SPECIAL PUBLICATION LABOUR AND SOCIAL JUSTICE A MAJORITY WORKING IN THE SHADOWS A six-country opinion survey on informal labour in sub-Saharan Africa Rudolf Traub-Merz, Manfred Öhm, Julia Leininger, Florence Bonnet and Henrik Maihack(eds.) September 2022 Most workers in sub-Saharan Africa are informally employed, which often leaves them without access to health care and social protection. The surveys show that the demand for access to health services is the most important expectation of the informally employed. Interest in health insurance schemes is strong and there is a clear willingness to pay. There is, however, little trust in government to provide such access. The Covid-19 pandemic has further increased this demand. Because of the insufficient access to social protection and health services among the informally employed, they have been particularly hard hit by the pandemic. The surveys reveal that pandemicrelated income poverty is a key concern of informal labour.

A MAJORITY WORKING IN THE SHADOWS A six-country opinion survey on informal labour in sub-Saharan Africa • The Friedrich-Ebert-Stiftung(FES) in cooperation with the International Labour Organization(ILO) and the German Institute of Development and Sustainability(IDOS) launched a research project on»Informal Employment, Social Security and Political Trust in sub-Saharan Africa« which includes an opinion survey of views on access to health services, on political trust and reasons for joining groups, including interest in trade union membership. The opinion surveys were conducted as country-wide representative surveys with a uniform research protocol to allow cross-border comparisons. The report presents findings from surveys carried out in Kenya(2018), Benin(2018), Senegal(2019), Zambia(2019), Côte d’Ivoire(2020) and Ethiopia(2020). • The survey results suggest that the establishment of comprehensive social protection floors, including universal health care(UHC), could be crucial in achieving the UN»Sustainable Development Goals«(SDG). The survey results suggest that one key challenge for policymakers in Africa is to ensure that access to social protection delinks social security from employment. Moreover, the results also suggest that such access is desired by the majority of informally employed, who in many countries in sub-Saharan Africa account for 80 to 90 per cent of all employment. Access to health care for them would mean the opportunity to join insurance schemes or have free or subsidised health services. • Among the most important survey findings is the importance the informally employed attach to access to decent health services. They also show their interest in joining health insurance schemes. What also comes to light is the level of trust the informally employed have in governments to improve social service delivery and the extent to which political regimes are deemed to be legitimate. Left without any public assistance the survey also finds evi­ dence of an increase in income ­pover­ty, especially among informal workers due to pandemic-related measures. It also explores the willingness of the informally employed to join groups for interest representation, asks their reasons for doing so, and establishes an understanding of the extent to which trade unions are accepted by informal labour to organize and represent their interests. Further information on the topic can be found here: www.fes.de/en/africa-department

LABOUR AND SOCIAL JUSTICE A MAJORITY WORKING IN THE SHADOWS A six-country opinion survey on informal labour in sub-Saharan Africa

Contents FOREWORD 2 Cynthia Samuel-Olonjuwon 1 INTRODUCTION 3 Manfred Öhm, Julia Leininger and Henrik Maihack 2 USE OF MEDICAL CARE: HOW PEOPLE IN INFORMAL EMPLOYMENT RATE ACCESS TO HEALTH SERVICES 8 Rudolf Traub-Merz and Manfred Öhm 3 FINANCIAL RISK PROTECTION – HOW DO THE ­INFORMALLY EMPLOYED PAY FOR MEDICAL TREATMENT? 19 Rudolf Traub-Merz 4 HEALTH INSURANCE COVERAGE AND NON-COVERAGE 25 Florence Bonnet 5 ECONOMIC EFFECTS OF THE COVID-19 PANDEMIC ON THE INFORMAL ECONOMY: WHO HAS FALLEN INTO POVERTY? 41 Christoph Strupat 6 TRUST IN INSTITUTIONS AND PERCEPTIONS OF THE TAX SYSTEM AMONG INFORMALLY ­EMPLOYED PEOPLE 49 Armin von Schiller 7 IS THE INFORMAL ECONOMY IN SUB-SAHARAN AFRICA ORGANIZED? AND IF SO, HOW? 58 Jürgen Schwettmann and Rudolf Traub-Merz 8 ORGANIZING IN THE INFORMAL ECONOMY – THE INTEREST OF INFORMAL LABOUR IN TRADE UNIONS 70 Rudolf Traub-Merz 9 TECHNICAL NOTES ON THE SURVEY 79 Michael Frosch, Richard Houessou and Abel Oyuke List of Figures 81 List of Tables 82 1

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows FOREWORD The transition from the informal to the formal economy is of strategic significance for hundreds of millions of workers and economic units around the world, which are working and producing in precarious and vulnerable conditions and are often facing poverty, insecurity and acute decent-work deficits. Policy makers, academics, workers’ and employers’ organizations now acknowledge that the high incidence of informality in all its aspects is a major challenge for sustainable development. From this perspective, the development of knowledge on the nature and characteristics of informal employment is essential to support national dialogue processes and action. Informality exists in all countries regardless of the level of socio-economic development, although it is more prevalent in developing countries. In most African countries, more than 80 per cent and up to 95 per cent of the working population make a living labouring in the informal economy. Informality reaches beyond the labour market and permeates many areas of an individual’s life. It shapes how people live and interact with the state and how they access public services. It governs people’s approach to solidarity when they do not benefit from social protection schemes that are mostly confined to the formal economy. In Africa, the Covid-19 pandemic affected societies where a majority of the population could not receive social support during lockdown periods and reduced economic activities. It worsened the conditions of many living in informality. When staff of the Friedrich-Ebert-Stiftung(FES) visited me in 2017 and proposed a joint research project on the nexus between informality and social protection, I immediately agreed as this is at the heart of decent-work challenges in the region. We quickly set up a team consisting of experts from FES, the German Institute of Development and Sustainability(IDOS) and the International Labour Organization(ILO). The team was joined by colleagues from AfroBarometer to support the design and elaboration of a questionnaire to assess the need for, and coverage, or lack of it, of social health protection among informal economy workers. The survey questionnaire also tackled existing coping strategies, perceptions, trust in institutions, and forms of organizations in the informal economy. Faster than what is usually observed in such inter-institutional arrangements, the questionnaire went for testing and the first survey was conducted in October 2018 in Kenya. Particular attention was paid to capture authentic views and to include those who often go uncovered by conducting face-to-face interviews in vernacular languages. By the end of 2020, five more countries followed suit(Benin, Ethiopia, Côte d’Ivoire, Senegal and Zambia). The surveys conducted combined a standard survey on the informal economy with an opinion poll, i.e., hard facts plus views, perceptions and opinions on social security and the representation of interests in the informal economy. The same research protocol was used in all surveys, hence it was possible to make cross-country comparisons of findings, and to show country specificities, as well as structural features across the continent. When the pandemic reached the countries pre-selected for surveys, some delays in conducting interviews were unavoidable. Yet, whenever circumstances allowed, surveys were continued, using all necessary precautionary measures. Understandably, some parts of the questionnaire were amended to keep the technical protocol in place while allowing some additional aspects linked to the pandemic to be added. The report presents the outcome of close collaboration between the FES as the lead agency, the IDOS and the ILO. It benefitted greatly from AfroBarometer, following their research protocol in designing the survey sample and the selection of households. Its findings are of immense value to both researchers and policy makers, and it is hoped that it will be widely consulted. I sincerely congratulate the team which worked together so meticulously and managed to overcome many impediments inherent in such a project. I also express my gratitude to many others who, in one way or another, helped to make this unique project a success. I look forward to continued collaboration among our organizations. Cynthia Samuel-Olonjuwon ILO Assistant Director General and Regional Director for Africa 2

Introduction 1 INTRODUCTION Manfred Öhm, Julia Leininger and Henrik Maihack Informal labour plays a vital role in African economies and societies. In most African countries, between 80 and 95 per cent of the working population make a living working in the informal economy. But informality goes beyond the world of work and permeates other areas of people’s lives. Informality thus shapes how people live and how they interact with the state. Typically, people working and living in informality have less access to public services than people earning their income in the formal economy. They seldom benefit from social protection and are often not protected by labour and social legislation. Furthermore, the Covid-19 pandemic has worsened the conditions of many Africans living in informality. Despite its vital role for African economies and societies, key aspects of the informal economy remain very much a black box. While statistics exist on the number of workers, who they are and where they work in more than 140 countries (ILO 2018), for several African countries little is known about informal workers’ expectations concerning state policies and how they organize to pursue their interests. The lack of empirical knowledge on the informal economy also blurs our view of public good provision, state–society relations in general and social protection in particular, as well as its relevance for people working in the informal economy. Although many African states’ capacities to provide public goods such as social protection are low when compared to other world regions(Clement 2020), governments still offer selective social measures to a minority of the population (Bhorat et al. 2019). Furthermore, for those who can afford it, private social protection schemes are an option. Numerous databases provide data on investment in social protection policies, but we know very little about the distribution or adequacy of these resources, not to mention the expectations, conditions and needs of people working in the informal economy in sub-Saharan Africa. The current unequal access to social protection is a question of distributional justice and has implications for social cohesion, that is, trustful and cooperative relations between citizens and between the state and society. While it is likely that the effects of the Covid-19 pandemic may erode social cohesion and increase the need for more reliable social protection schemes, we have very limited knowledge about the de facto needs and opinions of individuals – both in exceptional situations like a global pandemic and in day-to-day life. 1.1 RATIONALE FOR THE PROJECT The collaborative research project 1 »Informal Employment, Social Security and Political Trust in sub-Saharan Africa« attempts to open this black box by providing novel and substantial empirical insights into informality in African societies. It was initiated in 2018, before the outbreak of Covid-19 and offered an opportunity – from March 2019 onward – to include aspects of social protection during a pandemic in surveys. Overall, our surveys sheds new light on the broader political and societal context of informal labour, as well as specific questions related to the accessibility and existing social security programmes in the informal economy in sub-Saharan Africa. It offers previously unavailable comparative and detailed data on people’s expectations of and trust in states’ ability to provide public goods, as well as on how people organize to pursue common goals. In so doing, the project focuses on health care and social security, as well as the organizational and governance context of the informal economy in six East, West and Southern African countries between 2018 and 2021. The surveys cover Benin, Ethiopia, Côte d’Ivoire, Kenya, Senegal and Zambia. These six countries hosted 21 per cent of the African population in 2019(WB 2021) and 26 per cent of informal employment in sub-Saharan Africa(ILO 2022). The research project is motivated by the strong relevance of informal employment and its broader political and societal implications in Africa. The surveys were conducted against a background of social and economic transformation of African societies, which points to the increased importance of social protection. Rapid population growth and urbanization, combined with a lack of industrialization result in a lack of sufficient well-paid employment. The lack of access to public services underlines the urgency of improving social protection and extending coverage to the informally employed. 1 Three collaborating partners – IDOS- the German Institute of Development and Sustainability(»German Development Institute[DIE]« before renaming in June 2022), the International Labour Organization and the ­Friedrich-Ebert-Foundation – have brought together their expertise and interest in the topic. The Institute of Development Studies, the University of Nairobi, and Innovative Research in Economics and Governance(IREG), Cotonou, conducted the surveys in the respective countries. The Friedrich-Ebert-Stiftung, as the main sponsor of the project, has received funding within the SEWOH – Special Initiative (»A World without hunger«) of the German Federal Ministry of Economic Cooperation and Development. 3

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows The Covid-19 pandemic has led to an increased need for and attention paid to the provision of health care and social protection in African countries. The necessity of an adequate policy response to the pandemic not only provides a political opportunity to strengthen social protection, but also requires a better understanding of the situation based on new empirical insights. Second, building on the empirical evidence from six surveys, we offer entry points for future policymaking to improve the provision of social protection for the large majority of African labour force. As the empirical survey comprises data collected just before the outbreak and after the first wave of the pandemic, the findings in this report have a very timely policy relevance. The survey project was undertaken on the understanding that social protection and other state services are public goods that citizens are entitled to as a matter of distributional justice. The state’s important role as a provider of public goods and regulatory actor has become more evident not only through the negative effects of a strong focus on private sector development in economic policies since the 1980s but most recently also because of the Covid-19 pandemic. The project was conducted on the assumption that a functional social contract shaping state–society relations is indispensable to the provision of public policies and income redistribution. The role of the state in(re)distributing income is particularly important when countering external shocks such as the pandemic. Until recently, the development discourse on Africa focused strongly on poverty alleviation and on job creation, often with a focus on the private sector. Both subjects are highly relevant. But more comprehensive and integrated approaches are needed to achieve structural transformation in African societies and economies building on Agenda 2030 and its 17 Sustainable Development Goals(SDG), in particular SDG targets 1.3(social protection) and SDG 3.8, which calls for achieving universal health coverage(UHC). Access to health for all is also a key element of the implementation of Social Protection Floors, called for by the International Labour Organization(Recommendation 202, 2012) and is central to the African Union’s African Health Strategy 2016–2030. Overall, there is a broad global consensus that more comprehensive strategies should be developed and implemented to include the informal economy in public health care schemes, which tend to be confined to the formal economy. Third, by investigating perceptions, expectations and facts about unequal access to and provision of social protection, findings about distributional justice in a given state or region can be established. The aim of the research project therefore goes beyond the idea of policy relevance with regard to social protection. The findings suggest that improved investment in social protection will have a positive impact on distributional justice in African societies. 1.3 DEFINITIONS Informal employment encompasses as status groups employers, own-account workers, employees and contributing workers. 2 This research is based on the assumption that access to social protection is a right spelled out in the International Covenant on Economic, Social and Cultural Rights(ICESCR) and an entitlement of people in the informal economy. When using the term social protection, we refer to the ILO’s Social Protection Floors Recommendation, 2012(No. 202), which provides a broad definition of social protection, as does the ILO World Social Protection Report:»Social protection, or social security, is a human right and is defined as the set of policies and programmes designed to reduce and prevent poverty and vulnerability throughout the life cycle.« 3 Given the research interest of the project, we cover access to formal and informal social protection arrangements and the relevance of traditional support systems for people living and working informally, the latter two referring to self-organized systems. Our focus enables us to capture different types of social protection, including people living in informality and making their living in the informal economy. 1.2 OBJECTIVES OF THE SURVEY PROJECT The aim of this six-country survey project is threefold. First, the findings of this research should substantially improve empirical knowledge of social protection in informality and its broader political and societal context. Perceiving informal employment as a living condition, which characterizes the economic situation of individuals, as well as broader state– society relations, provides unique insights into, for example, access to public goods and their provision, accessibility of social protection or the strength of social cohesion. New empirical evidence can inform research in different disciplines, in particular health studies, social protection, state-building and social cohesion. Evidence from the surveys may also shed light on the effects of the external shock of the Covid-19 pandemic on living conditions and the provision of public goods in informal settings. 2 The term»informal employment«, which is used in ILO labour surveys, includes people employed in a formally registered enterprise, but denied access to social security. The terminology thus links the formal economy with informal employment. 3 This understanding of social protection is more comprehensive than the concept of social security. The latter sometimes refers to more specific social services, sometimes it is more or less synonymous with social protection. See, for example, the definition in ILO(2017: 1) World Social Protection Report:»Social protection, or social security, is a human right and is defined as the set of policies and programmes designed to reduce and prevent poverty and vulnerability throughout the life cycle. Social protection includes benefits for children and families, maternity, unemployment, employment injury, sickness, old age, disability, survivors, as well as health protection. Social protection systems address all these policy areas by a mix of contributory schemes (social insurance) and non-contributory tax-financed benefits, including social assistance.« See also the new ILO World Social ­Protection Report 2020–22.»Social protection at the crossroads: Making a ­decisive turn for a better future.« 4

Introduction Social protection is a public good. Public goods should be generally accessible to all(national) citizens, independent of their residence or occupation(Wimmer 2018). 4 Such public goods should be characterized by so-called»non-rivalry« in the sense that one person’s consumption should not change the amount available for others to consume. They also should be characterized by non-excludability, meaning that nobody can be excluded from enjoying a public good. Ensuring access to public goods is the responsibility of governments at different levels. What constitutes a public good and how such goods can be provided is thus the outcome of – often contentious – political processes. Whereas certain public goods are mentioned and referenced in international declarations, covenants and agendas(for example, Agenda 2030), access to public goods depends on concrete policies, available resources and political economy in a particular context. For example, whereas in some countries the provision of comprehensive health care is the responsibility of the state, in others private health care providers play a more prominent role. 1.4 METHODOLOGICAL APPROACH The approach chosen in this research project on social protection in the informal economy follows the idea that people’s attitudes and perceptions matter for policymaking. Against this background, the cooperating partners the Friedrich-Ebert-Stiftung(FES), the German Institute of Development and Sustainability(IDOS) and the International Labour Organization(ILO), with the support of research institutes affiliated to the Afrobarometer network, agreed to design an empirical survey to investigate people’s perceptions, similar to an opinion poll. We selected the various countries to do justice to the diversity of situations in the informal economy in different regional and national contexts. They were selected with different sizes of population in mind. We chose a mixture of countries, ranging from small(population of 15–20 million: Benin, Senegal. Zambia), to medium-sized(population of 21–70 million: Kenya, Côte d’Ivoire at the lower end), and 4 Who is considered a national citizen is itself subject to and an outcome of political struggles(Kobo 2010). large(population of 71–120 million: Ethiopia). This selection also reflects different levels of average per capita income, varying between least-developed countries(LDC) and middle-income countries(MIC). All six countries score low on the human development index. In addition, we took care to include both, Anglophone and Francophone countries. A detailed description of the survey methodology is provided in Chapter 9, which describes the questionnaire, sampling, data collection and data processing. As the individual chapters of this report address different types of research questions, the authors apply different methods, which fit the respective questions. 1.5 KEY FINDINGS AND ENTRY POINTS FOR POLICYMAKING A key finding emerging from the surveys is that in all six countries there is extremely high demand for social protection, in particular health care, but also great opportunities to establish health insurance schemes. At the same time, people’s trust in their governments and other actors to provide social security is low. Moreover, people in informal employment seek to self-organize to obtain some of the social protection which they do not receive from the state. The findings from the survey analysis suggest a variety of entry points for states, international donors and other actors to advance the provision of social protection. And indeed, the expectations and concerns expressed in the surveys by people in informal employment can reorient public social protection policies. The survey findings indicate that social protection should not be linked to employment contracts alone but be organized outside labour relationships as well. When employment does not provide social protection through contributory social provisions, social inclusion must come without a labour contract. Improved access to health care is the state service most in demand among people working in the informal economy. Several chapters in this report focus on health care – in particular, demand for it – and people’s willingness to join and pay for a health insurance scheme. The timing of the surveys, which began in 2018 and continued up to 2021, enabled us to include data on the consequences of the Covid-19 pandemic. As we found that Table 1.1 Traits of selected countries Country Small(15–20) Population size* medium-sizes (21–70) large(71–120) Benin X Ethiopia X Côte d’Ivoire X Kenya X Senegal X Zambia X Note:* Millions. Source: Authors’ compilation based on World Bank Indicators 2021. Income level LDC+ low human development MIC+ low human development X X X X X X Colonial influence Anglophone Francophone X X X X X 5

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows the urban poor are the group most affected by Covid-19, it is likely that the demand for publicly available health care among those working in the informal economy will increase even further. Based on the findings published in this report, different entry points may be suggested for establishing and improving social protection in the informal economy in African countries. – For people currently not covered by social protection schemes: collective financing, broad risk pooling and rights-based entitlements are key conditions to ensure effective access to health care for all, including for the informally employed and their families and, more generally, people in the lowest income quintile. – For all(covered and uncovered): While population coverage should be increased further, the adequacy of health care benefits needs to be improved. Investing in the enhanced availability of quality health care services is crucial. This requires ensuring availability, accessibility, adaptability and quality of health care services, allowing effective access in practice or access to a certain level, and providing comprehensive benefit packages. The empirical findings suggest that people’s trust in state institutions and intermediary organizations are important for extending social protection to the informal economy. We however observe at present strong reservations in relation to the state and intermediary institutions. Although this needs further country-specific assessments beyond the six countries covered, it will be essential in the future to invest in trust-building between the state and people working informally when thinking about how to reform existing social systems with the support of the informal economy. 1.6 FINDINGS OF INDIVIDUAL CHAPTERS 5 1.6.1 Health care and social protection: access, willingness to pay and the Covid-19 pandemic A demand for improved and accessible health care is – as Rudolf Traub-Merz and Manfred Öhm show in their chapter – a priority demand among the survey respondents. The informally employed, representing a large majority of the labour markets in all the survey countries, expect their governments to improve access to health systems and call this their top priority. The demand for better health services cuts across all social and spatial cleavages, emerging as a national priority. The surveys also find that people unable to pay for medical costs shy away from looking for treatment when they fall sick. The interplay between medical care and income depends largely on respective national health policies. The chapter by Florence Bonnet addresses financial health protection more specifically through health insurance among workers in informal employment(whether employees or independent workers). Regarding the extent and nature of health insurance coverage the author – unsurprisingly – states that the right to social health protection is far from being universal. The countries present a critical health insurance coverage gap for the majority of workers in informal employment. In all six countries, a clear majority of currently uninsured respondents are interested in joining a health insurance scheme(from 58 per cent in Benin to 70 per cent in Zambia and 81 per cent in Ethiopia). Despite a strong willingness to join, a lack of financial means is the first reason mentioned by uninsured informal economy workers(ranging from 44 per cent in Ethiopia to 68 per cent in Kenya and Zambia). The high level of support for a fairer distribution of public resources represents a political opportunity. The informally employed express broad support for policies that target the poor. The survey also shows that people in the informal economy are ready to contribute to social protection via taxes and fees. The organization of interest representation emerges as a promising pathway to attenuate decent-work deficits in the informal economy. To tap this potential, governments should establish a conducive policy, legal, administrative and institutional environment that facilitates the emergence, official recognition and operational activities of informal-economy groups and opens up venues for engaging in regular dialogue, as is done with tripartite arrangements in the formal economy. In the past, trade unions have largely bypassed the informal economy for membership recruitment, but there is substantive potential in helping to make the voice of informal labour heard. If trade unions should intend to recruit, they have to be aware that in some regards the interests of informal labour differ from those of people in formal employment. Christoph Strupat shows that economic activity of the informal economy has been strongly affected by the Covid19 pandemic. In Côte d’Ivoire and Ethiopia, 60 per cent of households faced income losses and 20–30 per cent, on average, lost jobs and work opportunities. The most serious negative effects were found in urban areas. The profiles of those who have fallen into income poverty because of the pandemic differ substantially from those who were poor even before the pandemic. The new income-poor are more urban, younger and better educated, relative to the former poor, who were more rural, less educated and predominately reliant on agriculture. A higher share of household members of the new income-poor work as employers or informal employees in the manufacturing or service sectors. The larger negative income effect points to the devastating impact of value chain disruption that also affects people living in rural areas. As a consequence, income poverty has increased by around 20 percentage points. The pandemic has not 5 Chapter authors are responsible for the contents of their individual chapters. Their findings and recommendations do not necessarily ­reflect the viewpoint of the institutions that have cooperated in this ­research project. 6

Introduction changed the low level of social assistance and packages targeted at the informally employed are largely missing in both countries. Only 4–7 per cent of the households in our surveys have received state support. The main reasons for this are financing problems, but also a large imbalance in what support is provided in favour of the formal sector. This holds true for health protection as well. 1.6.2 Governance and organizational context: do you trust or do you organize? Armin von Schiller, in his chapter, shows that many believe that political and social institutions are not particularly trustworthy, and even more so, that these institutions do not care much about respondents’ priorities. Trust is higher, however, in relation to political institutions connected to the executive(in particular, the president), as well as religious and traditional leaders. Respondents’ top policy priorities include health care and education(over 50 per cent of the sample raises these issues among their three top priorities). It is also essential to consider how people employed in the informal economy perceive the existing fiscal system. Whereas in some countries a majority regard the existing tax system as fair, in other countries the majority deny it. Overall, views are divided. In terms of tax morale, the vast majority consider that not paying taxes is wrong, but a substantial 40 per cent consider it understandable. Overall, the informally employed show consistent support for raising taxes(even on them), particularly to improve health services. They also very clearly (over 90 per cent of the sample) expect the state to support those more in need and the poor, regardless of their(in) capacity to contribute to taxation. In terms of the distribution of the tax burden, there is wide support for expecting everyone with an income to contribute. Opinions are more divided on whether the wealthy already contribute enough. Jürgen Schwettmann and Rudolf Traub-Merz show in the chapter on organizing that slightly over half of all respondents are members of a group, mainly savings clubs, cooperatives, faith-based associations or neighbourhood groups. The organizational density in the informal economy varies considerably between countries, but does not differ significantly with regard to respondents’ gender, age, location, education, employment status or income level. Most joined a group in search of economic, financial or social services. It is important to note that such groups are often multifunctional, in the sense that social support for members in need is a cross-cutting concern that transcends the organizational dynamics in the informal economy. These findings suggest that many types and forms of informal economy groups can play a role, in conjunction with local and central authorities, in supporting social protection systems, including extending health coverage. Organizations in the informal economy exhibit a surprisingly high degree of formality. This enables them to connect their members with the formal private sector and with local authorities. The overwhelming majority of respondents expressed a high degree of satisfaction with the services their group provides, and with the competence and trustworthiness of the group’s leaders. Organization-building emerges as a promising pathway to attenuate decent-work deficits in the informal economy. 1.6.3 Informal labour and organized trade unions Rudolf Traub-Merz shows that the informal economy is a vast terrain with a multitude of employment relations and work situations, some of which appear more trade union-friendly than others. The survey is also possibly the first ever undertaken to shed light on how trade unions are perceived by workers in informal employment across a whole country. There is a low level of awareness of trade unions in the informal labour force. In all countries, a clear majority have never heard of trade unions or have no knowledge of what they do or stand for. Those who have heard about trade unions, however, perceive they play in their countries a fairly positive role. While interest in membership is substantial, readiness to pay membership fees is not. Trade unions are viewed as organizations with substantial financial resources and thus able to offer membership at a discount, if not free. It is wrong to call the interests of trade unions and informal labour as »naturally congruent«. There are great opportunities for the two to come together, however. REFERENCES Bhorat, H./Cassim, A./Ewinyu, A./Steenkamp, F.(2019): Social protection and inequality in Africa: Exploring the interactions, in: Income Inequality Trends in Sub-Saharan Africa. United Nations. Clement, J.(2020): Social protection clusters in sub-Saharan Africa, in: International Journal of Social Welfare, 29(1): 20–28; available at: https://doi.org/10.1111/ijsw.12378. ILO(2003): Guidelines concerning a statistical definition of informal employment; available at: https://www.ilo.org/global/statistics-and-databases/standards-and-guidelines/guidelines-adopted-by-international-­ conferences-of-labour-statisticians/WCMS_087622/lang--en/index.htm –(2012): Social Protection Floors Recommendation, 2012(No. 202); available at: http://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB: 12100:0::NO::P12100_ILO_CODE:R202 –(2017): World Social Protection Report 2017–19: universal social protection to achieve the sustainable development goals(Executive Summary); available at: http://www.ilo.org/global/publications/books/ WCMS_605078/lang--en/index.htm –(2021a): World Social Protection Report 2020–22: social protection at the crossroads: making a decisive turn for a better future; available at: https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/@ publ/documents/publication/wcms_817572.pdf Kobo, O.(2010):»We are citizens, too«: the politics of citizenship in independent Ghana, in: The Journal of Modern African Studies, 48(1): 67–94. World Bank(2021): World Bank Development Indicators; available at: https://data.worldbank.org/indicator/SP.POP.TOTL(accessed on 4.6.2021). Wimmer, A.(2018): Nation-building: why some countries come together while others fall apart. Princeton: Princeton Studies in Global and Comparative Sociology. 7

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows 2 USE OF MEDICAL CARE: HOW PEOPLE IN INFORMAL EMPLOYMENT RATE ACCESS TO HEALTH SERVICES Rudolf Traub-Merz and Manfred Öhm 2.1 INTRODUCTION Health protection is a key pillar of social protection. Access to health care is a fundamental human need, and having access to health services when needed is – as we show in this chapter – of paramount importance. The informally employed, representing the large majority of employers, employees and own-account workers in all six surveyed countries, expect their governments to improve access to health care systems, citing it as their top priority among improvements in state services. The central concern of government health policies should be to establish a system of medical care that is accessible to all who are sick and need treatment. Accessibility to medical services is determined by supply and demand. Without going into detail, access to health 1 is a multidimensional concept and can be measured along three main dimensions: (a) availability of medical services: services should exist in sufficient quality and quantity where they are needed;(b) affordability: the extent to which people are able to pay for medical care – if health care is provided for a fee and charges are above what poorer segments in society can afford, medical care may be materially available but remain financially inaccessible;(c) acceptability: measures the extent to which a medical service meets a community’s cultural needs and expectations. Within this framework, access combines the availability of health care with the capacity to pay for it and culturally accepted health norms. Some aspects of access are supply factors, while others reflect the societal conditions of demand. This chapter looks at the use of medical care by the informally employed. In this chapter we do not assess how satisfied patients are with the services they receive when visiting medical facilities. Instead, we inquire how important it is for people to have access to good health services and how often they go for medical care. We begin with an assessment of the extent to which medical care is available when needed. The views of respondents are examined in terms of location, income, age and gender. A picture is established concerning the extent to which the use of medical care is determined by 1»Access to health« is a term that goes beyond access to health services, and also includes the availability of clean water and healthy food. socio-economic disparities. The study then looks at respondents’ needs hierarchies. By asking respondents to prioritize from a list of state services we are able to build up a ranking according to urgency. From there we are able to show how respondents adjust their needs hierarchies in relation to the intensity with which they use medical care. While no detailed analysis is provided on the availability of medical services, some supply indicators are nevertheless considered which show that the demand for better health services is inversely related to the availability of medical care. 2.2 USE OF MEDICAL CARE To assess the openness of a health care system and obtain an understanding of the barriers that may prevent resort to medical treatment, we asked interviewees about the frequency with which they were able to access medical care. 2 The question was phrased as follows:»Over the past year, how often, if ever, have you or anyone in your family gone without medicine or medical treatment?« 3 4 The question referred not to the respondent alone but included all members of their family. The question does not ask about the frequency of illness or the number of visits or non-visits to medical facilities. Instead, it asked interviewees to weight the incidence of their need for treatment and whether they in fact received treatment. We thus obtain testimony of the respondent’s impression of whether medical care was available when needed. Respondents could voice their opinion on the actual number of cases of treatment or just articulate a gut feeling. Either way, the question casts light on their perceptions: do people believe that they are able to use health services when needed? 2 We limit our analysis here to only one aspect in evaluating the use of medical services. A more detailed paper that looks at the assessment of quality of services and compares public with private health facilities will be published at a later stage. 3 This question is part of a five-sequence question that includes food, clean water, cooking fuel and cash income and is used to construct an index for lived poverty. We follow here the AfroBarometer(AB) approach and acknowledge the work of AB in this area. 4 In our data analysis, we include only households that in response to another question reported cases of sickness over the past 12 months. The answers from households with no cases of sickness were ignored. See Appendix Table 2 for cases. 8

Use of Medical Care: How People in Informal Employment Rate Access to Health Services 2.2.1 Use of medical care – disparities between countries The answers to our question indicate major differences and allow us to rank the six countries according to the perceived availability of medical care(Figure 2.1). In Kenya and Zambia, a majority of nearly 70 per cent declared that medical care is always or mostly available when needed. Around 10 per cent still feel marginalised, however, in that they could hardly obtain medical care. In Senegal and Benin, 30 and 32 per cent, respectively, face a situation that prevents them usually or always from obtaining medical care when needed. In these countries, the share of those who mostly or always go without medical care is nearly three times higher than in Kenya and Zambia. The other two countries occupy a middle position, with Côte d’Ivoire leaning towards Benin and Senegal and Ethiopia being closer to Kenya and Zambia. Concerning the perceived availability of medical care we can talk of an East African/West African gap. Respondents from Benin, Senegal and Côte d’Ivoire rate the availability of medical care lower than respondents from Kenya, Zambia and Ethiopia. 2.2.2 Use of medical care – inequality by urban/rural residence The residence factor confirms the existence of a country ranking in relation to the use of medical care. In Kenya and Zambia, living location is of minor or no importance for the use of medical care. 5 Judging from our surveys, Kenya and Zambia appear to have overcome the rural/urban divide in this regard(Figure 2.2). In Senegal, Benin and Côte d’Ivoire, the living environment becomes a distorting factor. While in urban locations, some 44 to 56 per cent of respondents confirm regular use of medical care, this is the case for only 25 to 38 per cent of rural residents. Ethiopia reports a similar gap between urban and rural residents. Resort to medical care in Ethiopia, however, is ensured at a higher level for both groups, so that the relative use gap – which we could call the urban/rural discrimination factor – is lower. 2.2.3 Use of medical care – inequality by gender To obtain an understanding of gender sensitivity in relation to the use of medical services, we divided our sample into male- and female-led households. The head of household answered the question for all members without giving separate accounts for male and female members. 6 Figure 2.3 reproduces the responses according to the sex of the head of household. None of the countries shows a significant disparity in the use of medical care between maleled and female-led households. Senegal and Côte d’Ivoire just miss the significance level, while the other four countries show a high level of conformity. With the data generated from our surveys therefore we are unable to assess whether the sex of the household head influences the use of medical care within the household. 5 We have simplified the data appraisal by putting the five-answer options into two groups. If people have»never« or»only once or twice« gone without medicine or medical treatment, we call it»regular use of medical care«; if people have»mostly« or»always« gone without medicine or medical treatment, we call it»low use of medical care«. The middle category of»several times« is ignored. 6 We assume that both male and female heads do not apply a bias when assessing cases of sicknesses in their households. Thus, the ­responses can be taken as an approximation for assessing the ­gender-related use of medical services. Figure 2.1 Use of medical care 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 32.4% 28.0% 39.6% Benin 30.2% 34.8% 34.9% 23.9% 29.2% 46.9% 17.5% 19.6% 62.9% Senegal Côte d'Ivoire Ethiopia 12.5% 19.4% 68.1% Zambia 10.0% 21.7% 68.3% Kenya regular use sometimes low use Question:»Over the past year, how often, if ever, have you or anyone in your family ›gone without medicines or medical treatment‹?« Answer options: never; just once or twice; sometimes; many times; always. Note: Households are included only if they reported cases of sickness during past 12 months. Definitions: regular use of medical care=»gone without medical care«: never or just once or twice; low use of medical care=»gone without medical care«: many times, or always. 9

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 2.2 Regular use of medical care, by urban/rural residence 90% 80% 70% 60% 50% 40% 30% 20% 43.5% 26.5% 17.5% 50.3% 31.7% 18.6% 55.6% 37.9% 17.6% 10% 0% Senegal Benin Côte d'Ivoire 76.6% 59.6% 66.9% 69.8% 17.0% Ethiopia 3.0% Zambia Rural Urban use gap Note: Households included only if they reported cases of sickness during past 12 months. 67.8% 69.6% 1.8% Kenya Figure 2.3 Use of medical care, by gender 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 31.4% 28.7% male female Senegal 32.7% 31.9% male female Benin 25.7% 19.9% male female Côte d'Ivoire 17.5% 17.5% male female Ethiopia 10.7% 9.4% male female Kenya use regular use low Note: Households included only if they reported cases of sickness over the past 12 months. 12.8% 12.3% male female Zambia 2.2.4 Use of medical care – inequality by income A look at the impact of household earnings on the use of medical care reveals a slightly modified pattern of inequality. Respondents’ household income was grouped into four income classes: the»extreme poor«(monthly household income less than half the statutory minimum wage), the »moderately poor«(monthly household income between half and full minimum wage), the»non-poor«(monthly household income between minimum wage and twice the minimum wage) and the»well-off«(monthly household income higher than twice the minimum wage). 7 In terms of these income classes five countries indicate a positive correlation. In Senegal, Kenya, Benin, Côte d’Ivoire and Ethiopia, the use of medical care is higher when income increases. The effect is negligible in Zambia. Zambia is thus the only country in which income does not decide whether medical treatment is used or not(see Figure 2.4). The situation of many poor people in Senegal and Benin is miserable. Around 40 per cent of households with an income below 50 per cent of the statutory minimum wage – that is, 7 Ethiopia does not have a statutory minimum wage. Income classes were defined as a multiple of 1,500 Birr. 10

Use of Medical Care: How People in Informal Employment Rate Access to Health Services Figure 2.4 Use of medical care, by income 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% regular use low use regular use low use regular use low use regular use low use regular use low use regular use low use Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia less 0.5 MW 0.5 MW-1 MW>1 MW-2 MW>2 MW Note: Households included only if they reported cases of sickness during the past 12 months. As Ethiopia has no statutory minimum wage, we use 1,500 Birr/month as substitute for the minimum wage. those we call here the extreme poor – hardly use medical treatment(Benin 42.9 per cent; Senegal 39.5 per cent). For Côte d’Ivoire and Ethiopia, the situation is only slightly better. A quarter report that they have not used health care services. In Kenya and Zambia, this share is considerably less, at 13.7 per cent and 13 per cent, respectively. Higher income in itself is not a sufficient condition for securing medical care. Even in Kenya, only 90.5 per cent of the highest income segment report good use of medical treatment at all times. In Benin and Senegal, less than half of the»well-off« gave a positive response. The»well-off« in Côte d’Ivoire, Zambia and Ethiopia fare little better. Income clearly facilitates the use of medical care. But it does not do so for all, and a significant number of fairly well-off people are still left without medical care. We can therefore assume that supply shortages play a role here as well. 8 2.2.5 Discussion 9 All the surveyed countries still have large numbers of citizens for whom the use of health care can only be called precarious. While even the»best performers« among our survey group are far away from providing care for all, there are huge differences between the countries, which merits a com8 We have to emphasize again that the evaluation measures the use of medical care and not the quality of medical treatment or satisfaction with the medical care received. It may well be that the quality of medical treatment equally depends on income and people’s capacity to pay for lower or higher standards of treatment. If so, inequality arising from income differentials could be even higher. ment. Even though this is not the place to compare in detail respondents’ perceptions of health service availability, a simple look at WHO statistics immediately points to enormous national differences that mirror respondents’ perceptions. Kenya and Zambia fare better than Benin and Senegal on all indicators listed in Table 2.1, while Ethiopia and Côte d’Ivoire are positioned in the middle. Kenya and Zambia spend more of their GDP on health, which also translates into higher per capita expenditures. Importantly, the share of expenditure invested in primary health care is substantially larger, which increases the provision of health care for the poor. They employ more medical doctors and nurses to care for patients. Out-of-pocket payments(OOP), which measure the share of direct payment for health-care goods and services from personal sources, serve as an indicator of the burden health costs impose on a household. Here again, Kenya and Zambia fare much better than the other four countries. Clearly, Zambia 10 and Kenya provide significantly more funding for primary health care and appear to provide easier access to medical care with fewer social funding barriers than Benin and Senegal. Côte d’Ivoire and Ethiopia take a middle posi9 This chapter does not include an analysis of the use of medical care by age and education. The age of the household head is statistically not correlated with the use of medical care(see Appendix Table 2.A3) while the household head’s level of education is(see Appendix Table 2.A4). Education is closely linked to income, however, and provides the same insight into social disparities as what we get from income. 10 Zambia under the National Health Care Package(NHCP) offers basic health-care packages at the primary(district) level free of charge. Capacity constraints and funding shortages do not always allow unlimited access to medical care under the NHCP. For details, see: WHO (2017). 11

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Table 2.1 Health expenditures and availability of medical staff(various indicators, 2018) Current health expenditure(CHE) as% of gross domestic product(GDP) Current health expenditure(CHE) per capita in PPP Primary health care(PHC) expenditure as % of current health expenditure(CHE) Primary health care(PHC) expenditure per capita in PPP Medical doctors(per 10,000 population) Nursing and midwifery personnel(per 10,000 population) Out-of-pocket spending(OOPS) as% of current health expenditure Senegal 3.98 146.4 65.9 96.5 0.691* 3.127* 55.9 Benin 2.49 83.2 n.a. n.a. 0.791 3.888 45.5 Côte d’Ivoire 4.19 176.3 80.4 141.7 2.314*** 6.048 39.4 Note:*= 2017;**= 2016;***= 2014. Source: https://apps.who.int/nha/database/ViewData/Indicators/en; https://apps.who.int/gho/data/node.main Ethiopia 3.29 66.6 82 54.6 0.769 7.135 35.5 Kenya 5.17 179.2 72.5 129.9 1.565 11.656 23.6 Zambia 4.93 208.4 78.5 163.6 1.628** 13.376 10 tion. In some regards, they do relatively well, while in others they perform worse. 11 To that extent, the supply indicators for health services correspond to respondents’ views on the use of medical care. The data show similarities and differences between the countries and clearly allocate an outstanding role for» better health services« and» better schools and education« as top priorities for the informally employed. 2.3 DEMAND FOR BETTER STATE SERVICES For the purpose of ranking people’s expectations for improved state services, eight services are considered»key responsibilities« of the state to society and offered for selection:»better schools and education«,»health services«, »water supply«,»roads and bridges«,»electricity supply«, »pensions for the elderly«,»food supplies in times of crisis« and»police services«. 12 Respondents 13 were first asked to assess in each case whether the government should improve a respective service on the list. Thereafter, they were asked to rank the services as their first, second or third priority. 14 2.3.1 Priorities for better state services Figure 2.5 provides an overview of respondents’ answers regarding their highest priority in the list. – In all six countries,»better health« is ranked top or second; in Benin, Senegal and Côte d’Ivoire»better health« is the top runner by far; in Kenya, Zambia and Ethiopia, it is relegated to second place by a low margin; on average, in a six-country average, 31.5 per cent of respondents make»better health services« their top choice. – Similar prominence is given to»better schools and education«. In Kenya and Zambia, it comes ahead of health in top position; in Senegal and Côte d’Ivoire it is ranked second; and only in Benin and Ethiopia it is »downgraded« to third and fourth place, respectively. – The bottom end of the ranking is unambiguous. In all countries,» better police services« and» better pensions for the elderly» are found at the bottom of the demand hierarchy. 11 While some countries fare better than others in relation to a specific indicator, none of them is anywhere close to the 4.5 workers per 1,000 people that the WHO estimates are needed to meet the SDG target of universal health coverage. 12 See Traub-Merz and Öhm(2021: Appendix I) for more information on the construction of the service ranking. 13 The interviews were structured into two parts. The question on the use of medical care belonged to part one and was answered by the household head; the assessment of state services belonged to the second part and was answered by the»selected person«. The selected person was in some 60 per cent of the cases again the household head, but in 40 per cent of cases a different member of the household. We compared the answers of the two groups(household head vs other member) and found no statistically relevant difference. Obviously, members of the same household can be seen as a community of norms( Normengemeinschaft), exchange views on the quality of state services and»conclude a united household view«. We therefore found no reason to separate respondents into heads of household and other members. 14 We limit ourselves here to findings according to the first priority. – Five of the six countries produce a middle cluster consisting of various infrastructural components:» better water supply«,» better food programmes in times of crisis«,» better roads and bridges« and»better electricity supply«. The rankings of these services vary between countries, but none comes first or last. – Ethiopia is a case on its own. There, respondents put better water supply and better roads significantly higher in their needs hierarchy than do respondents from the other countries. Calls for improved infrastructural services are level with calls for better health services. 2.3.2 Demand for health services by residence, age, income and gender Demand for a service reflects the gap between needs and shortage of supply. Various factors may impact on percep12

Use of Medical Care: How People in Informal Employment Rate Access to Health Services Figure 2.5 Demand for better state services(top priority) 50% 45% 43.6% 40% 35% 30% 25% 20% 15% 10% 5% 0% Senegal 41.2% 33.5% Côte d'Ivoire Benin 28.4% 23.0% Kenya Zambia better health services better roads and bridges better pensions for the elderly better schools and education better food programmes in times of crisis better police services 19.7% 31.5% Ethiopia 6-country average better water supply better electricity supply Figure 2.6 Demand for better health services(first priority), by urban–rural residence Côte Benin d'Ivoire Senegal -3.0% Urban Rural Urban/rural-gap Urban Rural Urban/rural-gap Urban Rural Urban/rural-gap Urban Rural Urban/rural-gap Urban Rural Urban/rural-gap Urban Rural Urban/rural-gap –1.0% 7.0% 2.3% 5.7% 3.4% 9.5% 7.9% 17.0% 27.0% 25.4% 18.0% 25.9% 22.4% 23.4% Zambia Ethiopia Kenya 37.0% 47.0% 44.7% 42.4% 38.1% 43.8% 35.3% 31.9% 34.9% tions of a supply–demand gap. We look at four of them: urban and rural residence, age, income, and gender. A. URBAN–RURAL DIVIDE Rural and urban areas differ mainly in the provision of state services. Cities with high population densities are usually provided with physical infrastructure before remote areas. In general, the availability of public goods such as roads, water and electricity improve with population density, as is the case with health facilities. We do not look at absolute supply indicators, however, but at the relative positioning of the demand for better health services, which reflects the personal needs hierarchy and may differ from the factual supply side. The results are presented in Figure 2.6. In Kenya, Ethiopia and to some extent in Côte d’Ivoire, we observe some shifts in the demand for health services. In those countries, urban residents assign health a higher preference than do rural residents. In other countries the urgency with which health is identified as highest priority is not connected to residence. Overall, the observed differences in priorities do not allow us to talk of a substantial urban–rural divide. B. AGE It is difficult to predict the impact of age on people’s priorities concerning better state services. The older generation may be more interested in better police services, better health services, better water supply and availability of pensions. Younger age groups may prioritize education, electricity connection and transport infrastructure to improve their(social) mobility. Where old and young people live together in a 13

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 2.7 Demand for better health services and pensions(first priority), by age groups 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Senegal age 15-24 years Benin Côte d'Ivoire age 25-34 years age 35-44 years Ethiopia age 45-54 years Zambia age 55-64 years Kenya age 65 a.m. years household, discussions of state service deficits may balance people’s priorities over time. The findings are shown in Figure 2.7. Here, we compare the priorities of various age groups for improved health services with access to pensions. Our assumption that many older people shift their priorities from health care to pensions is confirmed for Kenya, Zambia and, to a lesser extent, for Côte d’Ivoire. It is not confirmed for Senegal, Benin and Ethiopia. Here, the demand profiles indicate a broad conformity between age groups. The effect of age-biased demand for pensions is particularly strong in Kenya, where elderly people 65 years of age or over shift their priorities towards pensions to an extent that it even outstrips demand for better health services. All other age groups in other countries have a clear leaning towards improved health services and find the introduction of pensions less urgent. C. INCOME Income is another factor that makes predictions of how it impacts on demand for better health services difficult. The poor are likely to depend on state provision for access to primary health care, whereas the well-off are more likely to be able to afford private health services. In-between are the »non-poor«, who may be able to tap alternative sources where the state fails to deliver in some service areas(such as primary health care), but may find it difficult to do so in other areas(such as secondary health services). There may also be people who classify some duties as moral duties of the state towards all citizens and designate a particular service a top priority independent of their personal situation. To assess priority voting by income we have grouped respondents into four income classes, which differ as a multiple of the statutory minimum wage(MW). The results are shown in Figure 2.8. Voting on health hardly depends on income. Health demand shows some income dependency in Zambia and the»welloff« in Benin are less interested in better health services. Overall, the impact is minor and all income classes express a fairly similar urgency to obtain better health services from the state. D. GENDER Could gender be a factor that affects respondents’ priority setting with regard to health care? Figure 2.9 provides an answer from our survey data. For Kenya, we observe that more male respondents than female respondents favour better health care as first priority. The strength of this effect is small, however. In all other countries, gender is not a variable that explains the demand for improved health services. 2.3.3 Discussion Reasons why»health service« and»school and education« top the list of desired state services easily come into mind. Education is identified as the primary vehicle for social advancement. Obtaining educational qualifications is the easiest way for many to improve their chances on the labour market and climb the social ladder. The strong demand for more educational infrastructure expresses the hope of large segments of the population that they may be able to improve living standards for themselves or for their children. 15 The urgency with which people call for» better health services« may equally be argued to express a call for improved living conditions. It is important to note that the call for » better health services« far outstrips the demand for» better pensions for the elderly«. On average, eleven times more people identify» better health« as their top need than the number who call for» better pensions«. One explanation may be that the pressure of current social problems exceeds 15 There is indeed a positive relationship between higher levels of education and access to formal employment. See ILO(2018). 14

Use of Medical Care: How People in Informal Employment Rate Access to Health Services Figure 2.8 Demand for better health services(first priority), by income 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Senegal Côte d'Ivoire less 0.5 MW Benin 0.5-1 MW Kenya >1 MW-2 MW Zambia >2 MW Ethiopia Note: MW= statutory minimum wage. Ethiopia has no statutory minimum wage. Categories for Ethiopia: less than 750 Birr; 750–1,500 Birr; 1,501–3,000 Birr; more than 3,000 Birr. Figure 2.9 Demand for better health services(first priority), by gender 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% Senegal 42.0% 45.4% Côte d'Ivoire 39.4% 43.1% Benin 34.5% 32.8% Kenya 25.9% 31.2% Zambia 22.4% 23.3% Ethiopia 19.5% 19.8% male female that of future needs. Sickness, its treatment and costs are daily problems for many people and draw more attention than future income after retirement. We tested the dependence of four factors on first priority voting and found a small to medium correlation with urban–rural living environment, no or only a small correlation with age, no correlation with gender, and no or only a small correlation with income inequality. As these factors are strong components of socio-economic class formation, internal stratification within the informal labour force has not emerged to an extent that it sets apart priorities for better state services socially. While we do not know what will happen if income disparities are further deepened, we can state that the current level of income inequality does not produce disparities in priority setting for health. We thus obtain an interesting result: the use of medical care is strongly linked to social disparities, whereas the priority demand for health is not. If socio-economic factors are not yet at work to develop a differentiated need structure, we may consider our ranking to be typical of the informal labour force as a whole. As this segment of the labour market dominates the economy and encompasses between 80 and 90 per cent of total employment, the vote profile can be generalized and assessed as a reflection of general significance within a country. Our argument that the availability of medical care and the urgency with which people demand improved services from 15

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 2.10 Demand for health as first priority and use of medical care 80% 70% 60% 62.9% 50% 40% 30% 20% 43.6% 34.9% 19.7% 30.2% 17.5% 10% 0% Health 1st. Priority medical care-regular use medical care-low use Senegal Cote d'Ivoire Benin Kenya Zambia Ethiopia the state are strongly connected can easily be observed in Figure 2.10. Health as a priority demand becomes less important if medical care is easily available, and goes up if access to medical care is blocked. While this general trend is confirmed, the three West African countries can also be grouped into a different camp from the three East African countries. For respondents from Benin, Senegal and Côte d’Ivoire, improved health services is their highest priority because access to medical care is difficult. Respondents from Kenya, Zambia and Ethiopia put the urgency with which they demand improved health services in second place because medical care is already more easily available. 2.4 SUMMARY AND CONCLUSIONS The interplay between medical care and income depends largely on the respective national health policies. If people are covered by health insurance or may obtain basic medical treatment at no cost, access to medical care may improve and income might lose its role of deciding whether people are treated or not. Demands for better health cut across social and spatial cleavages and can be called a national priority. The findings of this study confirm that universal health coverage should be prioritised on national agendas. National policy initiatives should embrace universal health coverage as a top priority, in particular with a view to those in the informal economy. The survey looked at the accessibility of health care services for the informally employed, from two sides: it tried to determine the importance of health care within a ranking of essential state services; and it provided an assessment of whether health care is used when needed. The two sides are interlinked: if respondents are confronted by a situation in which health care is not available when health problems arise, they demand improved health services with more urgency; if they find medical care to be always or mostly available, they shift their demand for better state services to other deficit areas. We can conclude with confidence that in identifying»improved health services« as the highest-ranking need, then this is the principal area in which the informally employed want the state or government to intervene and to improve the provision of services. REFERENCES ILO(2018): Women and men in the informal economy: A statistical ­picture. Geneva: International Labour Office. Traub-Merz, Rudolf/ Manfred Öhm(2021): Access to Health Services. A Key Demand of Informal Labour in Africa – Findings from Representative Country Surveys in Sub-Saharan Africa. Berlin: Friedrich-EbertStiftung. WHO(2017): World Health Organisation Zambia, WHO Country Cooperation Strategy 2017–2021; available at: https://apps.who.int/iris/ bitstream/handle/10665/273149/ccs-zmb-eng.pdf?ua=1(accessed on 23.5.2020) Access to health is no doubt an issue of social justice. It is linked to income disparities in two regards: the lower the income the less medical care becomes available when needed, and the stronger the likelihood that people will have to incur debt to finance health treatment(for more on this, see Chapter 3). People unable to pay tend not to seek treatment when they fall sick. 16

Use of Medical Care: How People in Informal Employment Rate Access to Health Services APPENDIX Table 2.A2 Households with cases of sickness and no cases of sickness over the past 12 months Households with cases of sickness Households with no cases of sickness Total households N Senegal 32.1% 67.9% 100.0% 1,193 Zambia 57.8% 42.2% 100.0% 1,195 Kenya 56.9% 43.1% 100.0% 1,188 Benin 41.8% 58.2% 100.0% 1,192 Côte d'Ivoire 54.2% 45.8% 100.0% 1,200 Ethiopia 64.2% 35.8% 100.0% 2,332 Table 2.A3 Use of medical care in households with cases of sickness by age of head of household Regular use Several times Low use All N SENEGAL 15–24 years 32.4% 35.2% 32.4% 100.0% 105 25–34 years 37.8% 34.7% 27.6% 100.0% 196 35–44 years 33.0% 38.5% 28.5% 100.0% 179 45–54 years 35.9% 33.3% 30.8% 100.0% 156 55–64 years 35.7% 35.7% 28.6% 100.0% 112 65 years> 31.7% 26.7% 41.7% 100.0% 60 ZAMBIA 15–24 years 71.3% 15.8% 12.9% 100.0% 101 25–34 years 74.0% 15.0% 11.0% 100.0% 127 35–44 years 66.4% 21.8% 11.8% 100.0% 119 45–54 years 65.4% 20.5% 14.1% 100.0% 78 55–64 years 61.4% 27.3% 11.4% 100.0% 44 65 years> 57.1% 25.7% 17.1% 100.0% 35 KENYA 15–24 years 79.2% 11.5% 9.4% 100.0% 96 25–34 years 70.5% 24.4% 5.1% 100.0% 176 35–44 years 61.5% 24.0% 14.6% 100.0% 96 45–54 years 64.6% 22.0% 13.4% 100.0% 82 55–64 years 64.1% 20.5% 15.4% 100.0% 39 65 years> 54.5% 36.4% 9.1% 100.0% 22 BENIN 15–24 years 41.0% 27.6% 31.4% 100.0% 105 25–34 years 42.3% 27.8% 29.9% 100.0% 241 35–44 years 32.4% 29.7% 37.9% 100.0% 182 45–54 years 45.3% 28.4% 26.3% 100.0% 95 55–64 years 32.0% 32.0% 36.0% 100.0% 50 65 years> 57.1% 4.8% 38.1% 100.0% 21 CÔTE D’IVOIRE 15–24 years 53.0% 25.3% 21.7% 100.0% 83 25–34 years 53.5% 28.3% 18.2% 100.0% 159 35–44 years 43.9% 30.4% 25.7% 100.0% 148 45–54 years 41.6% 30.3% 28.1% 100.0% 89 55–64 years 39.6% 27.1% 33.3% 100.0% 48 65 years> 33.3% 42.9% 23.8% 100.0% 21 ETHIOPIA 15–24 years 58.3% 21.3% 20.4% 100.0% 108 25–34 years 62.1% 19.1% 18.7% 100.0% 235 35–44 years 63.5% 20.2% 16.3% 100.0% 178 45–54 years 60.4% 22.5% 17.1% 100.0% 111 55–64 years 82.5% 10.5% 7.0% 100.0% 57 65 years> 50.0% 18.2% 31.8% 100.0% 44 17

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Chi²-testing for use of medical care by age group of head of household Senegal Zambia Kenya N sig. Cramer-V N 808 0.763 0.064 Benin 694 504 0.685 0.086 Côte d'Ivoire 548 511 0.036 0.138 Ethiopia 733 sig. 0.162 0.338 0.012 Table 2.A4 Use of medical care in households with cases of sickness by education of head of household SENEGAL No school Some primary Primary Secondary University ZAMBIA No school Some primary Primary Secondary University KENYA No school Some primary Primary Secondary University BENIN No school Some primary Primary Secondary University CÔTE D’IVOIRE No school Some primary Primary Regular use 27.7% 29.6% 49.6% 48.1% 65.1% 55.3% 65.2% 69.7% 75.3% 100.0% 43.4% 52.3% 74.3% 79.0% 90.9% 32.6% 43.7% 41.0% 57.1% 52.9% 38.2% 47.4% 49.7% Several times 35.9% 36.1% 31.9% 37.0% 25.6% 28.9% 21.5% 18.9% 13.0% 34.0% 29.9% 18.6% 16.6% 9.1% 28.8% 23.7% 31.1% 35.7% 29.4% 34.1% 26.0% 28.9% Low use 36.4% 34.3% 18.5% 14.8% 9.3% 15.8% 13.3% 11.4% 11.7% 22.6% 17.8% 7.1% 4.5% 38.6% 32.6% 28.0% 7.1% 17.6% 27.7% 26.6% 21.4% All 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Secondary 62.2% 27.0% 10.8% 100.0% University ETHIOPIA No school Some primary Primary Secondary University 64.0% 55.4% 60.5% 74.8% 75.9% 82.4% 20.0% 22.1% 21.3% 13.7% 15.2% 11.8% 16.0% 22.4% 18.2% 11.5% 8.9% 5.9% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Chi²-testing for use of medical care and education of head of household Senegal Zambia Kenya N sig. Cramer-V N 810 0.000 0.173 Benin 691 503 0.516 0.085 Côte d'Ivoire 548 511 0.000 0.212 Ethiopia 835 sig. 0.018 0.069 0.001 Cramer-V 0.101 0.101 0.055 N 220 385 135 27 43 38 158 228 77 2 53 107 183 157 11 267 215 161 14 34 173 154 159 37 25 294 314 131 79 17 Cramer-V 0.115 0.115 0.124 18

Financial risk protection – how do the informally employed pay for medical treatment? 3 FINANCIAL RISK PROTECTION – HOW DO THE INFORMALLY EMPLOYED PAY FOR MEDICAL TREATMENT? Rudolf Traub-Merz Financial risk protection is a key component of universal health coverage(UHC). It is defined as providing access to quality health services for all without suffering financial hardship.»Financial protection is achieved when direct payments made to obtain health services do not expose people to financial hardship and do not threaten living standards« (WHO). Government transfers, insurance contributions, foreign donations and out-of-pocket payments(OOPs) are the four main sources for financing health costs. High direct payments for health treatment can cause households to incur costs beyond their means and push them into poverty. Tax transfers to reduce health costs or finance free health and extended health insurance coverage are key strategies to reduce out-of-pocket payments at a time of sickness and are the main strategies to protect the poor from financial risks and keep them out of a vicious health–poverty circle. or payments are made though other sources, such as health insurance. If households have to pay directly, three main options are available: people may settle bills by using(part or all of) their savings. If own funds are not available, they may sell some of their possessions, such as cattle, tools, advance sales before the harvest, jewellery, household equipment, means of transport. Alternatively, they could approach friends, relatives, neighbours, money lenders, banks, or others to obtain a loan. A further but less important option is to seek monetary assistance in the form of a donation. One may question the wisdom of combining access to traditional forms of solidarity with mobilising funds from market economy operators. But traditional forms of solidarity are based on reciprocity and while they may allow more leeway in pay-back procedures than market operators provide, pressure to»return the favour« is nevertheless at work and such debts have to be settled when the need arises for others. 3.1 SOURCES OF FINANCING MEDICAL CARE The various ways of paying health bills are presented in Figure 3.1 Long travel distances, unavailability of qualified staff or medicine or services that cannot be afforded due to their high cost may be the major reason why people refrain from visiting medical facilities. But what happens if they do go for medical treatment? If health services are provided free of charge, there are still incidental costs, such as payments for transportation, time of absence from home, out-of-pocket payments for food and other factors, and other things. If health services have to be paid directly by users themselves, however, the availability of sufficient funding may become a major concern. To obtain an understanding of how burdensome the costs of medical treatment may become we asked heads of households»How did you or your family find the money to pay for this treatment?«. We did not ask for the exact amount and thus are not able to relate private health expenditures to household income. Identifying the source of funding allows us to draw conclusions about the financial hardship people have to cope with when going for treatment, however. We classified the ways of paying health bills in terms of various sources. Patients may enjoy health services without directly paying for them, either because no costs are charged Patients in Zambia appear to be in a»privileged« situation when falling sick. Nearly half of them receive medical care without having to worry about the costs. A further 26 per cent can mobilize savings, while some 12 per cent are»less privileged« in that they have to obtain cash by either selling some of their assets or taking a loan to be able to pay for treatment(Figure 3.1). Senegal is positioned at the other end of the scale. A mere two per cent are provided free health services, while 41.5 per cent use savings. This leaves a whopping 47.9 per cent – or nearly every second patient – who is forced to sell property or go into debt. Kenya, Ethiopia and Côte d’Ivoire hardly fare better. A fifth of patients or less enjoy access to health care free of charge, while between 34 per cent(Kenya) and 57 per cent(Ethiopia) use savings. This leaves between 27.1 per cent(Côte d’Ivoire) and 42.7 per cent(Ethiopia) to search for»external« funds by borrowing or selling assets. Our findings are supported by WHO figures for out-ofpocket payments(OOP), which group Zambia and Senegal at the positive and negative ends of the ranking order(see Table 3.1). 19

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 3.1 Sources of finance for medical treatment 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Zambia Kenya Ethiopia 49.1% 20.8% 13.0% 2.6% 1.2% 33.9% 38.2% 1.0% 26.0% 12.2% 31.0% 42.7% 11.7% 13.1% 3.4% Côte d’Ivoire Senegal 6.8% 0.4% 2.0% 1.1% 57.1% 27.1% 41.5% 47.9% No cost was charged Assistance(government, employer, co-workers) Sale of assets or taking loan(all forms) Other means Savings 8.6% 7.6% Note: No data available for Benin. Table 3.1 Out-of-pocket payments as share of total health spending(2018; per cent) Out-of-pocket payments(OOP) as ­percentage of total health spending Benin 45.0 Kenya 24.0 Senegal 52.4 Source: WHO, Global health expenditure data base. https://apps.who.int/nha/database/ViewData/Indicators/en Zambia 11.8 Côte d‘Ivoire 39.4 Ethiopia 35.0 Figure 3.2 Sources of finance for medical treatment by health insurance membership(two countries) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 33.5% 34.2% 17.3% No health insurance 22.7% 32.8% 32.8% Health insurance 44.7% 39.8% 9.7% No health insurance 39.2% 35.4% 18.8% Health insurance Kenya Ethiopia . No cost Savings Sale of assets or loans The high level of free services in Zambia reflects the free primary health care policy introduced a few years ago at the district level. Our finding that about 50 per cent of patients still pay for medical services points to medical treatment at secondary or tertiary level that is not provided free of charge or to capacity constraints and funding shortages at primary level which limit the use of free medical care. Despite these limitations, free primary health policy is a key reason why fewer patients in Zambia have to sell property or take out loans than in other countries. The positive impact of health insurance on burdensome forms of financing health treatment can be shown if membership is compared with non-membership. In Ethiopia and Kenya, the two countries in our survey group with significant health insurance coverage(35.8 per cent and 25.6 per cent, see Chapter 4) the availability of free health services doubles 20

Financial risk protection – how do the informally employed pay for medical treatment? when health insurance is concluded. At the same time, sale of assets and indebtedness shrinks. 3.3 FINANCING MEDICAL CARE: IS THERE A GENDER BIAS? At first sight, it may be surprising that burdensome forms of health funding do not fully vanish with the availability of health insurance and many are still forced to sell assets or go into debt. Two issues have to be mentioned:(a) health insurance is hardly all-inclusive but limited to a certain number of illnesses and forms of treatment; those not included still have to be paid out-of-pocket;(b) the survey asked categorically about the sources of funding and not about the magnitude of the amounts. If the scale of asset sale or loans taken out had been included, we may have found insurance card holders reporting lower levels of burdensome funding than non-card holders. Insurance membership may thus reduce liabilities from health treatment even more than our data show. 3.2 FINANCING MEDICAL CARE: THE URBAN–RURAL DIVIDE If respondents are grouped according to urban or rural residence the modes of payment for medical treatment vary (see Figure 3.3). Discrepancies can be summarized as follows: – In Zambia and Kenya, there are more rural than urban dwellers who benefit from medical treatment at no cost. While free treatment schemes established by the government may not give preference to rural people, NGO projects and confession-based initiatives may do so. – With the exception of Côte d’Ivoire, urban dwellers have a higher share in financing medical expenditures from own savings. The reason may be higher cash income and closer integration into the cash economy. – Mobilizing funds from burdensome sources concerns smaller groups in both urban and rural Zambia but in other countries both segments of the population are strongly affected. The sources for mobilizing discriminatory funds differ strongly, however. Patients from rural areas are more involved in selling assets, while patients from urban locations have a higher tendency to go for a cash loan. The key differences between urban and rural residents in settling their health bills appear to be linked to the mode of production. A majority of rural dwellers are peasants who own some land, farming equipment or agricultural produce which they may mortgage or sell to obtain cash for medical treatment, while urban residents have less property and are forced to go into debt when the need arises for medical treatment. Both modes of mobilizing funds put a financial strain on patients afterwards. With our data we are not able to follow up on the negative effects and conclude whether rural people have more aftermath stress than urban people do. However, our data clearly show that the high level of financing health costs from the sale of assets and obtaining loans is a heavy financial burden for both groups. Selling productive assets or plunging into debt to pay for medical treatment becomes an economic liability for the future and may chain households to a cycle of poverty. Are these threats contingent on whether paying for medical treatment involves a male or a female-led household? The breakdown of the sample into male and female-led households is shown in Figure 3.4. No significant differences are discernible in Senegal and Côte d’Ivoire. Zambia, Kenya and Ethiopia, however, appear to provide better access for female-led households to free medical care and make them to rely less on health funding from asset sales or loans. Our findings point to the existence of programmes and policies in three out of five countries which are aimed at reducing out-of-pocket payments and are more beneficial for female-led households. Confirmation of such effects would require a more detailed analysis than is possible here, however. 3.4 FINANCING MEDICAL CARE: THE INCOME FACTOR The level of income may be linked to the various modes of financing medical care in different ways. If medical care is provided free of charge, it may or may not take income level into account. If the scheme has universal coverage, however, households with higher incomes have the same entitlement as the poor. Where such schemes do not exist, patients from better-off households are more likely to visit medical facilities with insurance cover and have their bills paid. On the other hand, the better-off may opt against free services because of quality concerns and go for more elaborate medical care, for which they have to pay. The link between income level and the use of own savings is more straightforward. Households with higher per capita incomes have a higher capacity to save and are more likely to be able to pay for medical services from their own pocket. The same can be said about the use of incriminatory forms of financing. When earnings go up, the need to sell assets or go for a loan declines. The results are shown in Figure 3.5. To identify the effects of income inequality on the modes of payment for medical care, respondents are grouped into four income levels, which are defined as multiples of the statutory minimum wage. 1 No link is discernible between income inequality and access to free services, with the exception of Zambia. As free primary health care in Zambia has a universal character and does not prevent the better-off from accessing it, many appear to disregard it. We can therefore assume that patients with 1 Ethiopia has no statutory minimum wage. We used the 1,500 Birr as income reference and established similar income boundaries. 21

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 3.3 Payment for medical treatment by sources of finance and urban–rural residence 60% 50% 40% 30% 20% 10% 0% 2.7% 1.3% Urban Rural Senegal 42.0% 54.0% No Cost was Charged 16.6% 22.6% Urban Rural Zambia Urban Rural Kenya 7.5% 6.1% Urban Rural Côte d’Ivoire 80% 60% 40% 20% 0% 46.0% 37.3% Urban Rural Senegal 30.7% 22.8% Urban Rural Zambia Own Savings 38.5% 31.9% Urban Rural Kenya 56.8% 57.4% Urban Rural Côte d’Ivoire 30% 25% 20% 15% 10% 5% 0% 25.8% 8.2% Urban Rural Senegal 2.8% 4.8% Urban Rural Zambia Sale of Assets 6.5% 15.1% Urban Rural Kenya 0.8% 1.1% Urban Rural Côte d’Ivoire 35% 30% 25% 20% 15% 10% 5% 0% . 33.0% 28.4% Urban Rural Senegal 11.8% 5.8% Urban Rural Zambia Taking Loan 26.0% 23.8% 28.4% 15.3% Urban Rural Kenya Urban Rural Côte d’Ivoire 15.1% 12.6% Urban Rural Ethiopia 44.0% 36.9% Urban Rural Ethiopia 23.9% 8.2% Urban Rural Ethiopia 20.8% 21.8% Urban Rural Ethiopia higher incomes who are not satisfied with the quality of free medical services opt for other forms of treatment, for which they have to pay. In all survey countries, households use more of their own savings if they have more financial resources at their disposal and belong to a higher income class. The opposite effect can be observed for external funding. 22

FINANCIAL RISK PROTECTION – HOW DO THE INFORMALLY EMPLOYED PAY FOR MEDICAL TREATMENT? Figure 3.4 Sources of financing medical care by gender 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 48.5% 40.8% 1.9% Male 47.1% 42.4% 2.1% Female 15.0% 27.5% 45.0% Male 9.9% 24.7% 52.7% Female Senegal Zambia 33.6% 28.6% 35.7% 32.1% 16.4% 25.1% Male Female Kenya 28.2% 24.5% 57.1% 56.9% 6.0% Male 8.8% Female Côte d’Ivoire No Costs Savings Sale of possession or taking loan(all forms) Note: No data available for Benin. 44.2% 38.9% 40.0% 33.9% 10.5% 19.1% Male Female Ethiopia Figure 3.5 Modes of payments for medical treatment by income classes 60% 40% 20% 0% 2.4% 53.3% 2.1% 37.1% No cost 19.0% 30.2% 4.9% 8.2% Senegal Less 0.5 MW Zambia Kenya 0.5 MW–1 MW Côte d’Ivoire >1 MW–2 MW 15.0% 9.4% Ethiopia >2 MW 80% 60% 40% 36.8% 20% 54.1% 21.0% Savings 40.0% 30.4% 44.2% 50.0% 59.6% 33.7% 52.9% 0% Senegal Zambia Kenya Côte d’Ivoire Less 0.5 MW 0.5 MW–1 MW>1 MW–2 MW Ethiopia >2 MW 100% 50% 52.6% 35.2% 11.8% Sale of assets or taking loan 20.0% 32.0% 18.6% 33.6% 23.3% 46.6% 34.1% 0% Senegal Zambia Kenya Côte d’Ivoire Ethiopia . Less 0.5 MW 0.5 MW–1 MW>1 MW–2 MW>2 MW Note: Income means household income; that is, the incomes of various household members are added together. Each of the four income groups refers to the statutory minimum wage(MW). For Ethiopia with no statutory minimum wage, the MW reference is 1,500 Birr. 23

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 3.6 Use of medical care if funding has to come from asset sales or going into debt 70% 60% 50% 40% 30% 20% 10% 35.8% 59.8% 12.1% Sale of possesion or loan(any form) 45.5% 14.3% 27.1% 20.0% 37.3% 0% Regular Low use Regular Low use Regular Low use Regular Low use Use Use Use Use Senegal Zambia Kenya Côte d’Ivoire 40.1% 47.6% Regular Use Low use Ethiopia Income inequality thus works as expected: well-off households are better able to pay their health care bills from available savings, while poorer households are more often forced to sell property or go into debt. The negative effects are barely discernible for Zambia, however, where free primary health care clearly provides a shield against more burdensome forms of funding. Where there is no free medical treatment and medical bills cannot be paid from savings but only through either the sale of assets or from taking out a loan, the general reaction is predictable. People try to avoid medical treatment. This correlation is highlighted in Figure 3.6. With the exception of Zambia, where risky forms of financing medical treatment are less needed because of the free health care system, all countries manifest this link. If financial means have to be mobilised by selling assets or taking out loans, people tend to stay away from medical facilities and the use of medical care declines. In Senegal, 60 per cent of respondents who have to sell assets or go into debt declare that they rarely seek medical treatment. Kenya and Ethiopia report similarly high incidences of health service»boycotts« as a result of low incomes. 3.5 SUMMARY Our data provide insights into the self-assessed availability of medical care and the financial consequences resulting from having to pay health care bills. People who are forced to sell property or go into debt in order to source funds for medical treatment will avoid visits to medical services whenever possible to prevent dire financial consequences. With the exception of Zambia, where risky sources are hardly used for covering medical bills, there is a clear negative interrelation between the capacity to pay with one’s own monetary reserves and the frequency with which medical facilities are visited. If dependence on external financing(asset sales and loans) goes up, the use of medical services goes down. Quite clearly, many poor people are forced to go without treatment not because medical services are not available but because they do not have the means to pay for them. REFERENCES WHO(website) Financial protection; available at: https://www.who.int/health-topics/financial-protection#tab=tab_1. – Global health expenditure database; available at: https://apps.who.int /nha /database/ ViewData/ Indicators/en We have no documentation on how many times the informally employed or their family members were sick; how many times they visited medical centres; how serious the sickness was in each case; whether the kind of treatment actually received corresponded to the treatment needed; or whether patients»opted« for simple medical care because it was provided free instead of going for surgery and intensive but costly medical treatment. Such information would be needed to enable us to carry out a fully informed evaluation of the availability of quality medical care for all. Furthermore, we did not collect information about the level of health bills and how medical expenditures are related to income. Such information, again, would be needed for a detailed evaluation of the affordability of quality medical care. 24

Health insurance coverage and non-coverage 4 HEALTH INSURANCE COVERAGE AND NON-COVERAGE WILLINGNESS TO JOIN AND PAY FOR A HEALTH INSURANCE SCHEME – REASONS FOR NOT JOINING Florence Bonnet 4.1 INTRODUCTION Despite laudable progress globally, barriers to access health care remain in the form of out-of-pocket payments on health services, physical distance, range, quality and cultural acceptability of services provided, long waiting times, as well as opportunity costs, such as income loss. Collective financing, broad risk pooling and rights-based entitlements are key conditions to ensure effective access to health care for all. The principles provided by international social security standards on medical care are essential on the road to universal health coverage(UHC)(ILO, 2021a). Across countries there are a variety of mechanisms to this end: national health insurance and social health insurance mandated by the state; national health-care services guaranteed free of charge or for a small, regulated user fee; and targeted social assistance for health services(user fee waivers, vouchers, and so on), which may be categorical or be means-tested. The focus of the first section is financial protection against health-care costs through social health insurance among workers 1 in informal employment. 2 The second section presents a broad overview of coverage taking into consideration the multiplicity of mechanisms before focusing on the extent of social health insurance(SHI) 3 and private health insurance (PHI) coverage among informal workers. The third section focuses on the majority of workers in informal employment, who are not affiliated to a social health insurance scheme. It assesses their willingness to join such a scheme, and the preferred frequency and level of contribution they would be ready to pay. For those not interested in joining, it looks at the main reasons for this choice. The chapter closes with some concluding remarks. The accessibility, availability and quality of health care services form part of UHC and also influence the level of trust in formal institutions and, associated with that, the willingness to join of those not yet covered. Being covered by social insurance or by public tax-financed health services does not necessarily mean that one has effective access to quality health care services without hardship. The findings presented below should be considered in view of the findings presented in Chapter 2 concerning the need for and use of health services. 4.2 SOCIAL HEALTH INSURANCE COVERAGE AMONG WORKERS IN INFORMAL EMPLOYMENT 4.2.1 Financial health protection against health care costs ILO standards allow for a plurality of approaches to ensure effective access to medical care. 4 They recognize the diversity of arrangements that can exist for the financing, purchasing and provision of health care – social health insurance, national health service or a combination of such models – as long as they comply with key principles(ILO, 2020). In practice, most countries use a combination of such mechanisms to extend coverage. The survey carried out in the six countries provides an overview of membership in social health insurance or in private health insurance, including micro-insurance schemes. This will be the focus of the following sections. For a more comprehensive picture of health coverage in terms of health care costs, however, a rough estimate of the coverage of workers in informal employment by tax-financed(free) national health services(NHS) is added. 5 Furthermore, the 1 In line with statistical practices, the term»worker« is used here to refer to all people in employment: employees, employers(independent workers with employees), own-account workers(independent workers without employees) and contributing family workers. 2 Workers in informal employment, together with their families, represent the majority of the population in the six countries under review. The percentage of workers in informal employment ranges from 69 per cent in Zambia to 90 per cent or more in Benin, Senegal or Côte d’Ivoire(ILO, 2018). 3 The term»social health insurance«(SHI) corresponds to public health insurance schemes. 4 See ILO Medical Care Recommendation No. 69, 1944 and Recommen­dation No. 202, 2012. 5 Access to free medical care is only partially assessed through the available questions. It considers situations in which, during the past 12 months, at least one household member in need of medical care went for treatment and obtained it free of charge. It is based on two questions: Q49. In the last 12 months have you or has any member of your household received regular benefits in cash or in kind?, ­ selecting those who answered(f) free medical services; and Q57 How did you or your family find the money to pay for this treatment? No fee was charged/ paid by another source. The estimates provided in Figure 1 refer to the selected worker in informal employment and should be understood as:»workers living in households where at least one household member in need of medical care over the last 12 months had access to free medical care«, on the assumption that this free ­access could benefit them as well. 25

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 4.1 Membership of health insurance schemes, free access to national health services and direct financing by the employer % workers in informal employment 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% . 44.3% 1.9% 0.8% 40.2% 1.3% Zambia 2.3% 1.2% 1.1% Benin** 19.3% 4.6% 2.8% 6.5% 5.3% Senegal 22.2% 3.9% 2.9% 8.0% 7.3% Côte d’Ivoire 32.1% 3.5% 6.5% 22.1% Kenya 37.7% 1.0% 2.0% Employer pays for health expenditure* PHI- Private health insurance(including micro-insurance) NHS- National Health Service (tax-financed) SHI- Social(public) health insurance 34.8% Sum of SHI+ PHI+ NHS (Financial health coverage against health care costs) Ethiopia Notes: Proportions can be added as each additional mechanism is assessed to take into consideration workers not yet covered by other arrangements to avoid any double counting. The assessment of coverage by(tax-financed) national health services should be read as»workers in informal employment living in households where at least one household member, in need of health care, had access to free medical care«. * The question on direct payment of health expenditure by the employer was not available in Benin and Kenya and not usable in Ethiopia. ** No additional coverage by national health services was captured in Benin. direct payment of health expenditure by the employer is also considered for employees who do not benefit from any of the abovementioned public or private mechanisms. Taking these multiple mechanisms into account – social health insurance, private health insurance and national health services – Zambia ranks first with an estimated 44.5 per cent of workers in informal employment being either covered by social health insurance(1.3 per cent), accessing health care services free of charge(40.2 per cent) or being covered by private health insurance(one per cent) or another private mechanism, namely the direct payment of health expenditure by their employer(1.9 per cent). Taking all mechanisms into consideration leads to proportions that range from around one in five informally employed workers benefiting from some forms of health protection in Senegal and Côte d’Ivoire, to 32 to 38 per cent in Kenya and Ethiopia, and the abovementioned, 44.3 per cent in Zambia(see Figure 4.1). 6 4.2.2 Coverage by social health insurance(SHI) of workers in informal employment Focusing on social health insurance coverage, the six countries covered reveal a critical social health insurance coverage gap for the majority of workers in informal employment. 7 The proportions of workers in informal employment who are members of social health insurance schemes range from 1.1 to 1.3 per cent in Benin and Zambia, to five to seven per cent in Côte d’Ivoire and Senegal, and to as much as 22.1 per 6 The three countries identified in Chapter 2 as showing the most ­extensive use of health care services when needed are also the countries where the extent of health coverage(social insurance or taxbased) is the highest: Zambia, Ethiopia and Kenya. 7 This chapter draws on Traub-Merz, Öhm(2021: Section 6). cent in Kenya and 34.8 per cent in Ethiopia(Figure 4.1). The share of workers in informal employment affiliated to social health insurance tends to be higher among women than among men(Table 4.A2 in Annex); and with the exception of Ethiopia, to be higher in urban than in rural areas(Table 4.A3 in Annex). Informal workers with higher levels of education and higher individual labour income are also more likely to be covered than others(Tables 4.A4 and A5). Country schemes As far as affiliation to social health insurance is concerned, Kenya and Ethiopia are likely to be the runaway two countries among the six countries reviewed. The Kenyan government has made a commitment to achieve universal health coverage by 2022 based on the opening up of the long-standing National Health Insurance Fund(NHIF) to v­ oluntary affiliation with proactive measures to enrol workers operating in the informal economy(Health Economics Research Unit 2019). The country’s strong political commitment to universal health coverage is embodied in the government’s big four agenda, which includes health care for all as one of the key development priorities. The fairly high number of insured persons may be an additional factor helping the dissemination of information on joining the scheme and the procedures for doing so(Traub-Merz/Öhm 2021; S­ chwettmann 2022). The Community-based Health Insurance(CBHI) scheme was introduced in Ethiopia in 2011, aimed at people who work in the informal sector in both urban and rural areas. Like in Rwanda, this is a government-led programme with a national public agency in charge of it(the Ethiopian Health 26

Health insurance coverage and non-coverage Insurance Agency(EHIA)). 8 Membership of the CBHI is voluntary and enrolment is at the household level. The CBHI scheme started as pilot in 13 woredas and four regions back in 2010 to 2014 and has expanded substantially. In 2018/19 the programme covered 4.9 million households. Contributions from members are complemented by allocations by local government to cover the contribution for the 10 per cent of indigent members participating in the scheme. The EHIA has developed a second Health Insurance Strategic Plan(HISP II(2020/21-2024/25)) hoping to provide universal health coverage(UHC) for all by 2030. Over the five years of the plan, HISP II aims to reach 80 per cent of the population in the informal sector(ILO, 2021b). Universal health coverage(Couverture Maladie universelle (CMU)) in Côte d’Ivoire officially entered its operational phase on 1.1.2020. This is too recent to enable us to capture significant results regarding its effective implementation through the survey. The CMU is a compulsory national health insurance system for the population, starting with the most disadvantaged groups. Its objective is to guarantee all Ivorians residing in Côte d‘Ivoire access to quality health care services at low cost. Two schemes have been set up: a basic general contributory scheme(RGB), at the rate of 1,000 CFA francs per month per person and a non-contributory medical assistance scheme(RAM) for the poor, who until now have been mostly excluded from the health care system. At the time of the survey, after a three-year pilot phase, 1.55 million people out of the estimated population of 24 million, were enrolled in the CMU in October 2019, including below 200,000 poor people. And while the distribution of cards at enrolment sites began in February 2019, fewer than 632,000 had been distributed by the end of 2019(Jeune Afrique, 2020). In Senegal, extension to the informal economy is under way but was still moving slowly at the time of the survey. Health care coverage in Senegal is provided through a combination of social health insurance and social assistance mechanisms aimed at different categories of the population, including workers in the informal economy. The CMU is not yet legally anchored and currently operates on the basis of a programme approach, with the Agence nationale de la CMU (ACMU) 9 supervising the CMU’s insurance and assistance programmes. Participation in the scheme is based on a contribution of 7,000 CFA francs per year, subsidized at 50 per cent by the state. The CMU Agency had 1.38 million beneficiaries at 3,500 CFA francs per year in 2018, but only 8 EHIA was established in 2010 as an autonomous federal organ through Regulation No. 191/2010. The agency has a vision of seeing all citizens of the country covered by equitable and sustainable health insurance system. Both SHI and CBHI schemes are managed by EHIA but implemented independently. For additional information see: https://ehia.gov.et/Vision-Mission-Objective. 9 Created by Decree No. 2015-21 of 7 January 2015. A draft law on the CMU has been under consideration for three years. This law should define the modalities for implementing the CMU’s insurance and assistance systems, the set-up of a general scheme that would cover 80 per cent of health expenses and would be the basic scheme for all Senegalese. 734,000 were up to date with their contributions and are – theoretically – covered(ILO, 2021c). Differences in social health insurance coverage between countries result from a number of factors, starting from the health financial protection mechanisms in place. It reflects, at the time of the survey, the existence(Côte d’Ivoire, Ethiopia, Kenya, Senegal) or not(Benin, Zambia) of an established national social health insurance scheme, which may be complemented(or not) by an alternative mechanism and, in particular, by a national health service financed through taxes, providing access to free primary medical care, as in Zambia(as shown in Figure 4.1). 10 Other factors related to the use of the health care system include differences in the availability, accessibility and quality of health care services for all. A second category of factors refer to the intrinsic characteristics, awareness and perceptions of individuals as users of health services and potential members of social health insurance schemes, including individuals’ ability to pay their contributions, their awareness or understanding of what an insurance scheme is and how it operates; and the level of trust in institutions and social insurance schemes. Finally, the focus on workers in informal employment explains the results regarding the importance of the gap of social health insurance coverage. They are, by definition, affected disproportionately by coverage gaps, being»not or insufficiently covered – in law or in practice – by formal arrangements« (ILO, 2015), the latter referring to both protections and some obligations. 11 4.2.3 Coverage of workers in informal employment by private health insurance (PHI) The proportion of workers in informal employment covered by private health insurance(PHI) is still limited(ranging from around one per cent in Benin, Ethiopia and Zambia to 2.8 to 3.5 per cent in Côte d’Ivoire, Kenya and Senegal; see Figure 4.1). Even combining social health insurance coverage with private health insurance(including micro insurance but without direct payments by the employer), the proportion of informally employed workers covered increases only slightly, reaching two per cent in Benin and Zambia, eight to 10 per cent in Senegal and Côte d’Ivoire, 25 per cent in Kenya and 10 See Universal Health Coverage Partnership (https://www.uhcpartnership.net/country-profile/zambia/) and ­Zambia’s Ministry of Health(https://www.moh.gov.zm/?p=6229). 11 The affiliation to health insurance(a fortiori voluntary affiliation) was not retained as part of the criteria for defining informal wage employment. In the six countries, in line with international statistical standards on informality, the operational criteria used to define informal employment differ for employees, independent workers or contributing family workers. Employees are considered to have an informal main job, if their employer does not pay contributions(at least partially on their behalf) to the pension scheme. The payment of contributions by the employer to the health insurance scheme is not considered a criterion that defines whether employees have a formal job, even if mandatory. Independent workers(with or without employees) are in informal employment if their business(activity) is not registered with the national competent authority(the National Register of Companies in Kenya, the Registry of Commerce and a NINEA number in Senegal). Finally, all contributing family workers are considered, by definition, to be in informal employment. 27

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows 35.7 per cent in Ethiopia. This is without including direct health expenditure by the employer, which plays a minor role but provides some kind of protection to four to five per cent of workers in informal employment in Senegal and Côte d’Ivoire, 12 although without any guarantees and of course outside the scope of social health protection. Direct support from employers covers primarily health expenditure related to work injury and more rarely, non-job-related health expenditure. 4.2.4 Main features of affiliation Regarding the minority of workers in informal employment who are members of a social health insurance scheme or have private health insurance, their affiliation is more often than not voluntary(Figure 4.2.1). Nearly all affiliated informally employed members in Ethiopia(99 per cent) and 65 per cent or more in other countries are affiliated on a voluntary basis. For most workers in informal employment who report being affiliated on a mandatory basis, the affiliation is actually indirect, benefiting from the affiliation of another member of the household. Irrespective of the voluntary or mandatory nature of the affiliation, indirect affiliation is the most prominent form of affiliation of workers in informal employment to health insurance(considering social health insurance and public health insurance together). Indirect affiliation concerns 40 per cent in Kenya and as much as 93 per cent in Ethiopia(Figure 4.2.2). 12 This type of support could not be assessed in Benin, Kenya and ­Ethiopia owing to the lack of the corresponding question or sufficient answers. The incidence of indirect affiliation highlights the importance of the household dimension. The distinction between informally employed workers living in»fully informal households«(all household members employed being in informal employment) and those living in households with at least one household member who is formally employed(»mixed households«) points to higher affiliation rates among those living in»mixed households«, benefitting from indirect coverage(Figure 4.3.1). Considering the same indicator of social health insurance and private health insurance coverage, with the exception of Ethiopia, coverage among workers in informal employment is higher among those living in households some of whose active members hold a formal job. Considering only coverage of workers in informal employment by social health insurance, the proportion is at least two to three times higher for those living in households with at least one member in formal employment. However, the proportion of households concerned and the overall effect of health insurance coverage of workers in informal employment remain limited. Including the household dimension as part of the analysis points to characteristics of the households of informal workers that may mitigate, or accentuate, the impacts of some of the vulnerabilities associated with informal employment, including levels of insecurity of income(level, stability, assets), with social protection playing an important role(OECD-ILO 2019). The proportion of informal workers affiliated with social health insurance or private health insurance schemes is positively associated with levels of household income(Figure 4.3.2) or household assets. Figure 4.2 Distribution of workers in informal employment who are members of health insurance schemes according to: 4.2.1 The voluntary or mandatory nature of enrolment 4.2.2. Direct or indirect coverage(through) another household member Eth i op i a(35% coverage) 1% 99% Eth i op i a(35% coverage) 7% 93% Ke n ya(26% coverage) 29% Mandatory Voluntary Côte d’Ivo i re (10% coverage) 32% Se n ega l(8% coverage) 35 71% 68% 65% Ke n ya(26% coverage) Côte d’Ivo i re (10% coverage) 52% 40% Main contributor Another household member Other 30% 68% 8% 3% Se n ega l(8% coverage) 59% 33% 8% Be nin(2% coverage) 48% 52% Be nin(2% coverage) 26% 61% 13% 0 20 40 60 80 100% 0 20 40 60 80 100% . % by type of affiliation% by type of affiliation Note: These results should be considered in light of the proportion of workers who are members of a health insurance scheme as provided in Figure 4.1 and in brackets(close to country names). In Benin, for instance, these results concern only two per cent of all workers in informal employment. 28

Health insurance coverage and non-coverage Figure 4.3 The household dimension of health insurance coverage 4.3.1 Health insurance coverage by level of informality within the household: the importance of indirect affiliation 60% 4.3.2 Level of health insurance coverage by household income 80% 54.4% 35.8% 34.6% % workers in informal employment covered by health insurance 50% 5% 70% 40% 30% 20% 10% 0% 1.5% 13.5% 1.4% 9.2% 6.4% 19.7% 8.1% 30.2% 20.7% 1% 3% 16% 48% 3% 6% 35% 32% 1% 12% 6% 2% 13% 1% 14% 18% 1% 3% 4% 7% % informal covered 60% 50% 40% 30% 20% 10% Kenya Ethiopia Cote d’Ivoire Zambia Senegal Benin Fully informal HH Mixed HH Fully informal HH Mixed HH Fully informal HH Mixed HH Fully informal HH Mixed HH Fully informal HH Mixed HH Fully informal HH Mixed HH 0% Less than 1/2 MW 1/2 MW – Minimum wage [MW] MW – 2 MW 2 MW – 8 MW More than 8 MW Benin Zambia Senegal Côte d’Ivoire Kenya Ethiopia PHI: Private health insurance PHI: microinsurance SHI(public health insurance) Coverage(SHI+ PHI) . Note: The reference is the selected household member in informal employment, taking into consideration the composition of the household in which they are living(Figure 4.3.1) and household income(Figure 4.3.2). In Figure 4.3.1, the selected worker in informal employment is in a »fully informal household« if all other household members in employment are also in informal employment. They are in a»mixed household« if at least one household member is in formal employment. 4.3 PERCEPTIONS OF AND WILLINGNESS TO JOIN A SOCIAL HEALTH INSURANCE SCHEME This section looks at the majority of workers in informal employment who are not yet members of a social health insurance scheme – ranging from 64 per cent in Ethiopia to 98 per cent in Zambia and Benin – and assesses their interest in joining a health insurance scheme, the preferred modalities in terms of frequency of payment of contributions, and the amount they would be ready to pay. Those not interested in joining were asked directly for their main reason. It should be noted that some respondents may not have been aware of the concept of health insurance. 4.3.1 Willingness to join Among workers in informal employment who are not covered, a clear majority expressed an interest in joining a social health insurance scheme. The proportion ranges from 56 to 58 per cent in Benin and Zambia to 71 to 74 per cent in ­Ethiopia, Kenya and Senegal, and 81 per cent in Côte d’Ivoire. There is no significant difference between women and men (Table 4.A6 in Annex), while living in urban or rural areas is relevant in three countries(Kenya, Senegal, Côte d’Ivoire) but plays no role in the others(Benin, Ethiopia, Zambia). By contrast, the capacity to pay contributions obviously influences people’s willingness to join a health insurance scheme, as can be shown if individual or household income is considered(Figures 4.4.2 and 4.4.3). This positive correlation affects all six countries, though with some reservations among informal workers at the upper levels of income, especially in Senegal and Ethiopia, where a lack of trust prevails. 13 To a lesser extent, the willingness to contribute also increases in most of the six countries with the level of education(Table 4.A7). The level of education is often associated with higher income, but possibly also with a higher level of awareness about rights, obligations and services. Finally, employers seem more likely to be interested in joining a health insurance scheme than own-account workers and employees(Table 4.A8). Whether it is for personal and family purposes or whether their interest in covering their employees matters cannot be assessed here. 13 In those two countries the CMU and CBHI are associated with coverage of poor people, and therefore people in higher income quintiles, regardless of employment formality, may not have a positive perception of the schemes. 29

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 4.4 Willingness of workers in informal employment(non-members) to join a health insurance scheme 4.4.1 By urban–rural location Benin 55.9% 61.1% 49.9% Zambia 58.3% 51.7% 66.6% Kenya 70.6% 70.4% 71.1% Ethiopia 72.9% 75.2% 65.9% Senegal 74.1% 74.4% 73.8% Côte d’Ivoire 80.6% 80.7% 80.5% 0 50 100% Total Rural Urban % of informal workers . interested to contribute % informal uncovered who wants to join Less than 1/2 MW 1/2 MW – MW MW– 2 MW More than 2 MW 4.4.2 Depending on individual labour income 90% 85% 80% 75% 70% 65% 60% Zambia Côte d’Ivoire Benin Kenya Ethiopia Senegal 55% 50% 45% Individual labour income % informal covered 4.4.3 Depending on household labour income 100% Kenya 90% 80% 70% 60% Côte d’Ivoire Benin Zambia Senegal Ethiopia 50% 40% Less 1/2 M 1/2 MW W–M wag in e im [M um MW W] – 2 MW 2 MW –8 MW More than 8 MW Household income 4.3.2 Reasons for not being willing to join Another way of assessing the factors influencing people’s willingness to pay contributions is to ask directly those who do not want to join a scheme for the main reason for their lack of interest or indifference. Awareness is a precondition for assessing the influence of other factors, such as poverty or mistrust, and is strongly related to knowledge of health insurance. In fact, only a minority lack awareness of how health insurance works and how to proceed: 15 per cent or less of informal workers not interested in joining a scheme in Senegal and Kenya, around 20 per cent in Zambia or Benin, and over 25 per cent in Ethiopia and Côte d’Ivoire. The lack of financial capacity is the main reason mentioned by those not interested in joining. This category is the largest group in Ethiopia(43 per cent) and the majority(more than half) in all the other five countries(53 to 55 per cent in Côte d’Ivoire and Senegal, 60 per cent in Benin and close to 70 per cent in Kenya and Zambia). Women consider the financial issue as the main obstacle even more than men(Figure 4.5.1) and so do less-educated informal workers, for example, in Benin, Ethiopia and Senegal(Figure 4.5.3), and employees, contributing family workers and own-account workers more than employers(Table 4.A9 in Annex). The differences by area of residence do not reveal any consistent pattern across countries(Figure 4.5.2). 14 Obviously, the lack of financial capacity is mentioned as the main reason by those with lower levels of labour income. Looking at the average in the six countries: 66 per cent of informal workers earning less than 14 The absence of urban-rural differences in Zambia and Kenya to some extent confirms the absence of significant differences in access to health care depending on the area of residence, as shown in Chapter 2. half of the minimum wage mention financial barriers as the main reason for not joining compared with less than 30 per cent among those earning more than twice the minimum wage(Figure 4.6.1). This sheds light on the importance of carefully designed policies to expand coverage to the informal economy, and to ensure that one does not overestimate contributory capacities when it comes to contributory schemes. What is required is an in-depth understanding of the diversity of situations in the informal economy, and accordingly of realistic options. Lack of trust in health insurance schemes is mentioned by less than 10 per cent of respondents in Zambia, but by as many as 24 to 26 per cent in Senegal and Ethiopia, and 15 to 19 per cent in Kenya, Côte d’Ivoire and Benin. Trust declines as the level of education increases(Figure 4.5.3). In Benin, Ethiopia and Senegal, a lack of trust in health insurance schemes is mentioned by more than 40 per cent of those with at least a secondary level of education. Mistrust also tends to gain importance with an increase in the level of income(Figure 4.6.2). If the lack of awareness is primarily a matter of information and education(confirmed by data from Benin, Côte d’Ivoire, Kenya and Zambia, as shown in Figure 4.5.3), then more transparency and well-designed public information campaigns to ensure that informal workers are aware of their rights and entitlements could reduce this group significantly (Traub-Merz/Öhm 2021). Poverty and mistrust appear to be the main obstacles to extension of the coverage of social health insurance schemes. In an effort to combat poverty what is needed is a combination of financial mechanisms relying on non-contributory schemes and/or partial or full subsidization of social insurance contributions for low-­ income groups. Building trust calls for institutional action to 30

Health insurance coverage and non-coverage enhance transparency and accountability, hand in hand with actions on the part of citizens to enhance awareness and the culture of social security, including health and a culture of formality(ILO, 2021d). What is required is a combination of approaches to address the prevalent causes of the gaps identified for different groups. Thus, the assessment and understanding of this diversity of needs calls for different solutions. Figure 4.5 Reasons not to join a health insurance scheme 4.5.1 By sex Senegal Zambia Kenya Ethiopia Total 68% 9% 19% 4% Women 71% % 17% 3% Men 65% 9% 23% 4% Total 55% 24% 14% 7% Women 57% 20% 18% 5% Men 53% 27% 11% 9% Total 69% 15% 14% 2% Women 67% 14% 17% 3% Men 70% 17% 11%2% Total 43% 26% 26% 6% Women 51% 21% 20% 8% Men 37% 29% 30% 4% Total 53% 17% 27% 3% Women 51% 20% 29% 1% Men 56% 14% 26% 5% Total 60% 19% 21% 1% Women 64% 14% 21% 1% Men 54% 25% 21% 1% 0 20 40 60 80 100% % workers in informal employment not willing to join Côte d’Ivoire Benin Insufficient financial resources Don’t trust health insurance schemes Don’t know anything about how they work . Other 4.5.3 By level of education Zambia Secondary+ Primary None 74% 67% 70% 8% 13% 6% 9% 19% 4% 7% 23% Senegal Secondary+ Primary None 29% 52% 64% 41% 18% 12% 27% 12% 8% 15% 16% 4% Kenya Secondary+ Primary None 70% 68% 67% 24% 5%2% 15% 14% 3% 3% 31% Ethiopia Secondary+ Primary None 27% 36% 57% 44% 26% 22% 7% 31% 7% 19% 20% 4% Côte d’Ivoire Secondary+ Primary None 69% 52% 52% 19% 14% 23% 26% 32% 8% 4% 2% Benin Secondary+ Primary None 44% 61% 62% 41% 21% 12% 15% 18% 26% 1% 0 20 40 60 80 100% % workers in informal employment not willing to join Insufficient financial resources Don’t trust health insurance schemes Don’t know anything about how they work . Other 4.5.2 By rural-urban Rural Urban 68% 70% 8% 21% 4% 11% 16% 3% Zambia Senegal Rural Urban 61% 49% 18% 17% 4% 31% 11% 10% Kenya Rural Urban 68% 69% 15% 15% 2% 17% 12%3% Côte d’Ivoire Ethiopia Rural Urban 48% 33% 23% 30% 23% 5% 30% 7% Rural Urban 50% 56% 16% 18% 33% 1% 22% 5% Benin Rural 55% 16% 28% 1% Urban 64% 21% 15% 0% 0 20 40 60 80 100% % workers in informal employment not willing to join Insufficient financial resources Don’t trust health insurance schemes Don’t know anything about how they work . Other 31

% informal who don't want to join: Lack of resources Less 1/2 MW 1/2 MW – w M age [M inimum W] MW – 2 MW More than 2 MW % informal who don't want to join: Lack of trust Less 1/2 MW 1/2 MW – w M age [M inimum W] MW – 2 MW More than 2 MW FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 4.6 Main reasons for not joining a health insurance scheme by level of income 4.6.1 Lack of resources 80% 70% 60% 50% 40% 30% 20% 10% 0% Zambia Kenya Senegal Ethiopia Average (6 countries) Benin 4.6.2 Lack of trust 80% 70% 60% 50% 40% 30% 20% 10% 0% Benin Ethiopia Côte d’Ivoire Kenya Senegal Zambia % informal who don’t want to join: Lack of awareness Less 1/2 MW 1/2 MW – w M age [M inimum W] MW – 2 MW More than 2 MW . Individual labour income . 4.6.3 Lack of awareness 60% Côte d’Ivoire 50% 40% 30% 20% 10% Zambia Benin Ethiopia Kenya Senegal 0% Individual labour income . Individual labour income 4.3.3 Preferred modalities of payment among those expressing an interest in joining Returning to those willing to join and their preferred modalities regarding the frequency of payment and the level of contributions they would be willing to pay, the survey highlights some common trends. 4.3.3.1 Preferred frequency of contribution payments In line with the irregular and unpredictable income that charact­erizes most workers in informal employment – with the exception of Kenya – close to or more than half of informal workers(who are not yet members of a health insurance scheme) would prefer to pay contributions once a year. The proportion reaches 42 to 47 per cent in Benin and Côte d’Ivoire, 52–59 per cent in Zambia and Senegal and 79 per cent in Ethiopia. The strong preference for payments that are not monthly confirms the need for flexibility of legal 32

Health insurance coverage and non-coverage Figure 4.7 Preferred frequency of contributions among those willing to join a health insurance scheme Zambia 52% 12% 10% 25% Senegal 59% 17% 6% 18% Kenya 23% 12% 13% 49% Ethiopia 79% 14% 6% 1% Côte d'Ivoire 42% 10% 9% 37% Benin 47% 0 10 14% 16% 20 30 40 50 60 70 80 % informal non-member of SHI but willing to join Once per year Once a month Once every six months Once a week Once every three months Daily 22% 90 100% arrangements compared with monthly payments, which usually apply to formal regular employees. The payment of contributions once or twice a year is the preferred option, in particular in rural areas(especially in Zambia and Côte d’Ivoire) and among informal workers living in»fully informal households«. 4.3.3.2 Level of contribution workers in informal employment would be ready to pay The findings on the level of contribution informal workers would be ready to pay to join a social health insurance scheme are based on a direct question calling for a spontaneous response, without reference to a particular package of covered health benefits. These amounts should therefore be taken as a rough estimate of the level considered reasonable. In the context of this multi-country survey, the objective is first to confirm that people are willing to join a health scheme, even if it comes at a cost, and second to establish a vague level of contribution that respondents would be ready to pay. With the exception of a few individuals who declared an unwillingness to pay anything, all respondents provided an amount that they were willing to contribute. Despite the limitations associated with this method, once converted into a common currency, there is little variation among the six countries in the proposed per capita amounts. On average, respondents were prepared to pay the following monthly contributions: Benin, 628 CFA francs; Côte d’Ivoire, 1,171 CFA francs; Ethiopia, 32 Ethiopian Birr; Kenya, 235 Kenyan Shillings; Senegal, 967 CFA francs; and Zambia, 20 Zambian Kwachas. Converted into US dollars per month, workers in informal employment interested in joining a scheme would be ready to pay the equivalent of around one US dollar per month in Zambia, Benin and Ethiopia and up to around two dollars in Senegal, Kenya and Côte d’Ivoire(Figure 4.8). Considering the proportions of workers willing to join and ready to pay, between 8.2 per cent in Ethiopia and 30.2 per cent in Kenya would be ready to pay the equivalent of at least two US dollars per month per person, while between 1.3 per cent in Zambia and 9.7 per cent in Côte d’Ivoire would be ready to pay the equivalent of at least five US dollars per month. Acknowledging that such»declared amounts« do not refer to any specific package of benefits but rather to»an expected package« that may vary from one respondent to another, the comparison of»intended contribution levels« with the contributions of existing schemes enables a rough »reality test«. Such a comparison is provided below for three of the six countries: Benin, Kenya and Senegal. In Benin, the amounts charged by the three reference schemes(mutual health insurance; la Caisse Mutuelle de Prévoyance Sociale(CMPS); and the Régime d’Assurance Maladie Universelle(RAMU)) vary between 850 CFA francs and 1,200 CFA francs. On average, respondents in Benin are willing to pay 628 CFA francs, which corresponds on average to more than 50 per cent of the premiums of health protection schemes already available. However, this average value is shaped by a limited number of respondents who provide high values. In terms of effective distribution it appears that close to 20 per cent of the respondents would meet the full amount of the premium in CMPS or RAMU and over a third would be close to half the actual value of contributions in existing schemes. In Kenya, the NHIF has a detailed contribution scale for employees that starts at 150 Kenyan Shillings(monthly) for salaries of no more than 5,999 Kenyan Shillings and goes up to 1,700 Kenyan Shillings for salaries of 100,000 Kenyan Shillings or more. Contribution levels differ between employees and independent workers. The self-employed are not grouped into income classes but are charged a flat rate of 500 Kenyan Shillings. Applying the different criteria to respondents according to their level of income and em33

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Monthly amount of premium (USD current) % uncovered ready to pay at least X for a monthly premium (SHI) Figure 4.8 Level of contributions suggested by non-members in relation to various benchmarks 2.5 2.0 1.5 $0.91 1.0 14.9% 0.5 7.5% 1.3% 0.0 Zambia 20 ZMK/mth $1.78 $1.15 $1.15 8.2% 3.7% 2.0% Ethiopia 31 ETB/mth 13.8% 12.7% 4.0% Benin 628 FCFA/mth 18.3% 17.6% 8.9% Senegal 967 FCFA/mth $1.95 30.2% 22.8% 3.5% Kenya 235 KES/mth 50 $2.15 45 40 35 30 21.8% 25 20 19.1% 15 10 9.7% 5 0 Côte d'Ivoire 1171 FCFA/mth Average monthly premium(Current USD) Proportion ready to pay at least USD 3 /month Proportion ready to pay at least USD 2 /month Proportion ready to pay at least USD 5 /month ployment status, 20 to 25 per cent of workers in informal employment are not yet covered, but are interested in joining and would be ready to pay the full amount of the National Health Insurance Fund(NHIF). Furthermore, 32 to 40 per cent provide a spontaneous amount corresponding to half of the current value of contributions. In Senegal, 57 per cent of informally employed workers suggested an amount above the current annual contribution of 3,500 CFA francs 15 to join the CMU. The readiness and ability to pay a certain level of premium raises the issue, for the extension of coverage to employees, of the shared contribution of both employer and employee, complemented – or not – by the state. Employees are not supposed to cover the full cost of health insurance, including in the absence of state subsidies. 16 This is not the case for employees in formal employment, so there is no reason for those in informal wage employment to do so. It is essential that employers be involved and willing and able to pay, and that this be sustainable. Table 4.1 Values of contributions in existing schemes and com­ parison of»spontaneous amounts« provided in Benin BENIN Mutual health insurance Contribution Percentage ready to pay the full amount Percentage ready to pay half the amount 1,200 13.6 27.4 Caisse Mutuelle de Prévoyance Sociale (CMPS) 850 Régime d’Assurance Maladie Universelle (RAMU) 1,000 21.0 20.9 35.0 35.0 4.4 CONCLUDING REMARKS Workers in informal employment and their family members are disproportionately affected by health coverage gaps, even in relation to national health systems that provide access to free health care. On average, over the six countries, 14.5 per cent of workers in informal employment are members of a social health insurance scheme and an additional 1.9 per cent are covered by private health insurance, including micro insurance. Including access to free medical care, still more than 72 per cent of workers in informal employment are either not covered by health insurance or cannot access free medical care when needed. Informality is not desirable either for workers(no access to decent work), or for a country’s ability to generate sufficient revenues to actually redistribute wealth and counter inequalities. Facilitating the transition to formality should also be pursued as a means of ensuring better access to protection, including social health protection. The extension of social protection is a key element of gradual processes of formalization. It contributes – together with other measures – to ensure a basic level, stability and predictability of income and to improve productivity. By doing so, it helps to reduce exposure to poverty, enhance access to health care and enable the informally employed to seize economic opportunities(ILO, 2021d). In countries that 15 The contribution per beneficiary is 7,000 CFA francs per year, subsidised at 50 per cent by the state(ILO, 2021b). 16 See Social Security(Minimum Standards) Convention, 1952(No. 102), Art. 71 about cost sharing:»The cost of the benefits provided in compliance with this Convention and the cost of the administration of such benefits shall be borne collectively by way of insurance contributions or taxation or both in a manner which avoids hardship to persons of small means and takes into account the economic situation of the Member and of the classes of persons protected. The total of the insurance contributions borne by the employees protected shall not exceed 50 per cent of the total of the financial resources allocated to the protection of employees and their wives and children.« 34

Health insurance coverage and non-coverage have opted for a national social insurance system for all social protection benefits, extension should be designed so as to ensure that there is solidarity in financing across the whole population, and facilitating the transition to formality forms part of that effort. Some progress has been made in recent years to extend financial protection against healthcare costs, including in the six countries surveyed. Some is captured in the results (Ethiopia, Kenya and Zambia), while in other countries it was still too early to identify any significant expansion(Côte d’Ivoire 17 and, to a lesser extent, Benin and Senegal 18 ). The results presented in this chapter highlight the widely shared interest in joining a social health insurance scheme, but also some of the barriers faced by informal workers regarding access to social health insurance and national health services alike. Out of the majority of the informally employed who are not yet covered, close to 69 per cent on average express their willingness to contribute. With a few exceptions, they are ready to pay but most would need at least a partial subsidization of contributions. Importantly, results highlight differences in the main determinants of the gaps, calling for a variety of approaches, each adapted to the relevant country’s trajectory and national policy. In practice, countries that initially adopted a(tax-financed) national health service on attaining independence backtracked on it during the structural adjustment period of the 1980s and 1990s and started to introduce user fees. On that basis, considering the social unrest and adverse health outcomes this gave rise to, countries have adopted a series of measures(usually combining social health insurance limited to the formal sector, or even restricted to civil servants and the military, and free care programmes either only for the poor or categorical for maternity and infant care). These programmes have left large pockets of the population uncovered, especially the»missing middle«, notably those in the informal economy not covered by job-related health insurance and not poor enough to be targeted by social assistance. For those countries that have adopted an NHS model, whereby coverage is secured through tax-financed free care in public facilities, the present study offers limited findings. Only one of the survey countries(Zambia) falls into the category of tax-financed free primary health care. But even here, only 40 per cent of workers in informal employment live in households that had access to free services, and more has to be done to improve health service capacities at primary level and to include finance protection schemes for secondary health services. 19 For countries that have opted for a social health insurance model, an in-depth understanding of the informal economy is needed, as well as a differentiated approach to different groups, depending on their capacity to contribute and their willingness to do so. 20 The latter is influenced not only by financial capacity but also by awareness and trust in institutions. On the basis of capacity and willingness to contribute, three main groups can be identified: – Group 1. With no capacity to contribute. A first group is composed of workers and families without the capacity to contribute, at least in the short term. What matters is their access to health protection despite their inability to contribute and independently of their willingness to do so. For them, social health protection should come in the form of non-contributory health benefits provided through fully subsidized contributions paid by the state for all low-income earners. As for other groups mentioned below, those measures should be complemented notably by interventions to reduce decent work deficits and, beyond that, vulnerabilities in order to support the gradual process of transition to formality; to raise awareness of rights and entitlements using the right channels in terms of support for communication and means of conveying the message and reaching out to them; to address administrative and geographical barriers to registration and effective access to benefits; and, on the part of institutions, to enhance accountability and the effective availability, accessibility and adequacy of health services. – Group 2. With capacity to contribute and a willingness to join. This second group is composed of workers in informal employment with some capacity to contribute, even if only partially, and who expressed their willingness to join a social health insurance scheme. According to the data, on average 54 per cent of workers in informal employment willing to join earn at least half of the minimum wage and 24 per cent earn at least the minimum wage. Measures should be aimed at removing the obstacles to translating this willingness to contribute into effective coverage. Access to social insurance should be facilitated through the adaptation of modalities provided by law to join the scheme, including registration, paying contributions and accessing benefits and services. This includes partially subsidizing contributions for those with limited contributory capacities, who are willing to join but not able to cover the full cost, while maintaining incentives for formalization. It also includes adapting the frequency of contribution payments to their situations (for example, considering annual instead of monthly 17 Universal health coverage(UHC) officially entered its operational phase on 1 January 2020 in Côte d’Ivoire(Jeune Afrique 2020). 18 In Senegal, the UHC cover was an active programme at the time of the survey, but recently a specific focus on independent, own account and small enterprises in the trade and handicraft sector was enacted, enabling workers in the informal economy in those sectors to enrol (ILO, 2021b). 19 Zambia is currently considering establishing a social health insurance scheme for secondary health services, while maintaining the free cost approach for primary services. 20 For all workers in informal employment(and their families), the ­capacity to contribute and the willingness to do so supposes first that »any legal gaps« have been addressed, including extending legal entitlements to all workers and their families, independently of sector, type and duration of contract, thresholds in terms of size of enterprise, hours of work or earnings. Measures to support the extension of social health insurance should be complemented by measures to enhance the capacity to join in a sustainable way, including measures to enhance productivity(access to finance, markets, property, etc.). 35

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows income, allowing for greater flexibility regarding the scheduling of contribution payments), using broad contribution categories to determine earned income or alternative reference values(other than earnings)(ILO 2019, 2021e). Measures to support registration could also include collective registration or insurance agreements to facilitate administrative procedures by using a workers’ organization(such as a cooperative or rural producers’ association, or trade unions) as an intermediary between workers and the social security institution. This would allow, for instance, own-account workers (who still represent the majority of informal workers) to enter collective or group insurance agreements with a social insurance scheme, provided that they belong to an organization which has the capacity to be an effective partner in such an agreement(ILO 2019). Measures to address the lack of information – awareness raising about their rights and entitlements, and how to proceed to join health insurance schemes and obtain access to services and benefits should also cover this second group, together with measures to address geographical barriers that prevent access to benefits and services. 21 a»culture of formality«. It is probable that education and awareness-raising campaigns, which seek to change attitudes towards informal work, will be successful only if changes are made in formal institutions. For this last group, as for all other groups, reforming formal institutions and investing in client orientation and satisfaction are necessary to improve their accountability, transparency and good governance and thus to boost trust in institutions. Experiences from countries showed that solutions exist to extend social health protection in terms of coverage, adequacy and equity. Determinant factors include:(i) the respect of the ten Social Health Principles 22 throughout the design and institutional arrangements for health coverage(ILO, 2020); (ii) The adequacy of public financing levels and modalities, resulting from high level commitment to universal health coverage/ social health protection; and(iii) the capacities to operate social health protection systems in a cost effective manner, including linkages with broader contributory and non-contributory social protection systems and other policies, including facilitating transition to formality. – Group 3. With capacity to contribute but not willing to join. Close to half of the informal workers who are unwilling to join earn on average(over the six countries) more than half the minimum wage and close to 20 per cent earn at least the minimum wage. All measures proposed for Group 2 also apply to this group, but should be complemented by interventions to address the reasons for their unwillingness to join, namely a lack of awareness and a lack of trust. Raising awareness and building trust concern everyone, but are of particular relevance for the third group. Interventions to strengthen the ability of individuals and enterprises to join a health insurance scheme and more generally to enter the formal economy, together with measures on the part of institutions to facilitate and incentivize their participation(addressing excessive costs, complex legislation or lack of adaptation) are necessary, but not sufficient. Strengthening the willingness to comply is essential. Willingness is not only a matter of individual choice but is largely driven by perceptions of fairness and the accountability of institutions. The objective is to seek voluntary participation and/or compliance. This is not just about changing people’s attitudes but also about reforming institutions themselves. Thus, additional interventions in this group should aim at changing the perceptions of workers(as citizens) in the informal economy: informing workers and employers about their rights and obligations through media campaigns and partnerships with workers’ and employers’ organizations, including those representing people in the informal economy; enhancing availability and accessibility of information for all and promoting»health and social security literacy« and 21 Such measures include expanding physical access points by increasing the number of local offices, establishing mobile offices and self-service terminals, as well as introducing mobile, online and other remote access points(ILO, 2021d). 22 There is no one-size-fits-all approach that might enable governments to guarantee health protection for all. International standards provide guiding principles to ensure universal protection to reflect risk-sharing, equity and solidarity – between income groups, men and women, age generations – in a fiscally, economically and socially sustainable fashion: principle 1- universality of protection; ­principle 2 – diversity of approaches and progressive realization; ­principle 3 – risk-sharing and solidarity in financing; principle 4 – overall and primary responsibility of the state; principle 5 – adequacy of benefits; principle 6 – predictability of benefits; principle 7 – nondiscrimination, gender equality and social inclusion; principle 8 – fiscal and economic sustainability with regard to social justice and equity; principle 9 – participation, social dialogue and accountability; and principle 10 – integration within comprehensive social protection systems(ILO, 2020). 36

Health insurance coverage and non-coverage REFERENCES Health Economics Research Unit(2019): Policy Brief. Towards universal health coverage in Kenya: Are we on the right path? KEMRI Wellcome Trust Research Programme, Nairobi Kenya; available at: https://­kemriwellcome.org/zp-content/uploads/2019/04/200-MEASURING-­ PROGRESS-TOWARDS-UNIVERSAL-HEALTHCARE-COVERAGE.pdf ILO(2015): Recommendation No. 204 concerning the Transition from the Informal to the Formal Economy; available at: https://www.ilo.org/ilc/ ILCSessions/previous-sessions/104/texts-adopted/WCMS_377774/ lang--en/index.htm –(2018): Women and men in the informal economy: a statistical p­ icture. Geneva: International Labour Office; avaialbe at: https://www.ilo.org/ wcmsp5/groups/public/---dgreports/---dcomm/documents/publication/ wcms_626831.pdf –(2019): Extending social security to workers in the informal economy. Lessons from international experience; available at: https://www.social-­ protection.org/gimi/RessourcePDF.action?id=55728 –(2020): Social Protection Spotlight. Towards Universal Health C­ overage: Social Health Protection Principles; available at: https://www.ilo.org/wcmsp5/groups/public/---ed_protect/---soc_sec/ documents/publication/wcms_740724.pdf –(2021a): World Social Protection Report 2020-22. Social protection at the crossroads: making a decisive turn for a better future; available at: https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/@ publ/documents/publication/wcms_817572.pdf –(2021b): Mapping of the National Social Protection System in ­Ethiopia, Including Social Health Protection; available at: https://www.ilo.org/wcmsp5/groups/public/---dgreports/---ddg_p/ documents/pblication/wcms_821183.pdf –(2021c): Revue globale et analyse du système national de protection ­sociale du Sénégal; available at: https://socialprotection-pfm.org/wp-­ content/uploads/2021/04/WEB-Revue-globale-du-systeme-de-­ protection-sociale.pdf –(2021d): Transition from the informal to the formal economy – Theory of change; available at: https://www.ilo.org/global/topics/employment-­ promotion/informal-economy/publications/WCMS_768807/lang--en/­ index.htm –(2021e): Extending social security to workers in the informal ­economy: Lessons from international experience. Good practices guide(online); available at: https://www.social-protection.org/gimi/gess/gess/Emodule. action?id=59 Jeune Afrique(2020): Côte d’Ivoire: la couverture maladie universelle, un immense défi pour le gouvernement. Marchs 2020; available at: https://www.jeuneafrique.com/mag/902959/societe/Côte-divoirela-couverture-maladie-universelle-un-immense-defi-pour-le-­ gouvernement/ OECD-ILO(2019): Tackling Vulnerability in the Informal Economy, ­Development Centre Studies, OECD Publishing, Paris; available at: https://doi.org/10.1787/939b7bcd-en. Ortiz, Isabel/Anis Chowdhury/Fabio Duran-Valverde/Taneem ­Muzaffar/Stefan Urban(2019): Fiscal Space for Social Protection: A Handbook for Assessing Financing Options. Schwettmann, Jürgen(2022): Towards Universal Health Coverage. The Cases of Benin, Côte d’Ivoire, Ethiopia, Kenya, Senegal and Zambia. Berlin: Friedrich-Ebert-Stiftung. Traub-Merz, Rudolf/Öhm, Manfred(2021): Access to health services. A Key Demand of Informal Labour in Africa – Findings from Representative Country Surveys in Sub-Saharan Africa. Berlin: Friedrich-EbertStiftung. Websites Universal Health Coverage Partnership. Zambia; a vailable at: https://www.uhcpartnership.net/country-profile/zambia/  Ministry of Health Website available at: https://www.moh.gov.zm/?p=6229 37

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows APPENDIX Membership of social health insurance or private health insurance schemes Table 4.A2 Membership of health insurance schemes by sex(frequency) SHI Benin 0.4 Côte d'Ivoire 7.3 Ethiopia 34.8 Kenya 20.8 Senegal 3.2 Zambia 1.3 MEN PHI(including microinsurance) 0.4 2.8 0.7 3.2 2.2 0.0 Total 0.8 10.1 35.5 24.0 5.4 1.3 Woman SHI PHI(including microinsurance) 1.6 1.8 7.4 3.0 34.6 1.3 23.4 3.7 7.7 3.6 1.4 1.4 Table 4.A3 Membership of health insurance schemes rural/urban residence(frequency) Urban SHI PHI(including Total SHI microinsurance) Benin 1.8 1.8 3.6 0.5 Côte d'Ivoire 9.4 4.8 14.2 4.9 Ethiopia 21.5 3.4 24.9 38.3 Kenya 21.7 5.4 27.1 22.3 Senegal 5.3 4.0 9.4 5.2 Zambia 1.1 1.3 2.5 1.5 Rural PHI(including microinsurance) 0.6 0.7 0.3 2.6 1.6 0.4 Table 4.A4 Membership of health insurance schemes(includes SHI and PHI) by level of education(frequency) Benin Côte d'Ivoire Ethiopia Kenya Senegal Zambia Average (6 countries) No school 1.2 8.0 37.2 11.7 5.7 0.0 16.6 Some primary 1.5 6.5 34.7 15.4 7.5 1.2 14.1 Primary 4.3 12.0 35.0 23.2 10.5 2.0 14.2 Secondary 5.3 15.5 35.0 33.9 9.3 4.7 24.1 University 3.4 30.2 33.3 58.0 21.7 8.3 27.7 Table 4.A5 Membership of health insurance schemes(includes SHI and PHI) by individual labour income(frequency) Benin Côte d'Ivoire Ethiopia Kenya Senegal Zambia Less than 1/2 MW 1.6 7.9 35.3 17.8 5.1 1.3 1/2 Minimum wage – Minimum wage [MW] 2.3 7.7 36.6 29.9 7.4 3.0 MW – 2 MW 3.2 14.3 31.4 45.0 11.2 3.1 More than 2 MW 5.2 24.3 32.8 40.9 24.7 7.5 Note: MW=Minimum wage Total 3.4 10.4 35.9 27.0 11.3 2.8 Total 1.1 5.6 38.6 24.9 6.8 1.9 Total 2.3 10.3 35.7 25.6 8.1 2.2 16.4 Total 2.3 10.2 34.9 25.8 8.3 2.1 38

Health insurance coverage and non-coverage Willingness to join a social health insurance scheme Table 4.A6 Willingness to join by sex(frequency) Benin Côte d'Ivoire Ethiopia Kenya Senegal Zambia Average(6 countries) Men 59.0 81.3 73.2 70.8 73.3 57.4 70.0 Women 76.4 79.7 72.5 70.4 75.1 58.9 67.3 Total 55.9 80.6 72.9 70.6 74.1 58.3 68.8 Table 4.A7 Willingness to join and use by level of education(frequency) Benin Côte d'Ivoire Ethiopia Kenya Senegal Zambia Average(6 countries) No school 53.5 77.6 69.2 55.6 69.6 42.0 65.0 Some primary 58.0 82.9 76.2 63.4 76.8 55.7 70.5 Primary 57.8 80.1 76.3 74.1 73.4 57.1 68.2 Secondary 50.0 85.4 70.8 77.3 73.7 71.8 73.5 University 58.9 86.4 66.7 63.2 76.5 70.0 70.5 Table 4.A8 Willingness to join by employment status(frequency) Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia Average(6 countries) Employers 82.3 68.1 72.7 64.8 91.3 68.0 76.0 Own-account worker 74.3 56.7 72.1 55.0 80.5 73.3 68.2 Employees 68.6 64.2 67.9 55.5 76.4 66.0 67.7 Family worker 37.5 61.5 71.4 63.6 76.9 83.3 70.3 Total 55.9 80.6 72.9 70.6 74.1 58.3 68.7 Total 74.1 58.0 70.7 56.1 80.6 72.9 68.7 39

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Table 4.A9 Reasons for unwillingness to join a health insurance scheme by employment status(frequency) Benin Côte d’Ivoire Ethiopia Kenya Senegal Zambia Employers Own-account workers Employees Family workers Employers Own-account workers Employees Family workers Employers Own-account workers Employees Family workers Employers Own-account workers Employees Family workers Employers Own-account workers Employees Family workers Employers Own-account workers Employees Family workers Insufficient financial resources 54.1 60.6 64.2 87.5 37.5 51.4 61.7 50.0 31.3 43.9 53.3 0.0 42.9 66.9 76.3 66.7 55.6 53.2 64.6 20.0 66.7 69.0 67.6 53.8 Don’t trust health insurance scheme 13.5 Don’t know anything about how they work 32.4 17.4 21.2 26.4 9.4 … 12.5 … 62.5 19.9 26.7 12.8 19.1 … 50.0 31.3 31.3 25.3 25.6 20.0 20.0 50.0 33.3 35.7 21.4 15.8 15.0 11.8 9.7 … 16.7 33.3 5.6 23.4 16.6 20.8 6.3 60.0 … … 33.3 10.4 17.2 … 26.5 … 38.5 Other … 0.8 … … … 2.1 6.4 … 6.1 5.2 6.7 16.7 … 2.3 2.2 16.7 5.6 6.8 8.3 20.0 … 3.4 5.9 7.7 Total 100 100 100 100 100 100.1 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 40

Economic effects of the Covid-19 pandemic on the informal economy: who has fallen into poverty? 5 ECONOMIC EFFECTS OF THE COVID-19 PANDEMIC ON THE INFORMAL ECONOMY: WHO HAS FALLEN INTO POVERTY? Christoph Strupat 5.1 INTRODUCTION The Covid-19 pandemic has triggered a major global recession(World Bank 2020). Widespread fear of contracting the virus, combined with strict policy measures to contain its spread have caused severe disruptions in livelihoods. To understand the full impact of the pandemic on low-income countries, it is necessary to consider the informal economy. People in informal employment are particularly vulnerable to the negative health and economic effects of external shocks, such as the pandemic. In addition, they constitute 90 per cent of all economic units in sub-Saharan Africa(Bonnet et al. 2019). Critical and diverse vulnerabilities are harboured in the informal economy:(i) the informally employed are more exposed to health and safety risks because of the poor health and safety conditions of their working environment, and limited access to public healthcare services;(ii) they face higher economic risks resulting from a combination of low income, high dependence on daily earnings for survival, and difficulties in accessing credit and social protection. Given these vulnerabilities, it is important to examine the economic impacts of the Covid-19 pandemic on the informal economy and which types of households have fallen into income poverty. Furthermore, it is necessary to explore whether governments have responded well in supporting households working in the informal economy by expanding social protection. By answering these questions, it is possible to advise policymakers on the targeting of future economic support or social protection programmes in order to reduce pandemic-related income poverty more effectively. To identify the economic impacts of the pandemic, many research organizations working in low-income countries have switched to phone surveys. While phone surveys can be used to provide useful information about the ongoing situation, they cannot fully replace in-person surveys. The main limitations include sampling bias because they can only be administered to respondents with a working phone. This is, for example, a major concern in rural areas where only 40 to 60 percent of households have access to a phone(Wieser et al. 2020). Phone-owning households are likely to be more educated and wealthier, implying that rural phone surveys in low-income countries are likely to miss the most vulnerable rural households. In order to overcome the limitations of phone surveys, we conducted in-person surveys in the informal economies of Côte d’Ivoire and Ethiopia, right after the easing of lockdown measures. The purpose of this chapter is to present countrywide descriptive evidence from these countries in order to capture a complete picture of the economic impacts of the Covid-19 crisis on the informal economy and to show whether households have fallen into poverty, and if so, which types of them. Furthermore, the chapter explores whether the governments have also responded by supporting households working in the informal economy and whether the response has helped to mitigate income poverty. 5.2 SPREAD OF COVID-19 AND LOCKDOWN MEASURES The first Covid-19 case was confirmed in Ethiopia on 13.3.2020. By 19.2.2021, there had been 151,016 cases and 2,260 deaths. The overwhelming majority of these cases and deaths have been in the capital, Addis Ababa. The first case was confirmed in Côte d’Ivoire on 11.3.2020. By 19.2.2021, there had been 31,825 Covid-19 cases and 185 deaths. The overwhelming majority of these cases and deaths have been in the capital, Abidjan. The first policy measures to limit the spread of the virus in Ethiopia were declared on 16.3.2020, just three days after the first confirmed case. The government of Ethiopia closed schools, banned all public gatherings and sporting activities, and encouraged physical distancing. Travelers from abroad were put into a mandatory quarantine, bars were closed until further notice, and travel over land borders was prohibited. Several regional governments imposed restrictions on public transportation and other vehicle movements between cities and rural areas. In particular, these restrictions were effective in Addis Ababa. The federal level State of Emergency was declared on 8.4.2020. Land borders were closed, except for cargo. Facemasks became compulsory in public spaces. Restrictions on cross-country public transportation and city transportation were also declared, for example, by limiting the carrying capacity of public transport providers to half their regular capacity. Some administrative regions took even stricter measures, closing restaurants and limiting movement between rural and urban areas. Unlike some other countries in the region, however, the country never went into a full 41

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows lockdown that severely restricted movement, imposed curfews or fully closed all borders. In Côte d’Ivoire after the first cases were diagnosed in March 2020, the government implemented containment measures. Lockdown measures included traffic limitations between Abidjan and the rest of the country to contain the spreading of the pandemic. The government implemented measures to limit the spread of Covid-19 in three steps. First, on 16.3.2020, the government closed land, air and sea borders indefinitely. It also made information available to passengers entering the country, while improving contact monitoring and tracing. From 23.3.2020, schools were closed and events, ceremonies and gatherings of more than 50 people were banned. Non-essential retail businesses, as well as bars, restaurants and entertainment venues had to close. In addition, a curfew was established from 9 p.m. to 5 a.m. in Abidjan. Finally, on 31.3.2020, the authorities restricted travel between Abidjan and rural areas, with the exception of freight transport and basic services. Most measures were lifted successively in July 2020. As the level and intensity of containment measures were less stringent in rural areas than in urban areas and, in particular, in the capital cities, it is possible that the negative economic impacts on the informal economy are larger in urban areas. To examine whether this is the case, the analysis will differentiate between the capital city, other urban and rural areas. 5.3 RESULTS 5.3.1 Employment One of the main economic outcomes of the analysis concerns employment. As many members of the informal economy depend on their daily earnings for survival, it is important to examine whether the pandemic and the lockdown measures have affected opportunities to work or whether respondents lost their jobs because of the Covid-19 crisis. Most people in informal employment reside in rural areas and the majority are engaged in agriculture. Considering this, the pandemic’s relative employment impacts are likely to remain small in rural areas. In urban areas, the concern about negative employment effects is more real. Figure 5.1 shows the share of households in which one or several household members have lost their job/working opportunity because of the pandemic. 1 On average, 30 per cent of households in Côte d’Ivoire indicated that they had lost their job/working opportunity, and 20 per cent of households in Ethiopia. The differences between urban and rural areas are large in both countries. It seems that twice as many people in urban areas indicated a loss of work opportunities than in rural areas. 5.3.2 Household income In order to understand the effects of the pandemic and the lockdown measures on household incomes, household heads or the most knowledgeable person in the household were asked to compare recent household income to household income at the same time of the year before the pandemic. The response options are typically qualitative, for example: »incomes were lower«,»same« or»higher«. While these responses provide some idea of the direction of income trends, they are very difficult to interpret when it comes to the magnitude or severity of income loss(De Weerdt, 2008). Thus, household heads or the most knowledgeable person in the household were asked to provide an estimate of the amount of household income. This makes it possible to calculate the relative fall in household income as a result of the pandemic. All income indicators are self-reported estimates and retrospective, but responses may also be affected by expectations about future income streams amid widespread uncertainty during the pandemic. With these important caveats in mind, Figure 5.2 shows the shares of households in both countries that reported having experienced a decrease of household income because of the pandemic. Some 60 per cent of the households in Côte d’Ivoire and 56 per cent in Ethiopia indicated that they had had lower household incomes in recent months as a consequence of the pandemic/policy responses. While in Côte d’Ivoire we see no significant differences between urban and rural areas, rural areas in Ethiopia seem to have been much less affected by the negative income consequences of the pandemic compared with urban areas. This result is in line with the findings on employment losses showing the smallest effects in rural Ethiopia. 1 Please note that this analysis is focused on households in which at least one household member is in informal employment. Figure 5.1 Loss of work opportunity or job because of Covid-19 100% 80% 60% 40% 20% 45.1% 42.3% 22.7% 32.7% 32.5% 18.9% 0% Côte d’Ivoire Ethiopia Capital city Other urban areas Rural areas 42

Economic effects of the Covid-19 pandemic on the informal economy: who has fallen into poverty? Figure 5.2 Reported decrease of household incomes because of the Covid-19 pandemic 100% 80% 60% 40% 61.3% 59.6% 55.4% 69.8% 66.5% 42.8% 20% 0% Côte d’Ivoire Ethiopia Capital city Other urban areas Rural areas Figure 5.3 Income poverty before and during the pandemic 100% 80% 60% 40% 20% 27.9% 46.6% 0% Côte d’Ivoire Income poverty before the pandemic 44.5% 67.2% Ethiopia Income poverty during the pandemic 5.3.3 Income poverty In order to understand the extent to which decreases in employment opportunities and household incomes have led to changes in household income poverty, the household heads had to indicate first whether their»household incomes were lower«,»the same« or»higher« in the past month compared with the same month before the pandemic. After answering this question, they had to specify the amount of income they had gained or lost as a result of the pandemic. This information makes it possible to assess household income before and during the pandemic and to explore possible changes in household income poverty. Households are defined as income-poor when the per capita household income is equal to or less than the international extreme poverty line. In Côte d’Ivoire, households that earn less than the statutory minimum wage of 30,000 CFA francs (54 US dollars per month) are considered to be income-poor. In Ethiopia, households that earn less than 1,500 Birr(41 US dollars per month) are considered income-poor. 2 Figure 5.3 shows how the shares of households that are considered to be»income-poor« have changed as a result of the Covid-19 2 The thresholds in both countries are below the international extreme poverty rate of 1.9 US dollars per person per day. pandemic. In Côte d’Ivoire, household income poverty has increased by 19 percentage points. Before the pandemic almost one-third of those working in the informal economy were income-poor, while after the first wave of the pandemic this rose to one-half. In Ethiopia, income poverty levels are generally higher, but a similar increase of income poverty by 23 percentage points can be observed. Almost 70 per cent of people in employment earn less than or equal to the extreme poverty line. In order to explore whether these changes vary between urban and rural areas, Figure 5.4 depicts the change in income poverty for the capital cities, other urban and rural areas. Income poverty has increased to a larger extent in urban areas than in rural areas. The share of households considered to be income-poor increased by 16 percentage points in rural areas in both countries. In contrast, urban income poverty increased by 25 percentage points in Ethiopia, and in Côte d’Ivoire by a little less, at 21 percentage points in other urban areas. In order to explore what types of households face income poverty due to the pandemic, the perspective of the analysis changes, focusing on household characteristics. So what types of household have fallen into income poverty and do 43

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 5.4 Increase in income poverty by regions as a result of Covid-19 60% 40% 20% 15.2% 21.0% 15.8% 24.5% 26.3% 16.5% 0% Côte d’Ivoire Ethiopia Capital city Other urban areas Rural areas these»new poor« differ from households that were already income-poor before the pandemic? Differences between the two groups could point, for example, to an urbanization and feminization of income poverty due to the pandemic. Tables 5.A1 and 5.A2 in the Appendix show the means of various household characteristics for those that have fallen into income poverty and those that were already income-poor before the pandemic. We start by describing the general characteristics of the new income-poor households of both countries. As we have information on informal employment status from all household members of the new income-poor households, we can see that most of the members of the latter are working as own-account workers and are on average between 10 and 29 years of age. Household heads of the new income-poor households are mainly female and have completed primary education. Beside these general characteristics of the new income-poor households, it is important to understand whether the pandemic has pushed specific types of household into poverty, ones that differ from the profile of the already poor. As we have detailed information on the household head, but also on all other members of the new income-poor and pre-existing poor households, we can investigate the differences between them. Figure 5.5 presents the differences between the two groups. The detailed differences in means can be found in column 3 of Tables 5.A1 and 5.A2. A positive difference between the two groups indicates that the mean of the indicator is larger for the new income-poor households than for households already income-poor before the pandemic. A negative difference indicates the opposite. while the old income-poor have a larger share of uneducated household heads. The share of female-headed households seems to be slightly larger among the new households. Overall, the findings indicate that the profile of the new income-poor is more urban and better educated, relative to the already poor, who are more rural, less educated and predominately reliant on agriculture. More household members of the new income-poor work as informal employees in the manufacturing or service sectors. The second panel of Figure 5.5 gives the differences in means for Côte d’Ivoire. Several differences between new and old income-poor can be detected here, too. Households of the new poor have a higher share of employers and are on average younger and living in urban areas. Similar to Ethiopia, the share of household heads who work in the manufacturing and service sectors is larger among the new income-poor. Additionally, more household heads work in the trade/retail sector. Interestingly, differences between the two groups can be detected with regard to education. A higher share of household heads of the new poor have secondary or university qualifications, which indicates that more better educated households have fallen into income poverty because of the pandemic. Overall, the findings for Côte d’Ivoire indicate that the profile of the new poor is younger, better educated and more urban. More household heads work in the trade/retail or service sector. In contrast to Ethiopia the households of the new poor have a higher share of employers, while no differences in the share of own-account workers or informal employees can be detected between the two groups. 5.4 GOVERNMENT RESPONSES The first panel of Figure 5.5 gives the differences in means for Ethiopia. The new income-poor differ in several characteristics from households already income-poor before the pandemic. The households of the new poor have higher shares of informal employees and more of them live in urban areas. Furthermore, the share of household heads who work in the non-agricultural, manufacturing and service sectors is larger than among the old income-poor. Interestingly, a higher share of household heads with primary and secondary education can be found among the new income-poor, Both governments have responded to the Covid-19 pandemic by implementing lockdown measures and supporting their health care systems. Considering the large negative economic impacts, it is necessary to explore whether the governments have also responded by supporting households working in the informal economy. Survey respondents were asked whether their households had received any assistance from the government in terms of food, cash or other support since the outbreak of the Covid-19 crisis. Figure 5.6 depicts the shares of households that received assistance from the 44

Economic effects of the Covid-19 pandemic on the informal economy: who has fallen into poverty? Figure 5.5 Differences between the new and old income-poor Share of employers in household Share of own-account workers in household Share of informal employees in household Share of family workers in household Share of females in household Household mean age 10–29 Household mean age 30–39 Household mean age 40–49 Household mean age 50–59 Household mean age>60 Female(household head) Work in agriculture(household head) Work in manufacturing(household head Work in trade/retail(household head) Work in service sector(household head) No education(household head) Primary education(household head) Secondary education(household head) University education(household head) Live in Addis Ababa Live in other urban area Live in rural area Ethiopia –20% –15% –10% –5% 0 5% 10% 15% 20% 25% Share of employers in household Share of own-account workers in household Share of informal employees in household Share of family workers in household Share of females in household Household mean age 10–29 Household mean age 30–39 Household mean age 40–49 Household mean age 50–59 Household mean age>60 Female(household head) Work in agriculture(household head) Work in manufacturing(household head Work in trade/retail(household head) Work in service sector(household head) No education(household head) Primary education(household head) Secondary education(household head) University education(household head) Live in Abidjan Live in other urban area Live in rural area Côte d’Ivoire –20% –15% –10% –5% 0 5% 10% 15% 20% Note: The figure displays results from column 3 of Tables A1 and A2. It depicts the difference in means(filled circles) between those that have become income-poor due to the pandemic and those already income-poor before it. 95 per cent confidence intervals(horizontal line) are added for each of the indicators. 45

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows government. In Ethiopia, the support was concentrated in Addis Ababa, but covered only 15 per cent of households. In Côte d’Ivoire, the support was concentrated in Abidjan, but covered only 9.6 per cent of households. Overall, comprehensive government assistance targeted at the informal economy is largely missing in both countries. The limited support in Côte d’Ivoire can be explained by the large imbalance of economic support funds towards the formal private sector. The government there implemented four emergency funds up to July 2020. However, two of the funds were designed entirely for the formal private sector (large enterprises and SMEs) and disbursed 11,000 billion CFA francs to just 79 beneficiaries(Sanogo et al. 2020). The fund dedicated to the informal sector disbursed just 3.18 billion CFA francs to around 130,000 households, while an emergency cash transfer programme for vulnerable people reached 100,000 households. In Ethiopia, the low level of support shown in our survey can be explained by the fact that no economic support programmes or expansions of existing social protection programmes were implemented by the government, mainly because of financing constraints(Bischeler et al. 2020). The Ministry of Urban Development and Construction designed a temporary income support programme for informal workers, but this was never implemented due to lack of funds. The design was developed with technical support from the World Bank. The programme was supposed to provide three months of unconditional cash support to approximately 100,000 informal workers who had lost their jobs or who had gone out of business because of the pandemic. The program was to be financed via a World Bank emergency Development Policy Operation(DPO). However, due to debt and other macroeconomic issues, the DPO failed to materialize and the programme was not implemented in the absence of alternative financing. In order to examine whether the limited government support in both countries helped to mitigate income poverty, Figure 5.7 presents the share of households that are income-poor with and without government assistance. Interestingly, in Côte d’Ivoire the share of income-poor households is smaller among those that received government assistance than among those that did not. There are no differences in Ethiopia. It seems that even limited government support can partly limit the economic impact of the pandemic on income poverty rates, at least in Côte d’Ivoire. Figure 5.6 Government assistance received during the pandemic 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 9.6% 6.2% 6.1% Côte d’Ivoire Capital city 15.3% 4.3% Other urban areas Ethiopia Rural areas 2.2% Figure 5.7 Income poverty by the provision of government assistance 100% 80% 60% 40% 37.5% 46.7% 20% 0% Côte d’Ivoire Support by government 69.3% 67.1% Ethiopia No support by government 46

Economic effects of the Covid-19 pandemic on the informal economy: who has fallen into poverty? 5.5 CONCLUSION AND POLICY ­RECOMMENDATIONS Economic activity in the informal economy has been strongly affected by the Covid-19 pandemic. On average, 60 per cent of households in the two countries face an income decline and 20 to 30 per cent lost jobs and work opportunities. The negative employment effects are smaller than what was initially feared. The fact that most of the workforce remains in agriculture also meant that the jobs of most Ethiopians and Ivoirians were only partly affected by the crisis. However, the larger negative income effect points to the devastating impacts of the disruption of value chains that also affect those living in rural areas. As a consequence, income poverty levels have increased by around 20 percentage points. The largest negative effects were found in urban areas. The profile of those that have fallen into income poverty due to the pandemic differ substantially from those of the already poor. The new income-poor are more urban, younger and better educated, relative to the former poor who are more rural, less educated and predominately reliant on agriculture. A higher share of household members of the new income-poor work as employers or informal employees in the manufacturing or service sectors. This group of better-educated households may just need a monetary boost from a support programme to escape income poverty. In addition, further assistance needs to be provided(also with the support of external donors) to countries with large informal economies to prevent further impoverishment in the future. It is necessary to create adaptive social protection systems that can respond better to new vulnerabilities caused by covariate shocks, such as pandemics or the impact of climate change. Adaptive social protection systems make it possible to increase the benefit value for existing social protection beneficiaries(vertical expansion) and to enrol additional beneficiaries(»the potential new poor«) in existing programmes or to create new programmes(horizontal expansion). Development cooperation can support countries in establishing such systems and improving their administrative capacities to adapt social protection schemes with a uniform social registry of(actual and potential) beneficiaries, covering all workers, including those in informal employment. Countries with such registries could easily explore and close gaps in their social protection coverage in the short term. Unfortunately, social assistance packages that are targeted to the informal economy in order to support them during the pandemic are largely missing in both countries. Only four to seven per cent of the households in our surveys received support from the government. The main reasons for this limited support are financing issues, but also a large imbalance in the support towards the formal sector. REFERENCES Bischler, J./Asheber, T./ Hobson, M.(2020): Towards shock-responsive social protection: lessons from the COVID-19 response in Ethiopia. Shock-responsive social protection responses during COVID, 19. This economic support gap between formal and informal sectors now represents an enormous challenge for countries with large informal economies. The increase in overall poverty in many countries in Africa is potentially due to the increase in income poverty in the informal economy. Thus, in order to reduce pandemic-related income poverty it is vital to support workers in informal employment with cash transfers or further social assistance programmes. The first attempts to implement such programmes have failed in many countries due to financing problems. As programme financing will become more difficult in the future because of rising national debt levels, policymakers should prioritize their limited resources and redistribute assistance programmes in such a way that they are tailored and targeted to specific types of households in the informal economy in order to effectively reduce pandemic-related income poverty and to be prepared for future shocks. As the analysis of this chapter focuses on those who have fallen into poverty(the new income-poor) due to the pandemic, policymakers should target better-educated households with a high share of employers or informal employees working in the service sector and living in urban areas. The analysis has indicated that this group fell into poverty during the first wave of the pandemic and differs to a large extent from the original poor. Targeting them could contribute to an effective reduction of pandemic-related income poverty. Bonnet, F./Vanek, J./Chen, M.(2019): Women and men in the informal economy: A statistical brief. International Labour Office, Geneva. De Weerdt, J.(2008): Field notes on administering shock modules, in:­ Journal of International Development: The Journal of the Development Studies Association, 20(3), 398–402. Gevers, C./Buyoya, O./Moller, L. C./Coppola, A./Welmond, M. J./ Mathrani, S. W./Sanogo, T./Tiemann, A./Adoho, F. M./Coulibaly Y Epse Oyolola, J./Kim, M. D./Marchat, J. M. N./Rougeaux, S. M. P. (2020): Taking Stock and Looking Ahead: Cote d’Ivoire and the COVID-19 Pandemic(English). Washington, D.C.: World Bank Group; available at: http://documents.worldbank.org/curated/en/149401601444415154/­ Taking-Stock-and-Looking-Ahead-Cote-d-Ivoire-and-the-COVID-19-­ Pandemic Wieser, C./Ambel, A. A./Bundervoet, T./Haile, A.(2020): Monitoring COVID-19 Impacts on Households in Ethiopia; available at: https://openknowledge.worldbank.org/bitstream/handle/10986/33824/ Results-from-a-High-Frequency-Phone-Survey-of-Households.pdf? sequence=5 World Bank(2020): Press Release: COVID-19 to Plunge Global Economy into Worst Recession since World War II. Washington D.C.: The World Bank 47

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows APPENDIX Table 5.A1 Characteristics of the new and old income-poor in Ethiopia Share of employers in household Share of own-account workers in household Share of informal employees in household Share of family workers in household Share of females in household Household mean age 10–29 Household mean age 30–39 Household mean age 40–49 Household mean age 50–59 Household mean age>60 Female(household head) Work in agriculture(household head) Work in manufacturing(household head Work in trade/retail(household head) Work in service sector(household head) No education(household head) Primary education(household head) Secondary education(household head) University education(household head) Live in Addis Ababa Live in other urban area Live in rural area Number of observations (1) New income-poor (mean in%) 4.36 77.39 10.27 8.00 51.31 61.82 17.17 2.83 1.21 0.40 50.58 48.99 7.47 10.1 33.43 36.36 53.74 9.29 0.61 14.85 21.01 64.14 495 (2) Old income-poor (mean in%) 3.09 82.70 6.50 7.71 50.98 62.09 15.50 5.27 1.96 0.93 45.99 63.97 3.62 7.75 24.67 49.80 45.97 3.40 0.00 10.83 12.09 77.09 968 (3) Difference(1–2) 1.28 –4.31** 3.77*** 0.29 0.33 –0.27 1.68 –2.44** –0.75 –0.53 4.59 –14.98*** 3.86*** 2.35 8.76*** –13.44*** 7.77*** 5.89*** 0.61 4.02*** 8.92*** –12.95*** Std error 1.89 1.84 1.3 1.27 1.1 2.68 2.03 1.14 0.72 0.48 2.64 2.53 1.19 1.54 2.09 2.73 2.76 1.44 0.48 0.8 1.97 2.06 Note: The table shows levels and difference in means(filled circles) between those that became income-poor due to the pandemic(new poor) (earned equal to or less than the minimum wage) and those already income-poor beforehand(old poor).* p< 0.10,** p< 0.05,*** p< 0.01. Table 5.A2 Characteristics of the new and old income-poor in Côte d’Ivoire Share of employers in household Share of own-account workers in household Share of informal employees in household Share of family workers in household Share of females in household Household mean age 10–29 Household mean age 30–39 Household mean age 40–49 Household mean age 50–59 Household mean age>60 Female(household head) Work in agriculture(household head) Work in manufacturing(household head Work in trade/retail(household head) Work in service sector(household head) No education(household head) Primary education(household head) Secondary education(household head) University education(household head) Live in Abidjan Live in other urban area Live in rural area Number of observations (1) New income-poor (mean in%) 11.26 70.79 14.68 3.27 47.87 71.86 13.57 3.52 1.01 0.50 54.45 45.73 8.54 29.15 30.97 36.19 51.26 6.53 6.03 17.08 22.11 60.81 199 (2) Old income-poor (mean in%) 4.10 71.32 17.94 6.64 50.77 61.67 16.03 5.57 0.35 1.74 50.65 48.78 4.53 16.72 15.58 38.68 55.05 3.83 2.44 18.12 16.72 65.16 287 (3) Difference(1–2) 7.16*** –0.53 –3.26 –3.34** –2.91 10.19** –2.46 –2.06 0.66 –1.24 3.80 –3.05 4.10* 12.43*** 15.39*** –2.49 –3.8 2.69 3.59** –1.04 5.39 –4.35 Std error 1.99 3.36 2.8 1.49 2.23 4.36 3.3 1.96 0.72 1.02 4.32 4.61 2.22 3.79 3.9 4.49 4.61 1.92 1.78 3.49 3.61 4.44 Note: The table shows level and difference in means(filled circles) between those that became income-poor due to the pandemic(new poor) (earned equal to or less than the minimum wage) and those already income-poor beforehand(old poor).* p< 0.10,** p< 0.05,*** p< 0.01. 48

Trust in institutions and perceptions of the tax system among informally employed people 6 TRUST IN INSTITUTIONS AND PERCEPTIONS OF THE TAX SYSTEM AMONG INFORMALLY EMPLOYED PEOPLE Armin von Schiller 6.1 INTRODUCTION Getting to grips with citizen–state relations is an important part of any attempt to understand how states function. Any expansion of health and social protection systems requires social and political support as well as funding. It is thus crucial to understand what broad sectors of society expect from government in order to support this expansion. It is just as important to discuss the scope for economically efficient and socially fair funding reform. Many observers perceive financing as the main constraint on extending and deepening social protection systems in low- and middle-income countries(for example, Barrientos 2008). Variability is high among such countries, and reliable and comparable data is lacking. It is safe to say nevertheless that most of the relevant systems are underfinanced and poorly developed. In most cases, the bulk of the funding is generated domestically, predominantly through taxes and social insurance contributions, but in some areas the role of external donors is crucial(Niño-Zarazúa et al. 2012). been particularly vulnerable to the impact of the pandemic precisely because of their position outside formal regulatory structures. As a consequence of the nature of their relations with the state, informal workers have often not benefited from programmes and measures launched to fight the impact of the virus because they were invisible to the channels employed to channel that support and relief measures(Gallien and van der Boogard 2021; Strupat, Chapter V). The vision of more comprehensive health and social protection systems that benefit also the informally employed is widely shared. Agreeing and implementing the necessary reforms to make this vision a reality represents a challenge that requires acknowledging and potentially reshaping citizens’ vertical relations with the state and also the horizontal relationships between citizens. In a scenario characterized by more comprehensive health and social protection systems the state will need to play a role. As a result, any strategy to expand the health and social protection systems requires thinking carefully about how the relationship between the informally employed and state institutions is perceived and defined. The question of how the informally employed perceive the state and tax systems is crucial in this context. Depending on the definition, informal workers make up over 60 per cent of global employment and more than 85 per cent in sub-Saharan Africa and South Asia(ILO 2018). 1 One very common narrative suggests that the informal economy is the main obstacle to domestic revenue mobilization in Africa. This narrative has been questioned by many who indicate that, for instance, the contribution of the informally employed is heavily underestimated(for example, Meagher 2016). Moore(2020) even goes as far as portraying this narrative as »diversionary«, wrongly distracting attention from the real sources of uncollected revenues, which he identifies as rather low tax contributions by the wealthy and the predominance of ineffective tax exemptions(Moore 2020). The account of the contributions made by informal workers might be more positive if we took into consideration the fact that commonly they do not benefit from many state services. This dynamic has been even more clearly shown in the context of the Covid-19 pandemic. Informal workers have Moreover, funding constraints will remain significant. Lowand middle-income countries need to increase their domestic revenues in order to afford the expansion of their health and social protection systems. 2 Social insurance contributions are an alternative way of financing this expansion and will need to play a role as well. However, social insurance has its limits, especially to expand coverage to low-income groups who might simply be too poor to pay for their own social protection. The poorest of the poor can be reached only through tax-financed approaches(Bastagli 2013; Bonnet, Chapter IV). In addition, convincing(or forcing) the wealthier to pay more and to redistribute is politically complex and administratively demanding(Berens and von Schiller 2017). As a result, the challenge to develop more comprehensive health and social protection systems is not just technical in nature, but political and social. Finding the scope for a reform in this direction requires changes in the predominant fiscal contracts in these countries, understood as an implicit agreement in a society defining how much its members can expect to benefit from state action and how much they are 1 See definition used in our survey in Chapter 9. 2 See for instance the connected discussion on implementing the SDGs (Gaspar et al. 2019): sustaining increases in tax-to-GDP ratios over time is considered essential for achieving them. 49

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows expected to contribute through paying revenue in exchange (Burchi et al. 2000: 19ff). 3 To find sufficient scope for such reform there are three crucial questions we need to answer. How much are people willing to pay and what would they be able and/or aspiring to demand in exchange? Is the existing tax system perceived to be fair? Who is expected to contribute more and who is considered to deserve support? The results of the survey presented in this chapter provide some indications. Health and education are the top priority areas that informally employed people would like governments to focus on. Generally speaking, informally employed people are open to contribute more via taxes and fees if they themselves or even poorer people obtain better services in exchange. The vast majority also expressed support for a vision of the state that targets services at the poor, regardless of whether they are able to contribute to the fiscal effort. One major political and social obstacle to coordinating and implementing more ambitious reforms seems to be the informally employed’s lack of trust in various social and political institutions, especially so-called intermediary institutions, such as political parties and trade unions. This is problematic, especially if we envision expansion to be based on democratic and participatory processes, as intermediary institutions play an essential role in navigating reforms and sustaining consent during implementation. Overall, there seems to be scope for reform but much work is needed to create the socio-political conditions required for the vision to be implemented and become a reality. 6.2 PRIORITY POLICY AREAS FOR THE INFORMALLY EMPLOYED AND PERCEPTIONS OF VARIOUS POLITICAL AND SOCIAL INSTITUTIONS To identify the scope for reform in health and social services a crucial starting point is understanding the degree to which informally employed people consider these priority policy areas. The results of the survey clearly indicate that better health and schools are the most relevant issues for them (Table 6.1; see Chapter 2 for more on these issues). Hence, it is fair to say that improvements in these areas are of interest to the informally employed. The articulation of these expectations, as well as the definition and implementation of measures to achieve them definitely requires the intervention of various institutions. In fact, it is essential to understand how informally employed people assess different institutions 3 Note that a fiscal contract does not imply that benefits and payments have to be levelled out. It can be the case that poor people are considered to be entitled to some benefits without contributing any revenue. Also it can be the case that rich people accept paying more than they formally get back in the form of direct benefits(Berens and von Schiller 2017). At the extreme, consider so-called»rentier states«, in some of which there are high expectations of what the state should provide based on resource wealth without insisting that members should pay any financial contribution in exchange. In fact, one essential characteristic of rentier states is that they need not tax – or need not tax much(Herb 2005: 298). Table 6.1 Stated main priorities among informally employed people (unweighted average for all countries in the sample) Better schools and education Better roads and bridges Better police services Better health services Better electricity supply Better water supply Better pensions for the elderly Better food programmes in times of crisis Mentions among the top three priorities 56.2% 37.5% 14.0% 67.6% 30.8% 41.9% 12.4% Mentioned as first priority 22.5% 11.3% 2.7% 30.3% 7.3% 14.8% 3.0% 24.5% 8.1% Note: Exact wording of the question: Please rank the sectors you have highlighted above(first, second and third priority). The question is put after asking: If the government wants to improve services for people, what in your opinion are the sectors the government should focus on? In this question the respondent has the option of saying yes or no on each of the sectors above before they are asked to prioritize. to get a feeling of which actors they trust to take care of these demands and who therefore might represent potential allies. The data reveal that, in the countries analysed, trust among the informally employed is particularly high for the president (Figure 6.1). In all six countries included in the analysis, this is the formal political institution with the highest trust levels. The only group consistently close to and sometimes even above the presidency in terms of level of trust comprises religious leaders. Civil society organizations, and particularly political parties and trade unions, are the least trusted institutions among those discussed here. 4 It is also remarkable how these trends in trust are stable across countries and subgroups. We might expect that the age of respondents, their income and education level, as well as whether they live in rural or urban areas might have a strong effect on trust. However, this effect seems to be minor (see Figure 6.A7, appendix). Beyond trust as such, in order to understand which actors are seen by the informally employed as potential vehicles for 4 It is certainly relevant to consider that exposure to institutions and knowledge of them might affect people’s assessments. We added a question on people’s knowledge of trade unions that partly allows us to investigate this issue. In general, the level of trust changed only marginally. Among those stating that they have never heard of trade unions the mean value for trust is(2.11); this value is marginally lower for those stating that they»have heard about trade unions but do not know about them«(2.10). Trust increases among those stating »I know a little about trade unions«(2.21), but actually decreases for those who claimed to know most about trade unions(»I have heard about trade unions and I know what they are doing«). In general, however, there is considerable variance among countries regarding responses; among those declaring they know what trade unions do the assessment is particularly positive in Kenya, Zambia and E­ thiopia, but particularly negative in Côte d’Ivoire, Senegal and Benin. In any case, for all countries, the number of people who report knowing trade ­unions well is low. 50

Trust in institutions and perceptions of the tax system among informally employed people Figure 6.1 Trust in institutions among informally employed people(% of total, by country) The President Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia Religious Leaders Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 25% 50% 75% Civil Society Organisations 100% 25% 50% 75% International Donors 100% 25% 50% Not at all 75% 100% Just a little 25% Somewhat 50% 75% A lot 100% Note: Exact wording of the question: How much do you trust each of the following actors? – Response options: Not at all; Just a little; Somewhat; A lot; Don’t know/ Haven’t heard of them[not read out but coded] Figure 6.2 Perceptions of the informally employed of how much different institutions care about their main priorities (% of total, by country) The President Religious Leaders Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 25% 50% 75% Civil Society Organisations 100% 25% 50% 75% International Donors 100% 25% 50% Not at all 75% 100% Just a little 25% Somewhat 50% 75% A lot 100% Note: Exact wording of the question: How much do you think that the following organizations/institutions care about your top priorities you highlighted? – Response options: Not at all; Just a little; Somewhat; A lot; Don’t know/ Haven’t heard[not read out but coded] 51

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 6.3 Perceptions of the degree to which different institutions can make a major difference in solving informally employed people’s main priorities International donors Civil society organisations Trade unions Religious leaders Traditional leaders Political parties Local government National parliament National government The president Not at all Just a little Somewhat A lot Note: Values are calculated as means for all available respondents in all countries(no particular weights are used). Exact wording of the question: Assuming the following actors are willing to work towards tackling your top priorities you highlighted, do you think that they can make a significant difference and change the situation to better through their actions? – Response options: Not at all; Just a little; Somewhat; A lot; Don’t know/ Haven’t heard[not read out but coded] realizing their demands it is important to assess the degree to which different institutions are actually seen to care about the issues that are relevant for them. Our survey included a question on this topic. In general, the results referring to how much institutions are perceived to care about people’s priorities are a bit lower than those referring to trust(see Figure 6.2). The gap between abstract trust in institutions and the more concrete perception of the degree to which these institutions care may signal a disconnect between many institutions and the informally employed. In fact, the numbers suggest that a remarkable proportion of informally employed people consider that many institutions focus on other issues that are not particularly important to them. Finally, in terms of perceptions of different institutions, it would be interesting to understand which institutions informally employed people expect to make a real difference to their situation. This certainly affects the degree to which they have an incentive to cooperate, engage and lobby different institutions, but also gives an indication of where they perceive that the power to change things lies. In line with the results described above the president is seen as the one person able to make a difference, while the trade unions and political parties are perceived as having less potential. The numbers are remarkably high for international donors, as well as traditional and religious leaders(see Figure 6.3). 6.3 PERCEPTIONS OF TAXES AND FEES ployed consider that non-compliance is an issue. Variation in the perception of non-compliance is relatively high, however. While in Ethiopia only 14 per cent of respondent considered that not paying income tax happens often, this number is above 40 per cent in Kenya and Côte d’Ivoire(see Table 6.2). Beyond the perceived occurrence of non-compliance, in order to understand the potential to increase tax collection as a step previous to expanding public social and health care systems, it is essential to get a feeling about whether the informally employed consider it legitimate for the state to extract revenue at all. The data presented in Table 6.3 indicate that there is a very strong awareness that not paying for services or not paying taxes is»wrong«. Still a remarkable proportion of respondents considered that not paying taxes is»wrong but understandable«(46 per cent). 5 This is a worrisome number in absolute, but also in relative terms. Afrobarometer surveys have asked the same question in several African countries over time. In essence, although with some caution, it can be argued that comparing the responses between our data and those of Afrobarometer amounts to comparing the perceptions of the population as a whole to those of the informally employed. For the countries in our sample for which Afrobarometer data on this issue are available(all except Ethiopia) the proportion of respondents who consider not paying taxes»not wrong at all« or»wrong but understandable« is clearly higher, on average. This implies that, on average, informally employed A large majority of the respondents considered that it is difficult or very difficult to avoid paying taxes. When confronted with a more abstract(and less self-referenced) question on the degree to which, in general, non-compliance happens and people do not pay the taxes or fees they owe, the numbers go up. Hence, a remarkable proportion of the informally em5 Details on variations between countries can be found in Table 6.3. The table does not include Ethiopia because Afrobarometer does not have data for it on the precise question. Interestingly, Ethiopia seems to be an outlier. In Ethiopia 12.3 per cent of the sample state that not paying taxes owed is»not wrong at all«, 16.9 per cent consider it »wrong but understandable«, 58 per cent»wrong and punishable« and 12.9 per cent do not respond. 52

Trust in institutions and perceptions of the tax system among informally employed people Table 6.2 Perceptions of compliance with taxes and fees among the informally employed How often do you think people do not pay the taxes they owe on their income? How often do you think people do not pay the fees for the services they receive from government? Never Sometimes Often Never Sometimes Often Senegal 28% 35% 37% 35% 34% 32% Zambia 16% 47% 37% 16% 48% 36% Kenya 6% 47% 47%  6% 53% 41% Benin 12% 47% 40% 17% 51% 33% Côte d'Ivoire 17% 40% 43% 22% 40% 38% Ethiopia 41% 46% 14% 41% 47% 12% Mean 20% 44% 36% 23% 45% 32% Note: Exact wording of the question: The question is asked as a battery. How often do you think that the actions we describe happen in general? For each of the following, please tell me: A. Not paying fees for the services they receive from government. B. Not paying the taxes they owe on their income. Response options. Never; Sometimes; Often; Don‘t Know[Not read aloud but coded] Table 6.3 Perceptions of tax evasion Not wrong at all Wrong but understandable Wrong and punishable Do not know Afrobarometer Data FES-IDOS-ILO Data Afrobarometer Data FES-IDOS-ILO Data Afrobarometer Data FES-IDOS-ILO Data Afrobarometer Data FES-IDOS-ILO Total 4.7% 9.1% 37.9% 46.3% 52.2% 39.1% 5.2% 5.5% Benin 3.8% 8.6% 47.0% 49.8% 48.9% 40.6% 0.2% 1.0% Côte d´Ivoire 2.0% 5.6% 41.2% 64.2% 54.2% 27.2% 2.6% 3.0% Kenya 6.9% 8.0% 36.5% 36.4% 51.3% 48.8% 5.3% 6.7% Senegal 3.4% 12.7% 36.4% 45.4% 58.1% 40.7% 2.1% 1.8% Zambia 5.2% 10.6% 29.7% 35.6% 49.2% 38.9% 15.9% 14.9% Note: Exact wording of the question: I am now going to ask you about a range of different actions that some people take. For each of the following, please tell me(…) B. Not paying the taxes they owe on their income. In this calculation the total does not include data from Ethiopia as Afrobarometer does not have any results on this country. Afrobarometer data correspond to Afrobarometer Data,[Benin, Côte d´Ivoire, Kenya, Senegal, Zambia], Round 6, 2014–2015, available at http://www.afrobarometer.org. people consider it to some degree justified to not pay taxes more often than the general population. The gap between the perceptions of the larger population and of the informally employed is particularly large in Côte d’Ivoire, Senegal and Zambia. 6 This tolerance for tax evasion might be connected to a perception that the use of the resources is corrupt and that the state has done too little with the revenues collected. Partly supporting this intuition, when asked about the perceived fairness of the fiscal system a majority of the respondents disagreed or strongly disagreed with the statement that »they get enough in exchange for taxes and fees«. Still, the result does not appear to be as negative as might have been expected; in fact, in two countries, the majority of respondents perceived the situation to be fair(see Figure 6.4). Considering alternative scenarios and how these are assessed, it is interesting to note what vision of the state people support. A clear majority of the informally employed support the idea that the state should support the poor, regardless of their capacity to contribute to the tax effort. Concretely, 69 per cent of respondents disagreed or strongly disagreed with the statement»The government should provide services only to those persons and households that pay taxes 6 No comparison for the question on fees is available. and fees«. The value is consistently around 75 per cent in the different countries, with Senegal as an outlier with a majority of respondents(55 per cent) actually agreeing or strongly agreeing with the statement. Still, it can be said that in general there is a consistent strong support among the informally employed for solidarity and redistribution towards the weaker and a recognition of the necessity to support those in need regardless of their capacity to contribute. The survey also gives us some clues about the level of support of the informally employed for measures geared towards activating certain sources of revenue, and how the existing and potential additional tax burden should be distributed. A remarkable 72 per cent agree or strongly agree with the statement that everyone earning an income should pay taxes, suggesting that the tax burden should be borne by as many as possible and the tax base should include everyone with the economic capacity to contribute. Some 35 per cent support the claim that only wealthy citizens should have to pay taxes and fees. If we cross-tabulate these two questions we get some interesting results(Table 6.4). Probably the most telling data are at the bottom left and upper right. The upper right data indicate support for putting more fiscal pressure on the wealthy without necessarily making everyone with an income contribute. This suggests a highly progressive tax system focused on a narrow tax base of high income earners. By contrast the bottom left data indicate reluctance to put all 53

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 6.4 Perceived fairness of public service delivery in exchange for taxes and fees Senegal Zambia Kenya Benin Côte d’Ivoire Ethiopia Strongly Agree 25% Agree 50% Undecided[DNR] 75% Disagree 100% Strongly Disagree Note: Exact wording of the question: Particularly in your case and that of your family, do you consider that you get a fair amount of public services considering the money that you contribute to the state in terms of taxes, fees and other payments? For the response options see the legend. The option undecided was not read out but coded. Table 6.4 Opinions on who should pay taxes and fees(unweighted average for all countries in the sample) Only wealthy citizens should have to pay taxes or fees to the government Strongly agree/ agree Strongly disagree/ disagree Total Everyone who earns an income, regardless of how much it is, should have to pay some taxes or fees to the government Strongly agree/ agree Strongly disagree/ disagree 21% 15% 51% 72%  8% 23% Total 36% 59% 95% Note: The questions were asked as part of a battery. Before mentioning the statement in bold in this table the respondent was asked: Do you agree with the following statement? The response options included» Strongly Agree«;» Agree«(both options, including»Agree«, are collapsed in the table);» Disagree«;» Strongly disagree«(both options including»Disagree« are collapsed in the table);» Undecided«; » Don’t know«. The options»Undecided« and»Don’t know« were not read out but coded. the fiscal pressure on the wealthy, and agreement with the idea that everyone who has an income should contribute to the tax effort. This might be interpreted as indicating support for moderate progressive taxation and a broad tax base. The difference in the proportion of respondents in those cells of the table is surprisingly large. A remarkable 51 per cent of respondents are located in the cell, suggesting support for a broad tax basis. By contrast only 15 per cent of the respondents indicate support for a highly progressive tax system centred on the wealthy. The data might be interpreted as suggesting that the informally employed broadly support taxing people with the economic capacity to contribute and having a regular income. But data are not particularly supportive for calls for tax measures very much geared towards the wealthy. It is true that 36 per cent agree with the statement that only the wealthy should pay but the majority of respondents also indicate that everyone with an income should pay. It appears that the informally employed are more open to support a larger tax base than is commonly assumed or claimed. In any case, this does not call into question potential support for progressive taxation and redistribution, as other data suggest strong support. Overall, it seems that redistribution is very much understood as focussing on addressing the needs of the poor and that there is openness to making large segments of society contribute. This reticence to put more pressure of the wealthy is also suggested by the fact that, with the exception of Kenya, the number of respondents who consider that the wealthy pay enough is pretty high (Figure 6.5). In three countries a majority support this idea (Côte d’Ivoire: 59 per cent; Ethiopia: 60 per cent; and Benin: 51 per cent), while in Zambia the figure is 47 per cent(with a high 16 per cent of respondents undecided). The clear outlier is Kenya where 75 per cent consider that the wealthy should pay more. 7 Also, those who consider the system to be unfair 7 In overall terms 49 per cent support the idea that the wealthy pay enough while 44 per cent believe that they should pay more. 54

Trust in institutions and perceptions of the tax system among informally employed people Figure 6.5 Opinions on the degree to which the wealthy contribute enough taxes Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia Strongly Agree 25% Agree 50% Undecided[DNR] 75% Disagree 100% Strongly Disagree Note: Exact wording of the question: Do you think that particularly wealthy regions and wealthy individuals contribute their fair share of taxes, fees and other payments to the development of the country? Response options: Strongly agree, they already pay more than enough; Agree, they already pay enough; Undecided[option not read out but coded]; Disagree, they should pay more; Strongly disagree, they should pay significantly more; Don’t know[Option not read out but coded] Figure 6.6 Openness to increase contributions if money is used for health 50% 40% 30% 20% 10% 0% Strongly oppose Somewhat oppose Undecided Somewhat support Strongly support If I benefit If people poorer than me benefit Note: Exact wording of the two questions compared in the figure: If the government decided to make you pay taxes or user fees in order to increase spending on public health care benefitting you, would you support this decision or oppose it? vS. Would you pay taxes or fees if the government would use this money to provide free health services for persons poorer than you? The option»Undecided« was not read out but coded. The values represent means for all observations. appear to attach much heavier blame to the contributions of the wealthy. These results suggest that there is a widely shared agreement that the poor have to be supported, regardless of their inability to contribute. They also indicate that there is a predominant expectation that if possible, everyone with the capacity to do so should contribute at least moderately to the tax effort. Also, the data suggest that there is no particular demand to focus much pressure on the wealthy. A willingness to contribute more is particularly high in relation to health care, which might indicate a need to find a mechanism to – de facto or de jure – earmark taxes for health spending. As Figure 6.6 shows, if money is used for health and to benefit the poor people’s openness to pay more is particularly high. 6.4 CONCLUSION At least in the short run, funding constraints will remain significant and limit the capacity of low- and middle-income countries to expand their social protection and health care systems. Such countries need to increase their domestic revenues in order to afford expansion. 55

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows How do the informally employed come into this picture? Part of the reform has to aim at improving how the informally employed and the state interact. There is a moral duty to improve access to these services for informally employed people. But also, more fundamentally, there is the potential (if not the need) to find ways to achieve more constructive relationships between these actors, as a necessary step, for any reform in this area to be successful and sustainable. The issue, as with any reform affecting the fiscal contract, concerns the degree to which proposed reforms are politically and fiscally feasible(Burchi et al. 2020). The results presented in this chapter give some indications that might be useful in assessing and potentially expanding the scope for these reforms. The first conclusion is that there is a demand for reform aimed at a more comprehensive social state: the informally employed consider that working on policy areas such as health care and education should be a priority. The second conclusion, however, is that trust in political parties and trade unions as particularly relevant intermediary institutions tasked with aggregating interests and coordinating collective action at the national political level, are rated badly in relative terms. There seems to be a feeling that these institutions care little about the policy priorities and that they have little capacity to change things significantly. This is bad news for anyone supporting such reform(especially if conceived as a democratic and participatory process). Given these perceptions, it is safe to expect that there are only moderate incentives to engage with and support these institutions. Given that these intermediary institutions are essential to generate a socio-political environment in which to identify the scope for comprehensive reform towards more inclusive social protection and health systems, as well as to sustain any reform efforts, the current situation calls rather for pessimism. The data suggest that these institutions are far from being backed from the informally employed and therefore lack the power and legitimacy to shape the agenda. It is important to analyse in the individual context whether these perceptions are legitimate or simply negatively biased perceptions in order to discuss remedies in the countries. But, regardless of the reasons, it is crucial to acknowledge the reticence of the informally employed towards these institutions. strongly support a vision of the state that supports the poor, regardless of their heavily restricted or even inexistent capacity to contribute to the fiscal effort. All these results are important in terms of assessing the scope for different policy proposals to resonate with the informally employed. In any case, these are trends in the overall data, and any attempt to identify the scope for reform should be based on detailed analysis and highly contextualized. All in all, the predisposition of the informally employed to support reforms in this area, as well as to help finance them with tax contributions seems not to be the main obstacle. Far more problematic appears to be creating a socio-political environment in which to pursue and sustain a reform effort. Revenue and spending policy are not purely technical domains and mirror goals that are defined in the socio-political realm. Improving health and social protection systems is also not a purely technical discussion and in fact could be defined as a collective action problem. There seems to be an agreement that everyone would benefit from a situation in which services were better, even if some individuals would have to pay more taxes, but many countries end up in a»low-tax trap«(Mosley 2015). One of the major problems lies in the lack of so-called commitment-devices as well as institutional solutions to coordinate and represent interests in negotiating reforms. The low level of trust in political parties and trade unions, as well as the comparatively high level of tolerance for tax evasion are signals of a major problem concerning the vertical relationship between the informally employed and the state, as well as with many political and social institutions. Addressing these problems is not just desirable but a fundamental if not necessary step towards improving the outlook for any reforms in this area. The third main conclusion to be drawn from the analysis concerns the support for different policy options aimed at increasing state revenue. Revenue can be generated by different actors. For policymakers the key is understanding the support for different options, as well as specific actors’ capacity to resist. The results suggest that the informally employed are open to paying higher contributions themselves. Interestingly, and potentially against what might be expected, the responses do not blame too strongly the lack of contributions of the wealthy(with the exception perhaps of Kenya and Senegal). There seems to be an understanding that funding the state is a collective task and that everyone should contribute as far as possible. The data also send a clear message that, on average, the informally employed 56

Trust in institutions and perceptions of the tax system among informally employed people REFERENCES Berens, S./von Schiller, A.(2017): Taxing higher incomes: What makes the high-income earners consent to more progressive taxation in Latin America?, in: Political Behavior, 39(3). Burchi, F./von Schiller, A./Strupat, C.(2020): Social protection and revenue collection: How they can jointly contribute to strengthening social cohesion, in: International Social Security Review, 73: 13– 32. Gallien, M./van den Boogaard, V.(2021): Informal workers and the state: The politics of connection and disconnection during a global pandemic. IDS Working Paper. Gaspar, V./Amaglobeli, M.D./Garcia-Escribano, M.M./Prady, D./ Soto, M.(2019): Fiscal policy and development: Human, social, and physical investments for the SDGs. IMF Staff Discussion Note.(2019/03). International Monetary Fund. Herb, M.(2005): No representation without taxation? Rents, development, and democracy, in: Comparative Politics, 37(3): 297–317. ILO(2018): Women and men in the informal economy: A statistical picture. Third edition. Geneva: International Labour Office. Meagher, K.(2018): Taxing times: Taxation, divided societies and the informal economy in Northern Nigeria, in: The Journal of Development Studies, 54(1): 1–17. Moore, M.(2020): What is wrong with African tax administration? ICTD Working Paper 111. Brighton, UK: International Centre for Tax and Development. Mosley, P.(2015): Fiscal composition and aid effectiveness: A political economy model, in: World Development, 69: 106–115. APPENDIX Mean across all countries Figure 6.A7 Trust in government by subgroups 3 2 1 0 Rural Urban 1 By Rural/Urban and age quartiles Rural Urban 2 Rural Urban 3 Mean across all countries 3 2 1 0 Rural Urban 1 By Rural/Urban and education level Rural Urban 2 Rural Urban 3 Rural Urban 4 By Rural/Urban and wealth quartiles Mean across all countries 3 2 1 0 Rural Urban 1 Rural Urban 2 Rural Urban 3 Levels of education correspond to 1= No school education, 2= Some primary education, 3= Primary education, 4= Secondary education, 5= University Rural Urban 4 Rural Urban 5 Rural Urban 4 57

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows 7 IS THE INFORMAL ECONOMY IN SUB-SAHARAN AFRICA ORGANIZED? AND IF SO, HOW? Jürgen Schwettmann and Rudolf Traub-Merz 7.1 INTRODUCTION – HISTORICAL BACKGROUND This report explores, on the basis of a survey covering 8,300 households in six sub-Saharan African countries, whether, how and why people employed in the region’s informal economies are organized. This is an aspect that has long been neglected by governments, researchers and development partners. After the wave of independence movements in Africa in the 1960s, many African leaders invested considerable efforts and resources in the promotion of cooperatives, as a means of achieving their vision of African socialism as a home-grown»third way« between capitalism and communism. Targeting mainly rural areas, these»cooperatives« were provided with marketing and supply monopolies and de facto compulsory membership, and placed under stringent government control. All these experiments failed, however, as did the socialist cooperative model promoted in the 1970s and the 1980s in Ethiopia and the former Portuguese colonies in Africa. Government control over, and support for, cooperatives diminished substantially during Africa’s structural adjustment era, which lasted from 1980 to 1999 and affected almost all African countries. Structural adjustment programmes entailed the systematic withdrawal of monopolies, privileges and subsidies hitherto granted to cooperatives. At the same time, frustrated by the failure of state-sponsored cooperative organizations, international donors withdrew much of their support for cooperative development in Africa. All this led to the rapid collapse of most state-sponsored cooperative movements in Africa. The disappearance of the state-controlled cooperative model, however, went hand in hand with the re-emergence of myriad»non-traditional« types of self-help initiatives active in areas such as housing, energy, handicrafts, finance, culture, recycling, transport, marketing, mining and social services, including informal sector associations and other forms of member-based organizations. The latter, largely ignored until the end of the last century, have been moving, during the past twenty years, into the centre of interest of researchers and development practitioners(including the ILO), not least because structural adjustment has accelerated the informalization of African economies, while at the same time curtailing social services and weakening public services. Informal economy actors, overlooked by the state, trade unions and the professional associations of the formal economy have no alternative but to organize themselves in order to obtain a minimum of protection and representation. Until recently, research into the organizational aspects of the informal economy 1 has focused mainly on the role of trade unions in organizing the informal economy, or has analysed specific types of organizations, such as cooperatives, mutual benefit groups and and rotating savings and credit associations(ROSCAs). A comprehensive overview of the organizational strength and diversity of Africa’s informal economy is still missing; the present survey helps to close this gap. 7.2 INFORMAL ECONOMY ORGANIZATIONS: AN OVERVIEW Informal economy actors(independent entrepreneurs, smallholder farmers, the self-employed and informally employed workers) are excluded from much of the security and protection that the state extends to the formal economy; they are not covered by labour law and social protection schemes; they lack recognition, appreciation and representation; and they hardly take part in political and administrative decision-making. Forming associations, cooperatives and self-help groups may be the most promising way in which informal economy actors can obtain economic opportunities, social security, political weight and a shared identity. But do they actually come together to form such organizations? What types of organization do they prefer? And what motivates individuals to become members of a group? We address these issues, and a few others, in what follows. 7.2.1 Organizational density and types of groups The proportion of informal economy actors who are members of a group of any type is shown in Figure 7.1. In the six surveyed countries, about half(51.8 per cent) of all informally employed people are organized in a group. The degree of organization varies significantly between the six countries. Benin has an organizational density of 36.1 per cent compared with 70.5 per cent in Senegal. 1 See, for example: Lindell(2010), Bonner and Spooner(2011), Schwettmann(2018) or the numerous working papers published by WIEGO (Women in the informal economy – globalizing and organizing). 58

Is the informal economy in Sub-Saharan Africa organized? And if so, how? Figure 7.1 Organizational density 80% 70% 70.5% 60% 50% 40% 30% 29.5% 20% 10% 1.82 0% Senegal 55.5% 44.5% 58.0% 42.0% 64.0% 36.0% 1.41 Zambia 1.14 Kenya Organized 1.23 Benin Non-organized 55.2% 44.8% 1.63 Côte d'Ivoire 66.4% 33.6% 1.85 Ethiopia 10.00 9.00 8.00 54.3% 7.00 45.7% 6.00 5.00 4.00 3.00 2.00 1.51 1.00 0.00 6-country average Groups per person Many people are members of several groups at the same time. Multiple membership is particularly widespread in Senegal and Côte d’Ivoire, where over half of all respondents belong to more than one group. The informally employed show a distinct preference for certain types of groups. Figure 7.2 provides a breakdown by type of organization. 2 On average throughout the six countries, savings clubs are the most popular, at 26.7 per cent, followed by religious associations(25.8 per cent), neighbourhood groups(19 per cent), and cooperatives(including credit unions; 13.8 per cent). Those will be called the»Big Four« as they are the predominant type of organization in all six countries. Political parties, cultural groups and other types lag far behind, representing five per cent or less. While the Big Four overshadow the others by far, the relative importance of these four groups varies strongly. Savings clubs are most prevalent in Kenya and Benin, whereas in Zambia and to a lesser extent in Senegal, Ethiopia and Côte d’Ivoire, religious groups attract most members. Cooperatives are more prominent in Zambia than in Senegal and even in Kenya. 3 Côte D’Ivoire, Senegal and Ethiopia count more neighbourhood groups than Zambia, where they are almost non-existent. 7.2.2 A brief introduction to the»Big Four« A savings club or a rotating savings and credit association is an association whose members agree to make regular contributions to a common fund, which is disbursed, in whole or in part, to each contributor in rotation. If ten members contribute the amount of 10 US dollars each per month, then each member would receive 100 US dollars when it is their turn. More sophisticated ROSCA derivates exist, but do not need to be discussed here. In many instances, ROSCAs operate an additional social fund to assist members in an emergency. ROSCAs are widespread not only in sub-Saharan Africa, but in many other parts of the world. A religious group is, as its name implies, a group formed by members of the same faith, who, however, often share additional common bonds, such as gender, location or age group. The main function of these groups is the organization of collective worship, often followed by a social gathering. In addition, many religious groups operate social assistance funds, or organize care and support if a member falls sick. In many African countries, religious groups or faith-based organizations(FBOs) play a key role in providing health care; in Kenya, for example faith-based organizations operate 74 hospitals and 808 health centres(Oliver et al. 2015). As religious groups belong to a faith represented by a national umbrella organization(for example, the Presbyterian Church) they may play a role in local or national politics. Cooperatives correspond to a universal definition contained in ILO Recommendation No. 193(2002) on the Promotion of Cooperatives:»The term ›cooperative‹ means an autonomous association of persons united voluntarily to meet their common economic, social and cultural needs and aspirations through a jointly owned and democratically controlled enterprise«. Cooperatives create, first and foremost, economic opportunities; they are seldom involved in the provision of social services, unless specifically established for that purpose (such as health care or housing cooperatives). Cooperatives may play a societal/political role(empowerment) when organized in horizontal networks and/or vertical structures (see below). Credit unions(or savings and credit cooperative organizations, SACCOs), which are also cooperatives, have often evolved out of ROSCAs. 2 In the first two surveys(Kenya and Benin), sports associations were listed as an option; those were replaced for the remaining countries with neighbourhoods or residential groups as a more popular type. 3 This is in stark contrast to the data published by the Kenya Bureau of Statistics, which in 2016 reported a total of 10.8 million cooperative members in the country – equivalent to 39 per cent of the population in the age bracket 15 to 64 years of age. Neighbourhood associations are probably the most difficult to define because they exist under many different names (farmers’ associations, village associations, urban market associations and so on), and serve many different purposes, including solid waste management, community develop59

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 7.2 Types of groups Kenya 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group Social movement Professional organization Mutual health Trade union Sport association Any other type 13.7% 13.5% 1.9% 3.5% 2.1% 0.4% 2.1% 0.5% 1.1% 3.7% 57.6% Benin 0.0% 10.0% 20.0% 30.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group Social movement Professional organization Mutual health Trade union Sport association Any other type 18.4% 11.8% 2.8% 6.3% 5.9% 4.0% 0.6% 0.4% 2.8% 1.3% 40.0% 50.0% 45.7% Senegal 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 0.0% 10.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group 6.2% 6.3% 4.3% 19.8% 15.2% 30.5% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group 1.3% 5.7% 1.5% Social movement 3.7% Social movement 0.4% Professional organization Mutual health Trade union Any other type 4.6% 4.0% 0.9% 4.6% Professional organization Mutual health Trade union Any other type 0.9% 1.3% 0.4% 1.9% Zambia 20.0% 30.0% 18.7% 23.9% 40.0% 50.0% 44.1% Côte d'Ivoire 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group Social movement Professional organization Mutual health Trade union Any other type 3.9% 1.8% 3.9% 1.5% 1.8% 4.1% 8.2% 17.6% 22.8% 20.3% 14.3% Ethiopia 0.0% 5.0% 15.0% 25.0% 35.0% 45.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group Social movement Professional organization Mutual health Trade union Any other type 1.0% 2.5% 4.1% 5.8% 1.5% 2.4% 1.1% 3.4% 13.3% 25.6% 39.2% 6-country average 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% Savings club Religious group Neighbourhood group Cooperative+credit union Politcal party Cultural group Social movement Professional organization Mutual health Trade union Sport association Any other type 4.6% 3.9% 3.3% 2.5% 2.0% 0.8% 1.9% 3.2% 19.0% 13.8% 26.7% 25.8% Note: Figures for 6-country average do not addup to 100 per cent. This is due to the fact that»neighbourhood group« represents average of 4 countries and»sports association« average of two countries. ment, mutual social assistance, protection against criminals, involvement in local politics, self-regulation of markets, collective farm work such as land clearing and information-­ sharing. The identity element is central to these associations because the neighbourhood(which could be a street, a village or a district of a town) serves as the common bond uniting its members. The activities of neighbourhood associations are related mainly to security and empowerment. Ethiopia is a special case in that the neighbourhood groups – known as »Idir« – were set up originally to assist members in case of an emergency; many Idirs have evolved into general purpose savings clubs, which function similar to ROSCAs. 60

Is the informal economy in Sub-Saharan Africa organized? And if so, how? 7.2.3 Organizational features: size – age – geographic coverage Groups’ membership size and geographic outreach have a considerable influence on the way they organize their operations. A membership size of 100 to 150 individuals may still be small enough to ensure a sufficient level of social cohesion as a basis for democratic governance. Larger groups may have to rely on indirect governance structures (such as delegates elected by certain categories of members), which ensure efficiency, but weaken identity. Smaller groups can assess the needs of members by examining individual cases, while larger groups are more likely to enforce rulesbased regimes and may be less inclined or able to consider individual cases. Informal-economy groups are generally small in membership size. In all survey countries, 70 per cent or more of all groups count 100 members or fewer(Figure 7.3). No more than 3.3 per cent(Côte d’Ivoire) and 6.2 per cent(Kenya) of all groups are large organizations with more than 1,000 members. The extent of a group’s geographic coverage has an impact on the strength of the common bond and the vibrancy of the shared identity that unites its members. One can assume that this bond is stronger in groups that cover just a market or a village than in those that operate in a larger province or even the entire nation. Figure 7.4 shows that two-thirds of all groups cover just one market or a village. This is consistent with the groups’ relatively small membership size(see above). The great majority of informal-economy groups in the six countries are locally rooted. What is surprising is the relatively large share of groups claiming that they cover the entire nation(14.1 per cent on average, rising to 35.5 per cent in Zambia). Religious groups, neighbourhood associations, ROSCAs, cooperatives and credit unions characteristically cover no more than a certain administrative unit, or a population group within an administrative unit. Trade unions, professional organizations and political parties are likely to operate nationwide, but their cumulative share in the group typology is just 8.6 per cent(6.9 per cent in Zambia). The high share of groups operating nationwide in Zambia could perhaps be explained by the high share of religious groups(44.1 per cent) in this country; religious groups represent a particular faith that is present in most cases throughout the country. It might be possible that some respondents, when answering this question, considered the outreach of the faith rather than the coverage of the religious group to which they belong. One can assume that the longevity of an informal economy group can serve as a proxy indicator for its performance and stability; one would further expect that groups emerging in the informal economy have a rather low survival rate because they rarely get any government support, and they are often managed or led by individuals with a low level of education. The survey showed the latter assumption to be false, however: across all six countries and all types almost half(49.6 per cent) are more than five years old. Figure 7.5 provides further details. In all six countries, religious groups have the highest»life-­ expectancy«, followed by cooperatives and savings clubs (Table 7.1); the longevity of religious associations may be due to the fact that they are solidly rooted in their communities, based on a strong sense of shared identity, and often contribute to social security. 4 No explanation could be found for the exceptional longevity of religious groups in Zambia, however, compared with the other four countries. Almost 4 As Bompani(2011: 2) observed:»Religion-driven charitable contri­ butions have been and remain a critical source of welfare and religious organisations remain the most significant non-state providers of basic social services to the poor in Africa. Faith-based organisations (FBOs) expanded or proliferated as a result of economic neo-liberalism as the faithful responded to growing poverty, inequality and social ­exclusion.« Figure 7.3 Membership size of groups 70% 60% 64.3% 53.0% 50.8% 23.1% 24.2% 9.4% 15.4% 14.4% 23.1% 35.1% 19.3% 12.0% 24.6% 24.8% 30.7% 38.7% 34.8% 40.1% 50% 40% 30% 5.1% 4.1% 6.2% 3.4% 3.3% 3.2% 1.9% 12.4% 8.0% 3.6% 6.7% 4.3% 20% 10% 0% Less than 10 10 to 50 51 to 100 101 to 1.000 Above 1.000 Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 61

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows half(47.4 per cent) of the 194 religious groups identified by the survey in Zambia have existed for over 21 years. 7.2.4 Higher-level structures The term»higher-level structure« means a secondary organization, such as a union or a federation, that is formed by the primary informal-economy groups. In Kenya, for example, the Kenya National Alliance of Street Vendors and Informal Traders(KENASVIT) serves as an umbrella organization of urban associations of street vendors, whereas in Benin, a national informal economy association is in the process of being established. Cooperatives, credit unions and trade unions show a greater propensity than other types to form sectoral and/or national unions. Higher-level structures enhance the Figure 7.4 Geographical coverage of groups Ethiopia Côte d'Ivoire Benin Kenya Neighbourhood group 20.8% 57.6% 16.2% 5.2% Religious group 11.5% 61.6% 16.1% 1.0% 9.8% Savings club 0.0% 83.3% 16.7% Cooperative 15.7% 60.7% 13.5% 1.1% 9.0% Neighbourhood group 0.8% 83.5% 12.4% 2.5% 0.8% Religious group 0.6% Savings club 9.6% 52.0% 74.8% 17.1% 6.3% 24.0% 11.1% 3.7% 0.7% Cooperative 0.9% Religious group 3.2% Savings club 4.8% 66.7% 63.5% 68.0% 8.1% 21.6% 2.7% 6.3% 4.8% 22.2% 20.2% 6.1% 0.9% Cooperative 3.2% 61.3% 16.1% 19.4% Religious group 11.8% Savings club 23.8% Cooperative 6.7% Neighbourhood group 33.3% 68.6% 64.0% 20.0% 100.0% 26.7% 9.8% 3.9% 5.9% 5.5% 4.5% 2.3% 13.3% Religious group 5.9% 26.8% 7.2% 5.6% 54.5% Savings club Cooperative 2.3% 31.4% 78.2% 47.7% 4.7% 4.7% 11.6% 6.8% 5.3% 7.5% Neighbourhood group 0.0% 86.9% 9.1% 2.0% 2.0% Religious group 1.7% Savings club 9.5% 62.7% 68.3% 19.2% 7.3% 18.0% 9.0% 3.2% 1.1% Cooperative 4.0% 60.0% 24.0% 8.0% 4.0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% A certain market or part of a market A village or a quarter of a town A district or administrative department A province. county or federal state The entire nation Zambia Senegal Figure 7.5 Groups’ lifetimes 35% 30% 25% 20% 15% 10% 5% 4.5% 0% less 1 year Senegal 19.6% 26.3% 22.2% 14.7% 12.7% 1-2 years Zambia 3-5 years Kenya Benin 6-10 years Côte d'Ivoire 11-12 years 21 years a.m. Ethiopia 6-country average 62

Is the informal economy in Sub-Saharan Africa organized? And if so, how? Table 7.1 Median age of groups Country Senegal Zambia Kenya Benin Côte d’Ivoire Ethiopia Average Religious group 10 20 10  6  8 10 10 Median age of informal-economy groups in years Cooperative (not including credit unions) Savings club 5 4 5 1 8 3 6 4 6 2 5 2 6 3 Neighbourhood association  5  2 n.a. n.a.  4 10  5 All groups  7  6  4  5  5 10  6 Figure 7.6 Affiliation with higher-level structures 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 90.3% 9.7% Senegal 64.9% 35.1% Zambia 85.8% 14.2% Kenya 91.4% 86.2% 8.6% Benin 13.8% Côte d'Ivoire 93.8% 85.4% 6.3% 14.6% Ethiopia 6-country average Belongs to a higher-level structure Does not belong to a higher-level structure affiliated groups’ bargaining power and political clout; they can also generate economies of scale and scope when they carry out business operations. As shown in Figure 7.6, only a minority of primary ­informal-economy groups belong to a higher-level structure. In Kenya and Zambia, one can assume that most cooperatives and credit unions are affiliated with their respective apex organizations, namely CAK and KUSCCO(Kenya), and ZCF and NASCU(Zambia). Such apex bodies do not exist in Ethiopia and in the three French-speaking countries(with the exception of the national association of savings and credit cooperatives in Côte d’Ivoire, UNACOOPEC-CI). Trade unions and mutual benefit groups are typically affiliated with a national federation, but the percentage of these two categories among all informal-economy groups is very small (see Figure 7.2). 7.2.5 How»formal« are the groups? Informal-economy groups, associations and cooperatives can build a bridge between a country’s formal and informal economies, provided they are officially registered and/or recognized. In that case, the group enters the realm of the formal economy, while its members remain in the informal. 5 The survey concluded that well over half of all groups in the two English-speaking countries(Kenya and Zambia) possess an official registration certificate, while this proportion drops to one quarter in the French-speaking countries. The difference may be due, at least partly, to the fact that the French-speaking countries have inherited the centralized administrative system prevailing in France, which means that people living outside the national capital may find it costly and cumbersome to register their group. Moreover, the proportion of groups that are registered depends on their type: ROSCAs and religious groups may not need(or desire) to register, while cooperatives and credit unions must necessarily register so as to be able to function. And thirdly, some groups may avoid registering because this may entail the payment of taxes and other duties. Those groups are likely to fall into the grey area of being recognized, but not registered. Moreover, it can be assumed that informal economy groups that keep accounts are more»formal« than those that do not. Certain categories of groups, such as cooperatives and credit unions, are obliged by law to keep accounts, while one would assume that others, such as religious associations, which do not carry out any financial transactions, would not see the need to keep accounts. The survey revealed that in all six countries, over 70 per cent of the respondents de5 For example: most motorcycle-taxi drivers in Cotonou, Benin, are members of a professional association. The association’s chair represents the members vis-à-vis the police, local authorities and insurance companies, which are hardly accessible to the individual taxi drivers, many of whom are illiterate. 63

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows clared that their groups kept accounts. This percentage is significantly higher than the proportion of groups that are registered(see above), although it appears that countries with a higher percentage of registered groups also have a higher percentage of groups that keep accounts. A third aspect indicating a certain degree of formality is the stability of informal-economy groups. One would assume that informal-economy groups are relatively volatile, because their members frequently change location and/or occupation. It was found, however, that the great majority of groups operate permanently; this proportion is highest in countries with the largest proportion of registered groups. The number of groups that function only seasonally(for example, during the harvesting season), or those that come together on an ad hoc basis(for example, to assist a member in case of an emergency) is quite low in all three countries. The survey therefore provides three indicators which allow us to measure the groups’ degree of formality and stability. They are summarized in Table 7.2. The table allows us to observe a(i) rather high overall»formality indicator« of 73.0 per cent, on average, and(ii) that the two English-speaking countries yield higher scores in all three categories than the three French-speaking countries. Ethiopia, as an Amharic-speaking country, appears in-between these two groups. 7.3 INFORMAL ECONOMY ORGANIZATIONS – MEMBER MOTIVATION 7.3.1 A multitude of objectives Some informal economy groups are set up specifically for a single purpose, for example to collect and secure savings. Most groups pursue several objectives at the same time, Table 7.2 Formality indicator Indicator Registered or formally recognized Account keeping Permanent operations Average Benin Degree of formality of groups(%) Kenya Senegal Zambia 46.4 67.1 55.3 85.2 73.9 86.2 77.7 90.5 76.5 86.5 84.3 88.9 65.6 79.9 72.4 88.2 Côte d’Ivoire 39.1 71.1 78.1 62.8 Ethiopia 56.0 75.3 74.7 68.7 Average 58.2 79.2 81.5 73.0 Table 7.3 A typology of informal-economy groups ROSCAs Religious groups Shared identity Organize church services A typology of informal-economy groups Economic opportunity Social security Generate capital for investments Often operate social assistance funds May operate social funds Cooperatives Neighbourhood groups Political parties Cultural groups Professional organizations Social ­movements Mutual benefit groups Sports ­associations Trade unions Organize neighbours for a variety of communal services Formed to promote a certain political orientation Organized around common identifiers, such as customs, language, ethnic group All members exercise the same profession Members belong to a shared identity Organize collective sports events Organize economic support services May render certain economic support services May organize social services if established for that purpose May operate social assistance funds and mutual care May operate a social assistance fund Organize health insurance through mutuality Members belong to the same profession or employment status Societal empowerment Take part in politics if organized at the national level Gain bargaining power through unions and federations Can play a role in local or municipal politics Formed to gain political power Mostly formed as a lobby group to influence politics Formed primarily to push for political change Primarily engaged in collective bargaining 64

Is the informal economy in Sub-Saharan Africa organized? And if so, how? however – they are multipurpose in nature. To illustrate this fact, we have developed the typology of» opportunity – security – empowerment – identity« as an organizing structure to categorize the objectives of the different groups. In this typology: –»Opportunity« refers to the potential for improving the economic situation of a group member. –» Security« stands for the quest for social security and protection. –» Empowerment« encapsulates the desire to obtain voice and representation vis-à-vis the state and/or a powerful private sector. While the type of service is closely related to the typology above, the aspect»shared identity« does not appear in Figure 7, but can be subsumed under»others«. The survey revealed that: – 45.5 per cent of all respondents sought services to generate or expand economic opportunities; this included financial services(33.6 per cent) and other economic or business support services(11.9 per cent); – over one-third(38.2 per cent) joined a group in a quest for social services, that is, to enhance social security; – very few respondents(3.7 per cent) cited societal empowerment as the primary reason to join a group; –» Identity« means the pursuit of a collective cultural, religious or ethnic distinctiveness. – a relatively large share(12.6 per cent) reported having joined the group for»other reasons«. Table 7.3 shows that almost all types of groups fall under more than one of these categories. The cells shaded in dark-grey indicate the group’s primary function; those in light-grey show the secondary function. 7.3.2 Motives for joining a group Actors in the informal economy may join a group because they expect benefits, such as the provision of essential services, which may be of a financial, material or intangible nature. Respondents can choose from a list of services, as shown in Figure 7.7. There are considerable variations between the six countries, which depend to a large extent on the type of groups that predominate. For example, in Benin and Kenya, the two countries in which savings clubs predominate, a majority of respondents joined a group in quest of financial services. The share of»other reasons« is particularly high in Zambia and Senegal and corresponds to the prominence of religious groups. We can therefore assume that these»other reasons« are mainly of a religious or cultural nature, meaning that respondents wanted to strengthen their shared identity. In fact, the survey found that two features, namely ethnic origin and faith/religion, were the most important factors determining shared identity in all six countries. Figure 7.7 Reasons for joining a group 0% 10% 20% 30% 40% 50% 60% 70% 80% Social services 38.2% Financial servces 33.6% Others(mostly identity/religious) 12.6% Economic/business/professional services 11.9% Societal/political functions 3.7% Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 6-country average 65

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows 7.3.3 Group support for members in case of crisis Social security coverage includes the provision of financial support to people who are unable to obtain an income. Traditional social insurance schemes and institutions are, most of the time, confined to workers in the formal economy. Persons in informal employment, who are excluded from such schemes, must find alternative ways of sharing risks through collective action and mutual help. This is reflected in Figure 7.8. – On average, 61.6 per cent of all groups support their members in case of an emergency. This share is nearly double the share of members who reported social services as the primary reason for joining a group. This means that many groups provide social assistance even if their primary purpose is a different one. The average of 61.6 per cent, however, masks considerable differences: in Kenya, four-fifths(81.5 per cent) of all groups assist members in case of an emergency, whereas in Senegal this share drops to just one-third(32.7 per cent). – While on average, 60.8 per cent of all religious groups provide assistance to members in case of an emergency, this proportion drops to just a quarter(26.3 per cent) in Senegal. Further research would be warranted to better understand the reasons for this considerable difference. 7.3.4 Satisfaction with group performance Are members satisfied with the performance of their groups? This aspect was measured through six indicators, related to the usefulness of group services, their timely delivery, the honesty and competence of group leaders, and the appropriacy of member fees. 6 Figure 7.9 shows that, across all countries and indicators, a large majority of members expressed satisfaction with their group’s performance. This may not be surprising because(a) membership of groups is voluntary, meaning that unhappy members would simply leave; and(b) group leaders are elected democratically(in most cases at least), so that underperforming leaders would not be re- elected. In all six countries, over 80 per cent of all respondents were satisfied, if not very satisfied with the performance of their groups. An equally high share(86.8 per cent on average) of respondents expressed satisfaction with the timeliness of services rendered by their groups. An even higher proportion(90 per cent) of interviewees thought of their group leaders as being competent, while 89.4 per cent were convinced of their honesty. If we look at the Big Four, some slight variations occur. Religious groups across all six countries were considered by the respondents to be most effective(over 95 per cent answered»satisfied« or»very satisfied«), while cooperatives scored about 10 percentage points less than the religious groups. This notwithstanding, we can still conclude, with a few exceptions, groups meet their members’ expectations. 7.4 IMPACT OF SOCIO-DEMOGRAPHIC AND SOCIO-ECONOMIC FEATURES In addition to exploring functional differences between groups and the reasons motivating individuals to join them the survey also assessed socio-demographic and socio-economic variables, such as age, gender, education, employment status and income. For reasons of concision, we provide only an overview of these variables, except in cases where significant variations exist. Figure 7.8 Group assistance to members in case of an emergency 0% 20% 6 Data for this last criterion(membership fees) are not available for ­Ethiopia. ­ 40% 60% 80% 100% Cooperative Savings club Religious group Neighbourhood group national average(all groups) Senegal Zambia Kenya Benin 63.9% 68.4% 60.8% 56.9% Côte d’Ivoire 61.6% Ethiopia 6-country average 66

Is the informal economy in Sub-Saharan Africa organized? And if so, how? Figure 7.9 Satisfaction with group performance 100% 90% 90.5% 86.9% 90.0% 89.4% 87.4% 80% 70% 60% 50% 40% 30% 20% 10% 0% Services are useful Services are timely Leadership is competent Leadership is honest Membership fees are reasonable Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 6-country average Figure 7.10 Socio-economic drivers of group membership 80% 60% 40% 20% 0% Senegal Membership by gender Zambia Kenya Male Benin Female Côte d’Ivoire Ethiopia 80% 60% 40% 20% 0% Senegal Membership by location Zambia Kenya Benin Côte d’Ivoire Ethiopia Urban Rural 80% 60% 40% 20% 0% Membership by age Senegal Zambia 15-24 years Kenya Benin Côte Ethiopia d’Ivoire 25-34 years 35-44 years 45-54 years 55-64 years 65 years a.m. 80% Membership by income 60% 40% 20% 0% Senegal Zambia Kenya Benin Côte d’Ivoire less 0.5 MW 0.5 MW-1 MW>1 MW-2 MW Ethiopia >2 MW 80% 60% 40% 20% 0% Membership by employment status Senegal Zambia Kenya Employers Employees Benin Côte Ethiopia d’Ivoire Own-account workers Family workers 80% 60% 40% 20% 0% Membership by education Senegal Zambia No school Kenya Benin Some primary Côte Ethiopia d’Ivoire Primary Secondary University 67

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 7.10 indicates that: – Age does not significantly influence the decision to join a group, with the exception of the 15–24 years cohort, which seems less prone to affiliate with a group than others. – Education levels are a stimulus for group membership, especially in Kenya, where respondents with a university education are nearly twice as likely to join a group than those without any schooling(whereas the opposite is true in Ethiopia). – Income levels do not have a significant impact on organizational density, except in Zambia. – Overall, women are just as likely to join a group as men. However, gender influences people’s priorities in group preferences. In Côte d’Ivoire, Senegal and Zambia, women have a significantly higher propensity to join savings clubs: the proportion of female members exceeds that of men by 30 per cent or more. Men are keener to join cooperatives and neighbourhood groups. – Similarly, urban informal economy actors have more or less the same propensity to form a group as those living in rural areas. – Informal economy employers are more prepared to join a group than employees, with own-account workers falling in-between. It appears that the overall effect of socio-demographic and socio-economic variables on group membership is small. This does not mean that income and employment status, for example, do not play any role. It could rather mean that income inequality is not high enough to make a difference in group affiliation. 7.5 CONCLUSIONS The survey confirms that many informal economy operators in sub-Saharan Africa are indeed organized. They have established various types of groups and associations, in particular savings clubs, religious associations, cooperatives and neighbourhood groups. The motivation for organizing and the selection of groups is to a lesser extent influenced by gender, age, location, education, employment status or income levels. Most individuals joined their groups in quest of economic opportunity, social security or shared identity, much less so in search of societal empowerment. Informal sector groups are often multi-functional. Even those groups that focus on the provision of economic or financial services, or are bound by a common faith, frequently assist members facing a social emergency. The uniformity of group members’ socio-demographic profiles indicates that social polarization within informal employment is not a major driver for getting organized. The quest for social security through risk-pooling and mutual assistance is a cross-cutting concern that drives the organizing efforts in the informal economy. As demonstrated above, the organizations exhibit a high degree of formality, which enables them to connect their members(who remain in the informal economy) with enterprises in the formal economy, and with local authorities. The survey revealed that an overwhelming majority of respondents expressed a high level of satisfaction with the services their group provides, and with the competence and trustworthiness of the groups’ leaders. Informal economy groups seldom experience governance problems because they are based on voluntary membership, active member participation, and democratic management structures. The groups are generally small, and mostly confined to a village, a market or a district of a town. The survey exposed substantial differences between the countries with regard to organizational density – which is twice as high in Senegal as it is Kenya – as well as with regard to the popularity of different types of groups. These differences may stem from the countries’ respective historical backgrounds:(i) the influence of the French or British colonial regimes;(ii) post-independence regimes(various derivates of African Socialism in Benin, Senegal and Zambia, pro-Western policies in Côte d’Ivoire and Kenya); and(iii) the attitudes of more recent governments towards economic informality. To fully grasp the causes of the observed differences the present survey needs to be complemented by historical research into the origins and patterns of group formation in the informal economies of the six countries. The characteristics and operations of informal-economy groups are influenced by the essential features of informal employment: – The majority of informal-economy actors work on their own account; those who work for others, or employ others(including family members), are a minority. Moreover, individuals in the informal economy may frequently change their employment status when new economic opportunities arise, when their business falters, or when a social crisis looms. Consequently, labour relations in the informal economy fluctuate as well, which makes groups that draw members from a specific category only (employers, employed workers, own-account workers) less attractive to those who frequently change their employment status. Consequently, informal economy workers and operators rarely organize in accordance with employment relationships. – Informal economy operations take place outside the application of civil, fiscal and labour laws, and outside administrative regulations. Informal economy operators are generally excluded from formal social security schemes, which means that life risks can be minimized only through the establishment of self-help groups and mutual benefit associations. Consequently, social security emerges as a cross-cutting concern, permeating all types of groups and associations, including those that are set up for economic or financial purposes. 68

Is the informal economy in Sub-Saharan Africa organized? And if so, how? A few policy recommendations can be drawn from the centrality of social security, when situated in the context of the Chapters 2 and 4 on»use of medical care« and»health insurance« in this report: – Universal health coverage is an objective that can be pursued through the joint efforts of state-led social policy initiatives and informal economy organizations. The survey shows that(a) respondents identify improved health services as their most important demand for better government services;(b) the majority of those who are currently not members of a health insurance scheme would be willing to join such a scheme; and(c) a majority of informal economy actors organized in a group would be interested in joining a health insurance scheme. – All six survey countries have forms of community-based health insurance(CBHI) which link the self-organizing interests of civil society, the administrative duties of local authorities and the technical, organizational and financial capacities of central government together. Senegal, Benin and Côte d’Ivoire have embraced the French model of» mutuelles«(mutual health benefit groups), while Kenya, Zambia and Ethiopia follow variants of community development approaches. These CBHI groups are local structures that cooperate with local authorities. REFERENCES Bompani, B.(2011): Religion and Faith-based Organisations in Africa: the forgotten actors, 27 September; available at: https:// www.e-ir.info/pdf/14362(accessed on 24.10.2021) Bonner, D./Spooner, C.(2011): Organizing in the Informal Economy: A Challenge for Trade Unions, in: IPG,2, Berlin. Lindell, I.(2010): Africa’s informal workers, collective agency, alliances and transnational organizing in urban Africa. London: Zed Books. Olivier, J. et al.(2015): Understanding the roles of faith-based ­health-care providers in Africa: review of the evidence with a focus on magnitude, reach, cost, and satisfaction, in: The Lancet, 386(10005), 1765–1775. Schwettmann, J.(2018): Organizing[in] the informal economy – the role of trade unions, Berlin: FES. – Of particular importance is the design of a hybrid model of health insurance in which three layers cooperate: local member-based organizations collect fees and play a role in health communication and prevention; local authorities manage the local supply and demand for health services, and refer patients to higher levels of the health service hierarchy, if necessary; and central governments are responsible for risk pooling and the application of country-wide rules. In this regard, informal-economy organizations are useful not only to their members, but also to governments and civil society at all levels. The organizing potential in the informal economy is a precious tool for building universal health coverage. In order to tap this potential, however, governments should establish an appropriate policy, legal, administrative and institutional environment that facilitates the emergence, official recognition and operations of informal economy groups. 69

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows 8 ORGANIZING IN THE INFORMAL ECONOMY – THE INTEREST OF INFORMAL LABOUR IN TRADE UNIONS Rudolf Traub-Merz 8.1 INTRODUCTION In many African countries, a debate is gaining momentum on whether trade unions, in order to turn declining membership around, should open themselves up to hitherto ignored segments of the labour market and start organizing the informally employed. Is the informal labour force a future recruitment terrain for trade unions? No representative survey has yet been undertaken to inquire directly in face-to-face interviews what those who obtain their livelihoods from informal employment really think about trade unions. Our joint survey sheds light on the willingness of the informal labour force to cooperate with trade unions or even join them as members or affiliates. 1 8.2 TRADE UNIONS ARE UNKNOWN TO MANY, BUT THOSE WHO KNOW THEM HAVE A POSITIVE VIEW In general, members of the informal labour force 2 – which is made up of the four status groups employers, employees, own-account workers, and family support workers – have a low level of awareness of trade unions(Figure 8.1). In a six-country average, only 39.4 per cent reported having ever heard about trade unions. Knowledge of trade unions varies tremendously between countries. In the Côte d’Ivoire, Kenya and Benin, more than half of the people in informal employment have at least heard about trade unions. In Senegal and Zambia, this share is down to a third or less; for Ethiopia, we have to conclude that nearly the whole segment of the informal employed are unaware of trade unions’ existence. Certainly, urban dwellers are more familiar with trade union affairs, but despite their proximity to urban-oriented union activities, large segments of the urban informal labour force are again not aware of what trade unions stand for. 3 This low level of knowledge points to a structural distance between the informal labour force and trade unions. Only between 1.2 per cent(Ethiopia) and 5.5 per cent(Senegal) of the informal labour force have ever participated in trade union activities. Personal contacts with a trade unionist are similarly non-existent. The low level of interconnection is a strong indicator that trade unions have largely neglected the informal economy. Those segments of the informal labour force who have some knowledge of trade unions, however, have a fairly positive view of them. In all countries, a clear majority emphasize the importance of trade unions and reject the view that trade unions should be banned(Figure 8.2). With some reservations in Zambia and Kenya, most even believe that trade unions should become stronger. Views are more scattered with regard to performance indicators. Around half of the informal labour forces are convinced that trade unions improve the social situation of many, while somewhat fewer believe in their efficiency. In different intensities, these views also point to the need for trade unions to improve their performance, become more efficient and less corrupt. Overall, trade unionism as a principle is fully endorsed, but its practical implementation is not supported to the same extent. 8.3 INTEREST OF THE INFORMAL LABOUR FORCE IN TRADE UNIONS To assess the interest of informal labour in trade unions and to appraise the willingness to be organized by trade unions a four-step approach is applied. 4 (i) Respondents first identify those who they perceive to be the main beneficiaries of trade union activities. This helps to put expectations about trade unions into a socio-structural perspective. 1 For a detailed account see Traub-Merz(2020). 2 Use of the term»informal labour force« is not technically correct here, as it includes those who are unemployed. We use»informal ­labour force« synonymously with»informal employment« because (i) in countries with no unemployment insurance, unemployment ­remains unregistered, and(ii) the key problem is not unemployment, but underemployment. Because of the absence of unemployment benefits, people have to accept even minor jobs to make(a little) ­income. (ii) Respondents then appraise trade unions’ organizational independence from the government. 3 For an account of urban/rural differences in four countries, see TraubMerz(2020: 9). 4 Reference is always to those who have heard of trade unions. 70

Organizing in the informal economy – the interest of informal labour in trade unions Figure 8.1 Knowing trade unions and interacting with them 0% 10% 20% 30% Côte d’Ivoire 6.4% 4.8% Kenya 2.4% 2.1% Benin 1.4% 2.0% 6-country average 2.9% 2.9% Senegal 4.6% 5.5% Zambia 0.8% 1.6% Ethiopia 1.7% 1.2% 12.2% 23.0% 40% 39.4% 33.2% 50% 60% 70% 61.7% 55.3% 51.1% Heard about trade unions Contacted by a trade unionist Participated in trade union activity Question:»Have you ever heard about trade unions and what they are doing«. Answer options:»I have not heard about trade unions«;»I have heard about trade unions but do not know what they are doing«;»I know a little about trade unions«;»I have heard about trade unions and I know what they are doing«. Note: The questions»have you had contacts with trade unions?« and»have you participated in trade union activity?« were directed only to those who had at least heard about trade unions. We assume that those who have never heard about trade unions had equally never had contacts with trade unionists or participated in trade union activity and therefore refer the answer to the total sample, that is: all informal employment. Figure 8.2 Perceptions of trade unions(TU) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 79.9% 79.1% 68.9% 56.9% 45.2% 40.3% TU are important organisations TU should be banned – statement rejected TU should become stronger TU improve the social situation of many TU are inefficient – TU are corrupt – statement rejected statement rejected Senegal Zambia Kenya Benin Côte d'Ivoire Ethiopia 6-country average Note 1: Answer options were»fully agree«;»mostly agree«;»partly agree/partly disagree«;»mostly disagree«;»fully disagree«. Figure 2 shows only the answer options»fully agree« and»mostly agree«, which are added. Note 2: For the statements»trade unions should be banned«;»trade unions are inefficient« and»trade unions are corrupt«, the figure shows those respondents who disapproved of the statement. Note 3: Reference is made to those who have heard of trade unions. 71

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 8.3 Groups perceived to benefit from trade union activities 60% 40% 23.2% 20% 36.2% 34.9% 0% –20% –40% –29.2% –60% –80% –51.6% Senegal Zambia Kenya 43.3% 4.5% 28.4% 29.6% 25.7% –12.4% –69.0% Benin –48.3% –42.1% Côte d’Ivoire 5-country average (excl. Ethiopia) Ethiopia Employees of the government Workers in the informal economy Own-account workers Employers in the private sector The Poor Everybody The government Farmers The Unemployed Values express net agreement: agree(fully+mostly) – disagree(fully+mostly); answer option partly agree/partly disagree is not considered. Question:»Do you agree that trade unions improve the situation of the following groups: a. the poor; b. workers in the informal economy; c. farmers; d. the unemployed; e. employees of the government; f. the government; g. employers in the private sector; h. own-account workers; i. everybody«. (iii) Respondents then declare their interest in trade union membership. (iv) The interest in trade union membership is then put to a »reality check« by asking whether those willing to take up membership are equally willing to pay membership fees. 8.3.1 Who are perceived as beneficiaries of trade union activities? To avoid leading questions the list of potential beneficiaries of trade union actions has broad scope and includes groups not usually seen as being part of trade unions’»recruitment terrain«. The list includes employers in the private sector and the government. The findings are summarized in Figure 8.3: – In five out of six countries, informal labour has a clear understanding that trade unions act as interest-based organizations. This is the case when we can identify groups who benefit from their actions while others do not. Ethiopia is the exception as trade unions are perceived to help all groups. – In the five countries in which trade unions are perceived to act as interest-based organizations, some views are uniform across borders and always identify the same cluster of beneficiaries while others differ and talk of national features or distinctions. Because of the special findings, we exclude Ethiopia in the following remarks and return to it later. In all other countries, the informal labour force usually perceives two groups as being usually neglected by trade unions:»the unemployed« and the all-embracing»everybody«. A third group,»own-account workers« is equally not seen as a trade union clientele, but here respondents in Benin differ. Their views are split: about half see trade unions as taking care of own-account workers. Opinions diverge with regard to other social groups. Senegalese and Ivorian respondents reject the view that trade unions support farmers, the poor and workers in the informal economy. Respondents in the other three countries are mostly undecided and split their views into»yes« and»no« votes of equal strength. Those gaining from trade union activities are believed to be three groups in particular:»government employees«,»private sector employers« and»government«. While the levels of support for them differ the three are the front runners in each of the five countries or belong at least to the four highest-rated categories. What do these findings on the perception of informal labour imply? Clearly, trade unions are not seen as being primarily caretakers of informal labour interests. When employers, formal sector employees and the government are primarily picked as beneficiaries we can conclude two things: (i) Employers and employees are not viewed as being embedded in adversarial employment relations. Respondents rather subscribe to the view that trade unions help both sides. Instead of being perceived as conflict agents trade unions are seen as forming a common bond between actors on different sides of employment relations in the formal economy. 72

Organizing in the informal economy – the interest of informal labour in trade unions Figure 8.4 Trade union independence from government 70% 60% 50% 40% 30% 45.4% 36.1% 40.0% 37.1% 20% 44.5% 38.2% 48.0% 49.3% 48.2% 44.3% 64.3% 21.8% 10% 0% Senegal Zambia Kenya TUs are NOT independent Benin Côte d’Ivoire TUs are independent Ethiopia Question:»Do you agree with the following statement? Trade unions in your country are independent from the government«. Note: Figures add up to 100 per cent if answer option»partly agree/partly disagree« is added. (ii) The actual social divide is not perceived to exist between employers and employees, but between the formal sector 5 and the informal sector. A majority of the informal labour force regards trade unions as an organizational transmission belt for the benefit of groups in the formal sector. The orientation towards the formal sector should, however, not distract from the point that even though it does not garner a similar level of support, large numbers of respondents also credit trade unions with taking care of»workers in the informal economy«. In Benin and Kenya, such assessments are even put forward by a slight majority, while in Senegal and Zambia more than 30 per cent support such an appraisal. The positive views of so many on the interest of trade unions in the informal economy are hardly rooted in practices on the ground. In Benin, some unions have been recruiting»workers in the informal economy« for a number of years and the Union Nationale des Syndicats des Travailleurs du Benin (UNSTB) appears to be doing quite well. 6 But in Kenya, the Kenya Congress of Trade Unions(KCTU) declared its readiness to campaign for the membership of informal workers only in 2019. Here, the high numbers with which trade unions are accredited as already taking care of»workers in the informal economy« can hardly be related to actual trade union activities. Instead, trade unions appear to enjoy a leap of faith, which is founded on the hope that they may arrive and take care of informal labour interests. Trade unions can take advantage of such»trust credit« when entering the informal economy in earnest. 8.3.2 Trade unions – independent from government? The view expressed by so many, that trade unions tend to act in the interest of the government immediately raises the question of the organizational separation between them. Are trade unions viewed as independent organizations or perceived as bodies guided and controlled by the state? In a liberal market-based society organizational independence from the state is regarded as a pre-condition for assessing the quality of interest representation. Only an independent organization is believed to decide freely whether it wants to support government policies or oppose them; a dependent organization is constrained. Figure 8.4 shows the views of informal labour on trade union independence from government: – In five of the six countries the views of informal labour are split into two camps of similar magnitude. While one camp is convinced that trade unions are indeed independent, the other is not and subscribes to the view that trade unions are an organizational extension of the state. – Ethiopia represents a special case in that a strong majority declare trade unions to be an adjunct to the state. The predominant view in Ethiopia on the dependency of trade unions allows us to reconsider the previous views with regard to beneficiaries of trade union activities(see above). If trade unions are guided by the government, they cannot play the role of an interest-based actor who helps some but 5 The term»formal sector« is used metaphorically, not to refer to the formal economy. 6 The UNSTB reports their membership base for 2017 as follows: of the 84,526 members, 71 per cent belong to informally affiliated unions (Uhlandssekretariat(2018: 1-2). 73

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 8.5 Interest in individual trade union membership 0% 5% 10% Côte d’Ivoire 9.9% Kenya 7.8% Senegal 7.6% 6-country average 6.4% Ethiopia 6.3% Zambia Benin 4.7% 4.2% 8.2% 15% 20% 16.0% 14.2% 18.2% 25% 23.0% 20.4% 30% 27.1% Interest in membership of those who have heard of TUs Interest in membership of all informal labour Note: For Kenya and Benin, respondents who answered the knowledge question with option B:»I have heard about trade unions but do not know what they are doing« were not presented with the membership question. Some of them might also be interested in joining trade unions, so the figures for Kenya and Benin are on the low side and should be adjusted upward. In a four-country average, 17.7 per cent declared an interest in membership even among those who had just heard of trade unions, but did not know more. If this rate is used for adjustment, the interest in membership would increase in Kenya to 14.7 per cent of total informal labour and in Benin to 10.0 per cent. not others, but is supposed to perform a government proper task, namely, to assist all. 7 8.3.3 Interest in trade union membership When trade unions reflect on strategies to expand their organizational base to the informal workforce, they may either want to recruit individuals or look for affiliation from established groups. For this reason, respondents were asked to disclose their interest twice: first, whether they were interested in joining trade unions as individual members; secondly, whether they would be in favour of their own group affiliating to a trade union. The first question was directed at all those who had heard of trade unions, while the second was raised only to those who are members of a group. Individual membership Interest in individual membership is shown in Figure 8.5. On a six-country average, 18.2 per cent of those who have heard of trade unions reported an interest in joining one as member. The differences between the countries are pronounced. In Ethiopia, the interest level reaches 27.1 per cent, while in Benin, it is only 8.2 per cent. 7 The argument holds true for those in the other countries who doubt the independence of trade unions. If trade unions are perceived as state-controlled it is easy to identify them as organizations who care primarily for formal-sector groups. In many countries trade ­unions do receive some government budget support to finance some of their infrastructural expenditures. Such support may not auto­matically infringe on the autonomy of trade unions. In Benin, despite such government support, trade unions have mobilized times and again against anti-democratic tendencies, despite receiving an annual ­subsidy. There are various ways to read these figures. Interest in membership is more than twice as high(6.4 per cent) as the level of personal experience with trade union affairs, which only 2.9 per cent of respondents reported having(see Figure 8.1). Obviously, membership interest is not based on familiarity with unions alone but derives from the hope that membership will come with a personal benefit. If, however, interest in membership is related to the informal labour force as a whole, the 6.4 per cent interest appears low. There are fluctuations between the countries, but in no case does interest go above 10 per cent. Affiliation of groups There appears to be a high interest among the members of established groups in being affiliated to trade unions as a collective. On a six-country average, a third of all respondents (33.6 per cent) are in favour of such an idea(Figure 8.6). Ethiopia and Kenya lead the field, with over 40 per cent, while Senegal and Côte d’Ivoire follow from the end(28.4 and 24.2 per cent). The declared interest in group affiliation appears to be slightly lower than the declared interest in individual membership. This, however, stems from the fact that only about half of informal labour holds membership of a group(see Chapter 7). A cross tabulation would show that respondents in favour of individual membership are usually in favour of group affiliation as well and that only a small group speak out against group affiliation, but opt for individual membership. 74

Organizing in the informal economy – the interest of informal labour in trade unions Figure 8.6 Interest in group affiliation to trade unions 0% 5% 10% Ethiopia 1.6% Kenya 4.6% Benin 5.6% 6-country average 4.9% Zambia 3.2% Senegal 6.0% Côte d’Ivoire 8.3% 15% 20% 25% 30% 35% 40% 45% 41.3% 40.7% 34.4% 33.6% 32.8% 28.4% 24.2% Interest in group affiliation by group members Interest in group affiliation by all informal labour Discussion The interest in individual membership and group affiliation has to be put into a wider perspective. Informal employment is the dominant form of employment and makes up around 90 per cent of it(see ILO 2018: Table B1). Low relative figures thus turn into large absolute figures. In Benin, some 200,000 persons appear willing to join a union; in Kenya, with a much larger employment sector, one million or more declare such an interest. 8 In contrast to current union membership, the potential of informal labour to fill up union ranks becomes visible. Unfortunately, figures for trade union membership are notoriously unreliable and outdated. They are usually based on self-reporting by trade unions, which have a tendency to set figures on the higher side to indicate organizational strength, and on the lower side if they have to pay fees to other organizations. information on trade unions becomes available, these figures may shoot up and a much higher percentage of the informally employed may be interested in joining. – A declaration of interest is not the same as actually applying for membership. Triggers and opportunities matter. In the end, an organizing strategy that offers interested segments of the informal labour force opportunities to participate in trade union actions may be needed to transform declared interest into actual membership. – Asking respondents to disclose their interest in affiliation establishes a perception of individuals and not the will of a group. Conclusions on a»group consciousness« can be made if the interest of the group leadership is considered. By looking at members’ views only, we are still far from ascertaining what a group may decide. But a simple back-of-the-envelope calculation can help. If we assume a trade union density of 50 per cent for formal employment, we would arrive at some five per cent of all employment, a figure which just equals the 5.5 per cent interest of informal labour. A lower density rate would put the interest of informal labour beyond existing membership levels. Even if these figures are only rough approximations, the potential for membership through intakes of informal labour appears to be huge. Winning over the already motivated in informal employment could easily double, if not treble the current membership of trade unions. Viewed from this angle, trade unions’ organizational base could be substantially enlarged. Several issues should be emphasized here: – Recommendations to trade unions on how to recruit are not a purpose of this study. It is left to trade unions to balance the costs and benefits of different recruitment strategies. What this survey does points out, however, is the fact that despite the generally low level of awareness of trade unionism in many segments of the informal economy and the widespread lack of personal experience with trade unions there is a significant potential for recruitment. Sizeable groups of the informal labour force are sympathetic to trade unions and open to engage if trade unions»go for them«. 8.4 WILLINGNESS TO PAY MEMBERSHIP FEES – Declarations of interest in trade union membership are made within a social environment in which trade unions are not(yet) known or seen as a relevant player. If more 8 For a more detailed discussion see Traub-Merz(2020: 21). Self-financing is central to membership organizations’ sustainability and independence. Self-financing implies that the bulk of revenue is derived from membership dues. A trade union that offers services to affiliates without demanding monetary compensation in the form of affiliation dues or service payments risks draining its resources, which will lead to a 75

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Figure 8.7 Willingness to pay membership fees 70% 60% 50% 52.3% 52.1% 40% 30% 20% 21.7% 10% 0% my group should pay full membership fees my group should be granted a discount trade unions should provide services for free Senegal Zambia Kenya Benin Côte d’Ivoire 5-country average Ethiopia Question:»If your organization became an affiliate of a trade union, what should be the financial relationship between your group and the trade union? A. My group should pay membership fees to the trade unions and not demand a reduction. B. My group should be granted a discount as my organization is too poor to pay full fees to a trade union. C. The trade union should not ask for fees but provide us with services for free«. Note: Answers are overlapping and figures do not add up to 100 per cent. rearrangement of the financing of other activities. Providing services for a nominal or no fee at all may be acceptable for certain reasons or as an initial incentive. But rendering services without monetary compensation may in the long run challenge the financial viability of a union and may even threaten the fulfilment of its other tasks. How do members of a group regard possible financial contributions if they opt for affiliation with a trade union? We raised this question in three versions, each time lowering the monetary benchmark. First, we asked about people’s preparedness to engage in a»standard relationship« by paying full dues. Then we asked about views on a discounted fee; and finally, we asked whether union services should be provided free of charge. Ethiopia is again a case to be treated separately. For the other five countries(Figure 8.7) we can conclude that: – Full membership fees are rejected by a large majority. On average, only 21.7 per cent were of the view that their group should pay full membership once affiliated. In Kenya, the financial pressure experienced by service providers is better understood; but even here, a mere 35 per cent were in favour of full dues. – Demands for a discount received wide support. On average, 52.3 per cent responded accordingly. Most of those who did not demand a discount went one step further and opted for free services. Ethiopia is a different case in that each of the three options received support from about half of the respondents. Views thus overlap considerably and many who at first agreed on full fees, then switched in favour of a discount and finally requested free membership. While logically, the three fee options are mutually exclusive, in real life, the final outcome on affiliation fees is usually the result of bargaining. One hopes to get services for free but if rejected, may accept a pay with or without a discount. 9 Ethiopian respondents clearly switched their views more often than respondents in other countries, an attitude that may be linked to their views on trade unions being state-controlled. The high percentage of Ethiopian respondents who were willing in the first instance to pay full dues may reflect an authoritarian political environment in which it is best to pay what is demanded, even if it comes from a state-controlled trade union. While the views in the other five countries appear to reflect a genuine preparedness to pay membership fees, the Ethiopian figures should be treated with some caution and not taken at face value. For the five country block, we may conclude that interest in affiliation is based mainly on the expectation that an asymmetrical relationship will evolve between service provider and receiver. The asymmetry is articulated most strongly by those who would prefer»free-rider« status. It is also articulated by those who want a subsidy in the form of a discount. With varying intensity, respondents identified a disparity between trade unions and their own groups. Their own group is seen – There are two large camps of similar weight, which demand either a discount or free affiliation. Only a small group was ready to pay full membership fees. 9 It is rather difficult to replicate a bargaining arena for group affiliation in an interview situation. Overlapping may be avoided if respondents can select only one option out of the three. This again, however, does not correspond to a bargaining situation. 76

Organizing in the informal economy – the interest of informal labour in trade unions as too poor to pay(in full), while trade unions are perceived as»rich« enough to forgo monetary compensation. If informal labour groups look for membership in trade unions, most hope for services for which they do not have to pay(in full). Membership thus implies a social contract that allows for monetary transfers from the stronger to the weaker, from richer to poorer. 8.5 SUMMARY AND CONCLUSIONS In summarizing the empirical results, we find several overriding perceptions: – In general, members of the informal labour force have a low level of awareness of trade unions. In all six countries, a clear majority have never heard of trade unions or have no knowledge of what they do or stand for. The low level of understanding of what trade unions stand for points to a structural distance between the informal labour force and trade unions. It is a strong indication that the informal economy has been largely neglected by trade unions. – Those segments of the informal labour force who have heard about trade unions, however, perceive the role unions play in their countries as fairly positive. In all six countries, large majorities emphasize the importance of trade unions, believe that they help to improve the social situation for many people, and want trade unions to be strong and grow even stronger. In different intensities, however, the views expressed point to a need for trade unions to improve their performance, become more efficient and be transparent in their activities to avoid accusations of corruption. Trade unionism as a principle is fully endorsed, while practical implementation may not be supported to the same extent. – In five out of six countries, a majority regard trade unions as interest-based organizations engaging primarily for the benefit of groups in the formal sector, that is, governments, employers and employees. They are not seen as being embedded in adversarial employment relations but as benefiting equally from trade union action. The real social divide is perceived to exist between the formal and informal economies, or between formal and informal employment. – Views differ about the role of trade unions in relation to the informal labour force which in our list is represented by »workers in the informal economy« and»own-­account workers«. In Benin, a good half see both groups as trade union beneficiaries, while in Kenya, informal workers receive the thumbs-up, while own-account workers do not. In Senegal, Côte d’Ivoire and Zambia, the informal labour force is seen as being neglected by trade unions. The partly positive views in some countries can hardly be explained by current trade union engagement in favour of this labour segment and appears to stem more from a leap of faith than an expression of experience. – Views of the informal labour force in Ethiopia differ substantially from views in other countries in several regards. The key reason appears to be that trade unions are largely seen as an adjunct to the state, which does not allow them to act as interest-based organizations. – Some segments of the informal labour force vest high hopes in trade unions. This is confirmed by their interest in becoming members. As low as these shares may look at first glance, they become sizeable when expressed in absolute figures. Recruiting the»already motivated« of the informal labour force for membership could easily double, if not treble or quadruple current trade union membership. – Affiliating to trade unions is also prominent within the already-organized segments of the informal labour force. Even though our survey does not assess the views of groups but of group members, and only a minority articulated an interest in(affiliate) membership, hopes of closer cooperation are widespread. – Not surprisingly, interest in paying membership fees is less articulate. Trade unions are viewed as organizations with good financial resources and members of the informal labour force who are willing to become members are interested in a discount, if not free services. Implications for trade union recruitment strategies In what ways can our poll findings contribute to trade union reflections on organizing segments of informal workers? This largely depends on the direction unions want to embark on in order to gain new members. Basically, there appears to be a matrix of four different approaches. On one hand, unions have to decide whether they will opt for membership of individuals or for the affiliation of groups. On the other hand, they have to choose whether new members could come from all branches of the labour market(open recruitment) or whether certain niches or branches should be prioritized (strategic recruitment). If unions opt for open recruitment, they may be confronted with high and over-optimistic expectations. They may be unable to use their organizational weight for collective bargaining if just a few individuals are willing to join; and they may face unbearable costs to maintain communication with members living in dispersed settlements. Strategic recruitment makes it possible to focus on workplaces with agglomerated labour and on economic branches in which trade unions already command some socio-economic knowledge or even have an organizational footprint. Under such circumstances, group affiliation could be the result of a negotiated deal, which would make it possible to adjust expectations concerning gains and costs to a realistic level and build a common platform based on the strengths and weaknesses of both sides. Organizational synergies would become the overriding motivation. 77

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows The same may be the case if not groups but non-organized individuals were the target of recruitment strategies. Here again, trade unions may select economic sectors and focus on branches in which they are already present. Resources that are in short supply, such as sending activists to locations with a high number of workplaces, can be used more economically. Limiting recruitment targets to fewer locations with more workplaces may make it possible to come up with a unified set of demands to smooth the way for collective bargaining. REFERENCES ILO(2018): Women and men in the informal economy: a statistical picture (third edition). Geneva: International Labour Office. Traub-Merz, R.(2020): Interest of Informal Labour in Trade Unions. Findings from Representative Country Surveys in Sub-Saharan Africa. Berlin: Friedrich-Ebert-Stiftung. Uhlandssekretariat(2018): Labour Market Profile – Benin, 2018; avail­able at: https://www.ulandssekretariatet.dk/wp-content/uploads/ 2020/02/Benin_lmp_2018.pdf(accessed on 20.8.2020). 78

Technical notes on the survey 9 TECHNICAL NOTES ON THE SURVEY Michael Frosch, Richard Houessou and Abel Oyuke Project team The survey project is realized as a joint project between the Friedrich-Ebert-Stiftung(FES) as lead institution, and the International Labour Organization(ILO) and the German Institute of Development and Sustainability(IDOS) as cooperating partners. National survey institutes(NSI) that are part of the Afrobarometer network are the implementing partners in the survey countries. Additional technical support, including data management, is provided by the Institute for Development Studies(IDS), University of Nairobi. Members of these institutions met on various occasions to jointly develop the questionnaire and to agree on details of the survey protocol. Objectives of the survey The main objectives of the survey are to get a better understanding of the social situation of the informally employed with regard to health issues, views on trust in state and government, self-organization and interest in trade unions. Operational definition of informal employment The ILO(2019: 6) provides a definition of informal employment for three categories of workers: (i)»Employees are considered to have informal jobs if their employment relationship is, in law or in practice, not subject to national labour legislation, income taxation, social protection or entitlement to certain employment benefits(advance notice of dismissal, severance pay, paid annual or sick leave, etc.).« (ii)»Employers and own-account workers are considered to be in informal employment when their economic units belong to the informal sector. The informal sector is a subset of household enterprises(not constituted as separate legal entities, independent of their owners) that produce goods or services for sale in the market, and that do not have a complete set of accounts and/or are not registered under national legislations.« (iii)»Contributing family workers are, by definition, informally employed, regardless of whether they work in formal or informal sector enterprises.« The definition of informal employment rests on the definition of employment. The definition of employment was changed in 2013 with the adoption of the 19th ICLS resolution. Employment became more closely linked to remuneration and own-use production of goods was excluded from employment and recognized as one of five forms of work. Our survey is aligned to these changes, and own-use production of goods i.e. farming, raising animals or fishing intended to only or mainly be consumed by the household, is listed as subsistence production and excluded from the survey and the concept of informal employment. To identify informal employment and its various categories, the survey used the following operational definitions: Informal employees: paid job with reference to an employer’s contribution to a public or private pension scheme. Reference is to a main job. If employers did not pay contributions, employees were grouped as informal. Informal employers and own-account workers(including agriculture activities mainly intended for the market): informality is defined by non-registration in the national registry, which is used for company taxation. Contributing family workers: defined, by default, as having an informal job due to the informal nature of jobs held by contributing family workers. In the case of multiple jobs: the main job is defined as the job in which the respondent usually works the highest number of hours for pay or profit. Only the main job was considered for identifying informal jobs and secondary jobs were not considered. The questionnaire The questionnaire originally consisted of 143 main questions, which can be broken down into several sections. The key groups are: Personal and sociographic data, such as age, sex, status within the household, education, respondents’ employment situation and income, household assets. Health issues: respondents’ experience with health services; respondents’ resources in financing medical treatment; health insurance, including reasons for joining/not joining. 79

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows Trust in state/government: respondents’ expectations with regard to services provided by the state; respondents’ views on the state’s capacity and willingness to provide services; respondents’ views on paying taxes and fees in exchange for services; respondents’ views on social inequality, social justice and the role of social policy. Self-organization and interest representation: where, why and how do respondents organize themselves in groups? Do respondents feel that their interests are represented by their group? What are the respondents’ views on trade unions? With the outbreak of the Covid-19 pandemic, some questions were added on how people have been responding and on government lockdown policies. Sample design and sampling process The sample is designed as a representative cross-section of all informally employed citizens aged 15 or above in a given country. Every citizen who corresponds to the criteria of age and informal employment is selected randomly for interview. The selected sample is achieved by random selection methods at every stage of sampling and the application of probability sampling based on population size. The sampling process is based on stratification of the country into regions. Regions are further classified as urban or rural. Primary sampling units(PSU) – sometimes referred to as enumeration areas(EAs) – are the smallest geographical unit for which reliable population data is obtainable. The primary sampling units are selected from each stratum based on its share of the national population, and further allocated based on the urban/rural divide of each stratum. A total of eight households were clustered in each enumeration area for logistical efficiency and to lower the cost of contacting the sample. The national sample of 1,200 households is large enough to make inferences about all informally employed persons who are 15 years of age or above with an average margin of sampling error of no more than plus or minus 2.8 per cent at a 95 per cent confidence level. The sampling process is structured in four stages:(i) selection of enumeration areas;(ii) selection of sampling start points; (iii) selection of households; and(iv) selection of random respondents for interview. This sampling method is applied across all survey countries as a standard design. The survey uses a standard questionnaire that contains identical or functionally equivalent questions. Because of this standardization, responses can be compared across countries and over time. – Selecting enumeration areas(EA) Based on the latest and updated population census the national statistical offices randomly select enumeration areas for each stratum and respective rural/urban divide, based on probability proportional to size of population. For a sample of N=1,200 the statistical office randomly selects 150 enumeration areas for a given country survey – that is, 150 x 8= 1,200 interviews. – Selecting the sampling start-points(SSPs) for each enumeration area Across the survey countries, no complete lists of households were available from which the sample could be randomly drawn. The next best method therefore is to use physical maps(provided by the office of statistics). A sampling start-point(SSP) 1 is marked on the map and field teams travel as close as possible to it, or to housing settlements nearest to it. A second SSP is selected as a reserve or substitute in case the initial SSP is inappropriate or inaccessible. – Selecting the household – walking pattern of interview teams The interviewers start walking away from the physical start-point, with interviewer 1 walking towards the sun; interviewer 2 in the opposite direction; interviewers 3 and 4 at a 90-degree angle to the right and left. With this walking pattern, all four directions are covered. By counting households on both sides of the walking path, household No. 5 is selected as the first household for the interview and household No. 15 for the second interview. 2 If the interview cannot take place because nobody is at home or the interview starts but cannot be finished the walk continues to the next household on the same side of the road or opposite(household No. 6), while the second interview is done in household No. 16. 1 Random selection of a start-point uses a grid. A ruler is placed along the top of the map and another along the side. A table of random numbers is then used to select pairs of numbers, one for the top axis and one for the side axis, resulting in a random combination. A line is then drawn on the map horizontal to the number chosen on the side, and another line is drawn vertical to the number chosen on the top. The point on the map where these two lines intersect is the sampling start-point. Each X-Y pair of numbers from the random number table can be used only once. 2 Special rules were applied in the case of multi-storey buildings, widely scattered households and settlements within commercial farms. 80

LIST OF FIGURES LIST OF FIGURES Figure 2.1 Use of medical  9 Figure 2.2 Regular use of medical care, by urban/rural residence 10 Figure 2.3 Use of medical care, by gender 10 Figure 2.4 Use of medical care, by income 11 Figure 2.5 Demand for better state services(top priority) 13 Figure 2.6 Demand for better health services(first priority), by urban–rural residence 13 Figure 2.7 Demand for better health services and pensions (first priority), by age groups 14 Figure 2.8 Demand for better health services(first priority), by income 15 Figure 2.9 Demand for better health services(first priority), by gender 15 Figure 2.10 Demand for health as first priority and use of medical care 16 Figure 3.1 Sources of finance for medical treatment 20 Figure 3.2 Sources of finance for medical treatment by health insurance membership(two countries) 20 Figure 3.3 Payment for medical treatment by sources of finance and urban–rural residence 22 Figure 3.4 Sources of financing medical care by gender 23 Figure 3.5 Modes of payments for medical treatment by income classes 23 Figure 3.6 Use of medical care if funding has to come from asset sales or going into debt 24 Figure 4.1 Membership of health insurance schemes, free access to national health services and direct financing by the employer 26 Figure 4.2 Distribution of workers in informal employment who are members of health insurance schemes according to: 28 Figure 4.3 The household dimension of health insurance coverage 29 Figure 4.4 Willingness of workers in informal employment (non-members) to join a health insurance scheme 30 Figure 4.5 Reasons not to join a health insurance scheme 31 Figure 4.6 Main reasons for not joining a health insurance scheme by level of income 32 Figure 4.7 Preferred frequency of contributions among those willing to join a health insurance scheme 33 Figure 4.8 Level of contributions suggested by non-members in relation to various benchmarks 34 Figure 5.1 Loss of work opportunity or job because of Covid-19 42 Figure 5.2 Reported decrease of household incomes because of the Covid-19 pandemic 43 Figure 5.3 Income poverty before and during the pandemic 43 Figure 5.4 Increase in income poverty by regions as a result of Covid-19 44 Figure 5.5 Differences between the new and old income-poor 45 Figure 5.6 Government assistance received during the pandemic 46 Figure 5.7 Income poverty by the provision of government assistance 46 Figure 6.1 Trust in institutions among informally employed people (% of total, by country) 51 Figure 6.2 Perceptions of the informally employed of how much different institutions care about their main priorities (% of total, by country) 51 Figure 6.3 Perceptions of the degree to which different institutions can make a major difference in solving informally employed people’s main priorities 52 Figure 6.4 Perceived fairness of public service delivery in exchange for taxes and fees 54 Figure 6.5 Opinions on the degree to which the wealthy contribute enough taxes 55 Figure 6.6 Openness to increase contributions if money is used for health 55 Figure 6.A7 Trust in government by subgroups 57 Figure 7.1 Organizational density 59 Figure 7.2 Types of groups 60 Figure 7.3 Membership size of groups 61 Figure 7.4 Geographical coverage of groups 62 Figure 7.5 Groups’ lifetimes 62 Figure 7.6 Affiliation with higher-level structures 63 Figure 7.7 Reasons for joining a group 65 Figure 7.8 Group assistance to members in case of an emergency 66 Figure 7.9 Satisfaction with group performance 67 Figure 7.10 Socio-economic drivers of group membership 67 Figure 8.1 Knowing trade unions and interacting with them 71 Figure 8.2 Perceptions of trade unions(TU) 71 Figure 8.3 Groups perceived to benefit from trade union activities 72 Figure 8.4 Trade union independence from government 73 Figure 8.5 Interest in individual trade union membership 74 Figure 8.6 Interest in group affiliation to trade unions 75 Figure 8.7 Willingness to pay membership fees 76 81

FRIEDRICH-EBERT-STIFTUNG – A majority working in the shadows LIST OF TABLES Table 1.1 Traits of selected countries 5 Table 2.1 Health expenditures and availability of medical staff (various indicators, 2018) 12 Table 2.A2 Households with cases of sickness and no cases of sickness over the past 12 months 17 Table 2.A3 Use of medical care in households with cases of sickness by age of head of household 17 Chi²-testing for use of medical care by age group of head of household 18 Table 2.A4 Use of medical care in households with cases of sickness by education of head of household 18 Chi²-testing for use of medical care and education of head of household 18 Table 3.1 Out-of-pocket payments as share of total health spending(2018; per cent) 20 Table 4.1	Values of contributions in existing schemes and com­parison of»spontaneous amounts« provided in Benin 34 Table 4.A2 Membership of health insurance schemes by sex(frequency) 38 Table 4.A3 Membership of health insurance schemes rural/urban residence(frequency) 38 Table 4.A4 Membership of health insurance schemes(includes SHI and PHI) by level of education(frequency) 38 Table 4.A5 Membership of health insurance schemes(includes SHI and PHI) by individual labour income(frequency) 38 Table 4.A6 Willingness to join by sex(frequency) 39 Table 4.A7 Willingness to join and use by level of education (frequency) 39 Table 4.A8 Willingness to join by employment status(frequency) 39 Table 4.A9 Reasons for unwillingness to join a health insurance scheme by employment status(frequency) 40 Table 5.A1 Characteristics of the new and old income-poor in Ethiopia 48 Table 5.A2 Characteristics of the new and old income-poor in Côte d’Ivoire 48 Table 6.1 Stated main priorities among informally employed people (unweighted average for all countries in the sample) 50 Table 6.2 Perceptions of compliance with taxes and fees among the informally employed 53 Table 6.3 Perceptions of tax evasion 53 Table 6.4 Opinions on who should pay taxes and fees(unweighted average for all countries in the sample) 54 Table 7.1 Median age of groups 63 Table 7.2 Formality indicator 64 Table 7.3 A typology of informal-economy groups 64 82

 ABOUT THE AUTHORS IMPRINT Bonnet, Florence: Labour market and informal economy specialist, Department of Conditions of Work and Equality Department(WORKQUALITY), International Labour Organization(ILO). Frosch, Michael: Senior Statistician, Department of Statistics, International Labour Organization(ILO). Houessou, Richard: AfroBarometer, Project Manager for francophone countries. Published by: Friedrich-Ebert-Stiftung e.V. Godesberger Allee 149 53175 Bonn Germany Email: info@fes.de Issuing Division: International Cooperation Division Leininger, Julia: Head of programme, Transformation of political(dis-)order, German Institute of Development and Sustainability(IDOS). Maihack, Henrik: Head of Africa Department, FriedrichEbert-Stiftung(FES). Responsibility for content and editing: Henrik Maihack| Head of Africa Department Phone:+49-30-269-35-7440| Fax:+49-30-269-35-9217 www.fes.de/en/africa-department Contact/Orders: Janine.Kaliga@fes.de Öhm, Manfred: Head of Finance and Administration Division, Friedrich-Ebert-Stiftung(FES). Oyuke, Abel: AfroBarometer, Project Manager for East Africa. Schiller von, Armin: Researcher, Transformation of political (dis-)order, German Institute of Development and Sustainability(IDOS). Schwettmann, Jürgen: Consultant, informal economy and the social and solidarity economy. Strupat, Christoph: Senior Researcher, Transformation of Economic and Social Systems, German Institute of Development and Sustainability(IDOS). Traub-Merz, Rudolf: Consultant. Coordinator of FESIDOS-ILO survey project»Informal Employment, Social Security and Political Trust in sub-Saharan Africa«. Typesetting: Ludger Stallmeister Cover design: Werbestudio Zum weissen Roessl, Susanne Noé The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung. Commercial use of media published by the Friedrich-Ebert-Stiftung(FES) is not permitted without the written consent of the FES. Publications by the Friedrich-Ebert-Stiftung may not be used for electioneering purposes. This publication is printed on paper from sustainable forestry. ISBN: 978-3-98628-143-4 © 2022 ABOUT THIS STUDY This publication is part of the project»Informal Employment, Social Security and Political Trust in sub-Saharan Africa«. You can find further publications of this project at https://www.fes.de/referat-afrika/publikationen. www.fes.de/bibliothek/fes-publikationen