This paper examines voters’ perceptions of how high councillor turnover shapes the quality of local governance. Adopting a qualitative approach, the paper interrogates the impact of excessively high (82%) councillor turnover during the 2019 Local Government Elections on local governance. This study specifically asks: What is the effect of high councillor turnover on local governance performance from citizens’ perspectives? Data were collected from voters and members of Village Development Committees (VDCs) in Karonga, Salima, Balaka and Nsanje Local Government Councils. Two wards were selected from each council, yielding eight wards in total. Voters and members of VDC members provided information through Focus Group Discussions, while chiefs, council officials and civil society organisation representatives were also consulted through Key Informant Interviews. The paper finds that the perceived impact of councillor turnover on local governance was uneven. Governance improvements were observed across both high- and low-turnover wards, with varied opinion within wards. This suggests that turnover alone is not a panacea for improving governance unless it is supported by institutional reforms that regulate the behaviour of both councillors and secretariat staff. The findings reveal how systemic elements, such as institutional capacity and local government system design, and councillor-related factors, such as approaches to representation and information-management abilities, shape governance perceptions. The paper further reveals that voter-related factors, such as civic literacy, biased evaluations of incumbents' performance, and performance expectations, also shape voters’ governance perceptions. The findings provide insights into the limits of election-based accountability. They are useful for reframing policy debates that may guide development of governance frameworks for local councils in Malawi and other jurisdictions with similar contexts.

Local government in Malawi faces persistent challenges related to service delivery, citizen trust, and accountability (Centre for Social Research [CSR], 2017; O’neil et al., 2014). This has prompted debates about the most effective interventions for improving performance of the councils. However, these debates have long overlooked electoral dynamics, given the absence of regular local government elections (LGEs). Since the attainment of democratic dispensation in 1993 and the enactment of the Local Government Act in 1998, Malawi had not held consecutive LGEs until the 2014-2019 electoral cycle. The first consecutive LGE resulted in 82% councillor turnover (Chiweza et al, 2021), surpassing the African Average of 54% and Malawi’s legislative turnover of 55% between 1999 and 2014 (Sambo 2015). This LGE opened up spaces for scholarly debate on the bearing of electoral changes on the operations of the councils, effectively providing a basis for examining the impact of councillor turnover on local governance from a voter perspective. High councillor turnover attracted attention as a potential catalyst for improved local governance. However, despite this expectation, there is limited empirical evidence on voters’ perception of the impact of the electoral change. This paper contributes empirical evidence on how the local governance system responds to electoral changes. The paper approaches this debate with the following primary question: How does high councillor turnover during Malawi’s 2019 LGEs shape citizens’ perceptions of local governance quality? This question is supported by two specific questions: How did newly elected councillors respond to high turnover? Did replacing former councillors make incumbents more responsive and improve governance? The study primarily explores citizens’ perspectives on governance change by comparing governance practices before and after the 2019 LGEs. The paper specifically examines citizens’ perceptions of whether governance progressed, regressed, or stagnated in relation to three governance tenets: participation, transparency and accountability.

A thorough examination of these perceptions helps identify levels of citizen satisfaction with governance practices, structures, and institutions, which can inform the design of local governance frameworks. This study provides rare qualitative empirical evidence from Malawi’s local government context, where limited research has examined the implications of councillor turnover in a decentralised set-up. Beyond Malawi, the findings contribute to broader debates on accountability and the effectiveness of decentralised governance in sub-Saharan African countries whose local governments operate under fragmented accountability systems. By focusing on how citizens interpret governance outcomes after elections, the paper provides insights into the limits of election-based accountability in emerging democracies.

This paper is organised into five main sections. The first section introduces the study and provides its contextual background. The next section reviews literature on councillor turnover, its drivers, and its effects on local governance from empirical and theoretical perspectives. The paper then presents the theoretical framework and methodology of the study. This is followed by findings and discussion focusing on citizens’ perceptions of governance changes across councils and wards. The final section concludes by highlighting the implications of turnover and governance reforms, the study’s limitations. and areas of future research.

Studies highlight the importance of councillors in shaping governance. At the local government level, councillors play an intermediary role by translating community priorities into policy outcomes and providing oversight roles to monitor bureaucrats. However, they too are held to account by citizens, mainly through elections. Elections are a fundamental accountability mechanism in representative democracies and generate turnover among politicians, including councillors (Toral, 2024). Political turnover generally refers to the aggregate level at which elected representatives exit office from one general election to the next (Matland & Studlar, 2004), including the ways in which they are replaced. There are two forms of turnover among elected representatives: electoral and pre-electoral turnover. Pre-electoral turnover happens when politicians voluntarily or involuntarily leave office before a general election, whereas electoral turnover occurs through defeat at the ballot when incumbents are voted out. Both reforms contribute to incumbency changes.

2.1 Determinants of councillor turnover

Reasons for councillor turnover are diverse and context-dependent. Pre-electoral turnover can occur through retirement and resignation due to lack of motivation (Alfazema, 2009; Gouglas & Maddens, 2017; Sambo, 2015. When incentives are low, incumbents may leave politics for other, more rewarding endeavours or seek political offices perceived as more rewarding. Death, illness, and political pressure are other drivers of voluntary pre-electoral turnover.

Electoral turnover depends on voters’ decisions, which are mainly influenced by personal evaluations and endorsements. Political party identification is a major determinant of voting behaviour during elections (Antunes, 2010; Mahsud & Amin, 2020; Zeib et al., 2022). This demonstrates that voting behaviour is not entirely driven by performance but is also mediated by emotional attachment to the candidate and partisan identity. Therefore, turnover should not be interpreted as a reflection of objective governance assessments, but rather as an outcome of politically conditioned evaluations. These evaluations may be shaped by local political alignments, party loyalty, and campaign narratives. Voters in many contexts demonstrate responsiveness to a candidate’s party affiliation during elections. For instance, in Bangladesh (Uddin, 2016), Argentina (Sebastian, 2017), and Zambia (Mac Donald, 2014), political parties play a major role in shaping voter behaviour, usually through party-based endorsement. Performance evaluations are another dominant basis used by rational voters to decide whether to re-elect or vote out the incumbents (Kulachai et al., 2023; Zeib et al., 2022). This particularly aligns with the reward-punishment thesis within the agency framework, according to which voters vote out non-performing incumbents and re-elect those who demonstrate the desired performance.

2.2 Governance consequences of councillor replacement

Scholarly discourse demonstrates that councillor turnover has both positive and negative consequences for representation, accountability, and the consolidation of democracy at the local level (UEDE, 2019). On a positive note, high turnover is often regarded as a corrective signal of voter dissatisfaction with the incumbents and a desire for improvement. Alfazema (2009) observes that turnover creates an opportunity for the introduction of innovative ideas from new representatives, as new office holders may influence new policy direction. The exit of incumbents also opens career-progression opportunities for lower cadres of politicians. Hall and Snyder (2013) observe that high-quality challengers are often scared off, discouraged, and prevented from joining politics by incumbents. Therefore, incumbent turnover may attract high-quality challengers. At the global level, drawing on data from presidential and parliamentary elections, Marx et al. (2022) found that turnover results in improved economic performance.

Relating to governance, Marx et al. (2022) argue that incumbency turnover signals the effective functioning of democratic accountability. Accountability in the local government system should be understood from two perspectives: horizontal accountability, whereby councillors are accountable to citizens, and vertical accountability, whereby councillors hold the secretariat to account. Furthermore, councillor turnover has important implications for understanding representation dynamics, local-level accountability, and the extent of democratic consolidation.

Political science scholars such as Huntington, Przeworski, and Limongi view turnover as an indicator for democracy consolidation (Thomas 2015; Dulani 2011). However, Dulani (2011) contends that, for alternation to be an effective tool for democratic consolidation, political leadership should alternate between different political parties. Dulani further argues that alternation between members of the same party may help them amass benefits in their favour, sustain patronage networks, reduce prospects for leadership accountability, and hinder the promotion of policy.

An analysis of Brazilian mayoral turnover reveals the adverse effects of turnover on the quality of public services, especially when the bureaucracy is not shielded from political control (Akhtari et al., 2022). Electoral turnover has also been observed to be costly for governance and to disrupt the delivery of public services (Toral, 2024). In a study of local government turnover in Malawi, Chiweza et al. (2021) argue that, on the one hand, very high turnover weakens the development of experienced politicians who can overcome executive dominance at the local government level. Chiweza et al. also argue that excessive turnover weakens institutional memory, which is essential for preserving undocumented experiences. Extremely low turnover, on the other hand, entrenches incumbency, which fuels corruption as ruling elites attempt to control bureaucracy, thereby weakening oversight mechanisms (Chiweza et al., 2021). Although studies by Chiweza et al. have explained drivers of councillor turnover in Malawi, its influence on local governance remains underexplored, creating a gap in empirical evidence on the implications of 2019 councillor turnover for local governance. Existing scholarship on councillor turnover predominantly focuses on patterns of incumbent exit and replacement, but limited attention has been directed towards understanding how turnover shapes citizens’ lived experiences and post-election governance evaluations. Most studies conducted in Malawi treat turnover either as a threat to institutional continuity or as a corrective mechanism for under-performance. Rarely do these studies examine whether electoral replacements actually translate into governance improvements, particularly in relation to participation, transparency and accountability. Additionally, existing studies have inadequately examined how councillor behaviour, governance institutions and civic literacy interact to mediate local governance outcomes after high councillor turnover. This study addresses these conceptual and empirical gaps by examining voter perceptions of governance changes across councils and wards with different turnover dynamics.

This study is anchored in Principal-Agent Theory (PAT). PAT explains the contractual obligations between voters and political representatives. The theory focuses on the contractual relationship between the principal and the agent (Bernhold & Wiesweg, 2021). The principal delegates some decision-making authority to the agent to maximise the benefits of the contract in return for incentives. 

At the core of PAT is the notion that the voter, as principal, holds the politician as agent, to account under the contract. This entails that the theory focuses on agents’ responsiveness to the principal’s goals and on how their actions are mutually mediated to form an accountability framework in political institutions (Sean, 2014). According to Bernhold and Wiesweg (2021), PAT is premised on four assumptions: (1) actors behave as benefit maximisers; (2) conflicting interests exist; (3) actors have only limited rationality; and (4) information asymmetries exist between the principal and the agent. These parameters moderate the behaviour of principals and agents. This study examines how the councillor, as an agent, balances personal interests and the public good by ensuring good governance in response to voters’ expectations expressed during elections.

The study adopted a qualitative approach, targeting four local government councils (LGCs): Karonga in the Northern Region, Salima in the Central Region, Balaka in the Eastern Region and Nsanje in the Southern Region. The LGCs were selected along regional lines that have, over time, displayed partisan voting patterns, especially support for the Democratic Progressive Party (DPP) in the south and the Malawi Congress Party (MCP) in the Central Region. This selection helped to assess the influence of partisanship on voters’ governance assessments and perceptions. The LGCs were also selected based on high councillor turnover in Karonga and Salima and low councillor turnover in Balaka and Nsanje during the 2019 LGEs. This selection criterion helped to ascertain the association between governance perceptions and the extent of councillor turnover. While most studies analyse turnover at an aggregate national level, this study adopts a comparative approach by comparing high- and low-turnover councils. A further comparison is made by the type of turnover through an examination of electoral and pre-electoral turnover wards. This multilayered approach enables the detection of variations in governance perceptions across institutional and political contexts, rather than assuming that governance outcomes are uniform.

Two wards were purposively selected from each council: one ward in which the councillor did not seek re-election and another in which the councillor sought re-election but lost. Therefore, eight wards were included in the study. This arrangement helped to determine variations in voters’ governance perceptions between pre-electoral turnover wards (PTWs) and electoral turnover wards (ETWs). Eight focus group discussions (FGDs), involving 40 voters and 38 VDC members, were conducted across the councils. Four chiefs provided information through key informant interviews (KIIs). All participants were purposively selected. Constituents voted with expectations of the new representatives, and these expectations formed the frame of reference for their evaluation of governance. VDC members and chiefs are part of governance institutions that oversee village-level governance (Government of Malawi [GOM], 2013). Additionally, KIIs were conducted with four council officials and four civil society organisation (CSO) representatives to obtain insights into governance changes and to enable comparison between ward-level and council-level perspectives. The study included only constituents and VDC members who had voted in the same ward in both 2014 and 2019.

The study used thematic qualitative analysis, in which transcribed FGD and KII data were coded and organised into recurring themes reflecting participation, transparency and accountability. During the coding process, repetitive governance narratives across and within councils were identified. Further interpretation of the data followed a multilayered analytical approach in the sense that analysis involved inter-council, followed by inter-ward and intra-ward comparisons. Inter-council-level comparisons helped to ascertain variations in voter perceptions between high and low turnover councils. Intra-ward analysis revealed variation in governance perceptions among various social groups: constituents, VDCs and CSOs within each ETW and PTW. This approach enabled the identification of institutional, councillor-based and citizen-based factors that shaped governance perceptions within wards. The approach resonates with the PAT by demonstrating the existence of accountability relationships at multiple levels.

Evidence across the four councils and wards shows that the impact of councillor turnover on perceptions of local governance is uneven and highly dependent on councillor behaviour, the effectiveness of institutional frameworks, citizens’ awareness of governance, and citizens’ expectations. This contrasts with the normative expectation that turnover automatically leads to improved governance. The paper finds positive and negative governance perceptions in both low- and high-turnover LGCs, suggesting that high councillor turnover alone does not guarantee improved governance. Therefore, changing councillors alone is not a panacea for improved local governance. To achieve meaningful progress, turnover must be accompanied by robust governance frameworks across councils to minimise council-level inconsistencies. In the absence of strong regulatory frameworks governing the behaviour of councillors and council officials, electoral changes risk being more symbolic than substantive as governance improvement mechanisms. Case-based analysis of data reveals council- and ward-specific governance perceptions.

A comparison across the four councils does not show a clear link between turnover patterns and governance improvement from respondents’ perspective. Despite all councils operating under the same legal framework, they produce divergent governance outcomes from the citizens’ perspectives. The comparison across the councils reveals both negative and positive governance perceptions, regardless of the degree and nature of councillor turnover. For instance, both high-turnover Salima and low-turnover Balaka councils demonstrated negative governance perceptions. Similarly, Nsanje, which had low turnover, registered perceived governance progression, while Karonga, which had exceptionally high turnover, recorded both progression in PTWs and regression in ETWs. 

These findings suggest that high turnover alone does not guarantee strengthened transparency, accountability and participation. In this case, turnover only replaces political actors without transforming institutional conditions that moderate governance behaviour. The findings complicate the assumption in democratic accountability literature that electoral sanctions against incumbents induce cooperation among political actors (Alam, 2024; Balliet et al., 2011). Instead, the findings align closely with Akhtari et al. (2022) and Chiweza et al. (2021), who argue that turnover can disrupt governance where institutional accountability structures are weak. The result also implies that other factors, possibly institutional and contextual, influence the behaviour of councillors. Councils with weak monitoring capacity and poorly defined reporting lines are likely to sustain poor governance after turnover because newly elected councillors join the same compromised system. In the absence of strong institutions, councillor replacement only maintains the governance dilemma that new councillors inherit and may sustain the same weak accountability system. This calls for complementary accountability mechanisms to regulate agents’ behaviour, reduce information asymmetries and inconsistent governance practices.

5.1 Governance perception patterns within wards and across councils

A number of cross-cutting patterns emerged from the data, explaining governance perceptions among voters with regard to participation, accountability and transparency. Citizens’ perceptions of participation varied widely within and across councils, mostly tied to councillor accessibility and consultation. For instance, participation was perceived to have improved in wards where councillors were accessible, physically present and attended ADC and VDC meetings. This pattern was predominant in Nsanje’s PTW and ETW and in Karonga’s PTW:

Respondents in Nsanje perceived an improvement in councillor engagement compared with previous office holders, highlighting increased consultation and responsiveness. One participant observed that "the current councillor at least consults people" (Nsanje PTW, VDC Members 3 and 4). Similarly, a traditional leader remarked:

"The current councillor comes to get our concerns and attends VDC, ADC and council meetings." (Nsanje PTW Chief, 24 September 2023)

By contrast, a participant from Karonga reflected on the previous councillor's performance, stating:

"The former councillor was not available to people as a messenger of their concerns."
(Karonga PTW Participant 2, 17 May 2023)

Conversely, respondents in both ETW and PTW sites in Salima and Balaka, as well as ETW sites in Karonga, perceived a decline in citizen participation. Participants attributed this to limited consultation, the bypassing of formal governance structures, and absenteeism from Village Development Committee (VDC) meetings. One respondent explained:

"We hear the council outcomes through functions such as funerals or when people are called to participate in community activities." (Karonga ETW, VDC Respondent 2, 18 May 2023)

Similarly, respondents in Balaka noted the declining engagement of elected representatives. One participant commented:

"The former councillor used to attend ADC meetings." (Balaka ETW, VDC Respondent 1, corroborated by VDC Respondent 5)

Another respondent expressed concern about the motivations of the current councillor:

"The current councillor only comes when invited by NGOs, where they receive an allowance of MK6,000." (Balaka ETW, VDC Respondent 47, February 2023)

The general pattern shows that constituents often equated participation with employment in local development projects, while VDCs understood participation through formal governance structures, particularly regular meetings. This entails divergence in the understanding of participation between governance institutions and voters. For citizens, participatory governance means getting involved in community projects for tangible material benefits, while chiefs’ and VDCs’ view participation as a legitimised decision-making procedure. This means constituents perceive participation not merely as a declaration, but as behaviour demonstrated by councillors. Therefore, constituents prioritise substantive participation over symbolic participation.

Transparency is another governance tenet that this paper examined. Citizens’ perceptions of transparency were shaped by information flows and communication channels. The information-management approaches adopted by councillors strongly influenced transparency perceptions. Councillors who shared project information, disclosed contractor details, and provided timely feedback from the council received positive transparency assessments. Positive assessments of transparency were prevalent in Nsanje’s ETW and PTW and Karonga’s PTW and, some parts of Salima, as evidenced by the following reflection: “Unlike previously, we now know everything that is happening’ (Salima, Constituent 1, supported by constituents 4, 5, 7 and 8, 14 April, 2023).” Negative perceptions of transparency largely emerged in Balaka, Karonga ETW, and parts of Salima. These negative perceptions were associated with wards where councillors withheld information or relied on informal communication channels, such as funerals and other unofficial gatherings:

Current councillor does not report our challenges, unlike former councillor (Balaka, Constituent 18, February, 2023). When we give him our concerns, he should be arranging meetings and giving us feedback (Balaka, Constituent 8, 8 February 2023).

Across councils, constituents appeared to receive more timely information than VDCs because councillors used selective communication through ad hoc gatherings. This raised concerns about information asymmetry, which potentially weakened accountability and led to inconsistent evaluations. Overreliance on informal channels for feedback weakens information exchange, which in turn limits citizens’ capacity to monitor councillors and hold them accountable. Unofficial communication platforms also allow selective disclosure of information, creating informational advantages for the councillors.

Accountability perceptions were also varied across councils and wards. The perceptions reflected councillors’ willingness to engage with institutional platforms through attendance at VDC meetings and adherence to governance norms. Positive perceptions of accountability were highlighted in both Nsanje’s PTW and ETW and in Karonga’s PTW due to councillors’ regular attendance at VDC meetings and responsiveness to community concerns. These strengths were lacking in Balaka and Salima, where chiefs highlighted persistent non-attendance at meetings, lack of feedback, and self-interest, signalling regression or stagnation of governance.

The current councillor does not come to meet us so that we explain our concerns for him to take them to council (Respondent 7 VDC, Lipimbi Ward, 13 April 2023).

Current councillor does not report our challenges, unlike former councillor (Balaka, Constituent 18, February 2023).

When we give him our concerns, he should be arranging meetings and giving us feedback (Balaka, Constituent 8, 8 February 2023)

This finding shows that turnover generated gains in accountability only where councillors satisfied institutional expectations, underscoring the importance of institutional arrangements that advance the answerability of agents to principals. Furthermore, the findings demonstrate that governance improvement is not completely electoral-turnover-driven but also councillor-specific. Overall, the findings demonstrate that where one governance dimension fails, others also weaken. This underscores the interdependence among participation, transparency and accountability.

Variations in governance perceptions were also prevalent in the wards, where VDCs and constituents often demonstrated divergent views on the state of governance after turnover. This shows that governance perceptions varied between councils and wards, as well as within wards, signalling the existence of factors that facilitated divergences at each level.

5.2 Determinants of governance perceptions across councils and wards

Governance perceptions among citizens were shaped by council-level, citizen-based and councillor-based factors. Across-councils focus reveals that variations in governance perceptions largely emanated from a weak institutional framework and from the design and structure of the local government system. Malawi’s local government system rests on weak institutions that compromise local governance. In this context, an institution is viewed as a system of formal rules, informal norms, procedures, or shared values that constrain and prescribe political actors’ interactions for proper coordination (Bourguignon & Platteau, 2022; Farkas, 2019). Several institutional factors shape governance quality, including legal frameworks, economic policies, the rule of law, control of corruption, voice and accountability, and regulatory quality (Grabowski & Self, 2020; Poniatowicz eta lal., 2020). However, not all these factors apply in all contexts; some are more relevant to central government and less applicable to local government.

The institutional framework for Malawi’s local government system is anchored in legal instruments that contain several gaps, which contribute to variations in governance practices and perceptions. This is evident in the lack of institutional capacity to constrain councillors to expected governance behaviours and protocols, culminating in variable governance perceptions across councils. For instance, councillors are not restricted to an institutionalised approach to representation. Instead, it is left to each councillor’s discretion to choose among trustee, delegate, and gyroscopic approaches, whether these match or deviate from voter preferences. This stems from limited institutional clarity regarding how councillors should perform their roles and responsibilities, resulting in inconsistencies in representation approaches that shape voter perceptions. Furthermore, weak institutional prescriptions lead to poorly defined accountability relationships between principals and agents, allowing councillors, as agents, to reinterpret their mandates to suit personal or partisan incentives, especially in the absence of immediate sanctions between elections.

The current institutional framework, as expressed in the Local Government System Guide for Malawi (Government of Malawi, 2013), provides secretariat-level compliance mechanisms but weak compliance frameworks for councillors’ accountability to voters. The lack of such accountability obligations and extra-electoral sanctions leaves accountability largely at councillors’ discretion, leading to variable governance approaches. Furthermore, the Malawi Constitution does not expressly set a recall provision for councillors. The removal of councillors between elections is limited to only two mechanisms: the constitutional provision on the date of general elections, or a presidential order following a High Court decision if the councillor acts unlawfully. Neither option gives voters direct authority to recall the councillor. The absence of recall provisions weakens citizens’ enforcement of the principal-agent mandate and, in the process, undermines institutional checks and balances that ensure accountability.

The local government system is the administrative tier closest to citizens and is directly involved in the delivery of public services in many countries (Riaz et al., 2020). It therefore plays a significant role in shaping local governance. A well-designed local governance system enhances participation, transparency and accountability, which are expected to produce consistently positive perceptions across governance levels. Malawi’s local government system is single-tiered and comprises 35 local government councils that operate independently within the same decentralisation framework. However, the effectiveness of decentralisation as a participatory structure varies across councils, depending on councillor competence, civic awareness of citizens and engagement, which deepen governance variations. Furthermore, Malawi’s decentralisation process is affected by policy incoherencies, which fragment local governance by creating blurred mandates and information asymmetries (O’Neil et al., 2014). These ambiguities make the system vulnerable to manipulation and inconsistent decision-making across councils.

Overlapping mandates create institutional coordination challenges (Matovu et al., 2010). O’Neil et al. (2014) further observe that the informality of Malawi's local government system gives rise to ad hoc systems adopted by district and city executives. This contributes to inter-council governance disparities, influencing differences in governance performance and perceptions. This finding is consistent with a study conducted in Norway and Denmark, which revealed that institutional mandate overlaps undermine accountability and fail to constrain actors’ behaviour for effective local governance (Torfing et al., 2022). In such cases, Torfing et al. encourage changes to local political institutional design to enhance the political leadership of councillors. This view also aligns with research on Jordan’s National Assembly, which found that significant institutional changes affect governance performance (Al-Alagmeh & Megdadi, 2024). The study argued that revising institutional governance principles, rules, and controls improves the quality of oversight performance.

Similarly, Armstrong and Li (2022) observe that local governments face the fundamental challenge of determining the sustainable governance structures and practices that meet community needs while addressing governance concerns. A well-designed local government system fosters clear power divisions, strong communication between levels of hierarchy, adequate local capacity, and financial independence to improve public trust. Malawi’s local government system diffuses responsibility across multiple actors, consequently compounding the agency problem.

Beyond the council-level perspective, variations in governance perceptions existed both across and within PTW and ETW. Within the wards, constituents’ perceptions diverged from those of local governance institutions. Chiefs and VDC members were more likely to perceive governance regression, whereas constituents reported governance progression. Variations in local governance perceptions are common in many settings. For instance, Iran et al. (2018) and Mbandlwa (2024) observed conflicting perceptions between the people and members of local government regarding citizens’ preferences on intensity of participation. These differences reflect not only power balance relations but also judgement of what participation entails.

The variations also existed between wards; this is evident from the observation that governance perceptions were slightly more positive in ETWs than PTWs, particularly in Salima and Nsanje. This suggests that incumbent councillors in ETWs were more responsive to electoral sanctions, demonstrating alignment with the reward-punishment thesis, which suggests that politicians respond to electoral punishments (Balliet et al.2024). Polarised governance perceptions, whether within or between wards, reflect limitations in local governance. This paper reveals councillor-based and citizen-based determinants of divergent governance perceptions between and within wards.

5.3 Councillor behaviour as a determinant of governance perceptions

The governance perceptions were linked to the behaviour of councillors. Positive perceptions of participation dominated in wards where councillors were accessible and regularly consulted citizens but declined in wards with absenteeism and weak consultation. Likewise, transparency was perceived to be strong in wards where councillors disclosed project information but weak where feedback was channelled through informal structures. Attendance at official meetings, such as VDC and ADC meetings, improved perceptions of accountability, although this effect was weak where such institutions were bypassed. These observations suggest that councillors contributed to variations in governance perceptions in several ways.

Councillors contributed to variable governance perceptions through differing information management strategies, which produced information asymmetries. Such asymmetries were prevalent, with constituents often more informed than formal governance institutions such as VDCs. Councillors differed in their communication channels; some opted for direct links with constituents, while others engaged formal governance structures. This created differences in governance perceptions between constituents and VDCs. Bypassing formal governance intermediaries often fragmented accountability by reducing official bargaining power within the governance hierarchy. Information imbalances between constituents and VDCs can fragment collective negotiating power at the grassroots level because they stifle voters’ ability to monitor policy outcomes (Bravo & Silvestre, 2004) and participation (Clarkson et al., 2007). Lee (2025) observes that information asymmetry helps agents dominate the narrative and facilitate self-interest, thereby creating a principal-agent problem. The VDC is the most proximate negotiating institution for the voters in the local governance hierarchy. Choosing to inform constituents rather than local governance institutions could therefore be a strategic way of weakening collective capacity to hold incumbents accountable through a stronger institutionalised platform.

By creating information flow challenges, councillors sought to exploit informational advantages by informing citizens directly and manipulating voters’ perceptions for electoral advantage. Such strategic communication may be intended more to secure electoral support and show relevance within the political space than to promote accountability for good governance. Sidelining formal governance structures also reflects the strategic behaviour of councillors to garner support directly from constituents in search of favourable evaluations to maximise electoral advantage with strategic principles. Providing more information to voters could also weaken party influence, as information asymmetry that disadvantages voters increases reliance on party affiliation in political evaluations (Zabinski, 2017). This also reconfigures the agency accountability relations, whereby the councillor responds to the principal primarily for political survival rather than in accordance with formal governance norms. Furthermore, the findings reveal a deviation from the traditional expectation that a single principal holds the agent accountable. Instead, agents appear to navigate competing demands from multiple principals: voters and governance institutions. This suggests a shift in the principal-agent model from voter-representative relations to a multi-level accountability environment for councillors, which fragments governance.

Councillors also contributed to varied governance perceptions through the governance approaches they adopted. Some councillors adopted a delegate-oriented approach by consulting voters directly or indirectly through VDCs, whereas others adopted a trustee-oriented approach that involved limited voter consultation. Representation approaches that did not correspond with voter preferences were perceived negatively. This signalled variability in preferences for participatory decision-making processes. VDCs preferred active participation in decision-making through consultation, reflecting a delegate approach, whereas voters appeared more tolerant of being sidelined from participatory decision-making processes. This suggests that citizens tolerated a degree of trustee-oriented representation, thereby creating governance perceptions that differed from those of governance institutions. A study conducted in Haifa, Israel, by Forman-Rabinovici and Beeri (2023) indicates that representation contributes to citizens’ satisfaction and trust, thereby shaping governance perceptions. Therefore, deviation from the preferred representation approach generates negative perceptions.

Effectiveness of turnover in ensuring positive governance perceptions depends on the extent to which the incoming councillor engages with institutional processes and grassroots structures. Consultative and grassroots-oriented councillors are perceived more positively than elitist leaders. These findings suggest that turnover alone cannot deliver meaningful improvement unless accompanied by adherence to governance norms, including balancing voter preferences with institutional expectations. The findings further suggest that councillors’ institutional and community embeddedness is more important for improving local governance than electoral change alone.

5.4 Citizens’ expectations, civic literacy and governance perceptions

Some governance perceptions were shaped by citizens’ political orientations and knowledge of governance. One such citizen characteristic was their varied understanding of governance, which complicated uniform interpretations of councillors’ governance performance and highlighted the need for sustained civic education. For instance, constituents often equated participation with involvement in local projects and accountability to councillors’ availability. VDC members and chiefs associated participation with attendance at meetings for structured consultation and accountability with formal scrutiny processes. These differences in understanding contributed to variations in perceptions of councillors’ governance performance. Voters with high civic knowledge are more likely to demand structured accountability, whereas those with limited understanding of governance tended to equate it with handouts and mere presence of a councillor.

These findings align with the observation by Mohammadi et al. (2018) that, while citizens and local government officials agree on the importance of participation, contrasting views arise regarding the level and extent of citizen participation in power sharing and decision-making. Local government tends to prefer authority-centred participation over citizen-centred participation (Mohammadi et al., 2017). This depends on how people understand participation within a spectrum that includes eight characteristics: openness, being informed, being listened to, discussion, consultation, exchange, collaboration, and co-creation (Källström & Smith, 2022). From a principal-agent perspective, heterogeneity in principals’ understanding of governance and their capacity to monitor agents appropriately reflects associated civic literacy challenges. Therefore, strengthening standardised civic education through structured institutional frameworks can harmonise agents’ evaluations of governance.

Citizens demonstrated divergent performance expectations, which in turn shaped their perceptions of governance. Many studies found that expectations influence perceptions and shape evaluations (De Lange et al., 2018; Trafton, 2019). Expectations can be biased and may lead individuals to selectively focus on information that confirms those expectations while ignoring or downplaying contradictory information. Differences in expectations among individuals therefore create disparities in perceptions of issues. Prior expectations of councillors’ performance shaped differences in governance perceptions among citizens. Voter expectations are heterogeneous; some voters have higher expectations of councillors’ performance than others. Lower expectations can result in more positive evaluations, even when improvements are modest, whereas higher unmet performance expectations may lead to perceptions of regression. Similarly, Markwat (2023) observes that voters with lower expectations easily reward fulfilment more than those with higher expectations. Therefore, moderate performance can be evaluated positively when expectations are low, but less favourably when the expectations are high. In some cases, even poor performance can be rated positively when voter expectations are too low.

Loyalty-linked voter biases and the tendency to rationalise candidate choice were among the citizen-based factors that shaped variations in governance perceptions among constituents. This was evident when most constituents who supported the former councillor mostly perceived regression and stagnation, whereas most voters who opposed the former councillor perceived governance progression. These differences reflect the influence of emotional attachment to the candidate more than the actual state of governance. Some studies show that strong voter loyalty can potentially hinder transparency and accountability because highly loyal voters may be less willing to hold their preferred party or candidate accountable for poor performance (Adriënne de Geus, 2017; Mbandlwa, 2024). Voter loyalty to a party or candidate can create biased halo effects, leading individuals to downplay negative aspects of an incumbent’s performance and focus instead on positive elements. Such voters are therefore likely to perceive progression even when indicators suggest otherwise. By implication, principals are not always rational evaluators of agents; rather, they sometimes act under the influence of subjective frames of reference that shape their perceptions. Loyalty-based biases and poor understanding of governance make perception-based evaluations of governance subjective because the same governance conditions may be interpreted in multiple ways.

High councillor turnover did not automatically translate into good governance. In many cases, positive and negative governance perceptions were observed across both high- and low- turnover councils, underscoring the need for stronger institutional frameworks to strengthen consistency of accountability efforts. This means that high councillor turnover can have a positive bearing if accompanied by institutional reforms that enhance councillor accountability, improve information flows, and strengthen civic education. Complementing electoral changes with institutional mechanisms, such as robust accountability frameworks across councils, would help constrain opportunistic behaviour and minimise council-level inconsistencies. Effective leadership practices are essential for ensuring good governance across and within wards. Without these practices, change through turnover risks being symbolic rather than substantive in improving local governance. The findings point to the need to re-interpret the agency framework to accommodate the thinking that agency dynamics are shaped by many recurring sanctioning patterns. These findings point to the need to reinterpret the principal-agent framework to accommodate the view that agency dynamics are shaped by recurring sanctioning patterns. The study provides an additional dimension in which turnover is viewed not merely as an electoral event, but as a broader governance issue shaped by institutional and contextual conditions.

Despite its theoretical and empirical contribution, this study has some limitations. First, the research focused on only four local government councils; therefore, the findings may not fully represent governance dynamics across all councils in Malawi. Second, the study’s qualitative design prioritised in-depth interpretation over statistical generalisation, thereby limiting the generalisability of the findings. Third, the findings relied entirely on inherently subjective citizen perceptions. These subjective interpretations may have been influenced by personal expectations, partisan attachment, and recall bias. Finally, the findings are based on the first consecutive local government elections held since the introduction of Malawi’s multiparty democracy in 1993. Therefore, the perceptions expressed by citizens cannot be interpreted as evidence of inherent governance behaviour or a stable trend unless confirmed by trends in subsequent elections.

Given these limitations, future research could adopt a mixed-methods approach to involve a larger number of councils to enable quantitative assessment of the relationship between turnover patterns and governance outcomes. Further studies could also be conducted to assess the effectiveness of decentralisation in supporting electoral changes intended to improve local governance. Furthermore, research may be carried out beyond the Malawi context to examine how party competition and bureaucratic capacity interact with electoral turnover to shape governance outcomes across sub-Saharan Africa.