to seven years. The poverty line is based on infrequent international price comparisons that ensure that it reflects the same purchasing power everywhere.

Unfortunately, these updates often entail unforeseen changes. In 2008, as the USD 1.08 (in 1993 prices) poverty line increased to USD 1.25 (in 2005 prices), we learned that global poverty was about 50% higher than previously thought, even though both poverty lines were supposed to reflect the same purchasing power. With the latest increase in the poverty line to International \$1.90 (in 2011 prices), the global poverty rate would have been significantly lower (by about 20 percent in 2010), had certain adjustments to the data not been introduced (Klasen et al. 2016a).

Although these periodic updates do not create any doubt about the reduction in poverty over time, the fact that there is so much uncertainty about the prevailing *level* of poverty is a major problem. If we now want to eliminate poverty altogether through the SDGs, such fluctuations in the measured level of poverty, which also vary from region to region, are very problematic, since they imply that the goalposts are constantly shifting. Thus, it is high time to translate the international poverty reduction goal into national poverty lines, which are not influenced by international purchasing power adjustments. This was suggested in Klasen et al. (2016a); the World Bank's Commission on Global Poverty recently made the same suggestion. This would make it possible to better reflect the progress made in each country with respect to the SDGs.

The situation is even more muddled when it comes to measuring hunger. A fundamental problem is that hunger is, to a large extent, measured using two different indicators, which sometimes lead to contradictory conclusions when observed in tandem. Moreover, both indicators are prone to methodological weaknesses. The indicator for malnutrition used by the Food and Agriculture Organization (FAO) is based on aggregated data and estimates the average number of people without access to sufficient calories. It finds that hunger is most acute in sub-Saharan Africa. It should further be noted that this indicator is frequently revised, and historical estimates have often been adjusted. While estimates until 2012 assumed that little progress in reducing the hunger rate had been made since 1990, calculations made after 2012 suggested that there had indeed been a substantial reduction, since the figures for 1990 had been adjusted upwards—resulting in a reduction over time (Pogge 2016).

The second indicator, used by the United Nations Children's Fund (UNICEF) and the World Health Organization (WHO), measures the number of children suffering from stunted growth and low weight. The data suggest that these problems are most pressing in South Asia and have been gradually improving globally over the last 25 years, though progress has been significantly slower than poverty reduction. As explained in De De Haen et al. (2011), there are problems with

both indicators and the estimates based on them. The FAO figures are based on very rough estimates and highly simplified (and probably erroneous) assumptions, while the figures on the nutritional status of children, which are generally more reliable, yield biased results for (small) genetic differences in height (see, for example, Klasen 2008; De Haen et al. 2011).

Based on these key indicators of poverty and hunger, then, we cannot say with any certainty *how many* people in the world suffer from hunger and/or poverty, or *where*. There is a considerable need for action when it comes to tackling these measurement problems (all of which can, incidentally, be solved). These uncertainties do not make it easy to come up with a course of action in the area of poverty and hunger reduction.

#### **3. Challenges for Overcoming Extreme Poverty and Hunger**

Aside from these measurement issues, there are a number of substantial challenges to eliminate poverty and hunger by 2030. Figure 1 would suggest that "only" a continuation of the existing trend of poverty reduction is necessary to bring the goal of overcoming absolute poverty by 2030 within reach. Gill et al. (2016) therefore propose a strategy for overcoming extreme poverty that continues past approaches and extends them by one additional policy. In particular, the focus should be on the two pillars of "labor-intensive growth" and "investments in the education and health of the poor", supplemented by a third pillar, "improved social security systems". It is argued that the success of poverty reduction over the last 30 years was mainly driven by the first two pillars in many fast-growing countries of Asia. This was enough to lift the bulk of the population out of poverty. To support those who remain poor due to various simultaneous occurring challenges (such as geographic remoteness, lack of education, ethnic minority status), and to successfully insure against local, national, and global shocks, social security systems need to be systematically be developed to provide the poor access to a subsistence minimum and protect them from economic shocks.

In theory, such a strategy sounds promising, and if the three pillars could be fully implemented, extreme poverty could indeed be eliminated by 2030. However, it is unclear *how* labor-intensive growth, investment in the human capital of the poor, and the development of social security systems are to be secured in many of the countries on which the achievement of SDG 1 depends.

Table 1 lists the countries in which the majority of the poor lived in 2015, using both the USD 1.90 line and the higher USD 3.10 poverty line. First, it should be noted that data for very poor countries with a (likely) high number of the extreme poor

are missing, simply because data are unavailable for these countries (Gill et al. 2016). These include Myanmar, Cambodia, South Sudan, Afghanistan, Somalia, Eritrea, Yemen, and Somalia. These countries have a total population of about 150 million people, of whom certainly more than 50 percent (i.e., more than 75 million people) live below the poverty line of USD 1.90 and more than 80 percent (more than 120 million) live on less than USD 3.10 a day. Of course, these blind spots in our poverty statistics point to much more serious problems in these countries, which I will discuss below.

Table 1 also includes data for 1990, and shows the extraordinary success of the Asian countries in reducing poverty since 1990. It also shows how the geography of poverty has changed since 1990. Whereas in 1990, 50 percent of the extreme poor lived in East Asia—and mostly in China and Indonesia—that figure is now less than 10 percent (World Bank 2016, see Figure 2). More than 50 percent of the poor now live in sub-Saharan Africa. Only the share of South Asians among the global poor has remained relatively constant over time, at around 30 percent. This means that to overcome poverty by 2030, in particular, poverty in Africa and South Asia must be reduced.


#### **Table 1.** Total population (in millions) living in extreme poverty, 2015 (and 1990).


Note: <sup>1</sup> data from 1991, <sup>2</sup> data from 1992, <sup>3</sup> data from 1988, ◦ data from 1984, <sup>−</sup> data from

1998, \* data from 1996, <sup>a</sup> data from 1997, ˆ data from 2008, data from 2003. Source: Povcal, Gill et al. (2016), UNDP (2015).

**Regional Distribution of Extreme Poverty**

**Table 1.** *Cont*.

**Figure 2.** Regional Distribution of Extreme Poverty. Source: Povcalnet, accessed 3 February 2020. Note: Based on the 1.90 poverty line and 2011 PPP. Developing world excludes high income countries in Europe, North America, and Australasia.

The question that now arises is: which of the countries in Table 1 has the potential to follow in the footsteps of the Asian growth miracle countries and generate rapid and labor-intensive economic growth with high investments in education and health? It should first be noted that two elements were crucial for rapid, poverty-reducing growth in Asia. On the one hand, there was considerable success in increasing agricultural productivity, which raised the incomes of the poor rural population and freed up labor for industrialization; on the other hand, structural change towards industrialization, and above all, the development of a predominantly export-oriented manufacturing sector producing globally competitive products, was crucial (Gill et al. 2016; World Bank 1993).

Looking at the economic development of the currently poor countries over the last 10 to 15 years, very few are actually on track to meet these two components of labor-intensive growth. These include Bangladesh, Cambodia, the Philippines, and India. This is also true, to a degree, of Pakistan and Myanmar, and there is some evidence for this in Ethiopia and Rwanda, although it is still unclear how sustainable the improvements in agriculture and structural change really are. If structural change in these countries continues to progress and agricultural productivity improves steadily, there is hope that a large proportion of the approximately 230 million living in extreme poverty in these countries will have escaped poverty by 2030. Yet, as can already be observed in China and Indonesia (see Table 1), it is likely that a small base will continue to be unable to escape poverty by these two mechanisms alone.

There are also a number of countries on the list, including Mexico, Brazil, and South Africa, that have a major industrial sector that would meet the basic conditions necessary for generating labor-intensive industrial growth. However, structural change in these countries has slowed in the last 10 to 15 years, the industrial sector has dwindled, and these countries have become significantly more dependent on exporting commodities. This structural change in reverse has had much to do with China's role in the world market, which displaced these countries from global manufactured goods markets with their cheaper products on the one hand and significantly boosted demand for raw materials on the other. However, further poverty reduction through increasing commodity exports will become difficult, especially as this sector is now also ailing. In other words, we do not see any signs that these countries will be able to overcome their still-existent poverty in the near future.

The largest group of countries in Table 1 comprises African resource exporters, with an economy focused on a small number of agricultural or mineral commodities. These include Nigeria, Republic of Congo, Burundi, Tanzania, Democratic Republic

of Congo, Mozambique, Malawi, Uganda, Kenya, Zambia, Mali, Burkina Faso, Niger, Cameroon, Sudan, Senegal, Benin, Angola, and Southern Sudan. More than 300 million of the extremely poor live here, and their long-term fate depends largely on the development of commodity prices. Although there have been improvements in agricultural productivity in some countries, which also contributes to poverty reduction, these advances are often thwarted by the commodity sector, which leads to overvalued exchange rates and thus, cheap food imports. Additionally, the commodity sector often undermines good governance and leads to bloated government sectors. The high commodity prices of the last 15 years (and better management of the resource boom) have contributed significantly to poverty reduction, but without structural change we have reached "the end of the line", and indeed, there is now a risk of regression if prices fall. Here, lasting poverty reduction will only be possible if a sustainable structural change is initiated and, at the same time, further active investments are made in improving agricultural productivity.

Another group of countries with many extreme poor, which partly overlaps with the previous group, is in an even more difficult situation: here, the poor live in states where civil war is raging, autocratic dictators are in power, or the state is extremely fragile and thus, unable to carry out basic state functions. This group includes Nigeria, Ethiopia, Uganda, Burundi, Democratic Republic of Congo, Mali, Uzbekistan, Sudan, Somalia, Yemen, Eritrea, Afghanistan, South Sudan, and Myanmar. While in some of these countries, conflicts are regionally contained (as in Uganda, Ethiopia, or Myanmar, for example), in others they affect the development prospects of the entire country (see, for example, World Bank 2013; OECD and DAC 2009). In the so-called Fragile States Agenda, attempts have been made to identify the particular challenges for effective development cooperation in these situations, and to identify appropriate means of interaction. Designing better ways to channel aid will not be sufficient to overcome extreme poverty in these states. Without the pacification of conflicts and the establishment of functioning states, overcoming poverty will be impossible. Development cooperation can only play a very limited role here. Instead, the international community will need consistent political commitment, UN peace missions with sufficient resources and a robust mandate, and the long-term development of institutions.

Another group of countries must expect the effects of climate change to complicate efforts to overcome extreme poverty. On the one hand, it affects countries with large coastlines and low-lying areas threatened by rising sea levels and more intense tropical storms. The most prominent among these are Bangladesh, the Philippines, Indonesia, Myanmar, and Nigeria. For these countries, significant

measures will be needed to prevent and combat flooding. On the other hand, agricultural countries that are highly dependent on rainfall will be affected. These include almost all the African countries in Table 1. In these countries, it will be particularly difficult to maintain or improve agricultural productivity.

For the last group of countries, there is yet another challenge: drastic demographic changes. After 20 years of strong economic growth and a significant expansion of women's education across the globe, one would expect birth rates to have fallen sharply. In Asia, most of the Middle East, Latin America, and many countries in southern Africa, a considerable decline in birth rate can be observed. This has led to a favorable demographic composition, with a larger working-age population, fewer children, and a limited share of senior citizens (for now) (Bloom and Williamson 1998). In most of these countries, the number of children per woman has fallen to about two. However, in parts of Africa, this demographic transition has either hardly begun, or else has come to a standstill. Figures from the Demographic and Health Surveys show that in Angola, Burkina Faso, Uganda, Mali, Niger, and Mozambique, fertility has stabilized at the very high rate of six children per woman, and in Cameroon, Senegal, and Nigeria, it remains stable at over five. As a result, the population will double in the next 15 to 20 years in these countries, and quadruple in the next 30 to 40 years if fertility rates remain as they are. The pressure this will exert on ever scarcer land, scarce water resources, and public infrastructure such as schools and health care centers will be enormous, and will make poverty reduction much more difficult. It will be important to understand the reason for this slowdown in fertility decline and to take appropriate measures to support smaller families.

I have concentrated exclusively on overcoming extreme poverty in the preceding discussion and ignored the goal of zero hunger. However, the obstacles to achieving this goal are similar in many ways. It will be particularly difficult to meet this goal in fragile and conflict-affected countries; resource exporting countries will make conditions for agriculture more difficult due to elevated exchange rates, neglect of agriculture, and poor governance; climate change will make food security particularly difficult in affected countries; and demographic dynamics in some countries will make the hunger problem even more acute. At the same time, two additional aspects should be considered. First, economic growth alone is not enough to overcome hunger. Indeed, Vollmer et al. (2014) showed that economic growth only makes a very small contribution to reduce child malnutrition. Instead, it is more important to strengthen maternal education, strengthen local health care (especially in rural areas), and invest in access to clean drinking water and sanitation. Such interventions are generally difficult and are particularly complicated to implement in fragile and conflict-affected

countries. Second, agriculture plays a more direct role in a country's food security than in poverty reduction. The reason is not that hunger can only be overcome with higher domestic food production. A country can also reduce undernutrition through food imports, provided the population has enough purchasing power (Dreze and Sen 1989). Rather, agriculture is important because, in most countries where hunger is rampant, it is most dire among the poor rural population, who mainly derives their income from agriculture. Thus, if one succeeds in increasing agricultural productivity, especially that of small farmers, their increased incomes will provide greater food security and help reduce hunger overall.

#### **4. Conclusions for International Cooperation**

First, it is important to stress that, as with the MDGs, the achievement of the SDGs will depend first and foremost on the policies of the countries concerned. The economic success of the Asian tigers over the previous decades, and the poverty reduction that came with it, was mainly achieved through their successful economic policies. An increasingly open world market for manufactured goods has also helped, and development cooperation has certainly made a supporting contribution.

Since the goal has become "ending absolute poverty", poverty must now be eradicated in all countries worldwide. If MDG 1 could still be met thanks to certain "overachieving" countries offsetting persistently high poverty rates in other countries, this is no longer possible with SDG 1. Since only 11 of the 37 countries in Table 1 have achieved the MDGs (including only three from sub-Saharan Africa), it cannot be expected that a majority of the countries in Table 1 will be able to meet the even more ambitious goal of total poverty eradication. For this reason, the World Bank has decided that the goal of ending absolute poverty globally is to be considered attained once the poverty rate falls below 3 percent. This could mean that even after 2030, about 14 percent of the population in Africa would still be living in poverty (World Bank 2016). While less ambitious, this still means that countries which did not meet the poverty target for MDG 1 must reduce their poverty and hunger at a faster pace than during the last 25 years. A stronger international commitment will therefore likely be necessary.

But how can the international community and development cooperation contribute? I would like to distinguish here between development cooperation in a narrower and a broader sense. The distinctions between groups of countries that I have outlined above will also be relevant here.

Financial cooperation will no longer be relevant for many middle-income countries. Technical cooperation can, however, continue to play an important role in

these countries. In countries such as India, China, and Indonesia, the main focus will be on supporting the establishment of reliable, broad-based social security systems that succeed in reaching the extremely poor. In other middle-income countries, such as Nigeria, Kenya, or Angola, the main aim will be to create incentives for increased diversification of the national economy. This includes a macro policy that creates competitive exchange rates and mobilizes investment in infrastructure, agriculture, and small- and medium-sized enterprises in the manufacturing sector. It should also address governance problems in these resource export countries and strengthen civil society groups. However, this policy focus requires long-term commitment and whether development cooperation can be effective in improving economic and political institutions in these countries is far from certain.

Forward momentum must be sustained in those countries that have the basic prerequisites for labor-intensive growth. This includes the promotion of infrastructure, industrial and structural policy measures, investment in training and further education, and support for trade.

Finally, countries in which the demographic transition has come to a standstill need a health and population policy that favors a decline in birth rate. This includes universal access to reproductive health and family planning services, further investment in female education, the outlawing of child marriages, and support for national population policies. Countries such as Rwanda and Ethiopia show that a consistent commitment could certainly bear fruit there.

In my view, this has already pushed us to the limits of conventional development cooperation. For other groups of countries, it can provide support at best, but other international initiatives will be crucial. The many commodity-exporting countries, for example, need much more than trade-promoting development cooperation. Since they operate in a global marketplace in which they have to contend with extremely competitive suppliers, it will be very difficult for them to develop a competitive industrial sector through structural change. What can help, however, is the substantial expansion, broadening, and simplification of non-reciprocal trade preferences, which are intended to benefit least developed countries the most. As Klasen et al. (2016c) have shown, existing trade preferences have already had a positive impact on the exports of least developed countries. However, there are too many exceptions, too many restrictions, and too much bureaucracy. Changes in trade policy, which are set at the level of the European Union for European countries, are more important here than development policy.

Support for the poorest countries in dealing with climate change also calls for new—yet to be developed—approaches. On the one hand, the promised

funds for adaptation to climate change must be mobilized and made available on top of standard development cooperation. We also need to rethink how funds are used. For countries affected by rising sea levels and storms, comprehensive strategies must be developed, including infrastructure measures, resettlement and land reallocation, disaster prevention, and disaster control. For agricultural countries, the focus will have to be on international research to develop improved seeds and production technologies and the expansion of irrigation systems (where sustainable and practicable).

Fragile countries affected by conflict and dominated by anti-development dictators pose a particular challenge. Here, development cooperation alone can do very little, and conflict can quickly destroy decades of productive development cooperation. Unfortunately, the international community's involvement with these countries has so far been minimal. Involvement in the form of UN peace missions in many countries listed in Table 1 has often prevented worse things from happening, but is characterized by inadequate force and a lack of mandate and resources. Increased political and military commitment will be necessary if we are serious about overcoming absolute poverty. As demonstrated by events in the Balkans in the 1990s, but also in Sierra Leone in 2003 or in Cambodia in the 1990s, a robust political and military commitment can make a major contribution to peace. A similar commitment will be necessary for countries such as the Democratic Republic of Congo, Mali, Niger, South Sudan, and Burundi. The situation is even more difficult for countries afflicted by state collapse or authoritarian dictatorships. To bring about conditions for poverty-reducing policies, the international community will first have to make a stronger and more consistent political commitment. This would include, for example, international condemnation, entry bans, and sanctions concentrated on leadership cliques for governments such as those of Eritrea, Sudan, or Zimbabwe, strengthening opposition forces, and much more. Of course, this alone does not guarantee success, but it could increase the chance of change.

Lastly, reducing inequality can make an important contribution to poverty reduction. It directly reduces poverty and increases the poverty-reducing effect of economic growth. This requires comprehensive approaches: for instance, inequality can be reduced through targeted investments in health, education, and social security systems for the poor. Tax and public expenditure policies can also play a role. For example, development cooperation measures can promote higher, more progressive tax revenues through advisory services and targeted investments, and simultaneously support the reform and monitoring of government spending. Moreover, a more concerted international crackdown on tax evasion and tax avoidance

by rich individuals and multinational corporations in developing countries will also play an important role.

#### **5. Outlook**

The international community has set high objectives with SDG 1 to overcome extreme poverty and hunger in the near future—which is welcomed in a world where there is abundance in many places and no one would have to suffer globally from extreme poverty and hunger. But, as I have tried to point out, business as usual, even in the context of a favorable international economic environment, can only mean that we miss this target. Furthermore, the current economic and health crisis caused by the COVID-19 pandemic demonstrates that years of progress in poverty reduction can be reversed in an instant.

To accelerate poverty reduction, we not only need to tackle the measurement problems of poverty and hunger as quickly as possible, but we also need a new international development policy. In particular, we will not reach SDG 1 if only ministries of development cooperation address the challenges that I outlined above. If we are serious about overcoming global poverty and hunger, it must become one of the main missions of international politics for *all* actors, including the UN System as well as the G7 and G20. Coordinated by these fora, there must be consistent political, economic, and occasionally, even military engagement to pursue these goals. Moreover, these actors need to help especially poor countries deal with economic and health shocks by providing concessional funds, technical support, and debt relief. Hopefully, governments are aware of this challenge (Klasen et al. 2016b).

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


Dreze, Jean, and Amartya Sen. 1989. *Hunger and Public Action*. Oxford: Clarendon Press.


© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **"Leave No One Behind" in Middle-Income Countries. A Review of Progress and Policies**

**Jann Lay and Jan Priebe**

#### **1. Introduction**

The principle to "leave no one behind" (LNOB) is a central commitment of the 2030 Agenda. The idea to "leave no one behind" is a key part of the Sustainable Development Goal (SDG) 1 and 2 regarding ending poverty and hunger in all its forms, and SDG 10 regarding reducing inequality. Moreover, the terms "inclusive", "for all", "equitable" and "equitable access" are used repeatedly in the 2030 Agenda in relation to health (SDG 3), education (SDG 4), gender (SDG 5), water and sanitation (SDG 6), energy (SDG 7), economic growth (SDG 8), industrialization (SDG 9), cities and human settlements (SDG 11) and, more broadly, societies and institutions (SDG 16).

The objective of this chapter is to identify the main conclusions of recent studies measuring progress in LNOB and to showcase examples of public policies that are or could be most successful at addressing and implementing the principle of "leave no one behind" in middle-income countries. LNOB is of particular relevance in these countries because they have a significant population living in poverty and because the gains of economic progress tend to be unequally distributed.

#### **2. "Leaving No One Behind" What Does It Mean?**

However, what does "leaving no one behind" actually mean? A comprehensive UN report, "Leaving no one behind: the imperative of inclusive development" (UN 2016), emphasizes the concept of social inclusion to operationalize LNOB. Social inclusion is understood as a "process of improving the terms of participation in society for people who are disadvantaged on the basis of age, sex, disability, race, ethnicity, origin, religion, or economic or other status, through enhanced opportunities, access to resources, voice and respect for rights". The report's definition of social inclusion thus explicitly refers to people who are disadvantaged on the basis of certain characteristics that increase the risk of social exclusion, defined as "a state in which individuals are unable to participate fully in economic, social, political and cultural life" (UN 2016).

Stuart and Samman (2017) provide a more comprehensive interpretation of LNOB and see the concept of LNOB as addressing three dimensions of social progress: (1) addressing discriminatory barriers, which could arise from geography or aspects of social identity, (2) ending extreme poverty (in all its forms) and (3) reducing inequalities. LNOB thus assumes that progress does not automatically trickle down to all groups of society. This implies that existing gaps between the worseand the better-off will grow without deliberate efforts targeting those left behind (Stuart and Samman 2017).

Following this approach, we also understand LNOB as a comprehensive principle that emphasizes social, economic and political inclusion and that extends beyond an anti-discrimination agenda. This interpretation of the LNOB principle implies that it is important to focus on development policy areas that are key for economic and political participation and on development policies that address areas where people are denied access to public services and are even discriminated against. Key arenas for LNOB-relevant policies are hence education, social protection (including universal health coverage), and the labor market. In all these areas, strong mechanisms of positive feedback are at work that keep people poorly educated, with low skills, bad health and in precarious jobs. In addition, inclusive institutions and corresponding cross-sectoral progress in legal frameworks are required, including constitutional anti-discriminatory provisions (gender, race, disabilities etc.), measures to protect minorities and affirmative action.

Further, we think it would be useful to stress even more than Stuart and Samman (2017) the dynamic element implied by "being left behind". "Being left behind" implies that the distance between the "left behind" and the rest of society grows larger over time, i.e., that future life chances are impaired. Examples include early childhood disadvantages, not being able to read and write, being in an informal job without training for too long.

Thus, LNOB is of particular importance in middle-income contexts, where inequalities are rampant and where the benefits of development are often unequally shared. In absolute numbers, many of the world's poor live in middle-income economies. India alone has about 175 million people living below the USD 1.90 (2011 PPP) poverty line (data for 2015<sup>1</sup> ). By definition, the absolutely deprived will be relatively more deprived than they are in contexts that are poorer on average. This situation provides a rationale to keep engaging with middle-income

<sup>1</sup> www.worldbank.org (accessed on 30 May 2018).

countries through development cooperation with a focus on LNOB. Yet, the European Consensus on Development (EU 2017) does not explicitly mention LNOB or related concepts, such as social inclusion or inclusive development. Instead, it refers to cooperation, policy dialogue and partnerships with middle-income countries on sustainable development and other shared interests.

#### **3. Measuring Progress in "Leaving No-One Behind"**

As has become apparent from the above conceptual considerations, even a somewhat narrower understanding of LNOB encompasses many of the SDGs. Counting the SDG indicators that are either directly related to the LNOB principle (such as SDG 1 or SDG 10) or that are to be disaggregated by groups amounts to 82 indicators (of a total of 232 indicators in the SDGs). The number of indicators to be disaggregated by sex is 33, by age groups 24, and by disability status 10 (own calculations). Interestingly, only two indicators are explicitly stated to be disaggregated by ethnic/indigenous status and so are two indicators by migrant status.

This vast number of often-disaggregated indicators presents an important challenge to measuring LNOB progress: the production of (micro) data required to provide the LNOB-relevant indicators.

To overcome this challenge, a couple of authors of the Overseas Development Institute (ODI) (Greenhill 2017; Stuart et al. 2016) have proposed a LNOB index that measures governments' readiness for implementing LNOB.


According to this LNOB index, about half (9 out of 17) of the middle-income countries are not on track to implement LNOB and this is despite the fact that the simple cumulative LNOB index is relatively generous in classifying countries as "ready".

Clearly, however, providing disaggregated data can only be a first step of addressing discrimination and inequality. Having policies in place does not yet mean that they are effectively implemented. Eventually, countries should not only be judged on their "preparedness" or their "intentions" to implement the LNOB principle, but on effectively delivering on it. Further, by not considering LNOB-relevant outcome variables, including relevant SDG indicators, the LNOB index falls short of providing information on the order of magnitude of the challenges, for example, the presence of majorly disadvantaged and discriminated-against groups and the extent to which exclusion and the resulting inequalities leave them behind.

We hence use the remainder of this section to present some key indicators for measuring LNOB progress—albeit data limitations do not allow us to say much about developments over time. We provide an exemplary selection of indicators (for selected middle-income countries) that also illustrate some of the challenges in measuring progress. We chose indicators that we consider to represent key aspects of the major LNOB policy areas identified in Section 1: Education, social protection (including universal health coverage), and legal frameworks and anti-discrimination policies.

However, before we turn to these indicators, we briefly discuss the progress in extreme poverty reduction in middle-income economies.

#### *3.1. Extreme Poverty in Middle-Income Countries*

Many of today's middle-income economies, i.e., middle-income economies as of 2015, have been very successful in combating extreme poverty. Figure 1 shows that all middle-income economies—except Yemen—saw extreme poverty decrease between 2005 and 2015. Yet, progress has been uneven: In East and South-East Asia, China and Vietnam stand out with poverty declines from levels around 20 percent to 5 percent or less. Considerable poverty reduction has also been seen in the middle-income economies of South Asia with India, Bangladesh and Sri Lanka starting from very different levels. In Indonesia (and the Philippines), progress has been somewhat slower with poverty rates still at about 10 percent (also down from 20 percent in 2005). Progress has been similar in many middle-income countries in Latin America, where the economic growth of this period also helped the very poor. Countries with initially lower poverty rates like Brazil, Mexico, or Nicaragua found it more difficult to further bring down poverty. In Africa, extreme poverty has proven to be much more persistent in both richer economies like South Africa and, even more so, in poorer ones like Cameroon, Nigeria or Zambia. Ghana stands out as the only African middle-income economy with recorded substantial poverty reduction.

**Figure 1.** Extreme poverty in middle-income economies, 2005–2015. Source: World Development Indicators, World Bank (accessed on 30 May 2018), used with persmission. Notes: All non-European middle-income countries with data from 2003 to 2007 and 2013 to 2017. Exceptions: Ghana (2012), India (2011), Nigeria (2009).

For monitoring LNOB, these poverty headcounts should be reported by age, sex, location and employment status (SDG 1, target 1.1.1). However, these disaggregated data are not yet available on the World Bank website (the World Bank is the "data custodian" for these indicators).<sup>2</sup> This already indicates that quite some efforts are required to improve reporting on LNOB-relevant indicators.

#### *3.2. Education*

Access to (high-quality) public services, and especially to education, is one key dimension of LNOB. Deficient education system puts many of the poor at risk of being left behind and not benefitting from growth in average incomes (World Bank 2017). If people lack education they are likely to end up in vulnerable employment that characterizes labor markets in many middle-income economies and it seems unlikely

<sup>2</sup> The World Bank (World Bank 2018a) has made some progress here, for example in its recent Poverty and Shared Prosperity Report (2018).

that a rapidly growing formal sector will absorb the informal workforce quickly (Stuart et al. 2018).

The fact that the lack of proper education impairs future life chances makes education central to LNOB and also explains why the indicators under SDG target 4.5 demand "parity indices" (female/male, rural/urban, bottom/top wealth quintile and others such as disability status, indigenous peoples and conflict-affected, as data become available) for all education indicators [. . .] that can be disaggregated" (UNESCO 2016).

Despite increasing enrollment rates, national education systems are failing worldwide to provide adequate learning outcomes for millions of children in low and middle-income countries. The (still patchy) data on learning achievements suggest that students in low-income countries and even middle-income countries often lack even the most basic skills in reading, writing and math. For instance, in rural India, about 75 percent of students in grade 3 could not solve a two-digit subtraction and by grade 5 half could still not do so. Students in Brazil, Indonesia, and Mexico clearly underperform their peers in the OECD, with the notable exception of China where students show learning achievements in both reading and in math above the OECD average (own calculations using data from www.oecd.org/pisa/data/). These poor learning outcomes hold in particular for disadvantaged children, be it because they are female, from poor families, disabled, or from rural or even conflict-affected areas (World Bank 2017). With the SDGs, the emphasis of measuring progress in education has therefore shifted from enrollment to learning achievement because "schooling is not the same as learning" (World Bank 2017).

Surprisingly, little data on learning outcome exist and the school enrollment rate can often not be disaggregated by the desired groups—except for gender gaps in education that are increasingly understood and addressed (Minasyan et al. 2019). As reported by UNESCO's Data Digest (UNESCO 2017) disaggregating education data by measures of wealth and disability status, for example, was possible in only 14% and 19% of countries, respectively.

Yet, some evidence for selected groups in selected countries is available. Figure 2 forcefully illustrates the extent of discrimination in access to education in some countries. In Latin America, for instance, the share of the indigenous population that has completed lower secondary schooling can be more than 30 percentage points lower than for other population groups. These differences can be observed despite major progress in increasing the school attendance of indigenous children between 2000 and 2010 in the region (UN 2014; UNESCO 2018). Figure 2 also shows lower secondary school completion by ethnic groups in Ghana, again showing considerable

differences across ethnic groups. For learning outcomes disaggregated data is even scarcer, but few studies exist. For instance, in Uruguay, poor children in grade 6 are five times as likely to be assessed as "not competent" in math than richer children. In general, the findings on learning differences tend to exacerbate the findings on school enrollment and completion rates, with disadvantaged groups seeing particularly large differences once learning outcomes are accounted for (World Bank 2017).

#### *3.3. Social Protection*

Social protection is an important target under SDG 1 and crucial for LNOB. Specifically, target 1.3 calls for the implementation of "nationally appropriate social protection systems and measures for all, including floors" and including "substantial coverage of the poor and the vulnerable". The corresponding SDG indicator 1.3.1 gives the percentage of the total population covered by at least one social protection benefit (effective coverage). The benefits considered include those for children, mothers with newborns, persons with severe disabilities, the unemployed, and older persons. Based on the sparse data that are available, among middle-income economies there has been quite some progress in achieving higher effective social protection coverage. One main reason for substantial progress in the past 25 years has been the rise of non-contributory cash transfer programs of various types across the developing world. In Latin America, the expansion of means-tested cash transfer program has improved coverage of child, maternity and family benefits among the poor. Yet, the situation has improved less for the somewhat better-off who are neither covered by cash transfers nor by contributory schemes.

**Figure 2.** Youth who completed lower secondary education in rural areas, by ethnicity, latest available data since 2011. Source: UNESCO (2018). World Inequality Database on Education at www.education-inequalities.org.

On the other hand, huge variation remains. Indicator 1.3.1 (all figures taken from ILO (2017a)) varies between about 40 percent in Bolivia and Colombia and 67 percent for Argentina. In Asia, coverage can be as low as 19 percent in India and reaches its maximum with only 47 percent in the Philippines. In Africa, the Southern African middle-income economies of Botswana, Namibia and South Africa today reach parts of their poorer populations, in particular those at old-age and with children, through a comprehensive system of social transfers. Yet, average effective coverage is relatively low in Botswana with 15 percent and reaches 48 percent in South Africa. In other African middle-income economies, for example Cameroon (9 percent) or Nigeria (4 percent), there is very low effective coverage, while Ghana takes a middle place with 18 percent. Finally, unemployment remains uninsured for most people in most countries of the world. Often, the differences between countries in social protection are not necessarily related to income differences between middle income countries. Colombia has double the per capita income of Bolivia, and South Africa is slightly poorer than Botswana in per capita terms. This implies a very important role for policies, specific institutions, and political commitment (see Section 3.3).

A key indicator for social protection is also the Universal Health Coverage index (UHC index, SDG indicator 3.8.1). This index reports the coverage of essential

health services in percentage (defined as the average coverage of essential services that include reproductive, maternal, newborn and child health, infectious diseases, non-communicable diseases and service capacity and access). The maximum value is 80. In Figure 3, we report this index for all non-European middle-income countries, for which the index is available, and, as can readily be seen, the index is available only for a limited number of countries (only 24 out of approximately 90). The Figure shows quite some variance among middle-income economies. While a number of countries, including Peru, Mexico, China, and Thailand are close to full coverage (index of 80), some countries have coverage sometimes well below 50 percent, including two major West African economies, Cote d'Ivoire and Nigeria.

#### **Universal health coverage index in 2015**

**Figure 3.** Universal health coverage index, 2015. Source: Authors' own compilation adapted from Development Indicators, World Bank (accessed on 30 May 2018). Note: All non-European middle-income countries with data.

According to the SDG indicator catalogue, the UHC index should also be reported for "the most disadvantaged population". In official databases, these indicators are not yet available by these groups, but Hogan et al. (2018) reports the UHC for the poorest wealth quintile (based on a DHS-based asset index) and the national average for few selected countries. In this sample of countries, it becomes apparent that middle-income economies do not only differ in their average level of access, but also very much in the equality of access. For example, Colombia and Peru come close to an index of 80 (the maximum) on average. Yet, while the UHC index is around 65 for the poorest wealth quintile in Colombia, this figure is well below 60 for Peru.

#### *3.4. Legal Frameworks and Anti-Discrimination*

Legal frameworks and anti-discrimination measures matter for many dimensions of LNOB. A stock-taking exercise of LNOB-relevant SDG targets and corresponding indicators reveals that gender discrimination is the only dimension of discrimination that is explicitly referred to in terms of legal frameworks. Key indicator 5.1.1 will be used to monitor "whether or not legal frameworks are in place to promote, enforce and monitor equality and non-discrimination on the basis of sex". The indicator, however, is (still) a tier III indicator, i.e., under development. Another gender-related indicator under SDG 5 monitors the "number of countries with laws and regulations that guarantee women aged 15–49 years access to sexual and reproductive health care, information and education" (5.6.2). Further, the means of implementation under the gender goal include references (and corresponding indicators) to women's equal rights to land ownership and/or control and the number of countries with systems to track and make public allocations for gender equality and women's empowerment. For all these indicators, no data are available to date, albeit the information can, in principle, be obtained (https://unstats.un.org/sdgs/metadata/).

With respect to the de-jure and de-facto discrimination gaps important regional patterns have been noted. While many countries in East Asia, South-East Asia and Latin America have often passed comprehensive legislation to protect women, there are still enforcement problems in a number of countries. In contrast, countries in the Africa, MENA, and South Asia regions even lack the legislative basis to protect women (Branisa et al. 2014). The Social Institutions and Gender Index (Branisa et al. 2014) that contains both legal (e.g., laws on marriage age, domestic violence, access to land, access to public space) and factual components (e.g., missing women, female genital mutilation incidence) shows the variance between middle-income economies also within the same region. In Africa, for example, Ghana and Nigeria are ranked as having high and very high levels of discrimination, while Namibia and South Africa are classified as having low level of discrimination.

With respect to national legal frameworks and policy documents (de-jure) it appears that developing countries are more ready to eliminate discrimination against women, children, the elderly, the disabled, and the rural population compared to discrimination against indigenous/ethnic groups, refugees/migrants or LGBT people. Stuart et al. (2016) show that almost all of the national development plans from 39 low- and middle income countries refer to the rural population and women as marginal group. In total, 79% of the plans explicitly mention the disabled as a vulnerable group, while a focus on ethnic and indigenous minorities can only be found in 50%; only 21% countries list refugees/migrants and only 15% of countries LGBT people as marginal groups.

Interestingly, these two groups also receive little (migrants/refugees) or even no (LGBT people) explicit attention in the SDGs to start with. While many indicators are to be provided by sex, age groups, geographical location, and disability. Migrants are only referred to with regard to employment. Migrants and LGBT people are also not formally defined as "vulnerable group" in the agenda document (in contrast to the children, youth, and disabled), albeit they would fall under some of the general provisions of the SDGs, for example target 16.b ("Promote and enforce non-discriminatory laws and policies for sustainable development"). Far from anti-discrimination measures for LGBT people in place, many developing countries criminalize homosexual relations by law. In 76 countries and territories, homosexual relations are illegal, in seven countries they are punishable by death (UN 2017).

To conclude, discrimination based on sex and age have been addressed (also not completely) over the last years by national policies and are mainstreamed across the SDGs. In contrast, discrimination against migrants/refugees, ethnic and indigenous groups as well as LBGT people does not rank high in national agendas. This is despite the fact that these groups tend to be at particular risk of exclusion

While legal frameworks adhering to the principles of non-discrimination constitute an important pre-condition for living in societies without discrimination they do not guarantee the de-facto absence of discrimination in real lives. The gap between de-jure and de-facto discrimination in the realm of LNOB can have multiple causes such as lack of legal access due to poverty and illiteracy or the lack of state capacity to enforce non-discriminatory policies effectively in the context of deep-rooted prejudices and cultural norms (Fredman 2013). A case in point is "equal [. . .] rights to economic resources, as well as access to basic services, ownership and control over land [. . .]" (SDG 1, target 1.4). The "access to basic services indicator is a tier III indicator still under development. For the tenure rights and perceptions indicator ("total adult population with secure tenure rights to land" (by sex and type of tenure)), a methodology has been agreed upon, but data (through survey instruments and a corresponding module) remains yet

to be produced (https://unstats.un.org/sdgs/metadata/files/Metadata-01-04-02.pdf). An index on perceived tenure security, the PRIndex, has been piloted for a number of developing countries including a couple of middle-income economies (https://landportal.org/book/indicator/la-pri-pts). The results indicate that between a quarter and third of respondents perceive themselves to have insecure tenure (http://www.prindex.net/analysis). Moreover, some differences between men and women arise. Women in India, for example, are only half as confident as men to have secure tenure rights if they own their dwelling. These insights are very instructive and may bear general relevance for measuring (progress in reducing) discrimination. Aggregate outcome variables (even though measured at the micro-level through surveys) may still conceal important discriminatory practices.

#### **4. Review of Key Policies to Implement LNOB**

Starting from these observations, we now assess which policies in education, health, labor markets, social protection, and anti-discrimination have been successful in achieving progress in implementing LNOB. We highlight specific examples and best practices and examine context factors and institutions that condition success in specific cases or in middle-income economies in general.

#### *4.1. Education*

Within the Asia region, for instance, the most successful countries both in terms of average learning outcomes and equity are from East Asia and they significantly outperform their peers from Indonesia, Malaysia, the Philippines, and Thailand. From this comparison, the World Bank (2018b) draws the general conclusion that the rigorous implementation of policies of "progressive universalism" with equal access at the primary and lower secondary school level has been key. While some lower middle-income countries, in particular in Africa, still need to address basic problems of education infrastructure (availability of and distance to schools, sufficient number of teachers, and school facilities; AAI (2015)), most of these policy challenges can be easily addressed in middle-income countries. They have achieved comparatively low fertility levels which puts less pressure on governments' budget to expand school construction and teacher hiring and training. Many crucial components of education systems in middle-income countries have already been established for longer periods. This and the fiscal space allows these countries to focus on fine-tuning these systems to focus more on the quality, universal access to secondary education, and reduction of educational gaps for vulnerable groups.

Replicating the success of some middle-income countries in implementing LNOB in education first of all implies to build-on and upgrade conditional cash transfer (CCT) systems that have considerably improved access to schooling among the poor in the past 20 years. Conditional cash transfer programs (CCTs) provide monetary transfers conditional on primary or secondary school enrollment and attendance (and/or other conditions). The beneficiaries of CCTs are usually the (very) poor with many programs focusing especially on women, mothers, children, and rural populations. Reviews on the impact of CCTs on schooling and learning find that these programs typically increase student enrollment (for example, Fiszbein et al. 2009). Schooling impacts are similar for men and women, at roughly 1.5 additional years (Parker and Vogl 2017). The success of the program has been attributed to conditioning money transfers to human capital investments. Likewise, the administrative capacity to build an accurate targeting database with moderate errors has ensured that the program was received by the poor (Parker and Todd 2017); www.gob.mx/prospera/documentos/ componente-prospera-educacion). The results on the effectiveness of CCTs are usually not available for certain vulnerable groups, such as children with disabilities or by religion and ethnicity. Over the years, however, many CCTs were re-designed to focus on specific vulnerable groups. For instance, Bolsa Familia in Brazil achieved a substantial increase of coverage of Afro-descent families in social protection schemes and an increase of educational outcomes of children in participating families. Familias en Acción in Colombia incorporated suggestions by local indigenous groups in its project design and locally defined education services and accountability criteria. Red de Oportunidades in Panama, a program targeted specifically to indigenous groups, has helped closing educational gaps (ECLAC 2015).

Second, the available evidence suggests that early child care and education has to rank high on the education agenda, since children tend to be left behind at very young age. Early childhood education presents a window of opportunity to address inequalities and closing gaps in physical, cognitive, linguistic, and socio-emotional development between children from richer vs. poorer and rural vs. urban background (Behrman et al. 2013; Engle et al. 2011; Glewwe et al. 2013). Developing countries have expanded access to preschool provision in the past decade with all regions in the world increasing pre-school enrollment. Yet, in developing countries less than 20% of children have access to early childhood education with facilities facing quality problems with children from poor families and rural areas often being excluded from these (World Bank 2016). Evidence from both developed and developing countries has highlighted investment in early childhood education as one of the most cost-effective ways of providing education (World Bank 2016). For instance, Gertler et al. (2014) find that in the context of an early childhood care and schooling intervention in Kingston, Jamaica, children later earned on average 25% more than children who did not receive the intervention.

Third, the quality of education and learning achievement have to improve not only on average, but in particular among the poor and disadvantaged. Here, teacher quality has been identified as a major cause of bad learning outcomes (World Bank 2017; UNESCO 2015) and its effect is exacerbated for poor children in poorer and disadvantaged areas and schools. Policies need to produce better teachers and distribute good teachers better. While most developing countries have professionalized their teacher management and training systems, almost all of these reforms have failed to improve teacher quality (Galiani and Perez-Truglia 2013; Ganimian and Murnane 2016). In general, it seems that in-service teacher training programs, which have been implemented in many middle-income countries in recent years, seem not be able to add important and relevant skills to teachers. One example is a two-year professional development program to pre-primary teachers in Chile that failed to improve student's cognitive abilities (Yoshikawa et al. 2015). Likewise, a national policy in Indonesia to double salaries of certified teachers found no effect of the program on student learning outcomes (De Ree et al. 2018). These disappointing results on in-service teacher training programs have highlighted the need for recruiting better and more suitable candidates in the first place (Pugatch 2017; World Bank 2018b). From a LNOB perspective, the improvement of education services in disadvantaged areas is key and many developing countries provide special incentives (financial, fast-track to become civil servant teacher, etc.) to locate to such areas. As pointed out in Luschei and Chudgar (2015) many of the existing teacher deployment schemes, such as Mexico's CONAFE, have only shown moderate benefits while cost concerns over these programs continue. Similarly, Gambia's hardship allowance which provided teachers with up to an extra 40% salary bonus for working in remote areas conditional on student learning improvements was found to have no impact on student learning outcomes (Pugatch and Schroeder 2018). On the other hand, Mexico's CONAFE program was reformed to focus stronger on hiring teachers who speak relevant indigenous languages which has shown some initial improvements in the effectiveness of the program in raising students' learning outcomes (Trevino 2013).

Finally, affirmative action programs have the potential to reinforce the effect of all these measures if successfully applied. Several developing countries have introduced some sort of affirmative action programs in education that directly address discrimination. The most studied country in this field is India, which

implements quotas based on caste or other social and income criteria in its educational system. Evidence on the success of the quota systems in education are mostly positive. According to Bertrand et al. (2010) India's affirmative action programs in tertiary education have led to substantial education increases for those low-caste families that benefited from the program while not lowering the quality of education provision. The same result was obtained by Cassan (2019) when studying primary school education. However, Cassan (2019) shows that among unscheduled caste families, only boys benefit. Similarly, success in raising enrollment and graduation rates was reported for Malaysia where various affirmative action programs are in place bumiputras, i.e., indigenous Malays. Enrollment in higher education of bumiputras has increased from 40% in 1970 to more than 80% in public universities (Marcus et al. 2016). We were not able to find examples of large-scale affirmative action programs in education for girls in middle-income countries, albeit the successes of CCTs reported above hold for boys and girls alike.

#### *4.2. Labor Policies*

LNOB-relevant labor market policies in middle income-economies will have to focus on generating decent employment, including social protection coverage, and better pay. Economic growth and structural change will drive employment generation, but complimentary policies can make sure that labor markets become more inclusive and protective. Examples are labor market policies that are relevant in addressing wage inequalities, under- and unemployment and labor market discrimination in middle-income economies, including minimum wages, public works programs and anti-discriminatory policies.

The available evidence suggests that formal employment cannot be easily increased by interventions aimed at formalizing firms. The effects of these policies have been shown to be negligible (Bruhn and McKenzie 2013). This is because the potential gains of formalizing—for example legalized access to infrastructure, legal processes, or access to social security systems—are not perceived by firms to outweigh the costs of registration and possible tax and social security contribution payments.

Public works programs have a long tradition in developing countries as a safety net instrument for the poor. Public works programs usually provide temporary employment at a wage rate below the minimum wage (often in the construction sector and sometimes only for rural areas) and have been shown to reduce poverty among those employed (Zimmermann 2014). Two well-known programs in middle-income countries are the Argentina's Jefes y Jefas program and India's Mahatma Gandhi National Rural Employment Scheme (MGNRES). Despite benefits for participants,

implementation problems regarding setting the right wage, rationing, corruption, and mismanagement have provided difficulties ensuring the success of public work programs. The programs have been shown to be extremely beneficial to the rural population and to post-conflict settings, which might render these programs a policy towards poorer and disadvantaged regions of middle-income countries (Subbarao et al. 2013). In most countries public works programs tend to attract mostly men. India's MGNRES, however, aims to foster women's participation by providing child care facilities, covered sanitary facilities, and equal wage policies. The evidence regarding female participation is mixed. While official figures put female participation rates above 50 per cent<sup>3</sup> , others claim that MGNRES experienced low female participation rates due to problems in the implementation of the stipulated measures (Subbarao et al. 2013).

Minimum wages have been a main tool in the formal sector to ensure that workers and their families can achieve sufficient income to avoid falling into poverty (ILO 2018). For middle-income countries compared to poorer developing countries minimum wage legislation is more likely to be an effective policy tool because in former countries a larger share of the labor force works in the formal sector and governments possess the capacity to enforce enacted legislation (better monitoring, higher fines for cheating, tax reliefs). According to ILO reviews of the literature, minimum wages have generally been shown to benefit minimum-wage earners and decrease the gender pay gap albeit the overall effects are small and the effects on employment remain controversial (ILO2016, 2017b). The most studied middle-income countries with respect to minimum wage legislation are Brazil (Cunningham 2006; Saltiel and Urzua 2017) and Indonesia (Driemeier et al. 2015; Hohberg and Lay 2015). In both countries, minimum wage policies were found to lead not only to higher wages among formal sector workers but as well for informal sector workers while no job displacement effects took place which seems to indicate that minimum wage legislation can be an effective policy tool to help the poorest workers.

Several developing countries have affirmative action policies in place that aim to increase inclusion of women, disabled, religious groups/castes, and ethnic groups. Affirmative action programs for women are increasingly adopted in developing countries but are mostly concentrated on specific vocational training and entrepreneurship schemes while quotas are still rarely used to enforce employment and wage equality except for special cases, for example parliament seats or positions

<sup>3</sup> See https://pib.gov.in/newsite/PrintRelease.aspx?relid=186368 (accessed on 20 May 2018).

in government (Marcus et al. 2016). With respect to race and caste, the most famous programs in this context come from India, South Africa and Malaysia. In South Africa the passing of Employment Equality Act in 1998, the Black Economic Empowerment Act in 2003, and the Codes of Good Conduct in 2007 set the objective to increase African descent's employment at all sectors and job levels. The policies were not found to have increased employment prospects as intended among the general black population. However, for women the policies seemed to have led to a small improvement in employment prospects for top positions. In general, weak enforcement, high unemployment rates and a lack of qualified applicants has been blamed for the limited success of South Africa's affirmative action legislation (Klasen and Minasyan 2017). In contrast, Malaysia's experience to empower the indigenous bumiputras starting in the 1970s by imposing employment and equity ownership quota has contributed to lowering wage, employment and asset inequality between the different ethnic groups. Similarly, positive impacts of quotas on civil and public service in India for scheduled and unscheduled tribes and other backward castes were found to have slowly increased the share of persons from scheduled and unscheduled tribes in the country's public sector (Marcus et al. 2016).

The most widely applied affirmative action instrument with respect to disability is the use of mandated quotas for formal sector enterprises in the private and public sector. Although very little evidence for developing countries exist, it seems that many countries do not fulfil these quotas. For instance, South Africa requires firms to employ 2% disabled persons and Tanzania's quota is at 3%. However, in both countries the actual employment share for disabled is less than 1%. Somewhat higher formal sector employment shares are reported for China which in addition to quotas provides tax incentives, and penalties for companies that fail the quota of 1.5%. Due to weak formal inspections, the target of 1.5% is not reached either, however. There is no reliable study yet that looks at the impact of disability quotas on wages and welfare outcomes of disabled persons (Marcus et al. 2016).

#### *4.3. Social Protection*

How to achieve universal social protection, including universal health coverage, is a key policy question to implement LNOB. In many middle-income countries, a bottom-up non-contributory approach targeted towards the poor and vulnerable has considerably advanced social protection coverage. The policies and programs (UHC, CCTs, social pensions, and unconditional family benefits) that have overcome the regressive nature of previous health and social transfer systems offer important insights for countries that have not yet introduced them and for further development

of these systems. With regard to LNOB, a first key remaining challenge in many middle-income countries is the significant lack of protection of the "informal non-poor" who are covered neither through social insurance nor through non-contributory programs. How large this group is depends on the size of the informal sector as well as the generosity and effectiveness of non-contributory programs (Bonfert et al. 2015; Cotlear et al. 2015; ILO 2017a). A second challenge remains equity of access despite the progress made, in particular for women, migrants and the disabled<sup>4</sup> .

We first examine UHC policies. McKee et al. (2013) points out that many countries achieved UHC while still being quite poor. Although legislation that guarantees UHC is important, it does not automatically translate into access to health services or insurance coverage. McKee et al. (2013) point out that of 75 countries that had de jure UHC, only 58 had de facto UHC in 2009. Countries committed to achieving UHC have adopted a multitude of strategies (focusing on supply-side vs. demand-side constraints, voluntary vs. mandatory health insurance, separate schemes for the poor and informal workers vs. schemes embedded in existing social insurance for the formal sector, financing through contributions vs. general revenues, universal vs. targeted schemes). In very general terms, most countries provide non-contributory health care to the poor and aim at collecting some contributions (often subsidized) from non-poor informal workers (Nakhimovsky et al. 2017). A synthesis of 24 case studies of health care expansion in developing countries (a sample of countries with a strong political commitment to achieving UHC) argues that the success in increasing health coverage of the poor and vulnerable is based on combining supply- and demand-side interventions that prioritize the poor (Cotlear et al. 2015). These interventions include upgrading health care services in poor areas, a focus on primary care, as well as removing economic barriers and providing incentives for the poor (for example, through conditions in CCTs). Cotlear et al. (2015) conclude, however, that there is no "best model". Success does not depend on whether programs target the poor or the entire informal sector nor whether programs are embedded in existing formal sector health insurance or operate autonomously. Rather, enrollment rates of the poor depend on quality of implementation and maturity of the programs. These findings indicate a strong role of political and institutional factors, a view confirmed by McKee et al. (2013). Moreover, effective UHC is highly political because it is redistributive. This implies

<sup>4</sup> Note that we do not address the interrelated issues of (minimum) benefits and financial sustainability, which have of course an important bearing on the viability of the efforts to achieve inclusive social protection systems in emerging countries.

that technical solutions hence need to be aligned with the national political economy context (Kelsall et al. 2016; Reich et al. 2016).

One option to extend the coverage of social protection is to accelerate formalization, but we have discussed the limited potential of deliberate formalization policies in Section 3.2. The second option is to open social insurance schemes to informal workers. In general, voluntary enrollment in social insurance schemes is very low if they are not adapted to the needs of informal sector workers. Many informal sector workers cannot (do not want to) afford even small contributions particularly when these are flat (as opposed to contributions that are sensitive to income). Their income may be irregular and they are often sensitive to time and administrative costs related to enrollment. This explains the low uptake of contributory schemes by informal sector workers (Bonfert et al. 2015). Programs that let beneficiaries choose which risks they want covered according to need and contributory capacity, allow for flexible contribution payments, and untie schemes from employment contracts and residency are more successful (ILO 2017a). For example, in several Latin American countries (Uruguay, Argentina, Brazil, Ecuador), a simplified tax and social security contribution collection mechanism called monotax has been introduced for small contributors. A single payment covers taxes as well as social protection contributions and has been found to facilitate formalization as well as social protection coverage. Indonesia is also set to expand health coverage using a contributory system (see Annex). A third option to reach informal workers is by expanding non-contributory systems to a broad population such as in the case of Mexico's and Brazil's CCT or South Africa's social pension (ILO 2017a). Albeit, potentially the most effective way to increase coverage the feasibility of such an approach depends, of course, on fiscal space. Further, the co-existence of a generous non-contributory system and contributory social insurance can generate disincentives to becoming formal (Aterido et al. 2011). Economic growth may help: The expansion of health coverage to the non-poor informal sector has been more successful in richer countries with a smaller informal sector, lower poverty ratios, and larger government revenues (Cotlear et al. 2015).

Women and migrant workers in informal employment are at particular risk to be excluded from social protection (Holmes and Scott 2016; Hopkins et al. 2016; Ulrichs 2016). They tend to work in the lowest paid, most vulnerable occupations (for example, women in domestic services) and social insurance schemes fail to cater their needs. In particular, social protection systems typically do not accommodate interruptions in employment, a problem for women who face longer periods dedicated to caring for others, and benefits are non-portable, which implies that migrant workers

cannot claim benefits when they return to their home country. Some policies address these problems. For example, in South Africa, labor regulations regarding maximum working hours, sickness benefits and annual leave have been extended in the early 2000s to cover agricultural workers as well as seasonal and temporary workers (Ulrichs 2016). The rights of domestic workers have been strengthened in the past 10 years in quite a number of middle-income countries, including Brazil, South Africa and Thailand (Ulrichs 2016). Bolivia has introduced child credits in its pension system that value the birth of a child with a certain number of months contributed towards pensions. In the same vein, the minimum vesting period to access pension benefits can be reduced to account for interruptions in employment histories of women (Ulrichs 2016). Moreover, women (widows, women in old age, pregnant and lactating mothers) are often specifically targeted by social assistance or UHC programs—they are usually among the first targeted groups before programs are extended to the wider population (Dodlova et al. 2018).

Many countries provide social assistance to disabled persons or families with a disabled member through specifically targeted programs or general social assistance. Yet, little is known about whether these programs reach the disabled in practice. The World Report on Disability (WHO 2011) acknowledges that there is little research on what works in providing safety nets for the disabled but cites anecdotal evidence that the disabled may face barriers in access to existing programs. The report notes that information on programs may be inadequate, welfare offices physically inaccessible, or the design characteristics of programs may fail to account for specific needs of disabled people (disabled children in CCTs, adjusted means testing formula, possibility to designate a person on behalf of the disabled person). A more recent systematic review implies that little progress has been made in the past 5 years, albeit some efforts can be observed in selected countries (for example, in Indonesia, see Annex). The study (Banks et al. 2017) finds that access to social protection appears to fall far below need with benefits from participation limited to maintaining minimum living standards. The review confirms the dearth of high-quality, robust evidence in this area.

#### **5. Summary, Conclusions and the Role of Development Cooperation**

We show that LNOB can be a meaningful guiding principle for national development policy as well as development cooperation in middle-income economies, as very unequal progress may threaten the gains for the poor of countries graduating from low to middle-income status.

There are two key lessons to be learnt from our review of measuring LNOB progress. First, the LNOB index is helpful to compare country commitments and performance, in particular with regard to data provision and policy formulation. It has, however, clear limitations. In particular, it is an index measuring preparedness rather than implementation. Further—and this also applies to an index that would focus more on outcomes—the condensation of multiple indicators into one number means that the same weighting is applied to all countries. Yet, country priorities for implementing LNOB will be and should be different because the challenges of left-behind groups and discrimination have very different causes and consequences. Second, our review clearly illustrates the huge data gaps that remain for key LNOB indicators. Disaggregated data are needed to identify the left behind, including children, women, youth, persons with disabilities, people living with HIV, older persons, indigenous peoples, refugees, internally displaced persons and migrants. For a number of these groups, in particular persons with disabilities and migrants, data are simply not yet available.

Our brief and selective review of LNOB-relevant policies and approaches in middle-income countries shows a very rich foundation for evidence-based policy-making (and this is only the tip of an iceberg). This is particularly so in education, health and social protection where major policy shifts and experiments, which were well documented and assessed, offer important lessons learnt. Importantly, these are lessons from middle-income countries for middle-income countries, which offers scope for more South-South knowledge-sharing and cooperation. In education, health, and social protection, the targets and (disaggregated) indicators of the Agenda 2030—aside the abovementioned problems—provide clear guidance to the objectives of implementing LNOB and thus the objectives of the relevant policies. This is much less the case for the labor market that is, however, equally relevant for implementing LNOB in middle-income countries. It is well understood that the labor market has a key role in how the poor benefit from growth. While our review of policies on labor markets illustrates some interesting findings on the effectiveness of selected labor market policies in terms of employment outcomes, the link of these policies to inclusive development remains vague. Partly, this is because developments on labor markets are driven by macroeconomic developments and structural change, which, in turn, are influenced by many factors and policies that we have not reviewed here.

With these caveats, a number of challenges and potential policy solutions stand out in terms of substantive sectoral policy issues: (1) In education, progressive universal policies should emphasize equity in learning achievements. There is an important role for data on learning achievements to uncover differences between the average (or the top) and those left behind. Our review clearly shows the limits to demand-side policies that remain ineffective if some supply-side factors, in particular teacher quality for disadvantaged students remain unaddressed. (2) In social protection, a very clear conclusion is that a universal approach to social policy is possible if there is the political will and commitment. Ideally, this approach is complemented by special or targeted measures to address the distinct obstacles faced by disadvantaged, marginalized or otherwise excluded social groups. Countries can choose different pathways towards expanding coverage and a key issue will be to address the needs of "informal non-poor". (3) With regard to labor market policies, no clear conclusions emerge. This is also since most policies tend to have modest effects. Clearly, labor market policies need to account for structural realities, i.e., in particular a (still) high share of informal employment.

Our sectoral approach in reviewing key policies should not encourage thinking in silos. In fact, the review forcefully illustrates that the challenges of implementing LNOB are closely interlinked, which calls for integrated solutions. Informality, for example, does not only affect and needs to be addressed by labor market policies. Policies to expand social protection coverage will influence formalization rates. Similarly, many social policies targeted towards to poor, in particular CCTs, combine education, health, and labor market components. Such integrated approaches also offer opportunities to mainstream anti-discrimination efforts, in particular women and girls. LNOB thus cuts across sectors and is not only, but also an anti-discrimination agenda.

This is why we support the idea of Risse (2018) who calls for countries to include a specific chapter on LNOB in VNR reporting. In such a chapter, it should be highlighted how policies and programmes are being adapted to reach the people who are furthest behind first. Further, the chapter should provide information on the status of data collection or plans to improve data availability.

Our review shows that anti-discriminatory measures can be found in all sectors, but that they tend to focus on discrimination of girls and women, in line with a review of affirmative action policies (AEPs) in developing countries by Marcus et al. (2016). According to Marcus et al. (2016), only few countries such as Vietnam, China and Malaysia have AEPs in education and labor markets for specific ethnic/indigenous groups. In light of the apparent discrimination against ethnic groups (for example in education), in particular in Africa, but also South-East Asia, this lack of anti-discriminatory measures is alarming. Similarly, discrimination against migrants/refugees and LBGT people is hardly addressed—be it in terms of

legal framework or policies. Some few countries start to address discrimination against the disabled. For all these groups, there is not only a lack of disaggregated data to assess the degree of discrimination in different dimensions. There is also very limited evidence on the policies required to implement this important aspect of LNOB.

Development cooperation for "inclusive development" faces important challenges as redistribution and in particular discrimination are highly political. Where groups are actively discriminated against (as opposed to "merely" not being prioritized), a powerful (political) majority or elite is likely to benefit from this discrimination (exploitation of cheap labor, social status). In such cases, a strong political will and a profound change in societal attitudes are prerequisites for inclusion of the left behind (Stuart et al. 2016). Sometimes, the role of cooperation may thus be limited to support to put in place these political and societal prerequisites.

These findings have important implications for development cooperation. First, cooperation in LNOB-related issues will be more political than in other spheres of cooperation. We have only briefly touched upon the challenges related to the involvement of external actors regarding the sensitive issues of inequality, redistribution, and discrimination, which will have to be kept in mind when engaging with middle-income countries.

A case in point is the lack of anti-discriminatory action for LBGT people. In our view, development cooperation should focus on the international level and encourage decriminalization of homosexuality and anti-discrimination legislation. At the same time, cooperation "on the ground" needs to adhere to the do no harm principle and abstain from donor-driven initiatives that may be incompatible with local priorities and values. Where feasible, development cooperation should support local advocacy groups through capacity building and finance (USAID 2013). A similar approach should be followed regarding ethnic/indigenous groups.

An important instrument for increasing awareness for not leaving behind these groups is the identification and quantification of inequalities and discrimination as an important first step in implementing LNOB. The deficits of measuring progress exist at the national level, but also at the level of the global data custodians. At the national level, new data collection efforts need to start which result in higher survey costs and investments into administrative processes that not all countries, especially poorer countries, might be able to afford (UNESCO 2017). Development cooperation can play a key role in supporting data collection and analysis at these various levels.

Building on the many substantive policy experiences from various middle-income countries, development cooperation can support South-South cooperation engage in triangular cooperation. This can take various forms, including knowledge transfer from middle-income countries to poorer developing countries or knowledge platforms where exchanges between middle-income countries are facilitated.

Finally, it should not be forgotten that a very large number of those left behind live in middle-income countries, and increasingly so. The "threshold" model of eligibility to receive development cooperation for certain types of support may not be suitable for this situation.

**Author Contributions:** Both authors have equally contributed to this work. All authors have read and agreed to the published version of the manuscript.

**Conflicts of Interest:** The authors declare that there is no conflict of interest.

#### **References**


© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **SDG 1 and Women's Work: Ignoring the Needs of Women and History—The Case of Sri Lanka**

#### **Megan Todd and Kanchana N. Ruwanpura**

#### **1. Introduction**

In this chapter, we assess SDG 1 from the perspective of working women in post-war Sri Lanka. As the breadth of SDGs may help frame policy and offer the possibility to look at the interconnectedness of its multiple aspirational goals, we want to evaluate its effectiveness with regards to labour reform policies under consideration in the country. Yet, we argue that, as SDGs remain grounded in the expansion of market interests, uneven development trajectories are a recurrent feature. Hence, attempting to meet SDG 1 without acknowledging these structural facets of uneven development are unlikely to bring meaningful changes to communities that need it. Furthermore, we emphasize that each SDG cannot be viewed in isolation. For their successful implementation, the SDGs must be viewed as interconnected; as we highlight via linking SDG 1 to SDG 5, SDG 8 and SDG 10.

Historically, Sri Lanka has been a paradigmatic case of high social development despite low economic growth, holding an enviable place within development circles as a positive outlier in human and social development. By 1980, Sri Lanka's life expectancy was higher than that of South Korea, a nation five-fold richer in terms of per capita GNP (Sen 1983). Within this backdrop, Sri Lankan labour was also favourably placed—benefitting from strong labour legislation and labour struggles from the late colonial period (Jayawardena 1972; Candland 2002; Ruwanpura 2016). During this time, capital made concessions, such as union rights and ameliorating hazardous work conditions, to manage labour unrest (Kearney and Robert 1971; Jayawardena 1972). Yet, the advent of open market policies combined with ethnic divisions from 1977 saw the country become embattled in a three-decade war—violently ending in 2009 (Venugopal 2018). It is also within this epoch that labourers initially experienced the unravelling of relatively secure labour rights that they had previously enjoyed.

As Sri Lanka emerges from three decades of war and violence, its integration into the global economy has proceeded undeterred. Venugopal (2018) notes that the Sri Lankan polity since 1977—and despite changing governments of various political hues—remained firmly committed to an open economy. It is within this context that there is a current attempt at consolidating its multifarious labour laws via reforming labour legislation. Using the on-going initiatives at labour reforms as a point of departure, we aim to question its relationship to SDG 1 through considering the specificity of Sri Lanka. In particular, we focus upon gender and ethnicity to illustrate how vulnerable groups may be left behind by efforts by labour reform initiatives that are pro-market friendly as a means to achieving SDG 1.

Arunatilake (2013) observes that women are placed in vulnerable positions in the name of enhancing 'national competitiveness' because they are overrepresented in labour intensive, low-paid work. Not long after the turn of the century, women constituted four-fifths of the garment industry, serving to propel it to the second highest source of foreign exchange in 2007 (Attanapola 2003; Hancock et al. 2015). The introduction of economic liberalisation policies within Sri Lanka served to drastically restructure the nature of employment, and thus workers' lives. Arunatilake (2013) notes that confronted with global competition, workers were faced with increased job insecurity and a work environment, which sought to increase productivity but decrease costs. The structuring of the economy is what matters for the type and quality of labour force (Seneviratne 2019). In 2017, women made up 53.1% of the informal work force, although labour force participation rate for women is only 36.6%, in comparison to 74.4% for men (Department of Census and Statistics 2018). Hence, Hettige (2017) notes that it is the informal sector that provides women with a primary source of income, which is concerning because the nature of work has direct bearing upon social security. In particular, the high proportion of women within the informal sector indicates their vulnerability and potential job insecurity.

Whilst Sri Lanka's 2025 'Vision' policy document considers the problems of social protection for the informal sector, this is done in isolation from the discussion of improving women's labour participation or their conditions of work (Vision 2025 n.d.). This can, in part, be explained by the infusion between patriarchal structures and ethno-nationalist facets that underline Sri Lankan women's labour market experiences (Withers and Biyanwila 2014). The evidence seems to suggest that Sri Lanka's adoption of open market policies only coalesced together with previous colonial projections of women as holding subservient roles (De Mel 2007).

The arguments that we present here are based on a decade of fieldwork on the topic by Ruwanpura (2012, 2016, 2018). These are supplemented with more recent interviews conducted during July 2019 with three officials from the relevant state bodies—the international labour office in Colombo, union representatives, the National Employers Organisation and labour rights organisations. This recent

empirical material is complemented by archival research done for this paper by Todd into pertinent policy document from the United States Agency for International Development (USAID), the Sri Lankan Wages Board Ordinance, Office of the Cabinet of Ministers and various newspaper sources.

We start this chapter by providing further historical contextualisation of Sri Lanka's development trajectory. It will help us to appreciate how extreme poverty was reduced considerably, although relative inequality has either remained the same or worryingly exacerbated. For instance, studies estimate poverty level to be around 52% in 1970 (Visaria 1979). We will then consider the place of women workers and ethnic minorities in the labour force, and how the vector of social justice continues to evade them. After this, we will briefly account for Sri Lanka's proposed labour reforms and underline how it may risk reversing the benefits that Sri Lankan workers have collectively fought for since colonial times and continued, despite the advent of market reforms. Or, to say the same differently, aiming to achieve SDG 1 of ending extreme poverty by 2030 without recognizing the importance of gender inequality and decent work is unlikely to offer "sustainable" poverty reduction (Rai et al. 2019). Negating that women's work is segregated within the labour market, or often does not contribute to GDP growth within labour reforms, acts as a critical barrier to realizing poverty alleviation for minority groups. In our paper, we want to highlight that objectives of SDG 1 are also intricately connected to other goals—and so any policies that worked through need to be cognizant of these connections. These concerns will be looked at through the perspectives of women and ethnic groups within the garment sector before providing some concluding thoughts. Considering that the case of Sri Lanka specifically offers insight into the troubling trend of external influence—further forcing the question as to whether the achievement of SDG 1 is attainable.

#### **2. SDG 1 and Sri Lanka's Development**

Sri Lanka's history can be loosely categorised into colonialism, post-1948 independence; development and under development in the 1960s and 1970s; and 'ethnic conflict' from the 1980s onwards (Gamage 2009 cited in Venugopal (2018)). Rhetoric may have morphed through time, but the ideologies which hinder poverty alleviation and inequality have, in large, remained—including in more recent times, a violent ethnic war. We will explore how ethno-nationalist politics have not only shaped recent Sri Lankan history, but also threatens to "define the parameters of a post-war future" (Venugopal 2018, p. 2).

The ethnic conflict in Sri Lanka has spanned nearly three decades in the country's political history, but it is impossible to separate this from uneven development processes, where divisions and social fractures are salient. The ascendance of the United National Party (UNP) in 1977, often composed of Western-oriented elite Sinhalese, catalysed the war between geographically concentrated ethnic divisions (Figure 1) (Venugopal 2018). In juxtaposition to the Sri Lanka Freedom Party (SLFP), who were defeated during this election and promoted strong social welfare, subsidies and trade protective mechanisms between 1956 to 1977, the UNP introduced swift neoliberal market economic policies after their ascension to power (Gamage 2009 cited in Venugopal (2018)). This shift in economic policies also had social consequences. Gunasinghe (1984) has noted that the opening up of the economy brought quick benefits to the capitalist classes, most of whom tended to be Tamil—at a time when subsidies were being withdrawn to the populace, who were predominantly Sinhala. In his view, instead of class solidarity—because all low-income groups across the ethnic groups were affected—this escalated to anti-Tamil sentiment and led to the 1983 ethnic riots that killed ~3000 Tamils. For Dunham and Jayasuriya (2000), it had been readily assumed that the greater inequality inherent within market economies would be accepted by the country's undergoing transition; for them, the case of Sri Lanka illustrates the significant impacts of such inequality, compounded by people's own perceptions of inequality.

The variety of theories attempting to link liberalisation to the civil war are a mix between contradictory to complementary that miss tracing the complete chain of events (Moore (1990) cited in Venugopal (2018)). Venugopal (2018) investigates the political history of the UNP itself. He argues that the ruling party aggravated ethnic divides by failing to offer concessions during the early stages of the conflict due to re-electability struggles. In his view, "they found a way to render the narrow economic interests of a party of traditional elites electorally viable by fusing it with populist electoral appeal on issues, such as imperial fervour or religious bigotry" (ibid., p. 86). They appealed to the Sinhalese masses via appropriated Buddhist tropes, and gained new support from rural hinterlands with their core free-market economic reform policies enshrouded in socialist discourse.

Within the context of liberalisation, public sector jobs, which are generally perceived as more secure and stable, became scarce and were instead replaced by the increased creation of jobs within the free trade zones, in particular within the garment industry. It was women garment workers and other poorly paid service sector work that largely filled these new positions (Lynch 2007; Dunham and Jayasuriya 2000). Due to the civil war, from the 1990s onwards, the military absorbed employment

creation for public sector jobs and, as of 2017, over 1.4 million people (out of the 8.2 million in the labour force) were employed in the public sector, compared to 1.03 million in 1977 (Venugopal 2018; ILO 2018). The military, along with garment sector work or informal sector employment, were the main job sources in Sri Lanka and accounted for ~15% of the labour force (ILO 2018; Department of Census and Statistics 2019).

**Figure 1.** This map illustrates the highly geographic ethnic demographics of Sri Lanka (South Asia Blog 2014).

Sri Lanka is also a curious exception for continued war-time economic growth (Stewart and Fitzgerald 2001). Between 1983 to 1998, real GDP growth averaged 4.6% per annum and Sri Lanka's garment industry developed into a two billion dollar export industry (World Bank (2001) cited in Venugopal (2018)). The geographical separation of the conflict can be offered as one explanation of this unique situation;

the North and the South experienced the worst political violence and turmoil, forcing these regions to be on the economic periphery (De Mel 2007). However, Venugopal (2018) contests that rather than 'separate spheres', conflict and growth "co-evolved, finding sustenance in each other" (p. 104). According to him, military life not only occupied an important space in the Sri Lankan economy but, by the mid 1990s, served to cushion the impacts from market reforms similar to that of the expansion of social programmes. Crucially, then, militarisation remains closely threaded to society in post-war Sri Lanka and has played a role in structuring state–society relations, becoming a hallmark of the country's political economy (De Mel 2007; Jayasundara-Smits 2018; Ruwanpura 2018).

#### **3. Vectors of Social Justice: Labour, Gender and Ethnicity**

Informal work represents two-thirds of the labour force within Sri Lanka (ILO 2018). To aspire towards SDG 1, and, in particular, the aspiration for the equal rights to access economic resources as ascribed in Target 1.4, it is crucial to consider women as informal sector workers and the barriers to protections that they encounter. Even though there is a conventionally held belief that creating opportunities enables empowerment, we want to illustrate that this may not necessarily be the case. We do this via the study of Sri Lankan garment workers. Despite women workers having high education levels, combined with over forty years of open economic policy, the lives of everyday women have yet to improve in any tangible way (Gunawardana 2016). This is especially the case for women in the North and East of the country who have been affected by decades of ethnic conflict, mirroring inconsistencies of the participation rates of women within other emerging economies (Seneviratne 2019). In order for SDG 1, within its relative income and regional inequalities, to be eradicated, Sri Lankan labour reforms must consider the gendered and post-war context under which they are actioned.

Whilst striving for SDG 1, women's challenges are often not contextualised within SDG 5 or SDG 10, both of which speak to gender equality issues. It has previously been identified that SDG 8 falls short of its own ambition by failing to address gender equality because of its inability to realise the interconnected nature of the SDGs (Rai et al. 2019). As such, for SDG 1 to be realised, it is critical that this minority group be situated within their specific circumstances. Traditional modes of development can fail to recognise the precarious place of women in the labour workforce, because their voices are neglected. Within the garment sector, they are never socially uplifted, despite the economic upgrading that the sector has witnessed (see Selwyn 2013; Ruwanpura 2016).

Sri Lanka again presents itself as a peculiar case as it embarked upon value added production within the garment sector by capitalising upon Western concern for ethical codes of conduct within the supply chain. The 'Garments without Guilt' campaign, for instance, has been a crucial process in upscaling the Sri Lankan garment sector. This process is highly intertwined with the country's history of strong labour movements and laws (Jayawardena 1972; Ruwanpura and Wrigley 2011). Trade unions and political parties grew from the colonial context and influenced regulations to varying degrees of success in the immediate pre-independence years. Sri Lanka was thus well positioned to participate in the upscaling of its garment sector due to the history of reinforcing and monitoring labour rulings (Ruwanpura and Wrigley 2011). However, discourses purporting ethical superiority have tended to ignore how gender divisions of labour impact labour practices. Within Sri Lanka, it has been shown that inequality remains rife, raising questions surrounding the ability of ethical rhetoric to actually challenge the structural conditions that disempower women and workers more generally (Ruwanpura 2016). Lynch (2007), for instance, illustrates the negative social effects upon young rural women looking for work in urban factories. Nicknamed *juki girls*, after the Japanese-made sewing machines that they used, these women were viewed as less respectable due to living away from their villages (ibid. 2007). Regulations may enable these workers to access the shop floor, but they do nothing to challenge the socially held views of what women should do.

The inability of codes of conduct to adequately consider gender dimensions is further emphasised by Gunawardana (2016). Based upon ten years of ethnographic fieldwork, the author emphasizes the hardships encountered within the workplace via investigating the depletion that women encounter through the market economy. Depletion occurs "when there is a critical gap between the outflows—domestic, affective, and reproductive—and the inflows that sustain their health and wellbeing" (Rai et al. 2014, p. 86). So, Rai et al. (2014) are concerned that there is not enough acknowledgement of the physical and emotional depletion experienced by workers as they toil and labour. These processes also occur within the factory setting due to increasing competitive pressures from the global market. Stemming from a gendered recruitment process that situates women in the lower-paying and physically demanding jobs, the study emphasizes the conditions accumulating in the inability to work anymore and "highlight the processes that maintain inequality in the global economy" (Gunawardana 2016, p. 862). Labour laws have thus far failed to ensure that women's work is structured such that they can be a force of empowerment and greater equality. Specifically, worsening inequalities and uneven development

clearly manifest in the failure to provide a living wage for garment workers, or this only being provided as a result of coercive overtime work; making moot the claim of 'garments without guilt' (Ruwanpura 2016; see also Gunawardana (2016)).

In the immediate post-war Sri Lankan context (circa 2009), the garment sector was offered incentives, such as tax holidays and fast-tracked development approvals, to bring branch factories in the war-torn North and East of the country. With poverty-ridden and unemployed youth abound in the region, they provided the surplus workers needed for an industry facing labour shortages (Goger and Ruwanpura 2014; Ruwanpura 2018). Yet, how women were affected in conflict-ridden regions is also important to understand.

Women-headed households are part of the social fabric in areas that faced severe conflict; though, matriarchal inheritance and households have predated the conflict, other factors, such as conflict-related deaths and disappearances together with male out-migration for work have precipitated the surge of women-headed households (Ruwanpura and Humphries 2004). For Eastern Sri Lanka, Ruwanpura and Humphries (2004) have documented how relations of dominance and inequality manifest within these household formations. Moreover, Thaheer et al. (2013) have recorded the considerable trauma and insecurity that affected communities went through. Alongside this, Sri Lanka's militarised landscape provided the support to expand the apparel sector (Goger and Ruwanpura 2014; Ruwanpura 2018). The creation of factory jobs had initially been aimed at young men to steer them away from involvement within political violence; contradictorily however, these positions attracted more women (Lynch 2007). In post-war Sri Lanka, given a greater awareness that women too were former para-military combatants, this contradiction became less pronounced and points to the continued military presence in the lives of women workers (Goger and Ruwanpura 2014). The post-war government then pursued their peace via development by getting industrial capital and the military to collude and, in doing so, ignored underlying tensions between war-torn communities and, in fact, may have potentially exacerbated these.

Post-war Sri Lanka's inability to acknowledge the need for political reconciliation and instead enact economic policies to redress political and ethnic grievances has meant human rights violations, in the form of interpersonal and structural violence, continue to be neglected (Thaheer et al. 2013). Examining the interaction between Sinhalese and Tamil women garment workers from various factories in the post-war region further illustrates how ethnicity, combined with the complex post-war context, impedes the realisation of social justice. This holds especially true as the entry of Tamil workers to the garment sector is primarily a post-war feature. A study conducted

by the Women's Centre (2013) details the difficulties prevailing ethnic relations and tension caused amongst women workers. They emphasize how each group perceives the other and this leads to unbridgeable tensions. A crucial perception for Tamil workers is that they do not receive as much leave as the Sinhalese. They also thought that factory managers preferred the Tamils, as they are not able to demand work benefits due to their low position of power (see Table 1). Derogatory language, such as 'Tigers' and 'Mongols', were reportedly directed at Tamil workers, representing just one type of the ethnic discrimination that was reported by almost 50% of the sample. These differences alone illustrate the prevalent differences that Tamil and Sinhalese women face in the workplace.


**Table 1.** Examples of the forms of ethnic discrimination experienced by the 139 Tamil workers interviewed by the Women's Centre (2013).

The Women's Centre publication also reports that 38% of the women remained silent to questions related to ethnic discrimination, hinting at the fear underlying ethnic relations. It appears that ethnic tensions have been exacerbated by the arrival of Tamil workers to garment factories considering Sinhalese responses: "We resent them because [ . . . ] they are willing to do limitless hours of overtime, we too are compelled to do so". This illustrates the cycle that fixes garment workers in a precarious position. Due to decades of ethnic conflict and discrimination this attitude continues in the factory setting, which negatively affects all workers, irrespective of ethnicity, to work collectively to champion their collective interests (see also Biyanwila (2011)).

It is evident, therefore, that gender and ethnicity have intersected with each other to cause clear challenges within the garment sector. Some problems exist within the garment industry as a whole, such as preconceived notions as to what women workers can do (Lynch 2007). However, it is evident that nuances exist within the Sri Lankan case. Whilst industrialists have successfully capitalized upon an increased concern for ethical supply chains, which are voluntarily governed, limits to corporate self-governance are revealed, as they have not necessarily protected women workers. The ethnic conflict and the legacies that remain have further added another layer to consider. It is within this context that we now examine Sri Lanka's proposed labour laws and their ability to contribute to SDG 1, especially with regard to Targets 1.3 and 1.4, given that they talk about providing greater economic security for workers.

#### **4. Proposed Labour Law Reform in Sri Lanka**

Sri Lanka is a signatory to multiple ILO conventions, of which 43 have thus far been ratified, including the eight fundamental agreements in which the Discrimination (Employment and Occupation) Convention, 1958 (No. 111) is a part. Tripartism is a central feature at the national level, with the National Labour Advisory Council being the primary platform that guides labour policy discussions and reforms. The ILO (2018) has, however, noted the importance of mechanisms being established for dispute resolution between workers and employers. The Ministry of Labour and Trade Union Relations joins with other ministries to guide labour governance, especially as unions within Sri Lanka are protected by the Constitution of Sri Lanka (1978) and the Trade Union Ordinance (1935). Whilst on paper establishing and joining trade unions is possible, the management has tended to react defensively to unionisation through harassment and/or ensuring loyalty through utilising worker's economic insecurity in much of South Asia (De Neve 2008; Hensman 2011; Ruwanpura 2015).

The Sri Lankan government provides several labour and social protections, such as determined holiday breaks, with these provisions accompanied by increasing reprimands for employers' lack of compliance, which are often viewed as superior to other low- and middle-income countries. Existing labour laws benefit those working in the formal sector, but the ILO (2016, p. xii) has viewed that "existing ordinances

lack compliancy" with relevant conventions. In particular, minority groups, including women, are inadequately protected within such schemes, given contentious gender and ethnic politics in the country, which may partly explain their low participation rates in the labour market.

With a pro-market policy environment in Sri Lanka, which is particularly pronounced at the moment, addressing entrenched inequities within the labour market has meant that policy reforms have given precedence to economic and corporate concerns rather than the labour standpoint. It has resulted in the involvement of the USAID who say that the current labour policy represents a "rigid, fragmented and outdated labour regime" (Zezulin and Stanislaus 2017, p. 7). In their view, it is such onerous labour laws that hinder the country's growth and, as such, they have been fundamental in instigating and drafting the currently proposed labour reforms (2019). It is not just such bare external influence that is troubling in any post-colonial country, within Sri Lanka decades of ethnic war and violence add another inflection to this interference. Whilst the government claims to be focusing on a reconciliation process for the benefit of the country, the fact that it is USAID who have been instrumental in the drafting process begs questions as to the motivations behind this.

Taken at face value, the proposed policy change appears to benefit informal workers, especially women, as it aims to consolidate prevailing fragmented labour laws within a single act. The official discourse also purports to get the "inactive labour population" to contribute to the economy (Office of the Cabinet Ministers 2018). Such rhetoric aspires to offer informal sector women workers greater prominence and protection, and there is also explicit reference to sexual harassment within the draft bill (Government of Sri Lanka 2019). The rejection of both physical and verbal forms of sexual harassment is a welcome signal, given that this abuse has been reported within garment factories and there are no explicit laws on curbing sexual harassment (Hancock et al. 2015). Safeguards against excessive overtime are similarly applicable to the situation of garment workers, with Part IV of the draft bill detailing protections and the enforcement of this. An employer must provide workers with a 12 h break in-between shifts, and overtime pay is one and a half times the base hourly rate. The proposed laws, however, provide employers with the power to ensure that these principles are upheld. In expecting them to be the arbiters upholding the law, crucial oversight is removed from the state or unions—i.e., the core constituents of tripartite mechanisms. In other words, two core institutions that provide workers with protections can now be ignored by employers. Hence, while it may appear that substantial steps have been taken to provide informal workers with greater

protections, by disconnecting other tripartite bodies that should be involved, workers' rights are to be wholly upheld by the employer—not the state or the unions.

Moreover, the draft act does not mention either the extent of external influence or the lack of ILO involvement—raising important questions on the impartiality of the process in preparing a draft act. Interviews from the ILO, unions, labour rights organizations and labour officials have all highlighted the ideological baggage attached to USAID's involvement. As summed up by one interviewee from the ILO, "the larger project is about to make Sri Lanka investor friendly". Such a perspective becomes clearer when realising that USAID did not want the involvement of the ILO or the unions or the relevant department of labour. The lack of ILO involvement runs deeper than just consultation, as one union respondent reasons, "they [the Sri Lankan government] have totally violated all" previously signed ILO agreements. It is thus necessary to consider the precedent this act may be setting, and future governments may have to sidestep ILO agreements, which are in place to protect the workers. Despite the emphasis upon tripartite values, in actuality, this draft agreement has not been created via a bottom-up process, but by vested external interests. The main focus, then, is to make Sri Lanka more accessible to foreign investors rather than improving labour rights, as evidenced by the emphasis placed upon removing the high costs employers have to bear due to Sri Lankan laws. Even an interviewee from the national employers' organization commented, "follow a normal process. Don't rush it"; thus, implicitly acknowledging that due process has been sidestepped.

Making Sri Lanka more accessible for foreign investment, however, neglects Sri Lanka's persistent inequality and the lack of living wages for the working classes. Within the Sri Lankan garment sector alone, workers toil under an uneven working landscape—some rights, such as no use of child labour or providing health and safety is upheld, while others, such as paying a living wage or upholding the rights of workers to unionize and collectively bargain, remain largely unmet. We have also traced how the movement of garment factories to the North and East of Sri Lanka, to ensure that the capital has a sustained labour force, has concealed ethnic tensions or discriminatory practices. To make employers the main arbiters within the context makes it questionable how this policy will manifest in terms of protection for workers or addressing post-war Sri Lanka's unresolved political grievances.

The proposed labour changes may have a flattening element to them but, as many of our respondents, including the ILO respondent records, it will "equalise downwards". Or, workers may be provided with equal rights, but such rights do not offer the same protections previously provided to formal workers. A recurring element found in the interviews was concern over the changes to how the minimum

wage is set. The proposed changes detail the disbanding of the Wages Ordinance Board, with this also comes the disbanding of a collective function to set wages. Employers are then able to dictate their own minimum wage, as there is no fixed wage in the draft act. A scenario may be encountered whereby the minimum wage is set lower to counter the time and a half overtime pay. No mechanism has been provided to revise the wage in line with current economic conditions, nor are there set wages for each industry. It is thus unlikely that women garment workers would encounter a higher degree of economic or social security. Such insecurity is compounded by loose overtime protections; for instance, employers will no longer have to go to the Labour Department every six months to authorize overtime work.

Whilst purporting to focus upon the inactive labour market, it is questionable whether the structuring of such laws truly understands the everyday challenges women workers face. The inability to understand the position of women can be seen by the way the act discusses sexual harassment, which has been critiqued for using language that is too restrictive and could prevent persons from being adequately guarded. For instance, when discussing what may be classed as sexual harassment, the draft act consistently relates this to work threats, rather than accepting complex forms of sexual harassment. It is thus unlikely that women garment workers will be offered more protections; in this regard, SDG 1 will continue to be hindered. Alongside this, and as we and others have shown, poor labour conditions are prominent features effecting women garment sector workers (see also (Ruwanpura 2016; Mezzadri 2017; Prentice and de Neve 2017)). As such, SDG 5 and SDG 8 are interconnected with the ability to realise SDG 1; these goals must be aligned within policy to ensure a genuine move is made to alleviate the position of women in poverty (Rai et al. 2019). The case of Sri Lanka's women garment workers then, connects to the questions raised by Lay in chapter five—without an inclusive labour market that focuses on wider caveats than just poverty eradication, these women will be left behind.

#### **5. Conclusions**

Through considering Sri Lanka's proposed labour reforms, contextualised by prior labour experiences, it is evident that the ability to end poverty for women is hindered by enduring gender inequalities in the labour force and lack of social protection in the labour market. Whilst Sri Lanka has been celebrated for its progressive social development and poverty reduction, we have shown that this has followed an uneven trajectory. Protracted ethnic conflict served to accentuate the injustices, which hold a complex history, felt by specific groups. This manifested

most evidently in geographic differences within the North and East of the country, with gender and ethnicity continuing to permeate socio-economic relations.

As a case study, we have investigated women garment workers and have shown the continued everyday barriers they face in realizing social and economic security. Ranging from innately held views, which place women in precarious work roles, to global demands which obligate unsafe work conditions, we have shown that social injustices remain rife. Ethnic discrimination adds another dimension for women garment workers, with perceived differences continuing to play a crucial role in workplaces and little attempt at resolving this by managers. We thus have shown that current labour laws are undermined by the proposed reforms without considering the context and differences that underline Sri Lankan society—especially as all oversight is to be placed on employers and not the unions or the state.

Taken at face value, the proposed labour reforms appear to consider gender and social protections for informal workers. As demonstrated, however, the reality of this act is balanced in the favour of the corporate sector and global investors due to the external influence of USAID and the pro-market emphasis of the proposed bill. The ability of the Sri Lankan government to ignore prior ILO conventions sets a troubling precedent for valuing market objectives over the lives of everyday people, contrary to the aims of SDG 1. Specifically, the ability of Sri Lanka to reduce their poverty levels by half (Target 1.2) and aim toward social protections for all (Target 1.3) is questioned in light of this analysis. If Sri Lanka is truly to make steps towards SDG 1, any proposed labour laws must recognize the contentious political terrain and long-lasting labour struggles which contextualize society.

Below we provide some policy recommendations to aid in the realisation of SDG 1:


have been pursued as a means of reconciliation, soft policies must be followed to invest in the welfare of people and communities, which may help alter harmful ethnic perceptions.

We now provide further recommendations specific to the case of Sri Lanka's labour reforms:


**Author Contributions:** Megan Todd undertook archival research into pertinent policy documents from the United States Agency for International Development (USAID), the Sri Lankan Wages Board Ordinance, Office of the Cabinet of Ministers and various newspaper sources, while the fieldwork for this paper was undertaken by Kanchana N. Ruwanpura in July 2019; it also draws on Ruwanpura's previous research. The final paper was co-written by both Megan Todd and Kanchana N. Ruwanpura. All authors have read and agreed to the published version of the manuscript.

**Funding:** Kanchana N. Ruwanpura would like to acknowledge a small grant received from her former employer, the Institute of Geography, University of Edinburgh that facilitated the research assistance and co-writing undertaken by Megan Todd.

**Acknowledgments:** Kanchana N. Ruwanpura would like to acknowledge the Humboldt Research Fellowship awarded to her during 2019, which provided her with the time and resources to do this research. She would also like to thank colleagues at the Centre for Modern Indian Studies (CeMIS), University of Gottingen, Germany for hosting her and enabling the friendships she developed with other Visiting Fellows at CeMIS and for the superlative mentoring she continued to receive during this time and after from the late Professor Stephan Klasen, who continues to be sadly missed.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


Visaria, Pravin. 1979. *The Incidence of Absolute Poverty in Sri Lanka 1969–1970*. Washington, DC: World Bank Staff Working Paper.

Vision 2025. n.d. *Vision 2025: A Country Enriched*. Colombo: Prime Minister's Office.


© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Social Protection in Ghana—History, Equity-Driven Reforms, Financing and Sustainability**

#### **Edward Asiedu and Anita Baku**

#### **1. Introduction**

Social protection, until recent decades, has been a critical tool for addressing vulnerability, poverty and inequality in many developed countries (see Norton et al. 2001). However, the adoption of social protection policies in sub-Saharan Africa has increased drastically in the past two decades, with Ghana leading as one of the dominant countries. Forty-three programs in Ghana are categorized loosely as social protection programs, with some of such programs touted as having their core mandate as social protection (International Labour Organization 2014). These programs include the Social Security and National Insurance Trust (SSNIT) (implemented in 1965); the National Health Insurance Scheme (NHIS) (implemented in 2003); the National School Feeding Programme (implemented in 2005); the Capitation Grant for Basic Schools (implemented in 2005); the Livelihood Empowerment Against Poverty (LEAP) program (implemented in 2008); the Labour-Intensive Public Works Programme (implemented in 2008); the Free School Uniform program (implemented in 2009); and the Free Exercise Books program (implemented in 2009). The increasing inequality both within Ghana (due to skewed sectoral growth) and across countries, increased unemployment in Ghana, demographic transition due to long-term changes in dependency ratios arising from an improving life expectancy but, at the same time, accompanied by a low pension coverage and the generally increased vulnerability of persons with low education and skills have necessitated the increased attention to social protection programs in Ghana.

Ghana is one of the fast-developing countries in sub-Saharan Africa, with real GDP growth of 6.3 in 2018. In 2011, Ghana attained status as a middle-income country (MIC) and, in 2018, achieved a per capita income of \$ 2201. Yet, the data and poignant news stories remind us that many Ghanaians still live in poverty and some face the risk of falling into poverty, despite the growth of the economy and the numerous social protection policies implemented. Specifically, the Ghana Statistical

Service (2018) categorizes poverty levels into two groups: an upper poverty line and a lower poverty line (poverty and extreme poverty, respectively). The poverty line is pegged at a level corresponding to an adult not being able to meet their essential food and non-food needs at an annual budget of GHS 1760.8 a year, translating into USD 1.15 a day. Extreme poverty is defined as a condition in which an adult is unable to meet their essential food items at an annual budget of GHS 982.2 a year, translating into USD 0.64 a day (Ghana Statistical Service 2018). The seventh round of the Ghana Living Standard Survey (GLSS7) indicates that about 6.8 million Ghanaians are poor, out of which 2.4 million are extremely poor (Ghana Statistical Service 2018). In terms of the effect of growth on poverty reduction, the poverty elasticity of growth, which was around 2 between 1991 and 2006, has, in recent times, reduced to 0.7 (Molini and Paci 2015), raising significant questions regarding the prospects of Ghana's growth for its poverty reduction strategy. Thus, although Ghana was able to reduce the two types of poverty between 1990 and 2006, in order to enable it to achieve the Millennium Development Goal of reducing poverty, the chances of alleviating poverty at the current rate appear impossible unless social protection strategies are designed and well targeted towards the purpose.

Bonilla Garcia and Gruat (2003) identified two types of risks that put people in a vulnerable state and create an environment of impoverishment. These are individual exposures and group exposures to economic and social risks. Individual exposures include social factors such as home-based violence, aging, disability and death (most often of the breadwinner). Group exposures include health-related epidemics, political uncertainty, environmental crises (as a result of climate change) or general economic downturns. The negative effects of globalization such as increasing inequities, high unemployment and underemployment levels among a certain group of people and the inability of governments to tax capital and rather tax labor and consumption have all compounded risks that have resulted in increased poverty levels (Bonilla Garcia and Gruat 2003). The extent of the impact of social protection policies implemented by Ghana to address poverty and vulnerability depends, to a large extent, on the type of policies designed and implemented, improved targeting and sustainability strategies implemented by the country. As it was well articulated by Molini and Paci (2015), tackling poverty and inequality in outcomes and opportunities is, therefore, a long-term development challenge for Ghana, but very critical for consolidating Ghana's middle-income status. Growth alone cannot be depended on to address poverty and vulnerability in Ghana. This chapter explains the history of social protection policies in Ghana, with a view to highlighting the various forms they have taken over time from the 1980s to the present, and broadly

discusses the sustainability of these policies in line with the key sources of financing for social protection in Ghana.

#### **2. Brief History and Categorization of Social Protection Policies in Ghana**

Social protection generally, and in Ghana, can be grouped broadly into three headings: social insurance, social assistance and labor market regulation. As defined and described by a number of researchers including Norton et al. (2002), Barrientos and Hulme (2008) and Devereux and Sabates-Wheeler (2004), *social insurance* includes programs that provide protection against risks arising from life course contingencies such as maternity and old age, or from work-related contingencies such as unemployment or sickness. *Social assistance* includes safety nets and policies that provide some level of support to poor people and also help reduce future risks of those in poverty. Examples of the latter also include policies that enable the poor to keep children in school. A broader concept of social protection is one that includes *labor market regulations* designed towards setting labor standards to reflect the needs of particular groups, such as women, children, the youth and disabled persons.

Prior to the colonial period, Ghana's social protection system was rooted in the Ghanaian culture where families, households and communities assisted in taking care of the less privileged such as orphans, disabled persons, mentally unstable persons, widows or widowers and old aged persons (Kalusopa et al. 2012). This system, although still practiced, has been weakened by urbanization and globalization (Abebrese 2011). The first form of social protection during the colonial period prior to independence in 1957 was the introduction of the social security system by the British for formal sector workers who were in the British Colonial government (Kuyini 2015).

When Ghana became a republic in 1960, the developmental strategy of the government took the form of social policy (de-Graft Aikens et al. 2015). It can be said that providing free education to the three northern regions was a social protection strategy right after independence. During Ghana's second republic (1969–1972), a number of social protection programs were rolled out. Notable amongst them were the Aliens Compliance Act which saught to free up the small-scaled retail business sector for Ghanaians and the introduction of student loans to university-level students who had, until then, received free education (de-Graft Aikens et al. 2015). After a period of five military governments between 1972 and 1979, the next democratically elected government focused its social protection programs on supporting farmers and developing cottage industries.

The economic downturn of the mid-1970s and early 1980s as a result of poor economic management, severe drought, increased oil prices, the return of over a million Ghanaians from Nigeria, the fall in world commodity prices and high international financial market interest rates (Baah-Boateng 2004) led to the military government at the time turning to the International Monetary Fund and the World Bank for assistance. This action led to the introduction of a structural adjustment program (SAP). The SAP was a program of fiscal discipline that witnessed government subsidies to various ministries decline drastically. Ghanaians paid more for health services and education. Subsidies to the agriculture sector such as those for fertilizers also declined, which led to a fall in agricultural production. This was followed by retrenchments and redeployments (ODI 1996). The effects of the SAP were instantaneous, with increased poverty levels to about 36% between 1987 and 1988, and extreme poverty was at 7% (Stewart and Van Der Geest 1995).

To put a 'human face' to the SAP, the government introduced a social protection program known as the Programme of Action to Mitigate the Social Cost of Adjustment (PAMSCAD) in 1987. According to the ODI (1996), this was a joint initiative by the Government of Ghana, the United Nations International Children Education Fund (UNICEF) and the World Bank. An amount of about USD 83 million was budgeted to be spent under the program to assist vulnerable households in various ways including helping them address nutritional shortfalls (supplementary food program) (see Barimah 1993). The five key action areas under the program were: to help retrenched workers, employment generation, education, provision of the basic need of vulnerable groups and community initiatives (ODI 1996). This program is considered the first social protection program designed to address poverty perceived to have been induced by SAP implementation in Ghana.

According to Stewart and Van Der Geest (1995), the action to help retrench workers involved training and provision of start-up advice. Employment generation activities involved food for job initiatives, public works, credit facilities for small-scale farmers, labor-intensive feeder roads, entrepreneurship training for women, support for small-scale mining and rehabilitation of school buildings. The employment generation activities primarily targeted the poor in the northern regions of Ghana and urban slums. The education interventions of that era focused on the provision of schoolbooks and food for boarding schools. The community-level initiative involved social and economic infrastructural developments within communities across the country. Basic needs provided under the program included the provision of primary healthcare services, deworming of school children, education on nutrition, food

supplements (food-for-work schemes), provision of water and sanitation services and rehabilitation of houses in rural areas. The program ended in 1990.

In 1996, the Government of Ghana introduced Vision 2020, a 25-year medium-term development plan (MTDP), aimed at raising the country to a middle-income country by the year 2020 with a focus on poverty reduction. The first part of this development strategy lasted four years (1996–2000) (Ewusi 2013). Under the plan, inequity in incomes and standard of living between urban and rural areas was to be bridged with public investment in rural areas. Human and infrastructural development was also undertaken during the period. Except for the service sector which experienced growth as a result of telecommunication, most growth in other sectors of the economy was eroded (Ewusi 2013). For instance, inflation increased three times from 13.8% to 40.5%, and the growth rate fell from 4.7% to 3.7% in 2000, exacerbating poverty and inequality. Typical of development plans in sub-Saharan Africa, Vision 2020 was curtailed after a change in government in 2000 and a decision to apply for debt relief by the new government under the Highly Indebted Poor Countries (HIPC) initiative.

Between 2002 and 2005, Ghana launched a strategy called the Ghana Poverty Reduction Strategy (GPRS I). The strategy was part of the requirements under the HIPC initiative where debts owed by Ghana were cancelled and monies meant to finance the debts were to be redirected into economic growth and poverty reduction activities (Osei and Quartey 2001). The decision to opt for debt relief was a result of the downturn of economic gains made from the first step of the Vision 2020 agenda. The strategies to combat poverty under this strategy included fiscal discipline, creation of an enabling environment for private sector growth and employment creation, assisting agricultural-based industries, provision of basic services to the poor and vulnerable groups, access to justice, government transparency and accountability and zero tolerance for corruption. The Government of Ghana (2007) report indicates that GPRS I was generally successful as there were macroeconomic stability, growth in agriculture and expenditure on education, and water and health were increased. The growth rate recorded for the period was 5.8%, suggesting steady economic growth.

Between 2006 and 2009, when GPRS I ended, the government decided to pursue GPRS II, with a focus on private sector development for poverty reduction and addressing the fragility in the social structure, focusing on women and youth empowerment. The pursuit of GPRS II was partly in recognition of the weak link between the attained growth and poverty reduction. Improvement in the social structure included activities directed at improving health, education and housing. GPRS II was also successful but faced some economic difficulties in the election year

of 2008. For instance, the inflation rate increased from 12.8% at the beginning of the year to 18.13% by the end of the year. Interest rates in 2008 increased, and the cedi depreciated against the US dollar (Ewusi 2013).

In 2010, a 'new economic growth poverty reduction strategy' was introduced known as the Ghana Shared Growth and Development Agenda (GSGDA) (Ewusi 2013). The main objectives of the GSGDA were to create macroeconomic stability, increase private sector participation and ensure oil and gas development, human resource and infrastructural development and agricultural modernization. Under macroeconomic stability, fiscal policy was to focus on critical infrastructural development, job creation and poverty reduction. Concerning private sector participation, policies aimed at reducing the risk of doing business, a more responsive financial sector to the needs of private businesses and developing human resources to assist in the growth were implemented. The creation of decent jobs and ensuring standards in the oil and gas industry were of international repute as the focus of strategies under oil and gas development. Under human resource development, improvements in education, health, income inequality, poverty reduction and social protection were the focus. The main focus of agriculture modernization was the use of scientific methods in food production, leading to employment creation and linking the agricultural sector to industry and marking of the produce. Under the GSGDA strategy, an increased growth rate of 14.4% was recorded in 2011, up from a rate of 7.7% in 2010, the inflation rate remained at 8.58 from 2010 and there was a lower budget deficit from 2010 (Ewusi 2013).

With the macro-based poverty reduction strategies not achieving their set goals, there was the need to strengthen bottom-up approaches to poverty reduction. In 2007, Ghana promulgated the National Social Protection Strategy (NSPS) to serve as an umbrella strategy for all social protection programs and provide a framework to ensure coordination and complementarity between programs. As it has been argued, a strong legal and institutional framework for social protection programs is a critical basis for sustainability. The development of the NSPS provided a strong institutional framework for the holistic implementation of Ghana's social protection policies. The NSPS mentions (1) the establishment of a new social grant scheme to provide a basic and secure income for the most vulnerable households, culminating in the introduction of the Livelihood Empowerment Against Poverty (LEAP) program, (2) better poverty targeting of existing social protection programs, which is still yet to be attained, and (3) packages of complementary inputs.

In more recent times (the 2000s), several programs were implemented by the government, some implemented prior to the NSPS coming into effect, whilst others

emerged out of the formulation of the NSPS. Social protection policies in Ghana in the 2000s—unlike those of the 1980s that were designed to address vulnerabilities created as a result of structural adjustment—were designed to address the widening gap in income and wealth due to skewed sectoral growth and inequality in opportunities, particularly educational opportunities. Growth in the agricultural sector took a dramatic dip in the 2000s, despite overall higher growth in the economy. Ironically, in the same year that Ghana attained middle-income status (2011), it also recorded one of the lowest agricultural growth rates of 0.8%. Many of the recent programs are designed to offset gaps within society. Thus, even though some sectors are not performing as well as expected, others such as the oil and gas sectors are performing extremely well, and therefore there is the need to support others, particularly rural dwellers who rely on agriculture as their main source of livelihood.

We now turn to the categorization of the post-2000s social protection policies in Ghana into the textbook classification of social insurance, social assistance and labor market interventions. Social insurance programs implemented in recent years include the conversion of the Provident Pension Fund to a fully fledged pension scheme in 1991 and the introduction of a three-tier scheme in 2008 that provides an avenue for informal sector workers who were initially not covered under the fully fledged pension to also contribute towards their pensions; the introduction of the National Health Insurance Scheme (NHIS) in 2005 to provide health coverage for the poor, the majority of whom under cash-and-carry did not have access to health services; and the introduction of the Ghana Agricultural Insurance Programme (GAIP) in 2011 to offset the debilitating effect of climate change on farmer yields and improve overall farm households' welfare.

With regard to social assistance programs, Ghana, in recent years, has introduced an array of policies to provide direct assistance to the poor. Notable among these policies are the Livelihood Empowerment Against Poverty (LEAP) program introduced in 2008 to serve as the country's flagship social protection program; the National School Feeding Programme introduced in 2005 to provide food supplements to children at school; the Free School Uniform program implemented in 2009; Free and Compulsory Universal Basic Education (FCUBE) implemented in 1995; and, lastly, the recent Free Senior High School Education (Free SHS) policy implemented in 2017. The evidence thus far shows that Ghana's social protection policies in recent times have been quite heavy on the provision of assistance, with little contributions from beneficiaries. For example, under Free SHS, in addition to the free tuition, the government provides books, meals and all other associated educational materials to senior high schools with very little or no contribution from parents.

Labor market interventions have also, in recent times, become a prominent component of Ghana's social protection policies. For example, Ghana established a government microfinance institution in 2006 called the Microfinance and Small Loans Centre (MASLOC) with seed money of about USD 50 million, with the object of providing microloans targeted at reducing poverty and creating jobs. The MASLOC also provides training programs to beneficiaries on financial management. The Youth Enterprise Support (YES) fund was established in 2014 to support the youth in creating businesses by the government. In 2017, the National Entrepreneurship and Innovation Programme (NEIP) replaced the YES fund, but with the same mandate of supporting youth entrepreneurship. As of July 2019, 1350 young entrepreneurs received funding for their business ideas with amounts ranging from GHS 10,000 to GHS 100,000). To (temporarily) address the increasing problem of graduate unemployment, the government established the Nation Builder Corps (NabCo) in 2018 to provide immediate temporal employment for graduates from tertiary institutions for three years after graduation. There are various modules under the NabCo focusing on specific sectors. The modules include: Educate Ghana, Heal Ghana, Feed Ghana, Revenue Ghana, Digitize Ghana, Civic Ghana and Enterprise Ghana. The expectation is that participants under the NabCo will gain employable skills to enable them to gain permanent positions elsewhere after the three years.

The labor market intervention directly linked to the poor is the Labour-Intensive Public Works instituted in 2011. This is a program aimed at providing temporary employment to the poor in rural areas during the off-cropping seasons to reduce the incidence of poverty as a result of seasonal unemployment. The program is implemented by the Government of Ghana, with assistance from the International Development Agency and World Bank as part of the Ghana Social Opportunities Project (Honorati 2015; Osei et al. 2015). An impact assessment performed by Osei et al. (2015) in 2013 revealed a reduction in poverty for beneficiary households from 62.4% before the intervention to 55.5% after the intervention. Thus, the 2000s saw an avalanche of social protection policies implemented in Ghana. The sustainability of these policies is broadly described in the next section.

#### **3. Financing and Sustainability of Social Protection in Ghana**

Whilst Ghana has an array of social protection interventions, expenditure on social protection as a percentage of the GDP has been, historically, relatively low. Ghana spent a total of 0.5% of its gross domestic product (GDP) on social protection programs in 2014 (World Bank 2016). As shown in Figure 1, that percentage increased to 0.76% in 2015, increased marginally to 1% in 2016 and decreased to 0.63% in

2019. This amount is less than half the average spent by other lower-middle-income countries in the same category which averages 1.6 (world) and 2.16 (sub-Saharan African countries). As it has been argued, more open economies call for more pro-active governments and higher social protection expenditure (see Rodrik 1997). However, the bigger question is not just how much is spent but rather what the source of the financing is. Thus, behind the façade of the number of social protection policies implemented by Ghana, the question is to what extent are these policies sustainable.

Unlike social protection interventions in the 1980s that were spearheaded heavily with funding mainly from international organizations, protection programs in the 2000s were mainly government-funded programs with support from development partners. The discovery of oil in Ghana placed the government in a better position to fund many of its social protection policies. Worth mentioning is the strong role of the government in funding the LEAP program as well all the post-2008 social protection programs. As it has been argued by Niño-Zarazúa et al. (2012), social protection programs that are largely supported by development partners may not be sustainable when the interest of these partners shifts.

Figure 2 shows the relative role of the government vis à vis donors in the funding of social protection in Ghana. The figure shows how the country's budgetary allocation to its flagship social protection program—LEAP—increased significantly from GHS 50 million in 2015 to over GHS 160 million in 2018, signaling an increased commitment to social protection. On the side of donors, we find a drastic reduction in monetary support for LEAP over the years.

Considering the traditional social protection programs implemented by Ghana (excluding the latest Free Senior High School policy), in the School Feeding Programme (SFP) budget allocation of the social assistance budget, in 2018, a whopping 64% of the social assistance budget went to the SFP, whilst 26% went to the LEAP program (UNICEF 2018). Thus, the data on the allocation of the social protection budget show a higher commitment to expansion in educational opportunities and increase in outcomes. The key theory of change of the SFP is the expectation that the SFP can lead to increased school attendance and retention of and improvement in children's cognitive skills, and for that matter, the government was willing to spend a big chunk of the social assistance budget on the SFP.

**Figure 1.** Allocations to social protection programs. Source: UNICEF Ghana Social Protection Budget Brief.

**Figure 2.** Source of LEAP funding, 2015–2018. Source: UNICEF (2018). Ghana Social Protection Budget Brief.

In 2017, the government implemented one of its latest large-scale educationbased social assistance policies—the Free Senior High School Policy (Free SHS). The program provides free tuition, textbooks and meals for high school children in all public schools throughout the country and lacks any form of targeting of beneficiaries. Unlike other countries where private secondary schools are considered superior to that of public, in Ghana, the reverse is the case. Ironically, private basic and junior high schools are considered better than public schools; therefore,

rich families take their children to the best private schools at the lower levels, and because of the quality of the schools, these students end up with the best grades, which allows them to enter the best free public schools. Thus, even though the government policy free SHS is targeting universal access to education for all children, it could indeed exacerbate inequality by allowing rich households to channel their savings on secondary education into investments and accumulation of land and other economically important assets. In 2018 alone, Ghana allocated GHS 414 million to the program, which is three times the allocation to the LEAP program. Improving the targeting of social assistance programs can lead to reducing costs and improving the overall wellbeing of those who really need support.

Some authors (see Azeem and Adamtey 2004; Stewart and Van Der Geest 1995) indicated failure for some social assistance programs implemented in Ghana in the 1980s. Azeem and Adamtey (2004), for instance, argued that poverty and unemployment continued to persist under the PAMSCAD program due to poor targeting. Evaluation of the program revealed that there was a high number of underserved beneficiaries and a high number of none-receiving poor excluded from the scheme (see Stewart and Van Der Geest 1995 for details). Whilst some 4% of redeployed households were impoverished before the introduction of the SAP, a review of PAMSCAD indicated the number of impoverished had increased to 22%. Azeem and Adamtey (2004) and Stewart and Van Der Geest (1995) were, however, of the consensus that lessons from PAMSCAD assisted in the implementation of subsequent social protection programs in the country.

Studies conducted on the LEAP program, for instance, have revealed immense benefits to beneficiaries, reducing poverty levels and improving their livelihood. For example, Fisher et al. (2017), in a study conducted on beneficiaries in two districts in two regions in Ghana, found that there was improvement in their general livelihoods in terms of preventing hunger, ability to make better job choices, investment in education for their children, reducing destructive coping strategies and improvement in the self-dignity of the beneficiaries. On the flip side, the study found that the amount given to beneficiaries was not enough to invest. However, heterogeneity in impacts was observed. Older beneficiaries under the LEAP program were observed to have engaged in petty trading.

Daidone et al. (2015), in their assessment of the LEAP program, observed that the program assisted in providing input for agriculture, improved on- and off-farm activities for males, increased the ability to hire labor for farm work, increased cash savings and improved paying off debts. The LEAP program, however, similar to the PAMSCAD, has also been fraught with targeting challenges of potential beneficiaries, with some reported cases of political interferences in the selection of beneficiaries (Jaha and Sika-Bright 2015; Ragno et al. 2016). Aside LEAP, evidence of political interferences has been observed in other programs. Alhassan et al. (2016) and Fusheini (2016) articulated evidence of political interference in the management of the NHIS, whilst Aryeetey et al. (2016) questioned the sustainability of the NHIS due to political commitment to reducing delays in payment to healthcare facilities.

The issue of the inability to properly target beneficiaries is still the main problem of social protection programs in Ghana. We argue that the wave of digital innovation can be harnessed to improve the targeting of social protection initiatives. Ghana successfully implemented the Taxpayer Identification Number (TIN) in 2018, which, unlike the previous tax regime, where people had to go to tax offices to obtain TINs, is easy to generate by individuals and corporate entities via a mobile phone. It has become mandatory to indicate one's TIN on all documents for a wide range of services. Thus, without a TIN, people cannot access a number of important services. The new TIN, by design, is required to register a business, open a bank account, register a vehicle, buy land, obtain a passport, collect goods from ports and airports, obtain payment from a government agency, bid for a contract, etc. Good data are required for improving the targeting of social protection interventions, and therefore having data on the economic activities of the majority of people in a database could be harnessed to improve targeting. For example, owning landed properties in certain neighborhoods in Ghana should exclude such persons from certain social assistance programs. Ghana has also introduced a large-scale National Barometric Identification System (Ghana Card) meant to improve the collection and use of administrative data; therefore, this system, together with the TIN, can help to improve the targeting of social assistance programs. Therefore, the integration of technology highlights the role of data science in transforming the targeting of social assistance programs, in line with the attainment of the SDGs of reducing poverty (SDGs 1 and 2) and inequality (SDG10). The next decade should see an improvement in the targeting of social assistance programs in Ghana if the technology is harnessed to generate improved data and therefore improved identification of the needy in society.

Table 1 provides a summary of the key social protection policies implemented in Ghana from the 1980s to the 2000s, highlighting the year of introduction and main source of financing.


#### **Table 1.** Summary of key social protection policies in Ghana.


#### **Table 1.** *Cont.*

Proper institutional coordination is needed for the success of social assistance programs. A better understanding and coordination of theories of change for the individual programs is needed considering the fact that some programs have similar and sometimes complimentary expected outcomes. Such coordination will reduce the overall cost of programs and bring about efficiency in social protection delivery in Ghana. What Ghana's experience has also shown is that a strong national social protection strategy which is linked with a national identification system is critical for the attainment of the social assistance objectives of reducing poverty and inequality.

Additionally, as noted by Banerjee and Esther Duflo (Banerjee et al. 2011), there will be a poverty trap whenever the scope for growing income or wealth at a very fast rate is limited for those who have too little to invest but expands dramatically for those who can invest a bit more. The numerous programs implemented by the government should be designed focusing not just on educated graduates, as in the case of the NabCo or NEIP, or wholesale, as in the case of Free SHS, but also the poor, as they should be one of the important focal points of these programs if the country is to make any significant gains in reducing the current level of poverty by 2030. One cannot also take out the effect of politics on the sustainability of social protection initiatives. In recent times (post-2000s), there has been a tendency for social protection initiatives to have been extracted from political parties' manifestos. Effort should be devoted to still link these initiatives to the national social protection strategy of the country. If not, the country runs the risk of introducing new initiatives every 4 or 8 years (based on the political cycle) and abandoning older programs by predecessor governments. Such a situation can erode the gains made on the social protection front. It is also important that the benefits promised under politically drafted social protection initiatives are reasonable considering resources and attainable over the long haul. There should be a national-level strategy to empower the youth devoid of political party programs embedded in bi-partisan youth policy.

#### **4. Conclusions**

Social protection is an investment in the social and economic development of societies and individuals. Investment in social protection is crucial for poverty reduction and can play an enormous role in reducing both economic and social vulnerability. Ghana's social protection policies have moved from an over-reliance on outside institutions and partners to self-reliance, which is needed for a sustainable social protection ecosystem. Donors' roles in funding social protection have reduced significantly over recent decades, with the discovery of oil and its associated revenues playing an important role in the government financing of social protection.

With the introduction of the National Social Protection Strategy (NSPS), the detrimental impact of political change in the commitment to social protection should be minimal, and if this is actually the case going forward, then Ghana's social protection should have a formidable future with increased assurance of reduction in future inequality and poverty consistent with the sustainable achievement of SDG1, SGD2 and SGD10. Thus, historically, Ghana is not short of poverty reduction strategies, as evident in the number of such strategies implemented from the 1980s. The country, however, must do more in moving from mere strategies to achieving outcomes, by guarding implementation and improving targeting.

Budgetary allocation to social protection programs in Ghana still lags behind that of other similar developing countries. To this effect, there is the need to increase social assistance to the very vulnerable and improve the targeting of programs, particularly wholesale education-based assistance. Precisely, there is a need to improve the targeting of the Free Senior High School program. As it stands now, children from rich homes are enjoying the same level of coverage as those from poor homes. With proper targeting, resources can be moved to programs that directly benefit the poor such as the LEAP program and the School Feeding Programme in deplorable schools. Lastly, to achieve adequate coverage of social protection programs in Ghana in the midst of reduced donor support, the private sector, if targeted and assured of transparency in targeting and implementing policies, could be an important source of financing for social protection policies.

**Author Contributions:** Literature search was carried out by both Edward Asiedu and Anita Baku. The final paper was also co-written by both. Final edits and response to editors was carried out by Edward Asiedu. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Edward Asiedu would like to thank the late Professor Stephan Klasen for all his mentorship, and the many illuminating discussions they have had on Africa's developmental strategies in general, and strategies to reduce poverty and inequality specifically.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

Abebrese, Joyce. 2011. Social Protection in Ghana: An Overview of Existing Programmes and Their Prospects and Challenges. Available online: http://www.fesghana.org/uploads/ PDF/FES\_SocialProtectionGhana\_2011\_FINAL.pdf (accessed on 9 September 2021).


© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Education Access and "Learning Poverty" in Seven Southern African Countries**

**Servaas van der Berg**

#### **1. Introduction**

There are many different ways to think about poverty and education. Education, as human capital enables individuals and societies to be more productive, and as an essential capability, in Amartya Sen's sense, is instrumental in improving economic prospects for individuals and societies. Education also has social benefits, such as fertility decline from initially very high levels, reduction in labour market inequality or contributing to political stability, that improve conditions for economic growth and thereby poverty reduction. Poverty can limit education's impact because of the strong association between educational outcomes and home background factors such as socio-economic status, home resources and parental support. The concept "learning poverty"—the inability to read and understand a simple text by age 10—is another conceptual link between education as capability and poverty.

The major success of the educational MDGs was the rapid expansion of access to primary and even secondary education. The primary net enrolment rate (NER, the number of primary-aged pupils actually in school per 100 of the primary-aged population) rose from 83 in developing countries in 2000 to 91 in 2015. Africa saw the most progress, with primary NER rising from 52 in 1990 to 60 by 2000 and a remarkable 80 in 2016 (United Nations 2015, p. 4). However, despite rapid improvement in access and children staying in school longer, progress in terms of how much children learn at school was disappointing. As a result, education quality became an increasingly important focus, as reflected in the 2005 Education for All Global Monitoring Report's focus on what they termed "the quality imperative" (UNESCO 2004). This view grew stronger as the extent of the learning deficit became clearer. The 2019 report on the Sustainable Development Goals (SDGs) (United Nations 2019, p. 30) pointed out that "an estimated 617 million children and adolescents of primary and lower secondary school age—more than 55% of the global total—lacked minimum proficiency in reading and mathematics in 2015." Two-thirds of this was not related to school access, but to low levels of learning in schools.

It is thus not surprising that there was a shift in emphasis, as reflected in Goal 4 of the new Sustainable Development Goals, to "Ensure inclusive and equitable quality

education and promote lifelong learning opportunities for all", and especially in sub-goal 4.1, by 2030 to " . . . ensure that all girls and boys complete free, equitable and quality primary and secondary education leading to relevant and effective learning outcomes" (United Nations 2015). The SDGs' emphasis on learning outcomes and equity were major departures from the MDGs (Crouch and Gustafsson 2019).

The countries of southern Africa considered in this chapter—South Africa and its six neighbours—share a common economic history. They are all affected by the recent weakening of the South African economy through their common links to the regional labour market and through the customs union that five of them belong to. They cover a wide economic development range: Mozambique is a low-income country, Lesotho, Zimbabwe and Eswatini (the former Swaziland) are lower-middle income countries, and Namibia, South Africa and Botswana are upper-middle income countries. Botswana's per capita income of \$ 8259 is almost 17 times Mozambique's \$490, and extreme poverty rates range between Namibia's 13% and Mozambique's 62% (see Table 1). Additionally, the expansion of school enrolment has taken different courses in these countries, reflecting their institutional backgrounds and history. This has given rise to varied challenges in the field of education and institutional responses to such challenges.


**Table 1.** The countries of southern Africa in a development perspective.

Where data for the year indicated were not available, data for the most recent year were used. Poverty headcount calculated as 2011 PPP dollars. Sources: Authors' compilation based on data from World Bank (2019b); United Nations Population Division (2019).

The central tenet of this short chapter is that educational policies and institutions in these countries overwhelmingly focus on expanding school and even post-school education. The institutional response to expanded enrolment or increased resources is typically a deepening rather than qualitative change in education policy, while there is a surprising absence of policies strongly focused on improving educational quality. By and large, this seems equally true for the education landscape in most developing countries: the emphasis has shifted little beyond those reflected in the goals of the MDGs, i.e., expanding access, while attention to education quality is seldom found in policies and debates within developing countries. This is the case despite efforts by international institutions such as UNESCO, the UN and the World Bank to improve and drastically expand the measurement of cognitive outcomes, in recognition of the fact that the MDGs largely failed to improve education quality.<sup>1</sup>

In the next section, I provide a very brief perspective on the economic and educational situation in these countries and discuss education access in the context of southern Africa. Section 3 then turns to case studies by discussing a few major education issues and the policies instituted to address them in some of these countries. Section 4 presents the conclusions.

#### **2. The Countries of Southern Africa in Perspective**

#### *2.1. An Economic Perspective*

The countries in this group share most of the features of developing countries. Extreme poverty is still above 40% in Eswatini, Lesotho and Mozambique, while the lower estimate for Zimbabwe of 21% is probably incorrect or dated, considering that the country has experienced great economic upheavals. South Africa and Namibia have extremely high levels of inequality, with Zimbabwe's being quite low. Total fertility rates in South Africa and Botswana are markedly lower than in the rest of the group, yet still high when compared to other upper-middle income countries.<sup>2</sup> Mozambique's total fertility rate of 4.9 is very high compared to countries in the region, though not much above the sub-Saharan average. The growth of the school-age population is high only in Namibia (1.5% per annum for the next decade)

<sup>1</sup> Spaull and Taylor (2015) suggest a way to measure both access and quality.

<sup>2</sup> The total fertility rate is the total number of children that would be born to a woman if current age-specific fertility rates were to hold. Conventionally a total fertility rate of 2.1 is regarded as replacement level, the level at which the current population would simply replace itself in the absence of migration. This 2.1 replacement fertility rate allows for some early mortality.

and Mozambique (2.3%) and quite low for the other countries, holding the promise of deepening investment per child.

#### *2.2. A Perspective on Educational Access*

Table 2 shows net and gross enrolment rates for different years, based on demographic and health surveys and AIDS indicator surveys.<sup>3</sup> No data are available for Botswana from these sources. The NER indicates whether children are enrolled at the right level of education. For the most recent years, the primary NER is only 74 in Mozambique, 84 in Eswatini and Namibia, 88 in South Africa, 91 in Zimbabwe and a surprisingly high 94 in Lesotho. Compared to the NERs, primary gross enrolment rates (GERs), the total number of children in primary school expressed per 100 children in the population that are of primary school age, are quite high—121 in Lesotho, 120 in Eswatini, 116 in Namibia and 115 in South Africa. This is mainly due to children who are too old for primary school still being enrolled at that level, either because they entered school late or because of high repetition rates in primary school. It is also indicative of limited drop-out in primary school.

<sup>3</sup> Net enrolment relates to the number of children of the appropriate age actually being enrolled in that phase per 100 population in that age group, whereas gross enrolment expresses the number of children in a particular school phase per 100 population in that age group.


**Table 2.** Net and gross enrolment ratios derived from surveys.

Source: Demographic and health surveys and AIDS indicator surveys, access through ICF/USAID (2019).

The secondary school NER is strongly influenced by the fact that many children have repeated or dropped out before getting to or completing secondary education. South Africa's secondary NER of 77 stands out, followed by Zimbabwe (50) and Namibia (46).

The earlier surveys shown in the table make it possible to discern some trends:


Bashir et al. (2018, p. 9) group 45 education systems in sub-Sahara Africa into four groups. They place six of the seven countries considered here in Group 1, the established group. In this group, "the primary GERs are high in the baseline year (2000) and close to 100% circa 2013; the out-of-school rates for children of primary school age are low in the latest year of available data; and primary retention rates are close to 100% in 2013". It is surprising that the group of southern African countries considered in this chapter constitute six of the 13 countries referred to as "established", given the wide differences within this group. Mozambique is the exception: it is categorised into Group 3, the emerging group, "where the GERs are low in the baseline year and high by circa 2013; the out-of-school rates for children of primary school age are high in the latest year of available data; and primary retention rates are low in 2013".

#### *2.3. Historical Patterns of Access*

Information on the highest grade completed by age/birth cohort from censuses and surveys allows approximations of historical patterns of educational access.<sup>4</sup> Table 3 shows that more than half of the 1930 birth cohort in Botswana, Eswatini and Mozambique never entered school. Access increased markedly in the subsequent four decades: for the 1970 birth cohort, only Mozambique had more than 20% that never entered school. In most countries, fewer than 5% of cohorts born in the late

<sup>4</sup> Though selective mortality introduces some sample selection bias, as higher income individuals amongst older cohorts would have had better chances of survival.

1990s did not enter school, with Mozambique still the laggard. South Africa's lead in primary school completion had largely disappeared among cohorts born in the 1970s due to rapid expansion of primary completion in other countries. For the most recent cohorts, Mozambique (49%) and to a lesser extent Lesotho (77%) show lower primary completion rates.


**Table 3.** Estimated proportion of birth cohorts that never entered school or that completed primary education (grade 7).

Sources: Authors' calculations based on data from censuses and household surveys.

#### **3. Issues and Policy Responses**

A number of education issues and policy responses in these seven countries are of interest from the perspective of inclusion. Because of space constraints I will discuss four, drawing examples from some of these countries:


#### *3.1. Resources and Priorities*

Resource constraints bind strongly in all seven countries, even in Botswana, which for a long time could expand education spending without fiscal stress. One prioritisation issue is the spending balance between school and tertiary education; these countries have high tertiary education unit costs and high tertiary spending levels. Within schools, the tendency to provide many electives at secondary schools

raises unit costs far above those in primary schools, shifting the spending balance towards secondary schools.

Fiscal constraints now bind so strongly in Zimbabwe that most schools have received no new textbooks since 2012. Adopting a new curriculum and a change in the language of teaching and learning in the first years of primary school from English to Shona or Ndebele, the two dominant indigenous languages, exacerbated textbook scarcity. Without appropriate textbooks geared at the new curriculum and the new language policy and with no funding to train teachers to deliver the new curriculum, learning necessarily suffers. Moreover, the strong emphasis on offering many electives raised unit costs in secondary schools to four times those for primary schools. High dropout also prevents economies of scale in some schools.

In Mozambique the big issue remains expansion of the school system. The civil war from independence in 1975 to 1992 caused death and destruction, including the destruction or closure of 58% of schools (Mozambique Ministry of Education 1996, p. 40). This affected educationally lagging regions most; more than 80% of all schools in the central region were destroyed or closed, but none in Maputo City, the capital (Van der Berg et al. 2017). Free primary education, greater financial support to schools, free textbook provision, appointment of more teachers and investment in classroom construction led to rapid expansion of enrolment, from 3.6 million in 2003 to 6.7 million in 2014 (UNICEF 2017). Clearly, supply side rather than demand side factors had been holding back this expansion. The expansion, though, had serious implications for teacher quality.

Today, Mozambique's school network is still inadequate. While 12,291 schools offer grade 1 classes, only 6,624 offer grade 7, the final primary grade. Much fewer schools offer secondary grades: 854 grade 8, and only 435 grade 12. In the Zambezia Province, this problem is most severe: 3252 schools offer grade 1, though only 85 offer grade 12. In the country as a whole, grade 8 enrolment is only 21% and grade 12 enrolment a meagre 12% of that in grade 1. This favours urban children, who make up 11% of students in grade 1, but 37% in grade 12. Hostels offer the only option for many rural students to attend secondary school (see Section 3.2 below).

Understandably, Mozambique's priority is to expand the school system's reach, but they face a lack of qualified teachers in rural areas. The junior primary pupil–teacher ratio was 75 to 1 in 2006, but the appointment of many thousands of new teachers reduced this to a still high 63 to 1 in 2011 (Mozambique Ministry of Education 2013, p. 27). It is especially difficult to entice qualified secondary school teachers to teach in rural schools.

Lesotho's secondary school network also has a limited reach. Primary NER is high, 95 in 2009, ranging between 97 amongst children in the richest quintile of households and 91 in the poorest. Secondary net enrolment is much lower at 34, both because many children are at earlier school phases than they should be due to late school entry and repetition, and because many children drop out before secondary school. There are ten times as many primary schools as secondary schools, which makes access to secondary schools difficult. High education costs to parents and distance to schools reduce access to secondary schools (see Section 3.2.1 below). The government offers some scholarships to poor secondary students, though these are inadequate in amount and in number.

For Lesotho's government, the costs of providing textbooks and teachers are already quite onerous, in part because of high teacher salaries. School feeding is provided with international assistance at primary schools, yet, despite high poverty levels, not in secondary schools. A new curriculum adds to training and textbook needs and makes it difficult for the government to provide textbooks, particularly at secondary level. It appears that the costs of curriculum change have not really been planned for.

The bigger problem, though, lies with the lack of secondary schools. Expanding the school network would be costly, with high costs of construction and teacher salaries. A recent report (World Bank 2019a) makes a case for extending the grades offered in primary schools as a way of retaining children who live far from secondary schools in school longer. Thus, instead of offering grades 1 to 7 only, some primary schools could evolve into combined schools, adding grades 8 to 10 to their offering. This would reduce the cost of expansion by utilising existing facilities better.

Botswana has more fiscal resources than other southern Africa countries, but fiscal constraints have become tighter. While educational outcomes are very unequal, resource allocation between schools is quite egalitarian. For instance, there is a close relationship between enrolment and teacher numbers in secondary schools: 93% of variation in teacher numbers can be explained by enrolment—i.e., the randomness index is only 7% (though data from TIMSS, the Trends in International Mathematics and Science Study, point to some differentials in availability of textbooks and school furniture).<sup>5</sup> The primary teacher–pupil ratio is 26 to 1 and the secondary school ratio an extremely favourable 12 to 1. Despite this, class sizes are often not small.

<sup>5</sup> This index is effectively simply 1 minus the coefficients of determination (R-squared) in a regression that fits the enrolment of a number of teachers employed per school. Thus, such a regression for Botswana had an R-squared of approximately 0.93.

Ninety percent of students in secondary school attend classes with more than 32 students in their core subjects and one-third are in classes of more than 40 students. In the 2011 TIMSS assessment of grade 9, the average core class size was 37.6, with a standard deviation of 8.4. The low pupil–teacher ratio is the result of a favourable fiscal situation and many electives being offered at upper secondary level, thus large classes are split for elective subjects. The average secondary school teacher teaches only two hours per day (World Bank/UNICEF/Ministry of Education and Skills Development 2020, p. 47).

While Botswana's pupil–teacher ratios at secondary level would indicate that fiscal resources are adequate, the full picture is not as rosy. There is a stark shortage of classrooms, 15% at primary level and even larger at the secondary level, especially for specialist classrooms (e.g., laboratories, libraries). Many classes take place under trees. This shortage is at least to some extent the result of the fact that the budget for teacher salaries and for classroom construction is split between ministries. There is currently pressure to appoint even more teachers in the face of rising numbers of trained teachers remaining unemployed. Moreover, there is even a shortage of textbooks: only two-thirds to three-quarters of the required textbooks are available (World Bank/UNICEF/Ministry of Education and Skills Development 2020, p. 81).

A report by the Ministry of Education and Training of Eswatini indicates that the average pupil–teacher ratio is only 31 to 1 in primary and as low as 13 to 1 in secondary schools. Yet, teacher allocation across schools is not particularly equitable: the pupil–teacher ratio at primary level ranges from 20 to over 60 (Ministry of Education and Training 2016, p. 14; Van der Berg et al. 2018) and Eswatini's allocation of teachers across primary schools had a fair degree of randomness in 2007, with a randomness index of around 20% (Bashir et al. 2018, p. 258).

#### *3.2. Serving Remote Communities*

Providing educational opportunities for children in remote rural communities is difficult—for instance, in large and sparsely populated countries such as Namibia and Botswana and in the mountainous areas of Lesotho. Botswana is larger than France or Thailand, but has a population of only 2<sup>1</sup> <sup>2</sup> million compared to 82 million and 66 million, respectively. This makes it difficult to bring schools close to the population and to attract good teachers to rural areas. In Kavango, a Namibian region, almost half the schools offer education only up to Grade 4, or even less (UNICEF 2015, p. 71), so children need to move to new schools as they progress to higher grades. Mozambique faces similar problems due to the limited accessibility of secondary schools.

3.2.1. School Hostels and Rented Accommodation versus Expanding the School Network

One option for serving isolated communities is to provide schools closer to the population, but that is not always possible. In some locations, the number of children may be too small; in other cases, the cost of constructing schools or providing teacher housing may be astronomical—e.g., in the desert regions of Namibia, where water provision is often unpractical or prohibitively expensive. At the secondary level, ministries often prefer larger secondary schools that make more electives viable. As transport (commuting) is costly or impractical, this necessitates school hostels, which are expensive for poor rural parents, and governments cannot always afford to subsidise them. Evidence to a Namibian Parliamentary Committee supported that "School hostels around the country are still faced with many challenges ranging from deplorable physical facilities, poor food preparation, lack of sanitation and poor hygiene" (UNICEF 2015, p. 73). In the absence of hostels, many secondary students are placed in rented accommodation near schools, but this is expensive and social conditions are sometimes appalling.

As discussed in Section 3.1, many children in Lesotho, especially those located in poor mountainous areas, live far from secondary school, and transportation options are limited or costly. The costs for an average secondary student have been estimated at between USD 650 and USD 900 for a child living in a hostel, with hostel costs constituting up to a quarter of these amounts. Costs include various fees, textbooks/stationery and uniforms (World Bank 2019a, p. 46). These costs are unaffordable to most parents in a country where 60% of the population lives in extreme poverty (see Table 1). Consequently, the net secondary enrolment rate in the richest quintile is 61%, as against only 11% in the poorest quintile (World Bank 2019a, p. 28). Currently, 11% of girls and 5% of boys are in school hostels, with a further 16% of both boys and girls enrolled in secondary school living in rented accommodation. School authorities try to vet the rented accommodation, but this does not completely prevent some serious problems accompanying such accommodation.

As discussed earlier, one of the options of dealing with the access issue at the secondary level is to extend primary schools to incorporate higher grades. This would make it possible for students to continue attending school to a higher level while still living in their own homes. Moreover, this would be cheaper than building completely new schools, and such capital spending would save the recurrent costs for maintaining school hostels. However, an issue that then would come into play is the availability of qualified teachers. This issue is addressed next.

#### 3.2.2. Teacher Assignment: Rural Incentives versus Deployment

Two alternative models of teacher assignment have vastly different consequences. Some countries, such as Botswana and Mozambique, apply a highly centralised system of allocation, often referred to as "deployment", whereas others, such as South Africa and Namibia, have a decentralised system where individuals apply for advertised positions at particular schools. In the latter case especially, incentives are often used to make remote schools more attractive for teachers to choose to go to. However, the fiscal constraints limit the size of incentives that can be offered, and in some countries incentives are regarded as taxable benefits, thus reducing their effect.

Teachers are often highly frustrated with forced deployment to remote areas. In Botswana, many rural teachers complain about what they refer to as "over-staying", a situation where they are initially deployed in rural areas and after many years still do not have the opportunity to move to more attractive locations. Such deployment is also sometimes socially very disruptive—e.g., cases where husband and wife have been deployed in different far-flung parts of this vast country.

A study of the effect of incentives in remote schools in Namibia concluded that incentives for teachers in remote schools may have helped to retain and attract qualified teachers and therefore proposed a modest rise in these incentives: "This places a small additional burden on the education budget, raising incentive costs... This may further improve teacher allocation, signal to teachers in remote schools their contribution is valued, and be a sign to parents in such areas that the education of their children is a concern for the government" (UNICEF 2014, p. 9). Another recommendation that was also accepted was to provide more teacher housing in the most remote areas.

#### *3.3. Repetition Policy, High Stakes Examinations and Dropout*

#### 3.3.1. Repetition in the Literature

The international literature is still divided over the issue of repetition as against automatic promotion. Much education research holds that grade repetition may increase the risk of negative outcomes for students. Individuals who repeat face potential stigma, which may contribute to dropout. There is no clear evidence that repetition improves learning, and it is expensive, as it increases the time it takes a student to progress through school. On the other hand, some see repetition as evidence that the school system is promoting learning. There is some evidence that automatic promotion reduces student efforts. A strong argument for repetition in developing countries is that large variation in student capacities in a class complicates

the task of teachers. Setting grade promotion thresholds ensures that all students meet some minimal standard. Yet, repetition tends to increase age ranges within grades. Repetition is much more common in developing countries than high-income countries, and in Francophone as opposed to Anglophone African countries (UNESCO Institute for Statistics 2012; Eisemon 1997).

Automatic promotion need not mean that learners are held to a lower standard, but it should be accompanied by increased quality in other inputs and remediation to support low achieving students (Van der Berg et al. 2019).

#### 3.3.2. Repetition in Southern Africa

Promotion policies in southern Africa vary widely between countries. Eswatini, Lesotho, South Africa, Namibia and Mozambique have high repetition rates, while there are stronger limitations on repetition in Botswana and Zimbabwe.

Zimbabwe formally has automatic promotion, yet their Annual Statistics indicate that more than 25,000 children repeated in primary schools (0.75% of enrolment) in 2018 and 12,600 in secondary schools (1.16% of enrolments) (Zimbabwe Ministry of Primary and Secondary Education 2019). Repetition rates have been declining. Especially in primary schools, boys are more likely to repeat than girls are. The Primary School Leaving Examination, which used to be a high-stakes examination that determined access to secondary school, no longer serves this function; all children taking the examination are promoted to grade 8 (Form 1), the beginning of secondary school, although many do not continue (only 80% of 2017 Grade 7 students continued to Form 1). The effect of the high-stakes grade 11 O-level examinations is much more severe, though: this determines access to senior secondary school—Forms 5 and 6 in Zimbabwe's 13-year school system—and also entry into some colleges. The majority of students fail this examination, and less than one-quarter proceed to senior secondary.

Botswana also practices automatic promotion, but with reservation. Officially, it is said that children only repeat when parents agree that they should. Grade repetition is concentrated in the earliest grades of primary school. In 2017, just over 14,000 primary (4.2%) and 719 secondary students (0.4%) repeated (World Bank/UNICEF/Ministry of Education and Skills Development 2020, p. 49). High failure rates in the Grade 10 Junior Certificate Examination result in a small number of students in senior secondary schools. In 2014, the senior secondary GER was 62, while the NER, the proportion of the senior secondary age group that was in fact in this school phase, was only 29 (World Bank/UNICEF/Ministry of Education and Skills Development 2020, p. 9)

Each year, about 15% of Eswatini's primary learners are not promoted to the next grade (UNICEF 2018, p. 11), despite official policy limiting repetition to 10% per grade per school (UNICEF 2016). Due to repetition and dropout, the number of students not overaged in grade 1 was just under 30,000 in 2011, while the number not overaged in grade 7 in 2017 was less than 10,000 (UNICEF 2018, p. 38). Low promotion rates (as low as 60% to 70%) are experienced in the grades before external examinations take place in grades 7, 10 and 12. Ministry officials ascribe this to accountability pressure on schools to perform well in the external examinations. This leads to weaker students being held back in the previous grade, something that also encourages high dropout rates in these grades. The effect of this high repetition rate in combination with high dropout can be seen in Table 4, which shows enrolment by grade for a number of years. By following diagonally downwards and to the left, one can observe a "pseudo-cohort", as, for instance, shown in the table by the highlights for the 2009 grade 5 cohort. These are not actual cohorts. For the highlighted diagonal, the number of students in grade 6 in 2010 was less than in grade 5 in 2009, due to dropout and repetition, but some of the 31,200 grade 6 students in 2010 were repeaters from 2009, so they were not part of the original 2009 grade 5 cohort. Following this diagonal further, it transpires that the Form 5 (grade 12) "pseudo cohort" in 2016 was less than 13,000, only 39% of the number of students in grade 5 in 2009. In a similar way, one can follow cohorts from earlier years in earlier grades, but they would not yet have reached the highest grades. The second panel offers another way of following a real cohort: the cohort of 11-year-old children in 2011<sup>6</sup> was reduced by dropout and repetition so that there were fewer 12-years-olds in 2012, 13-year-olds in 2013, and so on. Of the original age cohort, only 63% were still in school at age 18, but many of these would have repeated one or more grades. The 63% reflects the high dropout between ages 11 and 18; repetition cannot be observed in this panel.

<sup>6</sup> Age data are not available for the earlier years.


**Table 4.** Eswatini "pseudo-cohort" versus actual cohorts, 2009–2016.

Note: the highlighted figures show how one can track a pseudo-cohort across years and grades, as shown here for the 2009 grade 5 cohort. Source: Authors' compilation.

Due to late enrolment and early repetition, 29% of grade 1 students in Namibia are overaged. High repetition rates throughout primary and junior secondary grades raise the overage proportion to 72% in grade 9, but thereafter this proportion declines as dropouts exceed the high repetition rate, especially among the overaged. Thus, the overaged proportion declines, but is still as high as 51% in grade 12. As part of a major curriculum change, the grade 10 examination, which currently acts as a high barrier, is being abolished and replaced by a grade 11 examination, which is regarded as the end of secondary school for most children. Those who wish to continue further, especially those with university ambitions, have to write a new school-leaving examination at the end of the new grade 13.

#### *3.4. Cognitive Performance, Testing and Measurement*

To improve educational outcomes in line with Goal 4.1 of the SDGs requires measuring and monitoring of cognitive outcomes. Regular participation in international assessments offers a convenient way to do this, where performance of a representative sample of children in a country is measured against that of other countries. Further analysis of such data can also offer insights into shortcomings in learning, thus potentially allowing policy to address specific learning deficiencies. SACMEQ, the Southern and Eastern Africa Consortium for Monitoring Educational Quality, is the only such international assessment that all seven our countries participate in, along with eight other countries in southern and eastern Africa. However, SACMEQ only takes place at quite long intervals and the last data that have been released are for 2007. Botswana and South Africa on occasion participate in other international assessments, TIMSS (a mathematics and science assessment) and PIRLS (Progress in International Reading Literacy Study, a reading assessment).

Gustafsson (2012) used all international cognitive evaluations available in 2012 and a nonlinear programming approach to convert country average scores to a common "Pisa scale". These scores are shown for a selection of countries in Table 5, arranged from lowest to highest. Our seven southern African countries are marked with asterisks. The Pisa setpoint (average) was originally set to 500 with a standard deviation of 100. Lesotho and Namibia performed more than two standard deviations (roughly equivalent to 5 years of learning) below this Pisa setpoint. Even the two best performing countries in our group, Eswatini and Botswana, performed more than 1.3 standard deviations below the Pisa setpoint. Assuming a year's learning to be around 40% of a standard deviation, the average student in the seven southern African countries discussed in this chapter is between 3 and 6 years behind students in the average Pisa country. The extremely low scores of Botswana and South Africa on TIMSS and PIRLS confirm the large gap compared to developed countries. More recent international tests have given some cause for optimism that South Africa is making good progress, but even so it will take many decades for the countries of southern Africa to achieve an education quality similar to that in developed countries today.


**Table 5.** "Pisa scale scores" for a selection of countries based on Gustafsson (2012).

Note: The asterisks indicate countries discussed in this chapter. Source: Gustafsson 2012, Appendix 1.

In 2012, the South African Department of Basic Education introduced the Annual National Assessments (ANAs), which tested *all* children in identified grades (grades 1 to 6 and also grade 9) in both language and mathematics. Although nationally drawn up, the tests were administered by the schools themselves. This initiative offered the potential of providing information on student performance and on specific weaknesses in learning *by school*; thus, the National Development Plan saw the ANAs as a potentially important monitoring and accountability tool. However, there were reservations about the quality of the tests themselves, i.e., their calibration, an issue related to capacity of the education system. An even bigger problem was fierce teacher union opposition, especially to the accountability aspect of the tests, which led to their abolition in 2015. After long negotiations between teacher unions and government, there are now new plans for sample-based systemic evaluations to provide information about system performance in grades 3, 6 and 9, but these would

not provide the school-level results needed for accountability. This is particularly needed at the primary level, as South Africa is one of only a few countries in southern Africa without a standardised primary school examination that could provide some accountability pressure.

In President Cyril Ramaphosa's state of the union address in 2019, he committed the South African government to the goal of every child being able to read for meaning by age 10. This is directly in harmony with the SDGs and with international attention to eliminating "learning poverty", by acknowledging the reality of weak learning outcomes. Yet, thus far there is little indication that the government will be measuring and monitoring progress to this goal.

Although Botswana is one of the richest SACMEQ participants, it performs near the regional average. Poor children perform particularly poorly. The gap between the scores of the poorest quarter of students and the average is almost as large as in South Africa, a country known for its high inequality. Botswana's performance on other international assessments (TIMSS and PIRLS) is also far below the average of middle-income countries. The weak performance on Pre-PIRLS indicates that learning deficits start early. There has also been no improvement in performance on internal examinations in recent years (World Bank/UNICEF/Ministry of Education and Skills Development 2020, pp. 9–10, 12). A report for the education ministry found school monitoring and support services to be ineffective and stated that the country has made little progress in " . . . producing an assessment system that enables the government to monitor at a national level what the quality of the national education outcomes are" (Botswana, Ministry of Education and Skills Development 2014, p. 40). Botswana's three national examinations, as well as the international assessments it participates in, are not systematically used to provide information to assist teachers in their work or to direct in-service training. Moreover, there is almost no information on performance in grades that do not have standardised examinations (World Bank/UNICEF/Ministry of Education and Skills Development 2020, p. 72).

Eswatini scored higher than all the other countries considered in this chapter in SACMEQ 2007 in grade 6. High repetition rates may perversely have contributed to this by "weeding out" weaker students (UNICEF 2018, p. 12). It is interesting that the social gradient as measured by the difference in scores between the poorest and the richest quarters of children participating in the test was only 18 points in mathematics and 39 points in reading (the standard deviation across all SACMEQ countries is 100 points). In contrast, these differentials are a massive 119 and 159 in South Africa and 70 and 83, respectively, in Zimbabwe (Spaull 2012).

While Mozambique performed slightly above the SACMEQ average in 2000, in the 2007 version of the tests the average performance of Mozambican grade 6 pupils fell by more than 40 points (40% of a SACMEQ standard deviation) in both reading and mathematics. This deterioration is not surprising, given the massive enrolment growth that brought more pupils from poor backgrounds into schools and increased pressure on resources (infrastructure, textbooks and teachers). However, despite the large increase in students that performed at the bottom end of performance and caused the average decline, there was also growth in the numbers of students performing at higher levels, despite the pressures brought about by expanding enrolment (Taylor and Spaull 2015).

A service delivery indicator survey used by the World Bank for Mozambique points to severe teacher issues. It found that 45% of primary teachers were absent from school on the day the survey was administered, while another 21% were not in class teaching. Thus, only one-third of teaching time was available for teaching. Moreover, the absenteeism amongst school directors (principals), at 44%, is not significantly lower than amongst teachers. In addition, teachers were found to have poor subject knowledge (only 60% of lower-primary teachers could provide the correct answer to a simple subtraction question, 86 minus 55), and teacher pedagogy was found to be weaker than in other African countries where service delivery indicators had been measured (Molina and Martin 2015).

Absenteeism amongst children is also a grade problem. In a Portuguese comprehension test administered to a sample of grade 3 children in 2016, at least half of the students were absent in one-quarter of all schools tested (Mozambique, Ministério de Educação e Desenvolvimento Humano, Institute Nacional de Desenvolvimento da Educação (INDE 2017)). This was especially the case in the central region, one of the most disadvantaged areas. Performance in the south exceeded that in the other two regions—only 19% and 28%, respectively, of students (present or absent) in northern and central regions attained at least the top two performance levels, against 53% in the south.

#### **4. Conclusions**

The policy dilemmas and initiatives described above show a strong emphasis on issues of educational access, but generally speaking, cognitive outcomes have received little attention. Where these are discussed, there is often simply a presumption that other policies will somehow lead to better learning in classrooms: improved textbook availability, teacher training, information and communications technology (ICT), or better laboratories are regarded as the tools through which better learning will

happen, or even simply continuing with business as usual. There have been few attempts to see whether training actually leads to better teaching and learning, and whether textbooks are actually used; there is little enthusiasm for participation in international tests; available test data are hardly ever used to inform teaching practice or teacher training; internal school assessments or standardised examinations are taken as the measure of learning. The international community may have shifted their emphasis from access, as in the MDGs, to greater attention to cognitive outcomes and equity, as in the SDGs, but this is receiving scant attention in education ministries in southern Africa. "Learning poverty" is still widespread, but largely ignored compared to issue of access and performance on internal standardised examinations. Moreover, as ministries usually measure performance in terms of pass rates, limiting access to such examinations is widely practiced as a means of improving pass rates. The total number who pass has not received much attention.

While this is the case, it is unlikely that there will be much progress with learning, especially early learning. If the measure of learning simply remains how many children reach a certain grade or what proportion passes a standardised examination, too much attention is given to the wrong outcomes. Indeed, "schooling ain't learning", to paraphrase Lant Pritchett, but this does not appear to have sunk in yet in most southern Africa countries. This may be very similar in many other developing countries, especially poorer ones. There is a large schism between the insights of the international community (e.g., the UN organisations and the World Bank) and the policies and practices of many developing countries (often supported by some development partners). It is still easier to acquire support for implementing a new curriculum than to test regularly, and to use test results for accountability purposes. While 2030, the target date for the SDGs, is getting closer, it may still take some time for the fundamental shifts in the educational development goals from MDGs to SDGs to sink in as an integral part of education thinking, policy and practice.

In terms of policy, one would therefore want to see more emphasis on measuring education quality and introducing measures to improve this. Central to such improvements must be teachers—what they know, how they teach, how much effort they put into their work. Measures to improve teacher quality, such as both preservice and in-service training, should be evaluated regularly and rigorously. Participation in international educational evaluations should be expanded—not one of these countries participates in the Pisa for Development initiative—and results from internal evaluations need to be central issues in public debates. The countries discussed here have now fully engaged with the MDG objective of improving

educational access, and they continue to implement this. Similar enthusiasm is now needed for the SDG objective of improving quality. Only the combination of broadening access and improving quality can ensure that "learning poverty" is reduced—and this remains an essential component in all efforts to reduce poverty.

**Acknowledgments:** I wish to thank the government officials in the countries that I have been privileged to work in during the last few years for much of the insights contained in this chapter. Although I have greatly benefited from close interaction with government ministries, UNICEF and the World Bank, the views expressed here are my own.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


United Nations Population Division. 2019. *Population Prospects 2019 Revision*. New York: UNPD.


© 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Early Childhood Development: Current Status and Gaps**

**Günther Fink**

#### **1. Introduction**

The period between conception and a child's entry into the formal schooling system is increasingly recognized as critical for early and later life outcomes (Walker et al. 2011; Britto et al. 2016). The early childhood period is unique in terms of the pace and diversity of new skills learnt and acquired, and lays the foundation for later physical and mental wellbeing (Grantham-McGregor et al. 2007). Through an interactive and highly heterogeneous process, children acquire a remarkable share of their perceptual, motor, cognitive, language, socio-emotional, and self-regulation skills in the first years of life (Grantham-McGregor et al. 2007; Black et al. 2017; Harman et al. 2018). These skills are not only important for subsequent educational attainment, but also highly predictive of adult outcomes more broadly, including measures of general wellbeing, crime, marital status and income (Oreopoulos 2007; Kamal and Bener 2009; Heckman et al. 2013; Chan et al. 2017).

Recent work comparing long term costs and benefits of various policy programs suggests that the early childhood period may constitute the most cost-effective time interval to ensure that all children achieve their developmental potential (Heckman and Tremblay 2019), with estimated returns of up to \$17.7 for each dollar invested (Chan et al. 2017). Despite this, global gaps in early childhood development remain large. More than 250 million children under age five are currently estimated to not reach their developmental potential (Li et al. 2016; Black et al. 2017); 22% of children under age 5 in LMICs experience physical growth faltering (United Nations 2018), and 37% of 3- and 4-year-olds growing up in low- and middle-income countries display deficits in their cognitive or socioemotional development (McCoy et al. 2016). All primary risk factors for early development are directly linked to poverty: exposure to infectious disease, lack of access to early life health services, lack of access to early learning materials and personal support, as well as chronic malnutrition, have been consistently associated with poor developmental outcomes in the literature (Engle et al. 2011; Walker et al. 2011; Black et al. 2017).

Governments around the globe are increasingly recognizing the benefits of investing in early childhood development (Pelletier and Neuman 2014), and early childhood development (ECD) is now considered a foundation for sustainable development (Daelmans et al. 2017).

In this article, we first summarize the evidence on what is known regarding the current gaps in early childhood development, and then discuss the most promising intervention strategies. We conclude with some general reflections.

#### **2. Materials and Methods**

The materials presented in this article mostly represent previously published work. The primary data sources for each of the tables and figures are provided below.

#### **3. Results**

#### *3.1. Global Gaps in Early Childhood Development*

Data on children's early development at a national or global level remain remarkably scarce. In 2007, the Lancet launched its first "*Early Childhood Development Series*", which synthesized the then available evidence on child development (Grantham-McGregor et al. 2007). While it was rather obvious that large developmental gaps existed across socioeconomic groups within and across countries, comparable data on the actual skill sets or developmental levels of children under the age of three were not available. In order to create a first estimate of global gaps in early childhood development (ECD), the author team thus decided to focus on quantifying children's exposure to risk. In the ECD literature, the two most salient risk factors for healthy development are chronic malnutrition and exposure to poverty. Combining all available data, the author team estimated that 219 out of 559 million children under age five globally were at risk of poor development (Grantham-McGregor et al. 2007). With slightly better data, these numbers were revised to 249 million, or 43% of all children under the age of five globally at risk of not reaching their potential in the subsequent Lancet ECD series (Black et al. 2017).

Even though these numbers have been important in shaping global policy in this area, it is not clear how well stunting- and poverty-based estimates approximate the actual number of children displaying delays in their development. Among the primary challenges in the area of global ECD research is that, to date, "normal development" has not been defined at a global level. While the global nutrition community was able to establish a first set of reference growth tables in the early 2000s (WHO Multicentre Growth Reference Study Group 2006), the same still has to happen in the area of ECD. The lack of a global standard or reference norms makes an objective assessment of gaps challenging. There are two main reasons why global references for early development have been difficult to establish: First, there is a substantial body of literature suggesting that children's trajectories are highly context-specific, and thus should not be directly compared. According to this logic, comparing children across different settings does not make sense conceptually—if reference tables are needed (which is still questioned by many researchers), these should be country specific, and thus only compare children to other children from the same country. "Developmental delay" would then be defined as a child developing significantly slower than others in their own environment.

Similar arguments were made up to the early 2000s in the field of nutrition. Many researchers felt that children's early growth was genetically driven, and should thus not be compared across settings with different genetic pools. This notion was, however, soundly rejected in the original 2006 WHO study (WHO Multicentre Growth Reference Study Group 2006) as well as the larger subsequent INTERGROWTH 20th century study (Villar et al. 2013). Both studies followed groups of newborns in a very diverse set of high- and low-income countries. Rather than assessing representative samples of children, these studies focused on children that lived in comparable conditions that were assessed as suitable for healthy growth. While this definition of "healthy homes" varied slightly across countries, the idea was to compare children that did not have any obvious genetic defects and had access to a safe home, clean water and sanitation. Given that adult height varies substantially across those sites—British adults are about 5cm taller than Chinese adults and about 10cm taller than Indian adults—large differences in early childhood growth were expected. However, no such differences were found, suggesting that differences in adult height emerge only in later childhood or adolescence among children growing up in safe and supportive environments. While these findings were not necessarily anticipated, they were important for the field, since they meant that comparisons of population-level outcomes across countries or regions were valid, and that a standardized set of growth references could be applied to all children globally.

As shown in Karra et al. (2016), reference populations similar to those sampled in these two highly influential studies can be created from standard survey data by restricting the analysis to households that can offer similarly safe environments to children. In the pooled Demographic and Health Survey (DHS) used in their paper, the average height for age z-score in the general population is −1.41 (SD 1.66), which implies that on average children in low- and middle-income countries (at least those countries sampled in the DHS) had heights that were almost 1.5 standard deviations below the international reference median. The overall distribution of height for age z-scores is illustrated in Figure 1: both the reference population and the actual

distribution roughly follow a normal distribution—the entire distribution of height in the actual DHS sample is, however, shifted by 1.45 z-scores to the left, and wider, with an estimated standard deviation of 1.66.

**Figure 1.** Height-for-age z-score distribution among children in pooled demographic and health survey files used in Karra et al. (2016). Red dotted line shows standard normal HAZ distribution in WHO reference norm sample. Pink line shows empirical distribution of HAZ in the pooled sample.

This distribution of height in the pooled sample masks a remarkable amount of heterogeneity, however. When the analysis was restricted to children who live in households that match those of well-off Western families, these gaps disappeared. As Figure 2 shows, the distribution of HAZ in this selected subsample looks rather similar to the reference population, with slightly less density at the center of the distribution and slightly wider tails. These wider tails likely capture, at least partially, errors in height measurement; they may, however, also represent slightly larger (genetic) variance in the more diverse DHS sample compared to the six sites in the original WHO study.

**Figure 2.** Height-for-age z-score distribution among children with ideal home environments in demographic and health surveys. Red dotted line shows standard normal HAZ distribution for WHO reference norms. Pink line shows empirical distribution in ideal home environments. Children in ideal home environments were defined as children with (1) access to safe water and sanitation; (2) living in households with finished floors, a television, and a car; (3) raised by highly educated mothers; (4) single births; and (5) delivered in hospitals. More details on the sample are available in Karra et al. (2016).

While similarly comprehensive and normalized data are not available yet for any other domain of early childhood development, several new tools to assess child development in low- and middle-income countries have emerged and also been used in a growing number of contexts in recent years. Most tools focus on four main domains of development: (fine and gross) motor skills, cognitive skills, language skills and socioemotional skills. Using a sample of over 10,000 healthy children from four middle-income countries, Ertem et al. show that on average the age at which children achieve developmental milestones under the age of three varies very little across sites (Ertem et al. 2018). The only domains where substantial differences were found were "self-help" or "life skills", such as brushing teeth, using toilets or dressing, which tend to be strongly influenced by local customs and habits, and likely do not reflect true differences in children's inherent ability.

Using a more diverse sample of children under age 3 living in selected high-, middle-, and low-income countries, substantial developmental differences were, however, found across sites (Fink et al. 2019). In the six low- and middle-income

countries analyzed, average cognition z-scores for children under the age of 3 was −0.52, with an SD of 1.28 (Figure 3).

**Figure 3.** Based on a sample of 3447 children under the age of three used in Fink et al. (2019). Children in the LMIC sample are from Cambodia (N = 410), Ghana (N = 1512), Guatemala (N = 197), Lebanon (N = 376), Pakistan (N = 238) and Philippines (N = 714).

When the sample was restricted to children growing up with a mother who had completed high school or higher education, this gap once again disappeared (Figure 4, mean z-score −0.06, SD 1.27).

**Figure 4.** Based on 621 children with highly educated mothers from Cambodia (N = 20), Chile (N = 170); Ghana (N = 42), Guatemala (N = 19), Lebanon (N = 59), Pakistan (N = 11) and Philippines (N = 300).

Both the nutrition and the cognition analyses suggest that current developmental gaps in low-income settings are substantial. These gaps are neither driven by genetic nor geographical or climate-related reasons, but rather seem to be the result of (poverty-related) factors in the home environment. In general, for child health and development, maternal education is the single most important predictor of outcomes; once we compare mothers with completed secondary education or higher education, differences in health and developmental outcomes across countries become small; once differences in living conditions are also adjusted for, differences disappear completely.

Table 1 shows the global distribution of educational attainment according to the data collected by Barro and Lee (2013). While in developed countries (advanced economies), the large majority of women had at least secondary education in 2010, substantial gaps remained in several other regions, with particularly large gaps in South Asia and Sub-Saharan Africa, where secondary or higher education remained limited to a minority. Major efforts have been made in the past two decades to increase primary and secondary schooling enrollment; these efforts will undoubtedly improve children's home environment and support in the decades to come; for the time being, the lack of maternal education remains among the most critical concerns.


**Table 1.** Educational attainment in the female population age 15 and older in 2010. Source: Author's compilation based on data from http://www.barrolee.com/ (accessed on 9 September 2021).

Table 2 shows the percentage of 3- and 4-year-old children that benefit from interactions with their caregivers or other adults in their household. In most middle-income countries, adult engagement is substantial, with a large majority of adults engaging in at least four out of the six activities captured by the Multiple Indicator Cluster Surveys (MICS). In many low-income settings, and particularly in many African countries, that is, however, not true, with quite a few countries where less than one third of caregivers indicate that they engage with these child activities on a regular basis.

#### *3.2. Interventions to Improve Early Childhood Development*

Early life interventions are increasingly recognized as key for creating the environments that children need for a healthy development (Nores and Barnett 2010; Hoddinott et al. 2013; World Bank Group 2015; Richter et al. 2017), and for ensuring children's long-term economic and general well-being (Heckman 2006). In all likelihood, the most successful model for improving child health and well-being in low-income settings is home visiting programs to support parents; such programs have been successfully implemented in Bangladesh (Hamadani et al. 2006), Colombia (Attanasio et al. 2014), Peru (Hartinger et al. 2017), Jamaica (Grantham-McGregor et al. 1991; Walker et al. 2005), Pakistan (Yousafzai et al. 2014) and South Africa (Cooper et al. 2009). The main logic of home visiting programs is that trained child development or community agents meet with mothers or parents on a weekly, biweekly or monthly basis to discuss the child's overall well-being and needs, as well as to provide practical guidance for parents on how to provide a healthy and developmentally stimulating environment for the child. Home visiting programs are typically based on a tightly structured curriculum, which contains key topics of child health and development to be covered during each home visit. Through the regular interaction with caregivers, home visiting programs are designed to improve early childhood health and development (Engle et al. 2011; Yousafzai and Aboud 2014), but can also improve maternal well-being (Carta et al. 2013; Tandon et al. 2013).

Despite their remarkably large and consistently positive impact on child well-being (Yousafzai and Aboud 2014; Aboud and Yousafzai 2015), home visiting programs have not been adopted at scale by most countries, which is mostly due to the large cost and logistical efforts associated with these programs. To date, the only countries that have attempted the rollout of home visits are Brazil and Colombia; evaluations of both programs are still ongoing; given the scale and cost of these programs, and it not clear yet whether they will be sustainable in the long term.

**Table 2.** Percentage of children with appropriate home stimulation in low- and middle-income countries. 95% confidence intervals in parenthesis adjusted for two-stage cluster-sampling used in Multiple Indicator Cluster Surveys (MICS). Home stimulation was defined as adults engaging in the following activities with the child in the past 3 days: reading books or looking at pictures; telling stories; singing songs; taking the child outside; playing with the child; and naming, counting, or drawing with the child. Following previous work, using data from McCoy et al. (2018), we totaled the number of activities that adults engaged in with the child, and defined adequate home stimulation as exposure to at least four out of six activities.


From a conceptual perspective, home visiting programs were designed to support children in settings where they spend most of their time with their caregivers and families. In practice, the period in which this is true will likely shorten over the coming decade. In many middle-income countries, a large majority of mothers are engaged in the informal or formal labor market now, and often return to work within a relatively short period of time (typically within less than six months). After this period, children either have to be supported by other members of the family or by formal daycare centers. These centers are emerging rapidly around the globe now and should ideally provide suitable environments to children from all socioeconomic backgrounds. From a scientific perspective, the long-term impact of daycare and preschool programs remains somewhat unclear. Despite some very positive initial findings (Garces et al. 2002), the large US Head Start program has increasingly become criticized over time, with benefits rapidly fading over time as children enter primary school (Lee and Loeb 1995; U.S. Department of Health and Human Services 2010). From a societal perspective, the main benefits of government-supported early childhood programs go beyond long-term educational benefits, since they are essential for parental labor market participation; in the long term, early center enrollment may simply become the norm. Today, one or two years of kindergarten are becoming fairly standard in many settings; it will be interesting to see whether this will be further extended to even younger children over time.

#### **4. Discussion**

In this article, we reviewed the literature and data available on early childhood development globally today. The natural, medical and social sciences have created a rather substantial body of literature over the past twenty years to document and confirm the critical importance of the early childhood period for later life outcomes. The first years of life are not only the period when children learn to walk and talk but also the period when basic neurological, cognitive and social skills are developed that shape individual lifetime trajectories. As we have shown in this chapter, there is also agreement that the development of children in low- and middle-income countries is on average delayed compared to children in high income settings. The magnitude of these delays is well documented for physical growth, where more than one in five children growing up in LMICs continue to suffer from growth faltering. For other domains of child development, including language, cognition, socio-emotional and executive functioning skills, data availability is still very limited. The few studies that have attempted developmental comparisons across sites have often found sizeable gaps between high-, middle- and low-income countries. The

good news is that these gaps are generally not found when children from safe home environments are compared: several studies show that children growing up in similarly endowed households around the globe—children living in decent houses with access to decent water, sanitation infrastructure and benefitting from engaged parents—follow roughly the same trajectories everywhere. This is of course not to say that all children develop at the same pace; it is quite obvious that some children learn to talk or walk earlier than others. The main finding of the comparative literature is simply that the distribution of developmental outcomes is very similar across countries once analysis is restricted to children growing up in similar environments. While this is good news overall, large gaps do clearly remain today, with many children growing up in highly resource-constrained environments, without access to clean water or sanitation, safe play spaces or adults that can support them in their development. Global efforts to reduce poverty and to increase education will undoubtedly have large benefits in terms of the early environment of children growing up in LMICs. Additional government support for high quality early education as well as parental support programs may also help to bring us closer to a world where all children can fully reach their developmental potential, and become adults not facing extreme poverty in the future. In countries such as Brazil, home visiting programs supporting vulnerable families in the first two years of children's lives are now standard, and can hopefully inspire other countries. Government involvement in the pre-school and kindergarten period (ages 3–5) has also increased substantially over the past decade, with a growing number of countries offering free access to early learning opportunities. Even though the quality of some of these programs is certainly not ideal yet, overall policy efforts definitely appear be following in the desired direction at a global scale.

#### **5. Conclusions**

Large gaps in children's physical, cognitive and socioemotional development are common in low- and middle-income countries today due to poverty, lack of maternal education and lack of early learning opportunities. Major governmental efforts will be required to close these gaps and to offer all children globally a chance for a healthy and productive life.

**Author Contributions:** This article was designed and written by the first author alone.

**Funding:** This research received no external funding.

**Acknowledgments:** The results presented here build on several recent papers with coauthors cited in this report. This article would not have been possible without their contributions.

**Conflicts of Interest:** The author declares no conflict of interest. No funding was received in support of this project.

#### **References**


© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
