Walking rallies: opposition party’s new campaigning approach in Tanzania’s 2020 election

ABSTRACT

Rally-intensive campaigns have been one of the main characteristics of elections and political parties’ communication and organising practices in Tanzania (Paget, 2019). However, due to a changing political landscape – intensified authoritarianism from 2016 – opposition political parties have had to think of new ways of mobilising and organising. These new ways of political mobilising and organising had an impact on opposition parties’ election ground campaign. In this article, I explore and analyse the campaigns of Chadema, the main opposition party in Tanzania, in the 2020 general election. Through participant observation, personal experiences, interviews, and documentary sources. I examine the new campaigning activity, which I describe as a walking rally, deployed by Chadema’s parliamentary candidates. The article contributes to the studies on electioneering and campaigning in sub-Saharan Africa. Furthermore, the study adds to relatively sparse analyses of African opposition parties’ mobilisation, organising, and campaigning practices.

KEYWORDS:

• Rallies
• elections
• campaigns
• authoritarianism
• Tanzania

Introduction

Election ground campaigns, which means candidates’ direct and face to face contact with the voters (Enos & Hersh, 2015), are common in many countries in sub-Saharan Africa. These campaigns, often, take place within electoral-authoritarian regimes in which daily and small-scale oppression is increasingly intense and violent. These regimes, of course, oppress opposition parties and alter circumstances within which they campaign.

In this article, I ask, how do opposition parties adapt their ground campaigning practices in electoral authoritarian regimes? The implicit assumption is that ground campaign activities at large are activities to which opposition parties turn in electoral-authoritarian regimes. They offer a way for opposition parties to circumvent the challenges of regime oppression. However, I find that as oppression increased, the leading opposition party in Tanzania – Chadema – developed and adopted a distinct mode of ground campaign: the walking rally. In brief, a walking rally is a ground campaign activity whereby a candidate organises a public walk within the constituency area while attracting people to join the walk along the way. It is, thus, a moving rally, with the understanding of the rally as a ‘a public event at which speakers’ address an audience face to face for the ostensible purpose of politically mobilizing it’ (Paget, 2019, p. 451). While walking, the candidate does a number of other ground campaigning activities such as canvassing.

A critical look at scholarly literature shows that attention is turning to election ground campaigning. Most works identify and analysed the modes of face-to-face contact and associated campaign practices which are adopted in contemporary ground campaigns (for example, Foucher, 2007; Bleck & Van de Walle, 2018; Paget, 2019; Brierley & Kramon, 2020; Krönke et al., 2022). These include: rallies (Rauschenbach, 2015; Paget, 2019; Brierley & Kramon, 2020); door-to-door canvassing (Kramon, 2017); gift distribution (Brierley & Kramon, 2020); clientelism (Rauschenbach, 2015; Brierley & Kramon, 2020), pork barrel (Rauschenbach, 2015), ethnic appeals (Cheeseman & Larmer, 2015), valence appeals (Bleck & Van de Walle, 2018), voter intimidation (Rauschenbach, 2015), street marches (Foucher, 2007), and programmatic promises (Rauschenbach, 2015). Nevertheless, scholarship has not exhausted the topic, some studies are broad and there are still many countries that have not been adequately studied. This limits comparative examination, which is crucial for identifying key factors and causes informing election campaign strategies and practices in sub-Saharan Africa. In this paper, I argue that there is mutual production of various campaign activities – including phone canvassing, walking rallies, canvassing, and rallies. I do that through the analysis of 2020 election campaigns in Tanzania.

The context – the authoritarianism of the 2020 elections

Tanzania, specifically, is understudied with regard to election ground campaign. Moreover, most of the available analyses in the literature have focused on elections prior to 2020 such as those of 2015, 2010, and before (such as Morse, 2014; Makulilo, 2016; Paget, 2019). This paper focuses on the 2020 elections, which, I argue was significantly different from previous elections in Tanzania. This difference is due to increased authoritarianism ahead of the 2020 elections. Since 2016 the country has been governed by an increasingly electoral-authoritarian regime (Carothers, 2018; Campbell, 2020; Paget, 2021). Before the 2020 elections, the literature described Tanzania as a competitive electoral-authoritarian regime – that is a regime that combines democratic and authoritarian features (Ekman, 2009; Levitsky & Way, 2010; Morse, 2014; Makulilo, 2016). Similarly, prior elections reflected the same. There was space for opposition parties to operate amidst the dominance of the ruling party. Following the 2015 elections, a new administration by John Pombe Magufuli took a sharp turn towards authoritarianism (Paget, 2017; 2021; Carothers, 2018). Magufuli’s administration imposed impromptu prohibitions on opposition party activities, restricted civil society activities, stifled the media, and arrested opposition politicians. Critical journalists and activists were also arrested, and some disappeared. Altogether, in the 2020 elections, the authoritarian measures were so great that Tanzania was turned from a competitive electoral authoritarian regime into a hegemonic electoral authoritarian regime (Paget, 2021). In this section, thus, I present the context of the authoritarianism of 2020 elections. Understanding of the context is crucial to allow us to discover and explore the intricate ways in which ground-campaigns were organised and functioned (Mahoney, 2007).

The sharp authoritarian turn in Tanzania following the 2015 elections has been taken as a manifestation of a continuous and entrenched party state authoritarianism (Campbell, 2020; Boakye, 2021; Cheeseman et al., 2021). The 2015 elections were competitive, which prompted the moves to restrict political space in 2020. President Magufuli was determined to weaken and undo any progress that opposition parties had made (The Economist, 2017; Campbell, 2019). Going into 2020 election, the opposition political parties were repressed with the banning of political public rallies, restricted internal party meetings, censored media, arrest, kidnappings, and the disappearance of opposition politicians, activists, and journalists. In 2016 (Kabendera, 2016; The Citizen, 2016), political rallies were banned, which hindered political parties, and in particular the opposition parties from using their conventional mobilising and organising approach.

Consistent with trends since 2016, the 2020 elections in Tanzania were the most violent election (Human Rights Watch, 2020; Tanzania Elections Watch, 2020; United Nations, 2020; Minde et al., 2021) in the history of the country since the re-introduction of multi-party system in 1992 which emerged during the third wave of democratisation (Møller & Skaaning, 2013; Osaghae, 1999). The first election in 1995 heartened young opposition political parties,¹ with a 26% vote share for the opposition presidential candidate. Fast forward to 2015 – progress was not transformational but still significant – the opposition's vote share was at 39% and with over 100 members of parliament from the opposition wing. However, the 2020 election was a case of democratic backsliding in which the ruling party officially got 84% of the presidential vote and won 97% of parliamentary seats. Chadema secured only 1 seat – which had not even been held by Chadema before. These results were attributed to massive electoral fraud and electoral violence against the opposition. Thus, the 2020 election sharpened authoritarianism in Tanzania.

The case

In this article, I take the case of 2020 elections in Tanzania. The paper narrows its study to Hai District which is one of the opposition strongholds and the constituency from which the chairman of the leading opposition party – Freeman Mbowe (2020) – also comes from. Mbowe had served as its member of parliament for three terms (2000-2005;2010-2015; 2015-2020). From 2010, he served as the leader of the opposition in parliament. Freeman Mbowe was the parliamentary candidate for Hai in 2020. Given the profile of Hai and its candidate, the district was highly targeted by the regime. In light of this, the case of Hai generates observable inferences (George & Bennett, 2005; Mahoney, 2007) that can help us understand activities and new ways of ground campaigns in the context of electoral authoritarianism.

Through participant observation as the leader (Director of Hai District Chadema Campaigns) and therefore active member of the campaigning team, I experienced and observed the activities and entire campaign process for two full months between August 1st and October 31st 2020. I planned, listened, discussed, and participated in all campaign processes and activities – from inception to execution of all activities. Some of the activities occurred without planning while others were a result of planning. The findings from the observations were combined with data from other constituents in the form of videos and images. Throughout the campaign period and after, I followed what other Chadema candidates both in urban and rural constituencies were doing. By being a member and a leader of the party chairman’s campaign team – I had the advantage of meeting and discussing campaign activities and plans for other parliamentary candidates. There were times when candidates of nearby constituencies (in the northern Kilimanjaro and Arusha regions) would come to Hai to meet the chairman and discuss campaign issues. Through those meetings, I gained insights into ground campaigns in different and similar contexts. I also reviewed video clips and images from other constituencies such as Ubungo in Dar-es-Salaam, Mbeya Municipality, and Nyamagana in Mwanza. These videos and images were shared on social media platforms. I analysed those files during the campaigns (to learn about other candidates’ activities) and again when writing this paper. Interviews were carried out with candidates and other campaigning team members both in Hai and in other constituencies.

Ground-campaign practices and activities in Tanzania

Paget (2019) provides, arguably, the most thorough examination of a campaign practice in Tanzania – developing a theory of a ‘rally-intensive campaigns’. Rally-intensive election campaigns play host to ground campaigns with the following features: (i) local candidates convene rallies as frequently as national candidates do; (ii) local rallies are better attended in aggregate than national ones; (iii) the canvass pales in comparison to the rally as a means of campaign; and (iii) local campaigners do not direct most of their efforts to canvassing (Paget, 2019, p. 456). In brief, Paget defined rally-intensive campaigns as ‘the higher aggregate rally attendance climbs, the more rally-intensive the ground campaign becomes’ (2019, p. 454). In his analysis, he argued that Tanzania has the most rally-intensive campaigning in Africa and also in the world. The analysis goes beyond description to show how the rallies are produced at different levels. His study has added significant value to our understanding of ground election campaigns in sub-Saharan Africa. However, the study focuses on 2015 and preceding elections, which were not affected by the sharpened authoritarianism in the country, which led to the ban on political rallies in 2016. It is noteworthy, that rallies were a means through which opposition parties mobilised voters and members even when elections were not in progress. Chadema, for example, carried out a number of rally operations/movements across the country to build the party (Paget, 2019a; 2022; Kwayu, 2022). Therefore, from 2016 Chadema had to innovate new ways of mobilising, organising and party building in the context of authoritarianism. This was in line with the argument of Riedl (2014) and Riedl et al. (2020) that parties may institutionalise more under authoritarianism. Moreover, Paget (2019a) shows the ability of the party to mobilise was due to its extensive party building. The work done by Chadema between 2005 and 2015 had built party resilience that enabled it to innovate and mobilise even in the most autocratic periods. Chadema deployed a household level member registration known as Chadema Ni Msingi (Chadema is its foundation). This exercise, which involved registration of members, informed the 2020 election campaign at the constituency level. The exercise of member registration at household level – by door-to-door mobilisation – indicates that the party was able to adapt. Adaptation is one of the areas that Africanists studying political parties in Africa have overlooked (Paget, 2020).

This paper, thus, attempts to explain a new ground campaign activity that was observed in the 2020 election campaign by opposition parties in Tanzania. I call this activity a ‘walking rally’. The walking rally did not replace the other campaigning activities, but they defined a distinctive practice of ground campaigning that was not used in previous elections. It is another example of party adaptation to new contexts. In relation to the rally-intensive approach, Chadema still deployed the rally-intensive approach at the presidential (national level) campaign while having a mix of both approaches (walking rallies and conventional rallies) at the local level for parliamentary elections. The production and conduct of the national rallies conformed to the model outlined by Paget (2019). However, the parliamentary candidates deployed a different approach, a walking rally, that fed into a conventional rally at the end of the day. Critically, the walking rallies could be said to be part of the production of rallies and the rally-intensive campaign.

Through planning, execution, and observations of 14 whole day walking rallies in 14 electoral wards within the district, I gathered information and insights on preparations, activities, actions, reactions, practices, and conversations around the walking rallies. This was, thus, an ethnographic study as it involves all the features of ethnography including: exploring the nature of the walking rallies; unstructured data; a small number of cases; and an interpretation of meanings and functions of human actions (Aktinson & Hammersley, 1998). To seek further clarification, interviews, and regular conversations were carried out with the district parliamentary candidate and ward councillor candidates. Overall, 16 interviews were carried out: ten with campaign team members of Hai, and six with campaign team members of other districts – four urban and two rural constituencies. In addition, campaign teams members were also sought out for clarification. Social media posts (videos and images) of Chadema’s other parliamentary candidates’ campaigns including those shared in WhatsApp groups were also analysed for more insights. As Hitchen et al. (2019) and Cheeseman et al. (2020) show how WhatsApp has become a vital tool in election campaigns. Although the authors attribute the spread of fake news during elections to WhatsApp, it is also fair to argue that WhatsApp has played a positive role in mobilising and communicating to voters important party issues. Information shared includes content of speeches (audio and video) and the schedule for rallies and candidates’ plans. In the Hai district, for example, a Google spreadsheet with a voter database was linked to the web WhatsApp function through which party members and voters with WhatsApp accounts could receive daily updates on the campaigns events and a message from the candidate. The function allowed for interactions between campaign communication teams and voters. Information gathered by the candidates’ call centre team included: the number of voters contacted and key information such as membership status, branch and ward location. Knowledge about campaigns and candidates was also interrogated to provide further insight into walking rally production and preparations. I immersed myself in and analysed all this data and gathered new information on an ongoing basis to create room for further probing and analysis during the data collection period – which was the two month official campaign period. Following the end of the campaign, more systematic analysis was conducted against the background of literature on election campaigns in Tanzania and in sub-Saharan Africa at large. I also took time to acquire some analytic distance after my immersion in the campaign, and my position in it as a Chadema campaign manager.

With respect to the above, this article argues that the walking rally is a distinct activity in election campaigns that can stand alone as a campaign activity as well as feeding into rallies and other activities in the ground campaign. Thus, the walking rally can be analysed as an activity and a means for the production of rallies and other campaign activities. This paper makes a critical contribution to the literature, not only by exploring and examining a distinctive activity on election ground – campaigning in sub-Saharan Africa, but also by using the case of an opposition party. Studies on opposition parties in Africa are emerging but there is still a need for further research.

Electioneering and campaigns in sub-Saharan Africa

Analyses of election ground campaigns in Africa have identified a number of activities that are used either exclusively or inclusively. There are campaigns that use a mix of rallies, canvassing, and clientelism. Some focus on rallies only. However, there is a nuance that is still missing in these analyses. For example, how one campaign activity can be a means of producing another campaign activity without compromising its own independence. In this paper, I make a contribution by showing how walking rallies are a campaigning activity in their own right and also a means through which rally-intensive and other approaches can be produced. To understand that it is crucial to engage in the details of each of these campaign activities.

Rallies have been identified as the main election campaign approach in sub-Saharan Africa, yet they deserve further analysis as they can be affected by various factors including resources and contexts. Paget (2019) has so far provided a nuanced analysis of rallies in Tanzania by disaggregating the types of rallies, which led to the identification of the rally-intensive campaign. In his analysis, he examined the production of these rallies and how they are carried out. He emphasised the intensity of them by charting not only the frequency of the rallies but also the attendance. The analysis also shows the distinction between local and national rallies, in which Paget observed that local rallies are better attended in the aggregate than national ones. This analysis captures well the campaigning practices in elections prior to 2020 elections. Literature shows that often the rallies are accompanied by canvassing. But in rally-intensive, canvassing is minimal in that the ground campaign efforts are concentrated not on canvassing, but on a bundle of activities through which, Paget argues, are what produce rallies.

Nevertheless, in other sub-Saharan African countries such as Ghana and Senegal, canvassing has been described as one of the main means of campaigning. Michelson and Nickerson indicate that ‘personalised outreach is more effective at mobilising voters than indirect method.’ (2011, p. 230). Based on their argument, it is tempting to conclude that parties would, automatically, choose canvassing as opposed to rallies. However, Brierley and Kramon (2020), for example, show how campaigns in Ghana used door-to-door canvassing on the basis of a strategic decision, not necessarily for its theoretical capacity to mobilise voters. These strategic decisions, are based on the incumbency status of the party. From that, a party will choose a campaign activity or approach based on its confidence in voters’ behaviour. Due to the ability to connect directly with individuals or groups of voters, canvassing can be associated with clientelism as well as ethnic-based mobilisation strategies. Van de Walle (2009) defines political clientelism as the web of clientelist relationships that structure the relationship between the state and citizens. These relationships can come in various forms. In electoral politics clientelism is related to distributive politics and in election campaigns, clientelism has been related to vote buying and the growing role of businessmen in electoral politics (Pottie, 2003; Van de Walle, 2009). Clientelism is a campaign approach that is mostly adopted by ruling parties due to the link to state resources.

In relation to clientelism, there is the promise of giving public goods, which gives the advantage to ruling parties as compared to opposition parties. Programmatic promise giving and pork-barrel can go hand in hand with voter intimidation in particular with disbursement of funds and public services in their constituency following the election. Voter intimidation can also be done through threats of business closure or increased regulations. Moreover, campaigns can be accompanied by violence, disappearance, and arrests (Rauschenbach, 2015). Between 2016 and the 2020 election, voters and opposition politicians and candidates in Tanzania have faced intimidation and violence instigated by the ruling party (Cheeseman et al., 2021). Evidence of similar approaches has been documented in other sub-Saharan African countries’ elections (Basedau et al., 2007; Daxecker, 2014; Söderberg Kovacs & Bjarnesen, 2018).

The above discussed approaches and activities have all been practiced, to different extents, in Tanzania. The practice of one approach versus another depends on the political parties’ position and ability to execute an activity. For the ruling party, which has been in power in Tanzania for more than 60 years, there are many options for them given their access to state machinery, specifically security institutions and resources. Opposition parties’ campaigns since the 2005 election, and in our case Chadema, have mostly deployed rallies as a campaigning approach. In 2005, 2010 and 2015 Chadema carried out rally-intensive campaigns. In the 2005 elections, Chadema put up a presidential candidate – Freeman Mbowe – for the first time since it was registered. Freeman Mbowe used a helicopter to carry out rallies across the country. A means which was unprecedented in the country at the presidential level and that led to a nickname ‘kamanda wa anga’ or air commander (Mtatiro, 2015). The use of a helicopter gave more than logistical advantages – it also stimulated the crowd and attracting people towards the rally. Before landing the helicopter would hover around for a few minutes and people would run to the location where the rally was being held. This was a means to produce rally. At the same time, the party’s local candidate would organise rallies in their constituencies – both at district and ward levels. These conventional practices changed in the 2020 election.

Another interesting campaigning activity in sub-Saharan Africa that has been pointed out in the literature are the blue marches in Senegal (Foucher, 2007). In the 2000 election in Senegal, the opposition party deployed street marches as a campaign activity. These marches were called blue marches. They were distinct from conventional campaigning approaches in Senegal. Foucher (2007) provides a detailed description of these marches with some notable features. First and foremost, he shows how the approach was makeshift, due to the limited resources that the opposition had. This, again, exemplifies the ability of parties in Africa to adapt. Otherwise, the blue marches targeted the youth, and used local languages and entertainments. Moreover, Foucher (2007) insists the blue marches were also a message to the ruling party and the state that if elections are not fair, the opposition have the capacity to riot and cause disruptions. It was, thus, a ‘euphemised rehearsal for what would happen’ (Foucher, 2007, p. 125).

The blue marches are a type of campaigning that are similar to the walking rallies. Each involves a moving crowd that follows a candidate. With respect to that, both blue marches and walking rallies construct candidates (Paget, Beardsworth and Lynch, 2023). Candidates – in both blue marches and walking rallies – lead the crowd. In the walking rallies, candidates walk in front of the crowd that keeps gathering behind as they move from one place to another. If the candidate stops to speak to an audience – for example, in a bar or market – the crowd will surround him/her. These actions construct the candidate as the leader and the one with vision. The candidate also gets an opportunity to speak about his vision and why he is running for the office. The blue marches, also provides an opportunity for the candidate to show leadership. However, blue marches and walking rallies have differences too. While blue marches are mainly urban, the walking rallies are both rural and urban. Also, walking rallies intention was to connect to people from where they were – in their households, streets, shops, markets, village centres and bars so as to meet them and speak to them – before the final big rally at the end of day. The blue marches were partly a threat of potential violence and a statement of warning. Walking rallies were a product of four years’ ban of public rallies and restricted political space for opposition parties to mobilise, while blue marches – were an improvised approach, partly due to limited resources. Thus, walking rallies, as noted above, are still distinct ground campaign activity that needs to be studied.

In the next section, I present findings and analysis of the walking rallies in the 2020 election campaign in Tanzania.

The walking rally

The walking rally, which is distinct from the conventional rally, I argue, became a middle way between rallies and canvassing. In a sense, the walking rally embodied the culture of the party, given that Chadema Ni Msingi means ‘Chadema is its foundation’. This is confirmed as members were visited and re-registered at the household level. Chadema Ni Msingi was coordinated from the national party office. During its implementation from late 2018 to mid-2019, the national office trained and distributed officials in all districts to supervise it. It is important to note the practices that go hand in hand with Chadema Ni Msingi operations, which were also a manifestation of change within the party approach as reflected in the 2018 party policy that focused more on building the party institutions than just being a reactive opposition. Such practices include verification of voting registration status of members. Members were also sensitised to update their vote registration. ‘Chadema Ni Msingi is a way through which the party re-evaluate its status by knowing its members and their key information at the sub-village level.’ The member registration process and the consequent registering of members and leaders from sub village level facilitates the systematic follow up of members. According to the analysed registration books, Hai District was able to register 41,000 active members who also possessed voting cards. Chadema Ni Msingi prepared the party for different campaigning activities in the 2020 election, which was happening in an increasingly autocratic context. In the rest of this section, I present a detailed and systematic description of the walking rallies. There are three key aspects of a walking rally. The first, is the preparation and production of a walking rally. The second, are the practices incorporated in the walking rally and other practices and modes of ground activities into which they lead, for example, the rally. Third, are the performances evident in those practices.

Preparation and production of walking rallies

Walking rallies did not happen spontaneously. There were significant and systematic preparations behind the scenes that enabled each walking rally to occur. To be allowed to carry out any campaigning activities in a ward, there should be a ‘go ahead’ from the district election returning officer. To secure this, each political party submits its weekly or monthly (if possible) campaign schedule (locations and time) to the returning officer. The objective is to avoid having different candidates from different political parties in the same place at the same time. Thus, the programme is prepared, submitted to the election returning officer in order to get a stamped copy of the programme.

In Hai District, a walking rally would typically start at 11:00am. Thus, each ward, depending on its geographical size and shape, would be allocated either one or two days. The campaign team would prepare a ‘walking rally map’ with a starting point, route and the closing place – where there would be a rally-intensive event between five and six pm. On the eve of the next day’s walking rally, a PA system car would go around the targeted ward to publicly announce the candidate’s visit. The PA system is accompanied by music, which is meant to raise a positive mood for the rallies. The announcement will mention the time and the route – the starting village to the last village. The last village would often be at the centre of the ward. In the location of the closing rally, there would be a big stage with a music system and from around 3pm, there would be various performances on the stage as people flock into the rally.

Prior to the public announcement – a day or two before visiting the ward – the campaign call centre team would call members of the party – who had been registered through the Chadema ni Msingi exercise – to inform them about the walking rally and the importance of sharing the message to neighbours. All phone calls were registered in the database on a Google sheet – between September and October 2020. Based on the analysis of the database, 6,418 members were called and responded in Hai District, which included 65% males and 35% females. This number is 15% of the registered members in the Chadema Ni Msingi books. Out of the 6,418 called, 48% identified themselves as belonging to a party wing such as the women’s wing (BAWACHA) (15%); the youth wing (BAVICHA) (21%); and the elders’ wing (BAZECHA) (12%).²

The phone calls also helped the campaign team to pick up issues facing people in different villages in order to brief the candidate prior to the visit. Thus, agenda setting was partly informed by the information gathered from the phone calls. Nevertheless, the candidates for ward council as well as ward level party leaders were responsible for setting the agenda and bringing information that would inform the parliamentary candidate’s agenda in that ward.

Other arrangements included security details for the candidate as well as for the crowd and the route. This was crucial given the level of violence in Hai District that was instigated by district authorities (see Minde et al., 2021; Francis, 2023). For details of how President Magufuli used local authorities to oppress the opposition, see McLellan (30 November, 2018). A number of opposition party candidates and followers were tortured, kidnapped, and arrested during the 2020 election campaign. Given the violent environment in the district, it was important for the campaign team to ensure that there was adequate security. The walking rally was a safer way of campaigning due to the open crowd and changing locations. Moving made it difficult for the gangs to organise a target. Furthermore, the defiance and boldness of the candidate as well as the display of private security may well have deterred potential attackers. In the next sub-section, I describe the incorporation of other ground-campaigning activities into the walking rally.

Incorporating other ground-campaigning activities into a walking rally

As noted above, the walking rallies supported and helped produce a rally-intensive campaign. Walking rallies also facilitate other campaign practices such as canvassing, ‘meet and greet’, and phone campaigning. It is a way that the candidate connects with the voters in a more direct way than the traditional rally-intensive campaign. Walking rallies were a mix of rallies and canvassing in some ways. The authorities, in particular the police force, attempted to stop the candidates from canvassing. At one of the walking rallies, the district police in charge for Hai District was recorded on a phone stopping and harassing the candidate, an incident which made national news. He told him that the candidate could not win against the ruling party candidate in the election. The candidate, Freeman Mbowe responded to the police saying ‘how can I not speak to the people while they stop me and want to speak to me?’³

Overall, the walking rallies proved an efficient way of gathering support. Given the nature of the walking rallies, canvassing became almost spontaneous in that the candidate could speak with any voter who showed an interest along the way. Although the village and ward party leaders might have arranged the audience and specific spots where the candidate would stop and canvass, often people came towards the candidate to talk to them and then joined the walk. Some would call and invite the candidate into their homes. If the candidate had time he would go in for a few minutes. Chadema Ni Msingi registration also facilitated home visits by the candidate. Also, some members expressed their wish to meet the candidate. If time allowed, the candidate would stop at their homes to speak to them, and allow time for photographs to be taken. Such actions created a bond between voters and the candidate.

So, walking rallies are also a product of other campaign approaches – for example, the phone call campaigns, and canvassing – which made the walking rallies more effective. The rally planned for later that day serves as an implicit justification for the walking rally – in that the crowd follows in the hope that there would be a closing event – which is the rally.

Performances during the walking rally

The walking rallies were lively. There was dancing and other performances such as welcoming the candidate at the start of the rally with a traditional welcoming dance or performance. For example, on the first day of the walking rally in Hai District, which happened in a ward known as Narumu – women had prepared a traditional meal to welcome their ‘son’ (the candidate). Others – in particular youths and party members – called Mbowe ‘kaka’ or brother. This is another form of candidate construction – in that, Mbowe who is a national party chairman is being ‘normalised’ as a family and thus creating bonds and trust between the candidate and the audience. Welcoming Mbowe in a homely manner was symbolic since he had been arrested and jailed various times between 2017 and 2020 – and the people felt that their son/brother had been mistreated and taken away from them. Thus, the campaign was considered a home coming. Each ward prepared different welcoming performances at the starting point of the walk including using traditional Chagga leaf isale, which is a symbol of remembrance and rituals (Silayo, 2022). Other activities might include prepared ndafu (grilled whole goat) or mtori (banana soup) for the candidate and his team. This was usually sponsored by a voter or party member, who would have made a request to the campaign team to offer the meal.

Nevertheless, such performances were neither the main nor the only activity during the walking rallies. The walking rallies were a well-rounded canvassing exercise, whereby the candidate would have several stops in shops, markets, bars/pubs, and at times even in selected or prompted households to speak to people, take their questions, get some encouragement, and have other conversations. Depending on the crowd size, at times, the candidate might even stop and address a group of people, say in a pub, shopping centre, or market for a few minutes before continuing to walk and talk to smaller groups or individuals. The focus of these conversations often pertained to specific issues of the village or ward area.

Depending on distances and population density, at times the candidate and accompanying campaign team would get into cars to drive slowly to another village starting point (in the same ward) to start walking again. Often a crowd would follow the procession with musical entertainment. These processions finally ended up in the last location for a rally at the centre of the ward. The rally would usually take the form described by Paget. They were vibrant, crowded and with a typical rally schedule – whereby party leaders would start talking, with a break for entertainments, then the parliamentary candidate would speak – mostly addressing the public – and finally introducing the ward councillor candidate as well as requesting that the crowd vote for the councillor candidate, for himself and for the party presidential candidate. The rally would end with five to ten minutes or so of a ‘civic education’ lesson on voters’ rights and how to vote. On the last days of the campaign, there would be a demonstration of the ballot paper to show the right way to vote. These demonstration ballot papers would be simultaneously shown by young party members who circulated among the crowd showing people how to vote.

Conclusion

Using an ethnographic approach, this article has presented an analysis of a distinct ground campaign activity which was deployed by Tanzania’s opposition in the 2020 general election. Analysis of ground campaigns in sub-Saharan Africa has examined various campaign activities in relation to a number of factors: such as resources, voters’ behaviour, incumbency status, and the nature of political regimes. These activities include rallies, canvassing, and clientelism. In this article, I have presented walking rallies as a campaign activity which is mutually constituted in that it is informed by other activities as much as it informs them. In this respect, walking rallies are both a product and a producer of other approaches – including rallies and canvassing.

The paper has located the analysis within the electoral authoritarian context. The ethnographic approach of participant observation and collection of detailed information and insight enabled the author to capture nuances that underlined the sophistication behind walking rallies within that context. Evidence shows that this campaign activity was deployed in both rural and urban constituencies. Data from observation, interviews, documentary and image analyses, as well as party campaign strategies show that the walking rallies were systematically planned and executed. Other related activities occurred both spontaneously and after prior arrangements. For example, canvassing sessions during the walking rally were both planned and spontaneous determined by time and other factors on the ground. Given this uniqueness and complication of the walking rally in incorporation of other activities, the activity deserves to be classified as a form of ground campaigning. The analysis has located this activity within the wider literature with examples of the blue marches in Senegal and also the production process of rallies as described by the analysis of rally-intensive campaigns in Tanzania.

This article thus, has contributed to our understanding of not only different types of rallies, but also the overall ground campaign in an authoritarian context as well as in sub-Saharan Africa – a region where much still has to be studied. It has attempted to explore the question of ways through which opposition parties in sub-Saharan Africa adapt ground campaigning practices in electoral authoritarian regimes. Further, the case of opposition in Tanzania adds to the nascent literature on opposition political parties’ organising in the region. Walking rallies are a crucial aspect of campaign rallies and they deserve to be observed, studied, and analysed for the purposes of wider understanding of ground campaigns in sub-Saharan Africa.

Acknowledgements

Thanks to Dan Paget and the editors of this special issue for their comments on previous drafts of this paper. Thanks to all interviewees in Tanzania (particularly candidates and 2020 election campaign team members of Chadema Party in Hai District). Thanks to Shirumisha Kwayu for reading through the drafts and for asking relevant questions that helped me to think through some of the activities carried out during the election campaigns within and beyond Hai District. Thanks to participants of the Election Campaign Rallies in Africa Workshop for the critical comments on the initial draft of this paper.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 Mostly three years old by then – as registration for new parties started after the 1992 amendments.

2 BAWACHA stands for Baraza la Wanawake Chadema (Chadema Women’s Council), BAVICHA stands for Baraza la Vijana Chadema (Chadema Youth Council), and BAZECHA stands for Baraza la Wazee Chadema (Chadema Elders’ Council).

3 A video clip of the incident was made available online (Chadema Media TV, 2020).