Politicization of Non-state Cash Transfers: The Cases of Uganda and Tanzania

Abstract

Social cash transfers (CTs) have expanded considerably in Africa over the last decade. While the largest programmes are run by public institutions, non-state actors are increasingly becoming involved, but have attracted less academic attention. CTs are particular in that they are more fungible and liquid than other types of welfare provision, meaning that they are more difficult to control for both providers and governments. This article explores state strategies in relation to CTs by non-state actors focusing on when and how they are politicized. We analyse politicization from two intertwined vantage points. Firstly, we argue that the distributional logic in a society’s social contract contributes to understanding how and when cash transfers become politicized. We analyse the distributional logic along a continuum from universalist to partisan and apply this analytical lens to Tanzania and Uganda. Secondly, we suggest that politicization is intertwined with how the states coordinate non-state-CTs. We show that in Tanzania non-state-CTs are integrated into the state’s framework of social protection, while in Uganda they primarily operate on the margins of the state. In both cases, however, the state retains control over non-state-CTs and uses them politically, albeit in different ways, in order to maintain its relations with important constituencies. Hence, our analytical framework, which combines analysis of the distributional logic in the social contract and the mode of coordination, enables us to unpack when and how non-state-CTs are politicized, as well as how the tensions between governments and non-state-CTs have been handled differently in the two countries.

Keywords:

• cash transfer
• politics
• NGOs
• non-state
• Uganda
• Tanzania
• social contract

1. Introduction

Social cash transfers (CTs) have expanded considerably in Africa over the last decade. While the largest programmes are run by public institutions, international and national non-state actors are becoming increasingly involved but have attracted less academic attention.¹ This article explores governments’ politicization of CT programmes managed by non-state actors in aid-recipient African countries, based on how this is played out in Uganda and Tanzania.²

Cash transfers, i.e. direct transfers to households and individuals, are part of the broader package known as ‘social protection’, including social insurance, social assistance and labour regulation (Devereux and Wheeler, 2004; Lavers and Hickey, 2021), which in turn is part of social welfare provision, including education and health care.

While publicly managed social protection and CT programmes have been extensively scrutinized (see e.g. Beegle et al., 2018; Gliszcynski, 2015; Hickey et al., 2020), few studies exist of programmes managed by non-state actors. Studies of state-managed cash-transfer programmes show that they are often politicized by becoming part of political survival strategies by states, governments and ruling elites (Buur and Salimo, 2018; Hickey et al., 2020), which may negatively affect the programmes’ impact and ability to reach intended target groups. The other side of the politicization coin is that state-based social protection and CT programmes can have state-bolstering effects and improve (or at least change) state-society relations by creating new social contract dynamics even when based on patronage and clientelism (Porisky, 2024 forthcoming; Ulriksen and Plagerson, 2023).

How non-state CT programmes have been received or perceived by states and ruling elites has not been scrutinized systematically. Relations between NGOs and the state in aid-recipient countries were extensively studied from the 1980s and 1990s, when a substantial share of donor aid was shifted from governments to international and national NGOs in the wake of structural adjustment reforms. These studies tended to assume that the provision of public goods by NGOs would undermine the state (Watkins et al., 2012). However, as Brass (2016) showed subsequently, service delivery by NGOs can complement the state and, in some instances, even increase governments’ legitimacy in the eyes of their citizens, in particular in rural areas. Similarly, Cammett and Maclean (2014) have highlighted a broad range of political consequences, including both positive and negative effects on state capacity, of the last decades’ expansion of the non-state provision of social welfare in a wide range of countries. Ulriksen and Plagerson’s (2023) recent literature review highlights the differentiated and nuanced ways in which state-based social protection programmes affect state-citizen relations, complicating any overly optimistic assumption of state-bolstering effects. However, little is known about how these political and social contract dynamics are played out in relation to CTs by non-state actors.³

Although CTs are part of social welfare provision in a broader sense, we would argue that they are also particular in several ways. Firstly, they are more fungible and liquid than other types of welfare provision, as cash can be used for almost anything both by recipients and throughout the delivery chain, meaning that the provider of cash has less control over its use and effects, especially with regard to unconditional CTs. Secondly, CTs are still contested in many African countries. They have been promoted in Africa primarily by international donors, while ruling elites have often been reluctant to implement them more broadly, albeit with national variations (Hickey et al., 2020). This is to a large degree due to ideological ideas about deservingness, the perception that they lead to laziness being widespread among ruling elites (Hickey et al., 2020), in spite of overwhelming evidence that they can have positive socio-economic effects and do not have a negative effect on participation in work among adults (Bastagli et al., 2016; Beegle et al., 2018; Fisher et al., 2017). Thirdly, the fact that CTs are relatively new in Africa (as opposed to Latin America, for example) means that ruling elites have less experience of how to control them.

The difficulty in controlling the effects of CTs in combination with their potential to influence state-society relations mean that CTs by non-state actors can have political consequences, intended or unintended, that also might epitomize the political relations of state–non-state organizations more broadly. We argue that non-state CTs are politicized, in the sense that they are used by the ruling elite as one of their political survival strategies (Whitfield et al., 2015).⁴ Against that backdrop, this article explores state strategies in relation to CT programmes by non-state actors (non-state-CTs) in Tanzania and Uganda, focusing on when and how they are politicized.

Firstly, we suggest that to understand the politicization of non-state-CTs, it is important to appreciate the role of CTs and social protection in the country’s politics. We analyse this through the social contract lens, i.e. what role they play in the relationship between rulers and the ruled, focusing specifically on the distributional logic in the social contract. This analysis provides the background to understanding when and how CTs become politically important. We argue that the distributional logic in the social contracts can be analysed along a continuum from universalist ethos to partisan patronage, where the distributional logic in the Tanzanian social contract is leaning more towards a universalist logic, while the Ugandan social contract is rooted more in partisan, patronage-based distributions. We argue that these differences have implications for the role the state will allow non-state CTs to play and are important in understanding why some CT projects are publicly politicized while others are not.

Secondly, we analyse how the state coordinates or controls non-state-CTs, i.e. how their activities are allowed to operate in relation to the state’s social protection and/or CT activities. This sheds light on an important aspect of politicization, namely that states determine the room for manoeuvre for non-state-CTs, perhaps in particular for NGOs, and that they intervene when they find that the non-state-CT activities are not in line with their political needs and interests. We suggest that the mode of coordination between non-state-CTs and state-based social protection can be analysed along a continuum from integrated to parallel. We show that in Tanzania non-state-CTs tend to be integrated into the state’s framework of social protection, while in Uganda they tend to operate in parallel to the state.⁵ Furthermore, tensions between governments and non-state-CTs have been handled differently in the two countries and we suggest that the way conflicts are played out is related to the mode of coordination.

Our analysis of the distributional logic in the social contract is based on a literature review as well as our own political settlement analysis of the two countries (Nystrand et al., 2024, forthcoming; Mwinuka et al., 2024 forthcoming). The analysis of the mode of coordination is based on a mapping of non-state-CTs in Uganda and Tanzania by way of desk reviews and snowballing, complemented with information-gathering on details of the programmes through contacts with the identified organizations. The selection criterion was to choose organizations that were running or had run social CT programmes in recent years. In-person or online interviews were made with representatives of four organizations in Tanzania and four in Uganda, as well as with representatives of relevant ministries, donor agencies and CT and NGO coordinating organizations.⁶ All the identified non-state-CTs are funded by external donors. This was not a selection criterion, but a result of the mapping, indicating that only donor-funded organizations implement the type of formal CT programmes this study focuses on.

The next section presents our analytical framework and relates it to other studies of the role of non-state actors in social provision and social contract dynamics. The third and fourth sections analyse the distributional logic in the social contracts as they have evolved historically in Tanzania and Uganda and the role non-state-CTs have in each country and how they have been politicized. The article concludes with a summarizing discussion in which we draw theoretical and analytical implications of the analysis.

2. Analytical framework: distributional logic and mode of coordination

Our analytical framework for analysing how non-state-CTs are politicized consists of two parts: the distributional logic in the social contract and the mode of coordination between the state and non-state-CTs.⁷

2.1. The distributional logic in the social contract

The relationship between the state and the target populations may at first sight seem less important for understanding non-state organizations’ social CT projects. However, as indicated by the literature on non-state service-providers in general (Brass, 2016; Cammett and Maclean, 2014), NGOs and other actors can refocus state capacity, accountability mechanisms and access to services, thereby bolstering the state’s legitimacy among the population.

We argue that non-state CTs become inserted into relations between target populations and the state that are long-term in nature and embedded in often contested political histories involving the state and its relationship with different population groups (Buur et al., 2020) and at different levels of the state (Herbst, 2000). State–society relations are usually unevenly distributed, with great variation in state-ness across territories (Alik-Lagrange et al., 2021).

The social contract concept is old, having been developed by early political philosophers such as Hobbes, Locke and Rousseau. The central idea is that relations between rulers and ruled are based on mutual rights and obligations. The term ‘social contract’ does not refer only to legal contracts, although some of it can be codified in constitutions and laws, but also entails informal relational agreements or mutual expectations and actions that give legitimacy to the power arrangements in society. While the literature on state-based CTs has included perspectives on social contracts in its analyses (Hickey, 2011; Seekings et al., 2023), so far the scanty literature on non-state CTs has not discussed it systematically, beyond a few case studies (Grisolia et al., 2023). What is clear from the state-based literature is that, despite the political use of CT programmes for patronage or to mitigate social unrest, i.e. as a political survival strategy, such programmes can have an effect on state–society relations by creating new social contract dynamics (Ulriksen and Plagerson, 2023).

Social contracts can evolve in different ways. Nugent (2010) suggests that the types of social contract that evolve between African states and their populations can be placed on a continuum between ‘coercive’ and ‘productive’ relations. In principle a regime can uphold its power through violence and repression, or it can be based on creating legitimacy for the regime (Nugent, 2010, pp. 54–57), either through some ideology that legitimizes the ruler’s power or by providing for citizens’ needs. Our framework focuses on the distributional logic of the social contract, that is, the way in which a regime uses the distribution of resources in order to strengthen its social contract with its citizens, as distinct from using violence or ideology.

We suggest⁸ that the distributional logic in the social contracts can be analysed along a continuum from universalist to partisan, where the distributional logic in the Tanzanian social contract is based more on a universalist ethos and the Ugandan one on partisan patronage. ‘Universal’ and ‘partisan’ should be seen as endpoints in a continuum, or as ideal types, with actual distributional logics falling somewhere in-between.

The universalist distributional logic in the social contract⁹ means that all members of the polity have, formally and/or informally, the right to provisions and services and can expect to be included whenever possible, for example, when there are sufficient resources.¹⁰ We would usually expect to find this universalism, which could also be seen as a form of mass clientelism, in its strongest form in Western electoral social democracies (Hickey, 2011), but as Porisky (2024 forthcoming) suggests, this is also the case in Tanzania. This can be contrasted with a partisan distributional logic in the social contract whereby resources are distributed solely to those who support the regime, the governing elite or those they wish to favour (see van de Walle, 2007). Partisan distribution can be quite exclusive or narrow, for example, based on elite clientelism, in which patronage does not trickle-down patron-client networks to benefit non-elites. However, it can also take the opposite form: for example, when lower levels of the clientelist network or the ruling coalition are stronger, they can pressure the higher levels to deliver benefits, jobs and services in order to maintain their political support and remain in power. This kind of clientelism, which to some extent characterizes Uganda’s social contract, can be quite redistributive, even though it is still based on exclusion, and in our terminology it is still partisan patronage (Whitfield et al., 2015). This fluidity is consistent with our view of the universal versus partisan distributional logics as ideal types, with actual social contracts displaying traces of both logics.

We acknowledge the opportunity for further conceptual development, but for the purposes of our analysis of non-state-CTs in Uganda and Tanzania, it is sufficient to suggest that the distributional logic in the social contracts can been placed on a continuum between inclusion, which we refer to as universal, to exclusion, referred to as partisan. This novel terminology for social contracts allows us to: (a) capture important dimensions of the historical underpinnings of social contracts in Uganda and Tanzania; (b) identify more precisely how social CTs are related to developments in the social contract; and (c) thereby identifying when and how they might become politicized when non-state-CTs are involved.

2.2. Mode of coordination

The second part of our analytical framework concerns the mode of coordination between the state and non-state CTs, i.e. what space the state allows the non-state actors to occupy in relation to their own social protection activities.

Our analysis focuses on two dimensions of how the state’s and non-state actors’ activities are coordinated with regard to CTs: target groups and working modalities. We suggest that these dimensions of coordination can both be analysed along a continuum from integrated to parallel, where an integrated mode of coordination denotes that state and non-state activities are conducted jointly, or closely synchronized, and that they take place in the same localities and/or with the same target groups, while a parallel mode of coordination means that non-state-CTs are allowed to operate in areas or with target groups that the state is not focusing on, and/or in modalities not used by the state.

With regard to target groups, Herbst (2000) argues that a state’s ‘broadcasting of power’ over a certain geographical space determines its reach, i.e. the delimitation of the space within which authorities and people can make claims on each other. As Brass (2016) points out, it is common for the presence of weakly institutionalized states to be uneven within its territory and that NGOs fulfil roles in such places to substitute for the state. Additionally, the reach of the state can be uneven between the population groups within a locality. We therefore include in our analysis both the geographical spread of non-state-CTs and their programme target groups in relation to state-based social protection, in determining the degree of integrated or parallel mode of coordination.

With regard to working modalities, coordination between the state and non-state- CTs is determined by whether the activities are jointly coordinated (integrated), or whether the state allows NGOs or other actors to operate more or less freely within certain areas or with certain target groups (parallel).

Hence, both the target group and the working modalities’ dimensions of coordination in our framework focus on the space the state allows the non-state-CTs to occupy in relation to its own social protection activities.

Our empirical analysis, which we expand on below, shows that, with regard to target groups, non-state-CTs overlap with the state in Tanzania, while in Uganda they primarily operate on the margins of the state and with refugee and marginalized populations. Furthermore, with regard to the working modalities, non-state-CTs are integrated into the state’s framework of social protection and social CTs – for example, with joint work modalities – to a greater degree in Tanzania than in Uganda. Both these factors suggest that non-state-CTs are integrated with the state to a much higher degree in Tanzania than in Uganda, where non-state CTs operate in parallel to state activities. However, we argue that the state retains control over non-state-CTs in both cases, albeit in different ways.

Our argument that the state retains control in both the integrated and parallel modes of coordination differs in important respects from Brass’s (2016) analysis of the NGO-state relationship. She argues that the blurring of lines between the state and NGOs in aid-recipient countries means that NGOs also get involved in policy-making (‘steering’) and not just implementation (‘rowing’). We find this distinction useful, but as will be shown below with regard to CTs, the state is very unwilling to leave the ‘steering’ to non-state actors, illustrating the special status of this form of social provisioning.

Another way in which our framework differs from Brass’s (2016) framework is that we do not see conflict as a characteristic of the relationship, but as something that can occur in both an integrated and parallel mode of coordination, while taking on different contours. This is illustrated in the different ways in which the tensions between governments and non-state-CTs have been handled in the two countries. In Tanzania, projects that do not fit the government’s agenda are tacitly squeezed out, whereas in Uganda an open conflict erupted between the government and GiveDirectly, an international NGO providing cash to the poor. We suggest that the way conflicts are played out is related to the mode of coordination, and that the mode in which the state coordinates with non-state CTs (integrated or parallel) contributes to understanding how and when they are politicized by states and governments.

2.3. Model of our analytical framework

As explained above, our analytical framework includes two parts: (1) the distributional logic in the social contract; and (2) the mode of coordination between state and non-state-CTs, each with two end-points of a continuum: universalist versus partisan distributional logics, and integrated versus parallel modes of coordination.

Figure 1 illustrates how the two parts of the analysis can be combined into an analytical model that helps us analyse when and how non-state-CTs are being politicized. In our concluding discussion, we plot our analysis of the politicization of non-state-CTs in Tanzania and Uganda in the model, the aim being to clarify both the conclusions of the article and the usefulness of the analytical framework.

Figure 1. Model of the analytical framework used in this paper (created by authors)

3. State–citizen relations in Tanzania and social CTs

3.1. The social contract in Tanzania: a universalist ethos

Under the dominance of the leading party, Chama Cha Mapinduzi (CCM), Tanzania has been one of the most politically stable countries in East Africa, with a strong party-state presence in most spheres of society. Although Tanzania has been ruled by the same party since independence in 1961, the social contract has evolved over time (Whitfield et al., 2015).

The state–citizen relations captured in Tanzania’s post-independence social contract, as formulated in the Arusha Declaration adopted in 1967, can be described as based on the state’s control over associational life, which over time has generated a universalist ethos regarding access to free social services such as health, education and water for all citizens (Mwinuka et al., 2024 forthcoming). The universalist ethos or principle of providing for people in need merged with a general culture of supporting people in crisis in order to build national unity. As Kilonzo et al. (2019) have noted, in Tanzania citizenship or the social contract has strong roots in ‘contributing to the state [project of unity] rather than claiming from it [as well as not] contesting the authorities’ (Kilonzo et al., 2019, p. 152; see also Porisky, 2024 forthcoming, p. 231).

The universalist ethos came under pressure during the period after liberalization and the change from a one-party to a multiparty system during the 1980s and 1990s, partly because the state had fewer resources to provide it with, and partly because of a rise in partnerships between public and private actors. The liberal transition was conducted slowly and sequentially (Nyirabu, 2002), allowing the CCM to maintain its control, making sure that the party was reinvigorated along the way, and retaining the strong link between party and state.

The third regime under President Mkapa from 1995 to 2005 reduced government subsidies for public services and introduced cost-sharing in education and health (Shivji, 2021). However, it was also under this third regime that the Tanzania Social Action Fund (TASAF) was established in 2000, financed mainly by donors (Green, 2021), but increasingly taken over and promoted by the Tanzanian state as its programme. This was and is the key government instrument for the provision of social protection, including social CTs (Pedersen and Jacob, 2018; Ulriksen, 2016). TASAF was implemented in three consecutive phases, starting in forty districts with high poverty indexes, and later being scaled up to the national level. The main focus was initially on improving social-service provision through the construction of physical and social infrastructure like roads, schools, hospitals and irrigation schemes. Conditional CTs were piloted in three districts in 2009–2012 and subsequently, from 2013, scaled up in the form of the Productive Social Safety Net (PSSN). The core of the PSSN is an unconditional initial grant of TSH 10,000 (USD 4.5) per month for the poorest households, which can be topped up with various conditional transfers, typically linked either to children attending school or to health interventions (Jacob and Pedersen, 2018, p. 18) and seasonal public-works programmes (Ulriksen, 2016, p. 23). By now, TASAF covers almost all regions of Tanzania.

These policies were continued during the fifth regime under President Magufuli from 2015 to 2021, with an even stronger emphasis on ‘universalist’ ideas (Jacob and Pedersen, 2018; Ulriksen, 2016), although these were now based on individualism (each person’s own responsibility) and were introduced in parallel with reports of partisan distributions to individuals associated with the CCM (Mwinuka et al., 2024 forthcoming). TASAF’s expansion during the fourth, fifth and the current sixth political regimes saw new groups, such as the young, disabled and women’s groups, become integrated into the scheme.

The universalist ethos or principle of providing for people in need in order to build national unity has continued throughout different regimes, and it remains an ideal that affects expectations and relations between the state and its citizens. Social protection, including CTs, have been and still are used actively by the state as part of this overarching universalist ethos. Even though to a large extent TASAF is donor-funded, and from its inception was highly donor-driven, the CCM-state has appropriated it as its own programme. Hence, CTs are central to the distributional aspects of the social contract in Tanzania, and the distributional aspects are an important aspect of the state’s legitimacy. This, as we argue in the next section, is one of the reasons why non-state CT actors work in an integrated way with the Tanzanian state.

3.2. Social CTs: integration and co-production

Civil-society organizations and religious groups in Tanzania have operated within this universalist ethos, and with extensive state control, which essentially outlawed social, ethnic and religious associations from participating in the political sphere (Porisky, 2024 forthcoming, p. 213). In response, these actors mostly kept away from activities that could be considered political or that could be perceived as challenging state control (Wamucci, 2014). They focused instead on social and cultural development in the form of groups promoting women’s rights, children’s rights, land reform, constitutional reform, the environment and service delivery (Tripp, 2000). In Tanzania, as in many other African countries, the civic space expanded dramatically during the 1990s, following economic and political liberalization and structural adjustment programmes (SAPs) (Wamucci, 2014). During this period, the Tanzanian state had limited funding for the social services and welfare that had previously underpinned the social contract, leading to a decline in the government’s control over society. This led, as Porisky (2024 forthcoming, p. 233) argues, ‘to a ballooning of NGOs, religious organizations and local welfare organizations’ stepping in where the state had formerly been taking care of social services. The number of registered NGOs increased from 224 in 1993 to 8,360 in 1997 (Tripp, 2000).

Despite these changes, the CCM party-state never relinquished its control over NGOs. Not only did CSOs have to register formally and obtain approval, but a degree of adaptation also took place.¹¹ The limiting of the space for civil society was maintained by means of undemocratic laws, some of which dated back to colonial times, and there was constant pressure from the remnants of the party-state for co-optation or de-registration (Tripp, 2000), pressure that has continued until today.

Importantly, the increased involvement of religious and civil-society associations in providing local welfare, combined with an increase in support for the political opposition, did not undermine the universalist ethos or the basic tenants of the social contract. As Jacob and Pedersen (2018, p. 10) suggest, the electoral pressures placed on the CCM as opposition parties grew ‘helped bring the CCM back to its roots’ during the 2000s to focus on the provision of social services and welfare. In many ways, the ‘paternalism of the developmental state, evident throughout the ujamaa period, has survived to the present day’ (Porisky, 2024 forthcoming, p. 243). Now, however, it does not solely encompass the state, but also the diverse social, religious and civil-society organizations that work together with the state under the universalist ethos and actively formulate their intervention within state policies. This constitutes an integrated mode of coordination in our terminology.

As can be seen in Table 1, the 25 Tanzanian non-state-CTs we have identified are spread geographically over the country, with a slightly greater concentration in the Northern Highland region, which hosts refugee camps. The profiles of the Tanzanian non-state-CT programmes are mixed with regard to humanitarian and general development direction. Six of the organizations are national, while the other nineteen are international, but all of them receive most of their funds from international donors.

Table 1: Tanzanian non-state CTs by geographical spread¹²

Tanzanian regions	Organization
Coastal and Eastern	World Vision Tanzania, Tanzania Red Cross Society, Tanzania Education Fund*, ELCT*, Fistula Foundation, Ifakara Health Institute*, Compassion International Tanzania, Catholic Relief Services, Direct Aid
Northern Highland	World Vision Tanzania, REDESO*, Danish Refugee Council, Norwegian Refugee Council, UNHCR, Fistula Foundation, WFP, Kwa Wazee Tanzania, Farm Africa, Compassion International Tanzania, Path to Africa, Catholic Relief Services, HelpAge International
Lake	World Vision Tanzania, Fistula Foundation, SHDEPHA+*, Compassion International Tanzania, Catholic Relief Services
Central	World Vision Tanzania, Fistula Foundation, SHDEPHA+*, Farm Africa, Compassion International Tanzania, Village of Hope and Grace, Path to Africa, Direct Aid
Southern Highland	World Vision Tanzania, PADI*, UNICEF, Fistula Foundation, Compassion International Tanzania, Catholic Relief Services
Southern	World Vision Tanzania, Tanzania Red Cross Society, Fistula Foundation, Compassion International Tanzania, Save the Children

Source: Authors’ own mapping.

With regard to the target groups, non-state-CTs in Tanzania focus on the vulnerable, including children, orphans, women, the elderly, people with disabilities, the young, people affected by HIV/AIDS and the ‘most-at-risk populations’. With some exceptions for refugees, most beneficiary categories are the same as for TASAF, although TASAF has a much larger number of beneficiaries nationally.

The clear overlap in operations between TASAF and non-state-CTs, both geographically and with regard to programme target groups, suggests that the mode of coordination of Tanzania’s non-state-CT programmes is integrated in our conceptualization, in the sense that the state allows them to operate concurrently with state programmes, rather than primarily with marginalized populations, as in Uganda.

Moreover, with regard to working modalities, several of the non-state CT programmes operate within TASAF as ‘joint initiatives’, thus being formally integrated into the state-based CT programmes and strengthening them.¹³ Furthermore, the TASAF and non-state-CT programmes use the same local government mechanisms to identify suitable beneficiaries and monitor projects, thus actively coordinating all activities locally.

There is no significant difference between national and international organizations in the way they work with communities or in how they relate to the state, but there is a difference between activities in refugee camps and other activities. Refugee activities are considered a ‘security issue’, being more strictly regulated and coming under the Ministry of Home Affairs, while other activities are regulated by the Prime Minister’s Office. According to our sources, some organizations have attempted to get their activities re-classified to avoid the stricter regulations.

In sum, the involvement of non-state-CTs is closely related to and integrated with national programmes and involves a high level of collaboration and integration between non-state actors and the state.

3.3. Conflicts under the radar

Despite this overarching picture of close integration and collaboration, the cash-transfer programmes of humanitarian actors have been forced to reframe their activities or in some cases to close down in recent years. This mainly relates to a changed government stance on refugees in the form of an apparently reduced scope of the universalist ethos, i.e. a greater exclusion of refugees from the social contract.

Historically Tanzania’s universalist ethos also included taking care of refugees from neighbouring countries. However, under the Magufuli regime the approach to refugees hardened. The nationalist rhetoric portrayed them as a ‘burden’ on Tanzania, and refugee-deterrent policies were introduced (in line with those used in Europe and the USA), which also affected the possibilities for non-state-CTs in refugee settings.

Until 2015, CT projects were quite unrestricted in relation to refugees. As part of the changed stance on refugees, livelihood-enhancing support, including CTs, came to be understood as a ‘pull-factor’ that attracted more refugees, who were instead to be given a minimum of life-sustaining support and be strictly confined to their camps, not integrated into their surrounding communities. The change in policy came gradually and was to a large extent communicated in internal meetings between the state and humanitarian organizations, which means that it has been hard to substantiate what the policy change actually entailed. However, interviews with humanitarian organizations active in Tanzania during this period revealed that there was a strong perception on their part that CTs were no longer allowed in Tanzanian refugee camps. One organization, HelpAge, managed through negotiation with state institutions to obtain permission to continue with CTs focusing solely on the elderly and disabled, who were seen as particularly vulnerable groups and were aligned with TASAF priorities. Other organizations suggested that, if they did not comply with the political directions or could not align themselves with TASAF, they would not get their permissions renewed or new projects approved, and they would have to close their operations (interviews 2022 with international organizations). This would all take place without any public discussion or confrontation. The organizations interviewed indicated that there was more room for negotiation with the sixth regime of President Samia after 2021.

The discreet squeezing-out of non-state-CTs in refugee settings definitively involved tensions in relations between the state and humanitarian actors, but in our view, it also confirms the overall picture of close integration between the state and non-state-CTs in Tanzania. The non-refugee non-state-CTs under the universalist TASAF umbrella continued, and the lid was kept on the tensions related to non-state-CTs for refugees, with few open references to any challenges. The way the situation was handled, through continuous interactions between the parties, stands in stark contrast to the overt and media-covered conflicts that erupted between the CT organization GiveDirectly and the Ugandan state, as will be discussed in the next section.

4. State–citizen relations in Uganda and social CTs

4.1. The social contract in Uganda: patronage and exclusion

The National Resistance Movement (NRM) seized power in Uganda in 1986 after many decades of political instability, violence and war, and has kept power ever since. President Museveni, the leader of the NRM, is today one of Africa’s longest-serving leaders.

In the first decade after the 1986 takeover, the NRM regime enjoyed high rates of legitimacy and support in the southern part of the country, while a civil war raged in the north until 2006. The initial legitimacy was based to a large extent on its relative success in creating stability compared to previous conflict-ridden regimes (Nystrand et al., forthcoming 2024, forthcoming). The NRM initially created a ‘broad-based government’ system in which all political parties were banned, and the NRM became the umbrella under which all political activity took place. This was gradually replaced by democratic elections, first with presidential elections in 1996 and 2001, and then full multiparty elections from 2006. One legacy of the broad-based government system and the NRM’s long tenure of power is the overall fusion of the state with the NRM party, where ‘many Ugandans are unable to decouple the NRM from government, believing that the government of Uganda is the NRM and the NRM is the government of Uganda’ (Gibb, 2016, p. 98).

Even though the NRM has retained its grip on power, since the early 2000s it has witnessed increasing political opposition matched by increasingly open state repression of the opposition with each election cycle. This has gradually eroded the NRM’s initial legitimacy and led to a more selective and repressive social contract. However, the NRM regime uses both soft and hard power for regime survival (Golooba-Mutebi and Hickey, 2016), combining the repression of opposition with patronage-based responsiveness to popular demands and coalition-building with selected groups in society.

Central to the NRM’s soft power is the strong vertical relationship between the centre and local communities, particularly in rural areas. As Tripp (2010) has argued, relations between the NRM, the government and the local grassroots level are non-coercive and instead based on personalized patronage relations, including the use of state resources to convince voters to support the ruling party and its political candidates. This characterizes the way both the President and Members of Parliament interact with their constituencies, promising and distributing their resources to supporters, and threatening to withdraw resources from those who do not vote for the NRM (Titeca, 2018; Tripp, 2010). This is highly visible during election campaigns and institutionalized in local government structures.

Ugandan politics include elements of universalism, for example, in the form of the universal programmes in education and health that were introduced from 1996 onwards, and the more recently introduced Senior Citizen Grant (SCG), but these programmes have never become universal in practice due to resource constraints and their embeddedness in local patronage networks (Kjaer and Muwanga, 2019; Nystrand and Tamm, 2018).

Lavers and Hickey (2021) argue that state-based social protection, including CTs, is more recent, less widespread and less institutionalized in Uganda than in Tanzania, as well as compared to seven other African countries they analysed . The only large-scale CT programme is the SCG, which is an unconditional CT to senior citizens, piloted in 2010¹⁴ and then gradually expanded to national coverage by 2020 (Dietrich et al., 2020; Expanding Social Protection Programme, 2021). In the pilot stage the SCG targeted persons aged 65 and over, but in the national roll-out it was delimited to persons over 80 years of age, which in Uganda applies to very few people. Hence, even though the SCG constitutes an institutionalization of state-based CTs, the limited coverage means that its effect on state-society relations is very limited. In addition to the SCG, the government also used CTs to specific population groups to cushion the effects of the COVID-19 pandemic lockdowns (The Independent, 25 January 2022).

The main excluded groups in the current social contract in Uganda encompass not only political opponents but also people in parts of the country that have remained on the fringes of the state. This includes the Karamoja region in the northeast, a semi-arid pastoralist region affected for decades by poverty and conflicts related to violent cattle-raiding, as well as army-led disarmament interventions in the 2000s, which included a more or less forced settlement of the pastoralist populations (Catley et al., 2021). The other significantly marginalized part of the country is the north in a broad sense. The war that raged there from 1986 to 2006, and the way local populations were ‘incarcerated’ in internally displaced persons (IDP) camps for a decade, has contributed to the north being poorer than the rest of the country and to mutual distrust between the northern population and the ruling elite (Nystrand, 2014). While the north was once an opposition stronghold, the NRM has increased its electoral support in the region since 2006 through a combination of patronage to win voters, blocking alternative power-holders and taking back its role as a provider from non-state actors (Lie, 2017), as will be discussed further below.

In sum, Uganda’s social contract is based on a combination of using targeted patronage to galvanize support and increasing the repression of the opposition. The social contract is uneven, and there are still regions on the fringes of the state (Nystrand et al., 2024, forthcoming). In our terminology, the distributional aspects of the social contract are partisan in that support is used in a discretionary way to gain votes from particular groups, rather than universal as in Tanzania. This helps explain the pattern of non-state-CTs in Uganda, to which we turn in the next section.

4.2. Social CTs: patronage and parallel modus operandi

According to Twikirize (2017), 200 NGOs were registered in Uganda in 1986 and about 7,000 in 2010, mirroring the increase in NGOs in other East African countries (Wamucci, 2014). In 2019 a major review was conducted by the National Bureau for NGOs, which reduced the number of registered NGOs from 14,207 to 2,188 (The Guardian, 21 November 2019). This included suspensions of active NGOs, many of which were smaller domestic organizations.

In general, the Ugandan state, like the Tanzanian, has allowed NGOs to act in service provision, but it has restricted and/or co-opted those that are perceived to interfere with whatever has been considered political, in particular advocacy activities related to human rights, democratization and governance (Tripp, 2010; Twikirize, 2017; Wamucii, 2014).

During the last fifteen years, the regulation of NGOs has been amended and expanded several times, each time gradually increasing state control over the sector. An increasing number of NGO closures, suspensions, searches and seizures by state organs since 2011 have often coincided with the country’s electoral cycles. In many instances, these actions have also been at odds with government agencies working closely with the affected NGOs. Most NGOs that have large-scale operations are funded by international donors, whether they are nationally registered or not. The tougher regulations have affected domestic NGOs to a greater degree, but as is shown below, they affect international and donor-funded organizations as well, which is also linked to the general deterioration in government-donor relations. Since the late 1990s, international donors have channelled more funds through civil society in an attempt to avoid government corruption. However, the state has increased its control over these funds by regulating NGOs. When it comes to humanitarian aid and activities in relation to refugees, the coordination is placed directly under the Office of the Prime Minister. Uganda has a long history of receiving refugees from neighbouring countries, in addition to a history of Ugandans becoming internally displaced due to the country’s own internal conflicts. The role of non-state actors in these situations is important for understanding the pattern of current non-state-CTs in Uganda.

With regard to geographical coordination, Uganda shows a clear pattern of parallel operations between the state and non-state-CTs. Table 2 shows that 24 out of 32 identified non-state-CTs are located in the Northern, Eastern and Karamoja regions, which are parts of Uganda characterized by large-scale humanitarian needs.

Table 2: Ugandan non-state CTs by geographical spread.

Ugandan Regions	Organization
Northern and West Nile regions	Mercy Corps, DanChurch Aid, Catholic Relief Services, Lutheran World Federation, Danish Refugee Council, Action Against Hunger, Agricultural Business Initiative, Caritas, Malteser International, Oxfam, Send a Cow Uganda, Plan International, UNHCR, AFARD
Karamoja region	Save the Children Uganda, Adventist Development and Relief Agency (ADRA), BRAC Uganda, Restless Development
Northern, Eastern and Karamoja regions	Bantwana World Education Inc., Care International, Norwegian Refugee Council, World Vision International, ACTED Uganda, World Food Program
Rest of the country	GiveDirectly (entire country), Life Worth Vision (central region), Marie Stopes Uganda (Karamoja, Northern and Western Uganda), Compassion International (entire country), AVSI (entire country), International Rescue Committee (central region – Kampala), FinChurch Aid (Western region), Eight VZW Uganda (Western region)

Source: Authors’ own mapping.

The organizations that are active in the northern regions are almost exclusively involved in humanitarian work with refugees and associated activities in host communities. Currently the Karamoja region is also regarded as suffering a humanitarian situation, as will be expanded on below. Several organizations active in other parts of the country are also involved with refugees from Somalia, Eritrea, the DRC and South Sudan. All but two of the organizations are international, and all are funded by international donors. In sum, current non-state-CTs in Uganda are predominantly working within the context of humanitarian aid, hence on the fringe of the state.

The target groups differ between non-state-CT programmes, but in general the focus is on vulnerable groups. Since the SAGE programme currently targets primarily citizens over eighty years of age, the Ugandan non-state-CT programmes are parallel to rather than integrated with state programmes, including with regard to programme target groups.

State-non-state coordination with regard to working modalities includes a degree of coordination and control from the state, as described above, but due to the parallel form of operation, both geographically and with regard to programmes’ target groups, non-state-CTs are not integrated with the state’s policies for social provisioning as in Tanzania. This is partly due to the limited state-based social provisioning in Uganda, but it can also be seen as a state strategy, as until recently the Ugandan state has not used non-state-CTs to boost its social provisioning (this has started to change, as will be discussed below). Furthermore, we would argue that the parallel form of coordination does not mean that the Ugandan state has relinquished control to non-state actors. They retain the power of determining where they allow the them to work, and do not hesitate to change the conditions under which they operate when deemed relevant from a political perspective.

Lie’s (2017) analysis of changes in respect of who provides for the populations in the north and the Karamoja is a good illustration of the link between humanitarian NGOs, the social contract and the government’s role as provider. For decades, humanitarian NGOs and UN actors had been allowed to become the most important providers in the north and thereby take on the role of the state, but Lie (2017) shows how, from around 2006, the government actively took back control and reaffirmed itself as the provider of social services in the north. This came as a response to the low support for the NRM in the northern region in the first multiparty elections. It also coincided with the disarmament and forced settlement intervention in Karamoja, and although this region has always been marginalized, the situation in Karamoja was now discursively recast as a humanitarian situation, and humanitarian NGOs moved their operations from the north to the Karamoja region in the east (Lie, 2017). This example illustrates how the Ugandan state actively positions itself as a provider when a constituency becomes politically important, i.e. how the distributional logic in the social contract is partisan, and how the state directs which constituencies non-state actors are allowed to provide for.

From a social contract perspective, the current pattern of non-state-CTs in Uganda can be explained as being directed towards groups in society with which the government is presently not attempting to create a direct patronage link by way of social provision. The refugees are not Ugandan citizens, hence are not voters, and thus marginal within the social contract. At present the government allows humanitarian actors space in Karamoja and appears to see them as supplementing the government and not competing with it (as was previously the case in the north). Hence the majority of non-state-CTs in Uganda have a parallel role in relation to the state with regard to where they operate geographically, their target groups, state policies and their working modalities. But as Lie (2017) has shown, this relies on the state’s discretion: that is, if and when the state views these activities as being at odds with its interests, it has the means and power to discontinue any non-state activities. This is exactly what appeared to happen with GiveDirectly.

4.3. Open conflict and closure

GiveDirectly is an international NGO focusing specifically on CTs to reduce poverty, which started its activities in Uganda in 2013. In 2020 it launched a CT programme together with the Ugandan government and USAID, which aimed to help people hit hard economically by the COVID-19 lockdown. This programme came into the spotlight when GiveDirectly’s activities were suspended by the Ugandan government in October 2020 in the run-up to the January 2021 election. In September 2021, some months after the elections, GiveDirectly was permitted to continue with new projects (The Observer, 23 September 2021). The official reason given for the suspension was a query over the source of GiveDirectly’s cash. However, according to Kisakye (23 September 2020), the money came from USAID and DfID, which was well-documented in approvals signed by Ugandan government officials, and the programme had previously been cleared by the government.

At first glance, CTs are not overtly political in the same way as advocacy and human rights work. However, given that Uganda’s social contract is heavily reliant on partisan patronage distributions, including cash, non-state-CTs actually do compete with political actors when they disburse cash to specific population groups – or at least that is how the ruling elite appeared to interpret their activities. The timing of the suspension in the run-up to the 2021 election is important for understanding the dynamics at play. The government had increasingly come to see donor support from Western countries as support for the opposition and as interference in Uganda’s domestic politics, particularly support aiming to strengthen civil-society organizations in the name of democracy. A media interview with a donor representative who did not mention the governments involvement in the programme, and thus did not give them ‘credit’ for it, became a catalysing factor in the conflict erupting, according to some sources.

An interesting outcome of the suspension was that it eventually led to closer cooperation between the government and GiveDirectly. According to an interview with a senior representative of GiveDirectly in August 2022, the suspension period brought the organization into direct dialogue with several government ministers and the President. Furthermore, GiveDirectly became involved in trying to ‘save’ a not yet fully performing government CT programme targeted at the politically important group of privately employed teachers suffering from the pandemic lockdown (The Independent, 25 January 2022) as part of its closer cooperation with the government.

From the perspective of social contract relations, our interpretation is that the conflict between GiveDirectly and the Ugandan government was caused by the organization’s activities being seen as competing with government activities in an arena and at a time when this was perceived as a threat to the government. Ultimately, however, it led to closer cooperation, which moved the relationship between the government and the CT organization closer to an integrated mode of coordination as in Tanzania, albeit still within the partisan distributional logic.

5. Concluding discussion

This article has argued that the politicization of non-state-CTs can be understood in terms of the distributional aspect of the social contracts in which they become embedded and the mode of coordination between non-state actors and the state, in particular with regard to target groups and working modalities. We have developed an analytical framework suggesting that the distributional aspect of the social contract can be analysed along a continuum from universalist to partisan, and the mode of coordination along a continuum from integrated to parallel, and applied this analytical model to Tanzania and Uganda.

We have showed that Tanzania is characterized by what resembles a universalist ethos in its social contract, based on lingering traces of the socialist ideology implemented since independence, in which the state exercises strong control over society, which is legitimized by extensive (by African standards) provision for the needs of the population. When the state’s ability to provide was reduced in the liberalizations during the 1980s and 1990s, non-state actors became important as providers. However, the Tanzanian state did not relinquish control, nor did the universalist ethos disappear: instead the state integrated the non-state actors into the state’s social provisioning. This has come to include non-state-CTs, which are closely linked to and coordinated with the state’s social provision and CT programmes. This integrated mode of coordination is visible as pertaining both to target groups, in that non-state-CTs in Tanzania overlap to a large degree with the state-based cash-transfer programmes, both geographically and with regard to programme target groups, and working modalities, in that the state and non-state-CTs are involved in collaboration through formalized joint initiatives.

Conversely, Uganda is characterized to a greater degree by a partisan distributional logic in the social contract, i.e. distribution based on a personalized and selective use of state resources to convince voters to support the ruling party and its political candidates. Inclusive state-based social protection and CTs are not used to the same extent as in Tanzania. Current non-state- CTs in Uganda are primarily found among specific and marginalized groups in society, namely refugee populations and in the Karamoja region, none of which at present seems to be of interest to the government with respect to direct patronage in the form of social provision. Regarding working modalities, the Ugandan state definitively regulates and coordinates non-state actors’ activities, but it is not involved in joint coordination to the same degree as in Tanzania, partly because their own social provisioning is more limited and they do not require non-state-CTs to align with SAGE, for example. Historically, and in the case of GiveDirectly, the Ugandan state has shown that, although it allows non-state actors to operate in parallel to state activities, it can decide to intervene when the political situation warrants it or its attention to particular population groups change. Hence, although the mode of coordination between non-state-CTs and the state is primarily parallel, pertaining to both target groups and working modalities, the Ugandan state does not relinquish control to non-state actors, but preserves the power to define the space in which they are allowed to operate.

An important outcome of our analysis with regard to the politicization of non-state- CTs is that the state retains control over non-state-CTs and uses them politically in both countries, albeit in different ways. In Tanzania, non-state-CTs are integrated into the universalist ethos and the broad-based service provision, while in Uganda they are allowed to operate where the state is not currently forming direct relations with specific populations, allowing the state to focus its attention and resources on other constituencies. We therefore claim that the argument by Brass (2016) that NGOs in weakly institutionalized states tend to get involved in the governing of social provision, and not just its implementation, has to be modified in the case of CTs. Our findings suggest that, regardless of whether the mode of coordination is integrated or parallel, the state does not relinquish control when CTs are involved, but maintains control over how they affect the state’s legitimacy with important constituencies. We believe this might be more pronounced in relation to CTs than with other types of social provision, as the fungibility and novelty of this type of resource distribution compels the state to stay in control.

Our case studies also demonstrate that the politics of both non-state actors and social contracts are highly dynamic and that both countries move along a continuum between universalist and partisan distributional logics, as well as between integrated and parallel modes of coordination, thus shifting the pattern of politicization.

The example of GiveDirectly’s CT programme being suspended in Uganda when it was seen to interfere in the ruling elite’s patronage-based provision during election campaigns shows the potential political sensitivity of CTs in the context of a partisan social contract. However, this conflict also led to high-level dialogue between the government of Uganda and GiveDirectly which contributed to lifting the suspension and led GiveDirectly to engage more directly with the government in providing CTs. Hence, in this case the conflict led to a shift to a more integrated mode of coordination in Uganda as well. In Tanzania, the long-standing universalist ethos was challenged by the late President Magufuli, who became more hostile to the provision of services to refugees and emphasized that service delivery to specific localities would reflect loyalty to the state (i.e. to the party in elections). Hence, the social contract in Tanzania includes partisan dimensions, which also influences the role of the non-state actors, thus indicating movements along the continuum between universalist and partisan distributional logic in certain situations.

In Figure 2 we have plotted the most important conclusions of our analysis along the two axes of our analytical framework. The figure depicts Uganda as being characterized primarily by a partisan distributional logic in its social contract and a parallel mode of coordination between the state and non-state-CTs, and Tanzania as being characterized by a universalist distributional logic in the social contract and an integrated mode of coordination. We have also included the movements along the continuums in the case of GiveDirectly in Uganda and the issuing of CTs to refugees in Tanzania.

Figure 2. Conclusions plotted in the model of the analytical framework used in this paper (created by authors)

While recognizing that further conceptualizations might be warranted, we believe we have demonstrated the usefulness of understanding the historical evolution of a social contract, along the mode of coordination between non-state actors and the state as important analytical lenses when trying to understand politicization of non-state-CTs.

Acknowledgements

The authors wish to thank members of the CASH-IN research programme, who all have contributed with comments, suggestions and critical engagement and/or in data collection: Rehema Kilonzo, Raphael Macha, Elizabeth Nyaki and Mary Peter from University of Dodoma; and Fred Bateganya, Susan Kavuma, Brenda Birungi, Simon Peter Nsereko and Steven Alor from Makerere University. The authors also wish to thank Rasmus Hundsbaek Pedersen from DIIS and Jeremy Seeking from the University of Cape Town, as well as colleagues from the EPIC research group at Roskilde University for constructive comments on earlier drafts of the article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research forms part of the CASH-IN research programme located at Roskilde University in Denmark (see https://ruc.dk/en/research-project/cashin-privately-managed-cash-transfers-africa), which is funded by the Danish Fellowship Centre under the Danish Ministry of Foreign Affairs through grant number 19–05 RUC.

Notes on contributors

Malin J. Nystrand

Dr. Malin J. Nystrand is an Assistant Professor in Global Political Economy at Roskilde University in Denmark. She holds a PhD in Peace and Development Studies from University of Gothenburg, Sweden. She studies the intersection between economic, social and political rationales, drawing on political economy, economic sociology and economic anthropology. Her research areas include the political economy of investments and social cash transfers, social responsibility and social embeddedness of local business actors, and corruption and post-conflict reconstruction. Empirically her research has focused on Eastern and Southern Africa, and in particular Uganda.

Jacob Ulrich

Mr. Jacob Ulrich is a PhD fellow at Roskilde University and has a master's degree in International Affairs from Columbia University. He is a practitioner with +25 years of experience working with development and aid management in more than 30 countries worldwide.

Lars Buur

Prof. Lars Buur is Professor in the Political Economy of Development at Roskilde University. He has a PhD in Ethnography and Social Anthropology from Aarhus University. His key research interests include state formation and sovereignty, natural resource investments, industrial policy in Africa and social cash transfers. His key publications can be found in Extractive Industries and Society, Third World Quarterly, Geoforum, among others. He has also co-authored with Lindsay Whitfield, Ole Therkildsen and Mette Kjær The Politics of African Industrial Policy (Cambridge University Press 2015) and has edited several books and journal special issues on traditional authorities, security and sovereignty.

Bruno Yawe

Dr. Bruno Yawe is an Associate Professor at the College of Business and Management Science (COBAMS), Makerere University. He holds a PhD in Economics and Master's degree in Economics from the University of Dar es Salaam, Tanzania. He has 20+ years in teaching economics at Makerere University. He has also served as a Senior Research Fellow at Environment for Development and also capacity building of institutions like Uganda Management Institute (UMI), National University of Rwanda, Kenya School of Monetary Studies, Uganda Institute of Banking and Financial Services and many more.

Lutengano Mwinuka

Dr. Lutengano Mwinuka is an Economist who works at Department of Economics, The University of Dodoma, Tanzania, as a Senior Lecturer, Researcher and Consultant. His areas of research interests include value chain upgrading, livelihoods, innovative solutions, and the impact of social cash transfers on marginalized groups.

Emmanuel Keith Kisaame

Mr. Emmanuel Keith Kisaame is a PhD fellow at Makerere University and has a master's degree in Economic Policy and Planning from Makerere University. Emmanuel has 13 years of research experience in public financial management from a developing world context.

Notes

1 Non-state actors involved in cash transfers include international, national and local NGOs, as well as bilateral and multilateral organizations within the UN system. These types of organization often fund state-managed CT programmes, but here we are focusing on those who are managing such programmes directly. We use the term ‘non-state-CTs’, denoting CT programmes managed by non-state actors, but will, where warranted, distinguish between individual organizations, and types of organization.

2 This article is part of the ‘CASH-IN’ research programme located at Roskilde University in Denmark, with Makerere University in Uganda and the University of Dodoma in Tanzania as partners, funded by the Danish Ministry of Foreign Affairs.

3 A recent case study of a non-state-CT’s impact on state-citizen relations in Uganda (Grisola et al., 2023) shows mixed results in terms of citizens’ perceptions of the state and of non-state actors, thus highlighting the need for further exploration of these relationships.

4 We draw on the theoretical tenants of the political survival approach that is intimately related to the political settlement approach (see Behuria et al., 2017). This approach suggests that the “policy choice and implementation cannot be separated from ruling elites’ strategies to ensure political survival” (Whitfield et al., 2015, p. 11).

5 We acknowledge that no state is monolithic, i.e. different relationships exist between national and sub-national politics, where clear intentions across different scales cannot be assumed. Nonetheless, we posit that it is analytically possible and meaningful to recognize the distinct features of the distributional logics and modes of coordination in Tanzania and Uganda and to explore the patterns of politicization that emerge based on these differences.

6 In Tanzania, interviews were conducted with representatives of HelpAge International, SHEDEPHA+, the Danish Refugee Council and Kwa Wazee Tanzania, the Registrar of Societies at the Ministry of Home Affairs, the Registrar of NGOs at the Ministry of Health, and the Refugee Department with the Ministry of Home Affairs. In Uganda, interviews were conducted with GiveDirectly, DanChurch Aid, Lutheran World Federation, the Danish Refugee Council, and the Inter-Agency Cash Working Group.

7 We acknowledge that there exist many differences between the two countries, that CT programme features (conditional/unconditional; funder; type of non-state organization, size, time-frame, etc) might affect its politicization at the case level (Alik-Lagrange et al., 2021), and that it constitutes a reduction to analyse this along just two dimensions. However, we believe that this analytical framework captures important aspects of how non-state-CTs are politicized, which can be tested and explored in more detail at the case level in subsequent studies.

8 Although our framework is developed based on our analysis of Uganda and Tanzania, we suggest that it has a broader application, which can be explored in subsequent studies.

9 Seekings (2008) identified three ‘distributional regimes’: a labourist regime, i.e. giving social insurance etc. to formal sector workers; a pauperist regime, i.e. targeted social assistance schemes; and agrarian regimes, i.e. agricultural services to smallholder farmers. In Africa the agrarian regime is the most common, while social assistance is on the increase. While CTs fall into the ‘pauperist’ category in Seekings’ typology, our distinction between universal and partisan logic in how distribution is carried out could in principle be applied to the other two regimes as well.

10 A universal ethos in our definition does not necessarily mean that social provision takes the form of universal programmes (as distinct from targeted programmes), but that the distributional logic is based on an intention to include as many as possible, given the resources available.

11 An extensive web of regulations applies to various aspects of civil-society organizations’ activities, which are also monitored by various regulatory authorities and mean that each organization might need to relate to various government institutions, depending on its type of activity.

12 National organizations are starred and all the others are international, which in Tanzania means that the organization is registered in more than one country.

13 An example of non-state actors strengthening the state-based programme is the technical support provided in 2014–2016 by four UN agencies (the ILO, UNICEF, UNDP and UNFPA) to strengthen policy frameworks in implementing the PSSN (within TASAF), both at the national policy level and in implementation. An example of joint efforts between a non-state-CT and the state is that, based on the experience gained from the PSSN1 programme, TASAF, the Tanzania Commission for AIDS (TACAIDS) and UNICEF Tanzania, as well as other key stakeholders, worked together on a specific social protection and economic empowerment intervention targeting young people.

14 In the pilot stage a Vulnerable Family Grant (VFG) was included, but it was discontinued due to difficulties in implementation related to its embeddedness in patronage networks (according to Tran and Ghadially, 2021).