1 Introduction

In most African countries, social insurance has played a limited role in the provision of social protection.Footnote 1 Less than 10 per cent of the economically active population in Sub-Saharan Africa is covered by a social insurance scheme, most of which consist of old-age pensions and access to healthcare (OECD 2017, 11). While coverage rates in North Africa are higher, large parts of the population still remain excluded. The limited reach of social insurance is closely linked to low levels of industrialisation and high levels of informality that have persisted in African countries since the colonial period (Barchiesi 2019, 48). The bifurcation of labour markets led to the creation of two distinct segments of society with unequal social rights and entitlement to public welfare. In the light of these unfavourable economic and labour market conditions, what caused the introduction of social insurance in African countries?

In recent years, a certain consensus has emerged regarding the influence of the former colonial powers in introducing and shaping social protection schemes across Africa (Midgley and Piachaud 2011; Schmitt 2015; Künzler 2016). Taking this as a departure point, we argue that the introduction of the first social insurance programmes in African countries is however not the result of a simple extension of social policies from European countries to their colonies. The first old-age pension schemes in the form of provident funds emerged as a result of imperial staffing of public administration and distinguished between European and “indigenous” civil servants in colonial parlance. It was only with the new political and social arrangements that emerged during the historical period of decolonialisation that social insurance schemes turned into political reality. Specifically, we find their introduction to have been caused by appropriation of the colonial state by the formerly colonial subjects. As the latter gained in influence, they used and moulded colonial structures and policies along their domestic political constellation of forces. In this process, the pension schemes for domestic civil servants included more and more people as their relative number grew, while new insurance schemes were invented to cover permanent public employees as well as additional pressure groups.

Moreover, our analysis suggests the existence of two additional mechanisms that set in after independence. In industrialising countries with strong labour union movements, such as Tunisia, the new governments introduced broad social insurance schemes for private sector workers as a means of labour incorporation. In mainly agrarian societies with low levels of labour activism, such as Uganda, economic stagnation and low tax incomes seem to have motivated the top-down creation of pension funds for a small group of formal employees as a means of public resource accumulation, for example the mobilisation of capital that could be used for the purpose of state-directed development (cf. Kohli 2004). Overall, our hypothesised mechanisms underline the crucial importance of local agency, revealing both an adjustment to changing circumstances and clear intentions.

To demonstrate the above-mentioned causal mechanisms, we compare the most different cases of Tunisia and Uganda. Both countries have social insurance systems extending into the private sector, whereby more industrialised Tunisia has a diversified system providing old-age, disability and survivor pensions, healthcare and, for some professional groups, family, maternity, disease and death allowances, and additional cash benefits for survivors and in the case of work accidents (IILS 2011, 72). In 2019, 58 per cent of the Tunisian population were covered by at least old-age and health insurance, public and private sectors combined (CRES 2019). In agrarian Uganda, social insurance is limited to old-age protection only.Footnote 2 Here, 12.6 per cent of the estimated active labour force and 5.6 per cent of the total population were covered by a public provident fund in 2018 (Munyambonera et al. 2018).Footnote 3 As political entities, the two countries have little in common except a period of colonisation by a European power, during which political, economic and social structures developed that left their imprint on both as post-colonial states. Tunisia was a former French protectorate and a settler colony, economically relatively diversified (trade, agriculture, mining) and did not go through any major armed conflict following independence. Uganda, a former British colony, relied on enforced cash-crop production, which created a gentry of export-oriented farmers, contrasting with far less prosperous northern and eastern provinces. Huge differences in political traditions and the distribution of wealth paved Uganda’s way into an extended period of political violence between 1971 and 1987. In that period, practically all public institutions decayed.

The empirical analysis is based on a close reading of policy reports, official statistics and secondary literature, as well as information obtained during research stays in Tunisia (March–May 2019) and Uganda (November–December 2018 and October 2019). Our detailed study addresses two major stalemates when it comes to social policy research in African countries: first, its contextual focus on the more developed parts of the world, and second, the severe lack of case studies on the country level (Hickey et al. 2018, 6). We hereby hope to contribute to advancing theorising about social policy, which to this day remains largely based on West European and North American cases.

The remainder of this chapter is structured as follows: the next section discusses existing research on the introduction of social insurance with a specific focus on African countries. Then the causal mechanisms underlying the introduction and extension of social insurance will be specified. The subsequent case studies present the emergence of social insurance schemes in decolonising and early independent Tunisia and Uganda. Finally, the conclusion considers broader implications of the study for social protection in Africa.

2 The Emergence of Social Insurance in African Countries

There is general agreement today that the former colonial powers shaped the development of social insurance schemes in Global South countries. Specifically, it has been shown that the former French colonies introduced social insurance schemes resembling those in France, while the former British colonies adopted central provident funds as a transition to a pay-as-you-go insurance scheme (Hu and Manning 2010, 143). With the British Colonial Development and Welfare Act of 1939, the creation of the Fonds d’Investissement pour le Developpement Economique et Sociale in 1946 and the gradual extension of French legislation, including the Social Security Act of 1945, to French overseas departments after 1946, it indeed appears as if the introduction of public social protection in African countries was just a prolongation of politics in the metropoles. However, this perspective neglects that colonial expansion was initially a response to the social crises in the metropoles (Elwitt 1967) and that the first social protection schemes in the colonies were “for whites only” (Eckert 2019, 153).Footnote 4 Moreover, colonial arrangements and social legislation were far from homogeneous, even within empires. The French empire, often described as highly centralised and having “conceived colonies as extensions from France” (Künzler 2016, 4), spanned from French overseas departments to formally autonomous protectorates where French law did not directly apply. Many studies further tend to ignore the politics behind the introduction of social protection for the “indigenous populations”. Indeed, more detailed country studies have shown that European settlers used their political weight to veto more inclusive policies emanating from central government and that colonial officials boycotted or delayed implementation in the hope of saving the old power alliances (Lewis 2000; Eckert 2004; Künzler 2020). Overall, most “diffusion” studies underestimate the effects of local conditions and the impact of national political forces in shaping social protection, including the introduction and design of social insurance schemes (also see Seekings 2020).

In this regard, it seems useful to note that across the world, pensions and health benefits were first introduced for members of the military and civil servants, and only gradually extended to include more and more people, and to cover a wider range of eventualities (see, e.g., Hannah 1986; Turner et al. 2020). In their work on Latin American countries, Collier and Collier (1991) have convincingly linked this expansion to rising levels of industrialisation and labour market formalisation, which created pressure to integrate labour as a political and economic actor. Part of the process of labour incorporation was the expansion of social insurance schemes to the urban working class (Huber and Stephens 2005, 620). In his work on decolonialisation and the labour question in Africa, Cooper (1996) explicitly emphasises the role of social protection in the development and social integration of the new, “modern” urban African worker. Despite the proliferation of related ideas across the empires, more comprehensive social insurance schemes for workers emerged only after independence. This was not least due to the powerful combination of European employers, government, armed forces and white workers that hindered the organisation of African workers (Orr 1966, 73–74). Indeed, colonial governments ultimately considered more encompassing social protection too expensive (Eckert 2004, 475). Basic social services in health and education, for example, were delegated to a patchwork of religious institutions, in British colonies predominantly Christian missions (Scully and Jawad 2019, 557). These findings suggest that the introduction of more encompassing social insurance schemes depends not only on labour mobilisation itself, but also on the growing political influence of national labour unions after independence.

Still, the existing literature has difficulties explaining why social insurance emerged in countries with largely agrarian economies where labour never emerged as a pressure group. Particularly south of the Sahara, public social protection and social insurance achieve low levels of coverage because of the limited importance of formal, institutionalised labour markets (Eckert 2004, 468). This is particularly the case in the countryside, where most people continue to rely on informal and kinship-based forms of social protection. Scholars have insisted on the key role of the International Labour Organization (ILO) in spreading norms of full employment and universal social insurance (Hu and Manning 2010; Schmitt et al. 2015). Yet, related studies often remain silent on what made African governments invest in the introduction of social insurance, given that many agrarian countries were not pressed by the same social question as their industrialised counterparts.Footnote 5 Here, evidence from Sub-Saharan Africa and South-East Asia offers a promising perspective, which links the introduction or expansion of pension funds to their potential for increasing the capital available to the state (Gerdes 1971; Kuhlmann and Nullmeier, 2021).

In sum, the existing literature suggests that late-colonial actor constellations, including colonial authorities, domestic elites and emergent pressure groups, are crucial for the emergence of social insurance in African countries. However, related claims often remain restricted to a general level. Therefore, the next section will specify a number of causal mechanisms behind the introduction of social insurance schemes in decolonising Africa, which we derived from the literature and refined in the course of our case comparison.

3 Mechanisms Behind Social Insurance in African Countries: Imperial Staffing, Appropriation, Labour Incorporation, Public Resource Accumulation

In most African countries, social insurance was introduced in the global historical period of decolonialisation. The term usually refers to the three decades after 1945, but first signs of the declining legitimacy of the colonial project, including anticolonial unrest, the internationalisation of colonialism, and projects and expectations of reform, already emerged with World War I (Jansen and Osterhammel 2017, 38–42). It is therefore in this larger context that we situate our mechanisms.

While the course of decolonialisation was specific in each case, there is general consensus in the literature that the historical events of the time formed a mutually sustaining context for each single colony (see, e.g., Albertini 1971; Cooper 1996; Fieldhouse 1986). Since World War I, the League of Nations and later the United Nations had promoted self-government and controversies about the empires abounded, both outside and inside the colonies.Footnote 6 In the early 1950s, the idea of “self-rule” and “decolonialisation” as an engineered transfer of power to “’trustworthy’ indigenous leaders” (Jansen and Osterhammel 2017, 4) gained ground among imperial powers. This was not least because of US pressure in the Cold War context, as Soviet solidarity with the colonies and the Chinese revolution threatened to tip the balance between East and West. Conversely, African populations felt on the verge of achieving independence, and national liberation movements began to negotiate with the colonial governments throughout the continent. Their negotiating position improved as wars defending the status quo turned out to be ever costlier. This was the case in French Indochina (1946–1954), Kenya (1953–56), Algeria (1954–1962) and Cameroon (1955–1960). Following formal independence, domestic elites gradually took over while colonial officials, public employees and businesses prepared their departure. By explicitly situating our analysis in this period, we emphasise the presence of the past in the development of social policies in each country case but also on the international level.

3.1 Imperial Staffing Mechanism

Quite obviously, the imperial projects would have been impossible to realise without “indigenous” staff in the colonial administration, who were not only needed in terms of human resources but also as links and communicators between the imperial centre and the local populations (Darwin 2012, chapter 7). The World Wars increased the need for bureaucratic control and thus additional staff in the colonies. At the same time, anticolonial unrest and expectations of reform were on the rise, with French and English colonies—especially in North Africa—worrying about pro-German sympathies and the loyalty of their “indigenous” staff (Jansen and Osterhammel 2017, chapter 2). This shift in power relations motivated the colonial governments to make a number of concessions to strategically relevant groups, including the entitlement to pensions for African civil servants and soldiers fighting in the imperial armies. While these first pensions were accessible to only a small part of the respective populations and did not yet take the form of social insurance, they form the starting point for the introduction of the ensuing social insurance systems. Note that while in North Africa, nationalist movements and reform demands emerged by the First World War, many Sub-Saharan countries saw similar developments only shortly before or during World War II.

3.2 Appropriation Mechanism

Several authors have argued that the process of decolonisation must be understood in terms of an appropriation of colonial structures and policies, not a rupture (Bayart 1989; Cooper 1996; Bayart et al. 2007; see also Fanon 1961). Indeed, appropriation of governance forms facilitated cohabitation between colonial authorities and “indigenous” elites and, following independence, prevented violent resistance to the new order (Rinke et al. 2012). With regard to the introduction of social insurance, we find appropriation to have been triggered by the eroding legitimacy of the colonial projects, which motivated late-colonial governments to invest in the political and social integration of domestic populations into the colonial order. This involved partial autonomy and the “indigenisation” of the public sector, for example the inclusion of ever more “natives” into civil service positions. Domestic elites used the opportunities offered by the new structures, forming them to their own ends. This implied the inclusion of more and more people in the colonial pension schemes for public officials, the invention of social insurance for additional pressure groups and, in many cases, coverage of a wider range of eventualities. As a result, the emerging insurance systems were highly reflective of the social power relations in the late-colonial period.

Note that while the mechanism of appropriation may be generalised across different cases, the design of the resulting insurance schemes can vary. Both decolonialisation and the emergence of socialist governments in Africa produced considerable national variation within an imagined global pattern, hitherto dominated by the Bismarckian system of national and compulsory insurance (Hu and Manning 2010, 148). In the following, we propose two further mechanisms that we suspect to have influenced the introduction of social insurance beyond the public sector.

3.3 Labour Incorporation Mechanism

A crucial factor determining the introduction of social insurance for private sector workers is the weight of labour, itself a result of specific historical developments in the colonies, the degree of world market integration and capitalist differentiation (Collier and Collier 1991). Accordingly, we find the introduction of broad social insurance schemes in African countries to be related to the increased political influence of labour unions at independence. During the colonial period, African unions had begun to direct claims for social justice at the colonial authorities, which—given their repressive capacities—had nonetheless minimised their concessions to maintain cheap labour. Consequently, unions became important supporters of independence movements and a “standard component of African nationalism” (Orr 1966, 68). Once the new, independent governments were formally in charge, they needed to integrate their support base to maintain power. They entered negotiations with the unions, introducing comprehensive social insurance for private sector workers in exchange for political support. Note that this did not imply an extension of social insurance to the many informal workers without a lobby or mobilising power.

3.4 Public Resource Accumulation Mechanism

Many African countries experienced an economic downturn at independence, in addition to having been depleted of their resources for decades. Even today, the results of colonial exploitation remain visible in the overall low and inconsistent levels of capital accumulation (Nkurunziza 2019). At the same time, generalised expectations regarding improved well-being loomed high. Combined, these conditions appear to have triggered the introduction of social insurance funds, which not only projected the image of the aspired ideal of social progress, but also provided the government with a means of relatively rapid resource accumulation. The investment of capital from insurance funds in development projects or even defence efforts has been shown for both North and Sub-Saharan African countries (see, e.g., Gerdes 1971; Eibl 2020, chapter 6). Public resource accumulation can occur in combination with labour incorporation but can also unfold on its own: where not tied to the presence of a specific pressure group, the introduction and/or further extension of social insurance schemes are likely to have been driven by considerations of profitability.

4 The Expansion of Social Insurance in Decolonising Tunisia and Uganda

In the following, we will illustrate the above mechanisms with the most different cases of Tunisia and Uganda which, despite very different preconditions, have some forms of social insurance systems. As we will show, the mechanisms of imperial staffing and appropriation are common to Tunisia and Uganda. Labour incorporation paved the way for a more encompassing social insurance system in Tunisia, confirming the importance of labour history. The introduction of a limited social insurance scheme in Uganda is better explained by the need for public resource accumulation. The two case studies start with a general overview of the historical context and actor constellations. We will then sketch the emergence and expansion of social insurance in both countries with a view to the mechanisms outlined above.

4.1 The Step-by-Step Development of Social Insurance in Tunisia: From National Liberation to Labour Incorporation

Having been home to European settlers and transnational trade for decades, Tunisia became a French protectorate in 1881. Weakened by internal power struggles, state bankruptcy and pressure from European consuls, the regent (bey) was forced to accord the French responsibility over Tunisia’s foreign affairs and, in 1883, internal affairs. The country remained a de jure regency throughout the colonial period, but beylical authority was de facto exercised by the French resident general who relied on a highly centralised administration run by French bureaucrats (Nelson 1986, 33), supported by colonial intelligence (Safi 2020). Because the Regency of Tunisia was never formally abolished, French (social) legislation did not directly apply.Footnote 7 However, a number of French laws were subsequently promulgated via beylical decree. French colonial economic policy promoted trade, agriculture and mineral extraction, although the industrial sector remained underdeveloped in order not to create competition with the metropole. Politics and the economy were dominated by the French who settled and built their businesses particularly in the coastal areas. While only French citizens were represented in elected councils, a small Tunisian elite maintained sizeable influence. The protectorate status implied the coexistence and collaboration of French and Tunisian state officials, of which the latter were the first to benefit from colonial old-age pension funds. As decolonialisation set in, appropriation of the colonial state and its social policies occurred through two major groups: the nationalist conservative elites which also dominated the civil service and, following independence, the new leadership around Habib Bourguiba, head of the national liberation movement and first Tunisian president after the dispossession of the Bey in 1956.

The roots of Tunisia’s social insurance system, which formed throughout this process, can be traced back to the first decades of the occupation. In 1898, the colonial government charged a provident society with the management of retirement schemes for the French civil servants working in the colonial administration (Chaabane 2002, 3). During World War I, in 1915, imperial staffing prompted the creation of a similar pension fund for Tunisian civil servants (Destremeau 2010, 131), which was complemented by a system of family allowances in 1918 (Chaabane 2002, 3). These developments are closely linked to the enlistment of Tunisian soldiers in the imperial army. In addition to veterans’ pensions, they became eligible to family allowances, though inferior to those of French soldiers, after 1915 (Arnoulet 1984, 54–59). Upon their return, over a 1000 veterans obtained civil posts, including lower positions in the colonial administration. At the same time, the introduction of pensions did not mean equal treatment for French and Tunisian state officials. Not only were the schemes for Tunisians less generous, but in 1918, the introduction of the tiers colonial accorded one-third more pay to French nationals. This decision, alongside mounting calls for self-determination, triggered the creation of Tunisia’s first nationalist party, Destour, in 1920. Its programme demanded the restoration of the 1861 constitution (destour) as well as equal pay for equal work. At the same time, the party’s aims were far from revolutionary. Conservative and elitist in orientation, it limited its demands for equal treatment to civil servants and religious authorities, hereby setting itself apart from manual workers (Beinin 2016, 14). This increased its bargaining power vis-à-vis the French authorities, but also split and thus weakened the national movement.

In response to Destour’s position and its collaboration with French interests, a group of radical nationalists formed Neo-Destour in 1934, which combined demands for national sovereignty with social rights, hereby gaining the support of the labour movement. The colonial authorities responded through repression, jailing and exiling its leaders. At the same time, the short period of the Front Populaire in France led to the proliferation of ideas of “colonial socialism” throughout the empire. In 1936, parallel to the Matignon reforms in France, workers in individual industries were granted paid leave and a forty-hour week. In 1943, the authorities proclaimed a minimum wage for mine workers.Footnote 8 Family allowances for workers in the private sector, which had been demanded by the unions since 1932, were finally accorded in 1944, but even then, the Tunisian Caisses d’Allocations Familiales functioned differently to the one in France and its other colonies (Guelmami 1996, 99–102). Overall, social legislation remained far behind the workers’ demands.

Following World War II, growing aspirations for independence coupled with the proliferation of social rights gave rise to what Guelmami (1996) has termed a “colonial welfare state” (ibid., 77–131).Footnote 9 This implied a shift in colonial social policy, which now aimed at the social integration of the “indigenous” population into the colonial order.Footnote 10 Congruently, ideas of a reformist Tunisian government and stepwise internal autonomy began to take hold. In 1947 and 1951, France instated Tunisian governments with “indigenous participation” under the oversight of the resident general. This included, amongst others, the creation of a Ministry of Labour and Social Security under a Tunisian minister. Already in 1948, the new minister announced the creation of the Caisse Nationale de Retraite (CNP), a contribution-based pension fund for permanent employees in the public sector and concessionary enterprises (electricity, gas and transport). The scheme was extended to the banking and insurance sector in 1949 (Chaabane 2002). In 1951, a health scheme covering long-term sickness and surgery for permanent public employees was added via the Caisse de Prévoyance Sociale (CPS) (Ladhari 1996; Chaabane 2002; Ministère des Affaires Sociales 2021).

As a result, three categories of employees emerged (Guelmami 1996, 100): first, a superior category consisting of employees of the state, public services and municipalities, concessionary services, as well as employees in the banking and insurance sectors, who all benefited from social insurance. Many of these were, de facto, Europeans. The intermediate category regrouped private sector employees, mostly mine workers who were granted a minimum wage and could receive family allocations. Yet, even here, few Tunisians benefited compared to the total population: by 1952, there were 48,000 Tunisians formally employed in mining and industry (Murphy 1999, 81), and even among these, many did not benefit from their legal rights. Finally, the inferior category comprised agricultural workers, peasants and craftsmen who were not eligible for social benefits. Obviously, the late-colonial interest coalition consisting of French authorities, French-dominated business and a small Tunisian elite gave birth to a highly exclusive social insurance system.

Despite French efforts to hold on to Tunisia via political autonomy and expanded social protection, independence became inexorable. Neo-Destour grew into a powerful opposition, and in 1951, the political compromise between the French and the Tunisian authorities for reform instead of independence triggered a guerrilla war led by the armed wing of the liberation movement. Shortly after, Bourguiba entered negotiations with the French government who first granted internal autonomy (1955) and then independence (1956). The independence treaty of 1956 initially foresaw the instauration of a constitutional monarchy, preserving French and Tunisian elite interests. Following the first elections, from which Neo-Destour emerged as the single power, the new power-holders dethroned the Bey and proclaimed a Republic in 1957.

Under Neo-Destour, appropriation continued, though now accompanied by labour incorporation. First, the independent government took over the state apparatus and its personnel. One of the first steps consisted in the “Tunisifaction” of the state and the economy. This meant, on the one hand, replacing 12,000 French civil servants with Tunisians. Because of the shortage of qualified personnel, the Tunisian government agreed to keep 3500 French civil servants for a transition period of unspecified duration; this number had already dropped to 2000 by 1961 (Carter 1965, 26). In the light of overall low qualification levels, the new government relied on the Tunisian civil servants who had served in the colonial administration and who had not necessarily welcomed the abolition of the short-lived monarchy. To unify the state bureaucracy and secure support from public sector employees, Neo-Destour expanded social insurance benefits for the public sector, specifically by adding health insurance in 1959.Footnote 11

Second, the government needed to respond to the demands of labour, who had been crucial in achieving independence and bringing Neo-Destour to power. To understand why labour incorporation set in at this specific point in history, a more detailed look at the role of unions before and after independence is warranted. Already during the colonial period, Tunisia’s mining sector had produced an increasingly activist labour movement. The Compagnie des Phosphates et des Chemins de Fer de Gafsa (CPCFG) alone employed 20,000 miners in 1920, of which 3600 were Europeans (Beinin 2016, 13). In 1919, French and Italian workers established Tunisia’s first trade union, which also had a branch for Tunisian workers. Initially, the movement demanded that their working conditions be adapted to French standards. Joint mobilisation of European and Tunisian workers led to the introduction of the first social protection programmes, including work accident indemnity for mine workers (1921) and agricultural workers (1924). At the same time, the labour movement remained divided, as the French branch did not support wage equalisation demanded by their Tunisian co-workers. This prompted major strikes and led to the creation of the first Tunisian national trade union federation in 1924, the Confédération Générale Tunisienne du Travail (CGTT), that claimed to liberate Tunisian workers from the “union protectorate (protectorate syndical)” (Guelmami 1996, 85, own translation). Viewing this mobilisation as an anticolonial uprising, the French authorities banned the union and exiled its leaders.

This opened the door for the Neo-Destour party to emerge as a defender of the labour movement. In 1944, Ferhat Hached and Ahmed Tlili, both close to Neo-Destour, formed two regional federations of Tunisian trade unions with the aim of organising workers on a national basis. In 1946, the two federations merged into the Union Generale des Travailleurs Tunisiens (UGTT). The UGTT defined itself as a federator of national aspirations with the task of communicating “a progressive message aimed at the destruction of the colonial system and the foundation of a new society based on social justice and labour” (cited after Guelmami 1996, 87, own translation). Legalised in 1947, the UGTT grew to 80,000 members by 1952, henceforth constituting the most influential civic organisation and the main mobilising base of Neo-Destour. The establishment of an encompassing social insurance system for all employees, including health insurance and old-age pensions, figured among the core demands of its 1949 congress (ibid.).

Once independence was achieved and Neo-Destour took formal power, the new government was faced with the challenging task of fulfilling the high expectations held by its constituency and outcompeting the more radical socialist and dissident wings of the labour movement (Cameau and Geisser 2003). Social progress and economic development became the founding principles of the new independent state and its proclaimed “joie de vivre” (farhat el-hayat) (Catusse and Destremeau 2010). Important resources went into state-led industrialisation, involving capital-intensive projects such as steel mills, an oil refinery, a paper plant, several textile factories and an automobile assembly plant (Ayadi and Mattoussi 2014, 3). In parallel, the government negotiated the form and scope of the future social insurance system with the UGTT. Contrary to the draft proposal, the initial scheme did however not include pensions as the government sought not to overburden employers. Further, it first remained limited to employees in the non-agricultural private sector.

The implementation of the National Social Security Fund (CNSS), established by Law 60-30 of 14 December 1960, was supported by the ILO, which Tunisia had joined in the very year of its independence. In 1974, the scheme was finally complemented by pension benefits covering retirement, disability and survivors. Mirroring the UGTT’s socialist orientation, the resulting social insurance system featured aspects from the Soviet model insofar as it covered all the major social risks of injury, sickness and old age at once instead of introducing them separately (Hu and Manning 2010). At the same time, it followed the Bismarckian model promoted by the ILO as it bet on the expansion of paid labour in an industrialising economy. Mirroring the power constellations of the time, it was thus the “travailleur-citoyen” and not the “poor” that was put at the centre of the new social insurance system that took on paternalist, bureaucratic and corporatist structures (Catusse and Destremeau 2010). In exchange for the concessions made to workers, the UGTT was incorporated in a Single-Party, Single-Union system (Bellin 2002).

The CNSS, like the CNRPS for public sector employees, came under the auspices of a tripartite board consisting of the state, employers and employees (Chaabane 2002, 5). This meant that the social insurance funds were outside direct government control, attributing the Tunisian social insurance system a “purely social character” (Eibl 2020, 185). The insurance’s surplus was initially invested in bonds of banks and public infrastructure. Only after 1973 were these funds tapped to expand social welfare benefits, for example to cover low-income contributors and to finance social housing (Guelmami 1996, 159). The increase in public savings for development purposes occurred via alternative saving funds such as the Caisse Nationale d’Epargne and the Société Nationale d’Investissment. In this, Tunisia differed from Egypt, for example, where public resource accumulation was a major motivation (Eibl 2020, chapter 6).

Throughout the 1960s, the UGTT successfully pressured for the inclusion of further professional groups into the social insurance system. In 1965, it was extended to students (health insurance and family allocations only). Agricultural workers permanently employed for at least 180 days per year gained partial access in 1970; this was extended to agricultural workers employed for at least forty-five days per trimester in 1981. Independent workers have been included since 1982. Tunisian workers abroad gained access in 1989. Since 2002, the social insurance system also includes formally employed domestic workers, contractual workers, small-scale fishermen and farmers, ranchers, labourers and sharecroppers, as well as artists, and intellectuals, craftsmen and designers.Footnote 12 Today, roughly 25 per cent of the salary goes into the insurance fund (ISSA 2019, 29). While social insurance for all the above groups is managed by the CNSS, different schemes apply. These are the so-called general regime for non-agricultural workers (RSNA), two different schemes for agricultural workers (RSA and RSAA), a scheme for self-employed workers (RTNS) as well as a scheme for Tunisian workers abroad (RTTE).Footnote 13 In addition, several reforms of the various schemes occurred over time (see ISSA 2019). The stepwise inclusion of professional groups via different schemes resulted in a highly segmented social insurance system. There are, for example, eleven different pension schemes depending on the branch and nature of activity (IILS 2011, 74). Benefits vary not only between the public and private sectors, but also within the private sector (Ben Romdhane 2005, 65–68).

Overall, the Tunisian social protection system, today consisting of social insurance and social assistance, continues to privilege urban employees at the expense of those involved in small-scale agricultural or independent labour. Apart from general food subsidies, Tunisia’s many poor began to directly benefit from the proclaimed national solidarity relatively late. The Program for Aid to Needy Families (PAFN), created in 1986, constituted an important step. However, over 17 per cent of Tunisians are still estimated to be excluded from any form of public social protection (CRES 2019), which is not least due to problems in the PAFN’s selection process, which for a long time was guided by party patronage.Footnote 14 Further, even among those who are formally covered, problems of accessing adequate health services prevail, particularly in the country’s interior regions. Consequently, the social question remains highly politicised and potentially destabilising to this day (Thyen 2019).

4.2 How Ugandan Elites Appropriated the Colonial State and Its Old-Age Protection Institutions

Uganda’s colonial period began in 1888, when the British East Africa Company set up in the then Kingdom of Buganda after having helped to reinstate the Kabaka (king) after an internal uprising. It transferred its administration rights to the British government in 1894, and two years later, protectorate control was extended to Bunyoro, Ankole and Toro to cover what roughly corresponds to present-day Uganda. The colony had to be integrated into the Empire’s division of labour, with enforced cotton production as the first scheme for funding the costs of colonial rule without creating competition to British interests. Cotton production for export began in 1904 and was complemented, in the 1920s, by commercial production of coffee and sugar. In contrast to settler colonies, half of the land was reserved for the Bugandan gentry and could not be acquired by non-Africans. This was very much to the benefit of Bugandan chiefs who secured the most fertile land and would furthermore serve as tax collectors. This constituted an important difference to settler colonies such as Tunisia, where Europeans came to dominate the economy.

In this particular context, the British colonial state of the “Protectorate Uganda” was appropriated mainly by two powerful groups (Mamdani 1996, chapters 5–6). One was the group of chiefs and large-scale landowners of the colony’s core, the Kingdom of Buganda, which the British used as a textbook case of indirect rule. Large-scale Bugandan cotton and coffee producers benefited from the colonial economic policy and maintained considerable bargaining power throughout the period of colonial rule (1900–1962). However, formally, African actors gained access to colonial decision-making very late. Legislative and Executive Councils that were created already in the 1920s were opened up to African representatives only in the 1950s.

The other group that appropriated the colonial state was the African staff in the public sector, growing in number as the colonial state grew in size and extending into ever more spheres of life. When Uganda gained independence, the formal empowerment of these two groups was the main means by which Ugandans took over the state. This applied equally to the formal institutions of social protection that had emerged during colonial rule. The rudimentary system of free healthcare was massively expanded during the 1960s, as was the education system. The colonial state had already invested in both since the 1940s, but churches had remained the main players in both fields. Of course, in both fields, inclusion and output remained very low during colonial rule.

The mechanism of appropriation, however, becomes most visible in the field of old-age protection. This history can be told in three parts, with the most important background condition being that up until today about 85 per cent of the active labour force has never been involved in any formally institutionalised pension system. Uganda is still a largely agrarian society, with the majority of the workforce living from small-scale farming and informal economic activities.

The first part of this history started in 1921 when the first pension scheme was created for British citizens working for the colonial service in Uganda. In 1927, this tax-funded scheme was further extended to Asian employees of the colonial state. Thousands of South Asian migrant workers had been hired for the construction of railways in the early colonial period, and many of them decided to stay, forming an intermediary group between an African peasantry and a dominating white colonial class. In 1929, the growing number of African civil servants were included in this pension scheme (Bukuluki and Mubiri 2014, 37), which might be explained by the imperial staffing mechanism elucidated above. Formally, the year 1929 thus marks the introduction of a non-racially discriminating old-age protection system; yet for a long time it included only a tiny minority of the colonially subdued population.

With the extension of public service, more African Ugandans became beneficiaries of this pension scheme. In 1939, with the beginning of World War II, an Armed Forces Pension Scheme was created for the more than 70,000 African Ugandans who fought for the British Empire. This number amounted to a tenth of the male workforce of the time. Yet formally, these were pensions, so that the only real social insurance that came about during colonial times started in 1941 when African employees, including the growing number of teachers in public schools and clerks in the local administration, became beneficiaries of the Government Employees Provident Fund (Barya 2011, 9).

With the end of World War II, the days of colonial rule were numbered. In Uganda, several political parties mobilised through general strikes in 1945 and 1949, to which the colonial government reacted by repressing trade unions. By 1957, the Railway African Union was the only one left with more than 500 members (Mamdani 1996, 191). On the political level, colonial rule opened up slowly in the 1950s, including African representatives into the Legislative Council, who would make up 50 per cent by 1955. The Executive Council was, similar to developments in other colonies, developed into a Ministerial Council. This and the massive increase in African personnel in public service (see Table 6.1) indicated the encroaching appropriation of the colonial state by Ugandans.

Table 6.1 Africanisation of public service personnel in Uganda 1956–1959
Full size table

In contrast to other colonies where independence was achieved through violent struggle, the transition was comparatively smooth in Uganda. In 1961, one year prior to independence, general elections were held to determine the new government. Benedicto Kiwanuka’s Democratic Party led Uganda into independence but was defeated by Milton Obote’s Uganda People’s Congress (UPC) in the first general elections in 1962. The king of Buganda became the non-executive president. Up to this point, Uganda did not have any contribution-based social insurance apart from the provident fund for local government employees, of which the vast majority were teachers. And even this had been integrated into the tax-based pension scheme in 1953 for civil service. It was only in 1967, five years after gaining independence, that Uganda would have a social insurance system outside the public sector offering old-age insurance for a small group of private sector employees in the urban centre(s).

Part of the explanation for the long-term absence and limited significance of social insurance schemes for private sector employees is the mode of colonial world market integration. Right from its start, colonial economic policy aimed at the production of cotton and coffee, which were processed elsewhere, in the case of cotton, for example, in Great Britain. In Uganda, both major export goods were produced by farmers, who depended on kinship solidarity for their social protection. The political economy of the colonial state relied on rents from exporting agrarian production while industrialisation was avoided in order to protect industries in the UK metropole. This meant a lack of wage earners in the Ugandan economy (Elkan 1961). While Ugandan labour unions formed relatively early in comparison to other African countries, the labour movement thus never developed the strength and size it did in more industrialised countries, so that the labour incorporation mechanism cannot explain the Ugandan case.

Apart from a small urban service economy, wages and salaries were thus restricted to the public sector for which the tax-funded pension scheme mentioned above was entrenched. It was only in the late stage of colonialism of the 1950s when a stronger “development” orientation in colonial policies allowed for the creation of textile production in Uganda itself. Industrial production, however, never grew to an extent that organised urban labour became a powerful political force that could call and push for social insurance legislation. The actual birth of Uganda’s social insurance system was thus a top-down initiative in the form of the Social Security Act of 1967. In contrast to the development in Tunisia and other cases on the African continent, this innovation was thus not the result of social conflict.

Uganda, like other successor states of former British East Africa,Footnote 15 introduced a National Social Security Fund (NSSF) for private sector employees and non-pensionable public servants. It consisted of a savings scheme based on earnings-related contributions by workers and employers. The creation of this social insurance scheme can be seen as part of the social service expansion of the early independence period with its huge investments in health and education (Reid 2017, 244–67). Its introduction may also be explained by the government’s need for capital and its aspiration to build a modern public infrastructure with a strong welfare element (Gerdes 1971).

The fund was planned to cover an estimated 200,000 workers, whereby 15 per cent of their salary should go into the NSSF. One-third of the contribution was to be paid by the respective employee, and the other two-thirds by the employer.Footnote 16 The accumulated reserves of the fund were to be used by the Ugandan government to accelerate Uganda’s economic development (Nyakundi 2009). While the reporting US embassy saw at the time “few administrative difficulties” ahead (US Department of Labor 1967, 23), the fund was, however, entirely plundered during the rule of General Idi Amin (1971–1979), mostly for prestigious projects, for example, a conference centre for the OAU summit in 1975. This dissolution was part of a larger process of institutional decay in Uganda. During the ensuing years of political turmoil and civil war, almost all public institutions stopped working. It was only after 1987, under the enormous impact of donor initiatives and policy prescriptions, that state agencies slowly recovered.

The NSSF was reconstituted by a law in 1985 and started working again in 1987. Since then, it has been a relative success story, even if many of its directors were sued for embezzlement and mismanagement. In the view of contributors at least, it is of now “fairly working”.Footnote 17 In 2014, it covered about 1.45 million members of which 500,000 were working. Yet the fund is not really a solidarity fund as each contributor maintains her or his own account of assets on which an annual interest rate of 10 per cent is paid. It covers the average inflation rate in Uganda over the last fifteen years. Again, like in 1967, the NSSF is also the largest domestic capital formation institution. It has invested its USD 3.5 bn in government bonds and in a host of Ugandan enterprises (Kamukama 2019). Its emergence and revitalisation might, again, be explained along the public resource accumulation mechanism. But case literature does not exist, so a thorough historical reconstruction still has to be written. Its revitalisation was certainly driven by other needs than social protection, namely the perceived need for capital formation and the financial demands of the Ugandan state. Causally though, this growth has become possible only since the Ugandan workforce has grown considerably in services like construction, tourism, trade and transport. The business of development, with a multitude of NGOs, and the boom in private education are also part of this change in the occupational structure. The decisive pressure to create the NSSF, we assume, thus rather stemmed from the interaction between the government that came into power in 1986, starting with empty state accounts, and the International Financial Institutions aiming at reducing the share of grants and loans in the government budget (Schlichte 2021).

The system of old-age protection is currently generally politicised. Pensions for public employees as well as the special funds for armed forces and the police are perceived as socially unjust, because the pensions of former state employees are a tax burden for a largely informally employed or self-sustaining population. Pensioners have been mostly better educated and are rather urban males, so that the pension system is much less beneficial for rural areas and the less educated. The vast majority however, about 70 per cent of the workforce, does not benefit at all from these schemes. It consists of subsistence farmers who sell a little surplus on the market (Munyambonera et al. 2018, 14). In this regard, Uganda’s old-age protection system displays a continuity since colonial times: since the last decade of colonial rule, public employees have been criticised for their relative privileges and for demanding Western standard salaries in a society of poor peasants (Ehrlich 1963, 264).

5 Conclusion

As former colonial subjects took over the colonial state, they also appropriated its social policies, including the early pension and healthcare schemes that had been introduced for “indigenous” civil servants as a means of promoting imperial staffing. During this process, the first social insurance was granted to public employees and additional professional groups that had gained influence in the course of decolonialisation. Following independence, pressures for labour incorporation and/or public resource accumulation led to the invention of further insurance schemes beyond the public sector. By unravelling these causal relationships, we show that the introduction of social insurance in African countries is not simply a result of colonial powers extending their policies into their colonies, but strongly related to political contentions that emerged in the context of waning empires and persisted beyond formal independence. Specifically, in contexts where organised labour played a crucial role in national liberation, as in Tunisia, their increased negotiating power after independence led to a broad extension of social insurance into the private sector. The socio-economic trajectory of colonial Uganda, however, did not create a strong demand for social insurance and led to a top-down decision to extend social insurance with the motive of creating a public investment fund.

More generally, our analysis shows that the limited coverage and segmentation of present-day social insurance can be directly related back to colonial structures and divide-and-rule strategies. In Tunisia, social protection was—since its inception—organised via a discriminatory system that offered little protection to the larger population (Catusse and Destremeau 2010). The gradual, but unequal extension of social insurance mirrors experiences from Latin America, where the stepwise expansion also resulted in highly fragmented and generally unequal social insurance systems (Huber and Stephens 2005, 620). Needless to say, social insurance systems in Africa cover only the formally employed, thus excluding many, if not the majority of workers in the informal economy. Especially in the countryside, public social protection rarely exists. While this situation has been exacerbated by the cutbacks in public sector employment since the 1980s (Eckert 2019, 163), this particular constellation of the “social question” in Africa has a long history that goes back to early colonial times (Veit et al. 2017) and is particularly remindful of Mamdani’s famous binary of “citizens” and “subjects” (Mamdani 1996).

Our comparison shows quite clearly that the realm of possible “diffusion” of social policies, including the introduction of social insurance for broader parts of the working population, was heavily influenced by the structures set by the mode of colonial world market integration. In the case of Uganda, it has turned out to be extremely difficult to differentiate the colonially inherited economy as other countries had already occupied prospective niches in the world market. During the 1960s, this was a direct outflow of the colonial division of labour. Due to a period of civil strife and economic recession in the 1970s and 1980s, Uganda’s polity deinstitutionalised generally. Nor did the process of internationalised state formation after 1990, with the neoliberal age in full swing, offer a chance for economic diversification (Obwona et al. 2014). In contrast, Tunisia mobilised enormous resources for state-sponsored industrialisation after independence. Private sector development followed in the 1970s, whereby the Tunisian economy diversified into new sectors in manufacturing and services. These investments meant an expansion of formal labour, offering the base for a more comprehensive social insurance system that today covers more than half of the population. The map of how social insurance spread in space and time might thus ultimately mirror the history of the global division of labour and national political constellations.