PARLIAMENTARY PRIMARIES IN GHANA’S NATIONAL DEMOCRATIC CONGRESS: EXPLAINING REFORMS TO CANDIDATE SELECTION AND THEIR IMPACT Susan Dodsworth Seidu Alidu Gretchen Bauer Gbensuglo Alidu Bukari MIASA Working Paper No 2021(1) University of Ghana, Legon December 2021 MIASA Working Papers 2021(1) Edited by the Merian Institute for Advanced Studies in Africa (MIASA) The MIASA Working Paper series serves to disseminate the research results of work in progress prior to public-cation in order to encourage the exchange of ideas and academic debate. The objective of the series is to publish research findings from the work of the MIASA Interdisciplinary Fellow Groups (IFGs) and other MIASA fellows in an open-access manner. Inclusion of a paper in the MIASA Working Paper Series does not constitute publication and should not limit publication in any other location. Copyright remains with the authors. The research leading to these results has received funding from the Maria Sibylla Merian Centres Programme of the Federal Ministry of Education and Research, Germany, under grant no. 01UK2024A-D, with co-funding from the University of Ghana. The Merian Institute for Advanced Studies in Africa (MIASA) cannot be held responsible for errors or any con- sequences arising from the use of information contained in this Working Paper; the views and opinions ex- pressed are solely those of the authors and do not necessarily reflect those of the Institute. Copyright on this Working Paper is held by the authors: © Susan Dodsworth, Seidu Alidu, Gretchen Bauer and Gbensuglo Alidu Bukari WP Coordination: Agnes Schneider-Musah All MIASA Working Papers are available and free of charge on UG Space: http://ugspace.ug.edu.gh Merian Institute for Advances Studies in Africa (MIASA) P.O. Box LG 1075, Legon, Accra, Ghana Tel.: 030 290 7319 Email: miasa@ug.edu.gh Website: https://www.ug.edu.gh/mias-africa/ To cite: MIASA Working Paper 2021(1). Susan Dodsworth, Seidu Alidu, Gretchen Bauer & Gbensuglo Alidu Bu- kari. 2021. Parliamentary primaries in Ghana’s National Democratic Congress: Explaining reforms to candidate selection and their impact. Online: hyperlink. MIASA Working Paper 2021(1) 1 Parliamentary primaries in Ghana’s National Democratic Congress: Explaining reforms to candidate selection and their impact Susan Dodsworth (University of Queensland) Seidu Alidu (University of Ghana) Gretchen Bauer (University of Wisconsin-Madison) Gbensuglo Alidu Bukari (University for Development Studies, Tamale-Ghana) Abstract Candidate selection procedures play a crucial role in shaping parliaments and influence the quality of democracy. Yet our understanding of how and why political parties reform their candidate selection mechanisms over time is surprisingly limited – especially in sub-Saharan Africa, where a number of parties have begun to shift towards more inclusive procedures. To address this gap, we examine the experience of Ghana’s National Democratic Congress, which reformed its selection procedures in 2015 allowing all party members to vote in primary elections for its parliamentary candidates. We identify four motivations that drove these reforms: making the party more democratic by expanding participation, reducing the cost of the primary process, building the organizational capacity of the party, and keeping up with the party’s main competitor. Each motivation mattered more to some within the party than others; almost all ended up disappointed due to a substantial divergence between actual and intended effects that ultimately led to the reversion of the reforms in 2019. Our findings leave us better placed to understand both why political parties in sub-Saharan Africa’s more democratic regimes have shifted towards more inclusive candidate selection mechanisms over time, and why the pace of that change has been slow and uneven. Keywords: Democracy; Primary Election; Political Party; Candidate Selection; Reforms Résumé Les procédures de sélection des candidats jouent un rôle crucial dans la formation des parlements et influencent la qualité de la démocratie. Pourtant, notre compréhension des reformes des mécanismes de sélection des candidats par les partis politiques et des raisons pour lesquelles ils le font au fil du temps est étonnamment limitée, en particulier en Afrique subsaharienne, où un certain nombre de partis ont commencé à s'orienter vers des procédures plus inclusives. Pour combler cette lacune, nous examinons l'expérience du Congrès national démocratique du Ghana, qui a réformé ses procédures de sélection en 2015, permettant à tous les membres du parti de voter lors des élections primaires pour ses candidats parlementaires. Nous identifions quatre motivations qui ont été le moteur ces réformes : rendre le parti plus démocratique en élargissant la participation, réduire le coût du processus primaire, renforcer la capacité organisationnelle du parti et se maintenir au niveau de son principal concurrent. Chacune de ces motivations était plus importante pour certains au sein du parti que pour d'autres ; presque tous ont fini par être déçus en raison d'un écart important entre les effets réels et les effets escomptés, ce qui a finalement conduit au retour des réformes en 2019. Nos conclusions nous permettent de mieux comprendre à la fois pourquoi les partis politiques des régimes plus démocratiques d'Afrique subsaharienne ont évolué au fil du temps vers des mécanismes de sélection des candidats plus inclusifs, et pourquoi le rythme de ce changement a été lent et inégal. Mots-clés: Démocratie; élection primaire; parti politique; sélection des candidats; réformes MIASA Working Paper 2021(1) 2 Acknowledgement The research leading to these results has received funding from the Maria Sybilla Merian Centers Programme of the Federal Ministry of Education and Research, Germany under the grant no. [01UK1824A]. We wish to express our gratitude to MIASA for including us in the Interdisciplinary Fellow Group (IFG) on parliaments and democracy in Africa, September to December 2019, and to thank all the members of the IFG for inspiring comments and feedback during the fellowship. We also want to thank the reviewers for their constructive comments on an earlier version of the paper. Introduction In many African countries, the selection of parliamentary candidates by political parties is where the “real” electoral competition takes place. The reason for this varies between coun- tries. It may be because a ruling party is dominant – as in Tanzania,1 because voting is heavily influenced by ethnicity – as in Kenya,2 or because parties have clear regional strongholds that render certain seats “safe” – as in Ghana.3 In such countries, political parties become the gate-keepers to the political world, since voters can only select from the “menu” of can- didates they offer-up.4 Parties’ candidate selection procedures influence who has access to political power, and what they do with it; they shape the composition of legislatures and the nature of political debates. These effects have particular significance in sub-Saharan Af- rica, where the ideological divides that characterise inter-party competition elsewhere are often weak – or absent.5 Candidate selection procedures can also create barriers to the par- ticipation of women in politics,6 or allow parties to favour wealthy candidates over those who are better qualified but have fewer financial resources.7 Both effects may, in turn, have a corrosive effect on the quality of democracy. Across sub-Saharan Africa, political parties are increasingly making use of primary elections to select parliamentary candidates, particularly in the region’s more democratic 1 Melanie O’Gorman, “Why the CCM Won’t Lose: The Roots of Single-Party Dominance in Tanzania,” Journal of Contemporary African Studies 30, no. 2 (2012): 313–33, https://doi.org/10.1080/02589001.2012.669566. 2 Nic Cheeseman and Miles Larmer, “Ethnopopulism in Africa: Opposition Mobilization in Diverse and Unequal Societies,” Democratization 22, no. 1 (2015): 22–50, https://doi.org/10.1080/13510347.2013.809065. 3 Anja Osei, “Political Parties in Ghana: Agents of Democracy?,” Journal of Contemporary African Studies 31, no. 4 (2013): 543–63, https://doi.org/10.1080/02589001.2013.839227. 4 Gideon Rahat, “Candidate Selection: The Choice Before the Choice,” Journal of Democracy 18, no. 1 (2007): 157–70, https://doi.org/10.1353/jod.2007.0014. 5 Sebastian Elischer, Political Parties in Africa: Ethnicity and Party Formation (New York: Cambridge University Press, 2013). 6 Vibeke Wang and Ragnhild Louise Muriaas, “Candidate Selection and Informal Soft Quotas for Women: Gender Imbalance in Political Recruitment in Zambia,” Politics, Groups, and Identities 7, no. 2 (2019): 401–11, https://doi.org/10.1080/21565503.2018.1564056. 7 Dominika Koter, “Costly Electoral Campaigns and the Changing Composition and Quality of Parlia- ment: Evidence from Benin,” African Affairs 116, no. 465 (2017): 573–96, https://doi.org/10.1093/afraf/adx022. MIASA Working Paper 2021(1) 3 political regimes.8 In 2001, for example, the ruling Botswana Democratic Party (BDP) adopted new regulations allowing all registered members to participate in the selection of its parliamentary candidates in a reform that became known as ‘bulela ditswe’ (open to all).9 In Nigeria, parties have been legally required to hold primaries since 2011,10 though party elites have typically retained a substantial degree of influence over candidate selection. In most cases, Nigeria’s primaries remain indirect, in the sense that the selectorate is consti- tuted by delegates (rather than all members). However, in the lead up to the 2019 election, the All Progressives Congress (APC), one of Nigeria’s two major parties, piloted direct pri- maries (in which all registered members were able to vote) in 16 states.11 African scholars have analysed this trend in terms of its implications for intra-party democracy.12 Yet despite their prior work, our understanding of why political parties in Af- rica might choose to reform their primary processes at a particular point in time, and the challenges that they confront in implementing those reforms, is curiously poor. This is due to two weakness in the existing literature on candidate selection mechanisms which we elaborate in more depth below. The first of these is an over-emphasis on the structural de- terminants of candidate selection mechanisms, something that makes it difficult to explain change over time. The second is a tendency to deduce the motivations of party leaders from the impact of candidate selection mechanisms in practice – an approach likely to lead to inaccurate conclusions given increasing evidence that candidate selection mechanisms of- ten have unintended effects. These leaves us poorly placed to understand how and why the candidate selection mechanisms adopted by political parties across Africa are evolving. The primary goal of this paper is to expand our knowledge of what motivates parties to reform candidate selection procedures, and the ways in which reforms can diverge from expectations, in the context of Africa’s more democratic political systems. In order to do so, it presents a case study of Ghana’s National Democratic Congress (NDC). Specifically, we examine the changes that the NDC made to its primary process in 2015 and 2019, prior to 8 This has occurred primarily in countries with single member district electoral systems, as primary elections are rarely used to draw up candidate lists in proportional representation systems. 9 Zein Kebonang and Wankie Rodrick Wankie, “Enhancing Intra-Party Democracy: The Case of the Botswana Democratic Party,” Journal of African Elections 5, no. 2 (2006): 141–50. 10 Section 87 of the Electoral Act 2011 stipulated that “a political party seeking to nominate candi- date for elections under this Act shall hold primaries for aspirants to all elective positions.” 11 International Republican Institute and National Democratic Institute, “IRI/NDI Nigeria International Election Observation Mission Final Report” (Washington D.C.: IRI and NDI, 2019), https://www.iri.org/sites/default/files/nigeria_election_report_updated.pdf. 12 Mpho Mlomo, “The Roles and Responsibilities of Members of Parliament in Facilitating Good Gov- ernance and Democracy,” in Botswana, Politics and Society, ed. W. A. Edge and M. H. Lekorwe (Pre- toria: J.L. van Schaik, 1998); A. O. Omobowale, “Flawed Political Party Primaries In Nigeria’s Fourth Republic: The Case Of The Peoples’ Democratic Party (PDP),” African Journal for the Psychological Study of Social Issues 11, no. 1 (2008): 282–91; Marietu Tenuche, “The Albatross Called Primary Elec- tions and Political Succession in Nigeria,” Canadian Social Science 7, no. 4 (2011): 121–30, https://doi.org/10.3968/j.css.1923669720110704.063. MIASA Working Paper 2021(1) 4 the 2016 and 2020 elections respectively. In short, in 2015 NDC leaders dramatically ex- panded the selectorate (that is, the body of people responsible for selecting candidates) by allowing all party members to vote in the parliamentary primaries. In 2019, this change was undone, with the NDC reverting to a more exclusive process, under which a small number of delegates and party officials (rather than all party members) were entitled to vote in pri- mary elections. We investigate the motivations and impact of each of these decisions, ask- ing several related questions. First, to what extent did the impact of the reforms made in 2015 match the party’s intentions in making those reforms? Second, what explains any dis- crepancy between the two? Third – and finally – how did the NDC’s experience of reform in 2015 and 2016 shape its decision to reverse the reforms prior to the next election? We find that one of the main stated goals of the reform process – to democratize the party primaries by expanding member participation – was achieved in a numerical sense: the number of people who voted in the NDC’s primaries increased substantially. However, this increase in participation was not matched by the achievement of the other two goals that had motivated the NDC’s leaders to adopt the reforms. The reforms largely failed to reduce the overall cost of the primary process – the second goal – although there is some evidence that it did make vote-buying more difficult. Nor was the third goal, to build organ- izational capacity of the party, achieved. Failure on both these fronts was significant for the sustainability of the reforms, since these latter two goals played a more fundamental role in driving the reforms. In contrast, the expansion of participation – though often talked about – was more a convenient means of justifying the reforms to the public than a true cause of change. Senior NDC leaders were also motivated by a desire to keep up with – and prefera- bly outpace – their main competitors, the New Patriotic Party (NPP). However, this was a longer-term driver of change rather than a more proximate trigger for the reforms adopted in 2015. We also show that the gap between the party’s expectations and the outcome was not primarily due to the design of the reform, but rather the manner in which it was imple- mented: party leaders seriously underestimated the scale of the logistical challenge that a radical expansion of the primary selectorate entailed. This, together with a perception that the reforms contributed to the NDC’s electoral defeat in 2016, led to the reversion of the reforms in 2019. Despite that, there appears to be a large degree of consensus within the NDC that a more inclusive primary process represents “the future of the party,” providing it first builds a reliable mechanism for identifying its own members. These findings deepen our understanding of how and why the candidate selection mechanisms employed by African political parties are changing over time. In particular, our analysis draws attention to the complexity of the motivations behind these reforms, which are rarely driven by a single goal, much less one that is shared equally by all actors within a party. Our findings also suggest that, despite the challenges encountered by the NDC, we are likely to see more political parties experiment with inclusive candidate selection mech- anisms in sub-Saharan Africa. Many of the factors that motivated leaders within NDC – such as a desire to curtail the role of money and strengthen the party’s organizational capacity – MIASA Working Paper 2021(1) 5 are likely to be shared by the leaders of other parties in the region. Perversely, however, our case study also exposes the fact that poor organizational capacity, in particular a party’s limited ability to identify its own members in a reliable manner, can act as a significant bar- rier to the introduction of inclusive candidate selection mechanisms long after the moment of democratic transition has passed. Such a lack of capacity – which might seem exceptional to those who study political parties in the Western world – is by no means unusual across sub-Saharan Africa. Together, our findings leave us better placed to understand both why political parties in sub-Saharan Africa’s more democratic regimes having shifted towards more inclusive candidate selection mechanisms over time, and why the pace of that change has been slow and uneven. Before presenting our case study, Section 1 provides an overview of prior research on candidate selection, and explains why existing research leaves us poorly placed to un- derstand what motivates reforms to candidate selection procedures and the challenges of putting them into practice. Section 2 describes and justifies our research design, before we present and analyse our empirical material in Section 3. In the final section we reflect on the implications of our findings for future research, and for political parties seeking to make their candidate selection procedures more inclusive. 1 Understanding reforms and the challenges of implementation Candidate selection mechanisms, including primary elections, vary in many different ways. They can differ with regard to eligibility (who may be a candidate?), the inclusivity or exclu- sivity of the selectorate (who chooses the candidates?), the degree of geographic centrali- zation (is the decision made at the national, regional or local level?), the selection method (is it by voting or appointment?) and the extent of institutionalisation (does the process consistently follow clearly specified rules?).13 Prior research – much of it conducted in estab- lished Western democracies – identifies a wide range of factors that shape a party’s choice of candidate selection mechanism.14 These factors include variables internal to parties, such as their ideology and degree of organization, was well as external ones, such as anticipated electoral benefits, the degree of party competition, state or national laws, and contagion at both the domestic and international level.15 Some warn that the determinants of candidate selection mechanisms may differ in sub-Saharan Africa, because there is relatively little ide- ological variation among political parties – a fact that drives party leaders to prioritize the 13 Merete Bech Seeberg, Michael Wahman, and Svend-Erik Skaaning, “Candidate Nomination, Intra- Party Democracy, and Election Violence in Africa,” Democratization 25, no. 6 (2018): 959–77, https://doi.org/10.1080/13510347.2017.1420057. 14 For reviews see: Gideon Rahat and William P. Cross, “Political Parties and Candidate Selection,” in Oxford Research Encyclopedia of Politics (Oxford: Oxford University Press, 2018), https://oxfor- dre.com/view/10.1093/acrefore/9780190228637.001.0001/acrefore-9780190228637-e-625; Bonnie N. Field and Peter M. Siavelis, “Candidate Selection Procedures in Transitional Polities: A Research Note,” Party Politics 14, no. 5 (2008): 620–39, https://doi.org/10.1177/1354068808093393. 15 Field and Siavelis, “Candidate Selection Procedures in Transitional Polities.” MIASA Working Paper 2021(1) 6 attraction and retention of wealthy candidates, as well as the mobilization of grassroots ac- tivists.16 Others, such as Bonnie Field and Peter Siavelis, argue that choices about candidate selection mechanisms are likely to be different in new or transitional democracies.17 In such contexts, they suggest, four main factors push parties to adopt exclusive rather than inclu- sive candidate selection mechanisms. First, they argue, party leaders in transitional democ- racies face a much greater degree of political uncertainty – a product of the transition pro- cess itself. Second, leaders must work with party organizations that are weaker and less ex- tensive, a state of affairs that is often the result of their repression (or outright illegality) prior to the transition. Third, in a transitional democracy, party leaders must contend with strate- gic complexity generated by the interaction of certain types of electoral systems18 with frac- tionalized party systems. Fourth, and finally, party leaders in transitional democracies are subject to greater pressure to maintain elite alliances, since elite pacts are often critical for sustaining democratic transitions. Together, these constraints lead political parties to adopt exclusive candidate selection procedures, rather than inclusive ones, because they reduce uncertainty, are logistically less demanding, facilitate co-ordination among the political elite and make it easier to maintain elite alliances. Despite prior research, our understanding of why political parties – both in Africa, and further afield – choose to reform their candidate selection procedures at specific points in time remains surprisingly poor, as does our knowledge about the challenges that parties confront in putting reforms into practice. This is for two reasons. The first reason is that much prior research on the determinants of candidate selection mechanisms emphasizes the role of structural factors, things like party ideology, the electoral system, the territorial and constitutional structure of the state, the degree of party competition and mass political culture.19 These structural factors are undeniably important, but they offer little leverage in explaining the proximate causes of reform, save when that reform occurs at a moment of political rupture, such as a democratic transition. Very little work has focussed expressly on reforms that take place at other points in time. The work of Shlomit Barnea and Gideon Ra- hat provides one exception: they explain the adoption of reforms by reference to whether a party is in opposition or anticipates an electoral loss, and demonstration or contagion ef- fects when other parties in a system have successfully introduced reforms.20 The second reason that our understanding of reforms, and their effects, remains lim- ited is that existing work tends to deduce the intentions of party leaders from political out- comes – and vice-versa. If, for example, a selection mechanism appears to increase party 16 Nahomi Ichino and Noah L Nathan, “Primary Elections in New Democracies: The Evolution of Can- didate Selection Methods in Ghana,” in Routledge Handbook of Primary Elections, ed. Robert G. Boatright, 1st ed. (New York: Routledge, 2018), 369–83. 17 Field and Siavelis, “Candidate Selection Procedures in Transitional Polities.” 18 Specifically, small magnitude proportional representation or majoritarian systems. 19 See Table 1 in Field and Siavelis, “Candidate Selection Procedures in Transitional Polities.” 20 Shlomit Barnea and Gideon Rahat, “Reforming Candidate Selection Methods: A Three-Level Ap- proach,” Party Politics 13, no. 3 (2007): 375–94, https://doi.org/10.1177/1354068807075942. MIASA Working Paper 2021(1) 7 cohesion, then party leaders are implicitly assumed to have chosen that mechanism in pur- suit of this outcome.21 Rarely have researchers actually asked party insiders what motivated their decisions, and only fairly recently has research explicitly questioned whether the re- sults that party leaders expect candidate selection mechanisms to achieve are realised in practice. Yet divergence between intentions and outcomes appears to be a common prob- lem. Prior work by Nahomi Ichino and Noah Nathan illustrates this: examining the candidate selection mechanisms employed by both of Ghana’s major parties, they observe that party leaders’ often ‘adjusted the rules to address one set of problems only to inadvertently com- plicate another, creating a need to further adjust nomination procedures in the future.’22 It is particularly important to understand how shifts towards more inclusive candi- date selection procedures become possible over time, the more proximate drivers of reform and the ways in which the impacts of reforms can diverge from a party’s intentions in the context of sub-Saharan Africa. Across the region, many African political parties – including those in Ghana – operate in political regimes that are quite stable and so can no longer be termed transitional.23 This does not mean that those regimes are fully consolidated democ- racies, but rather that the “moment” of transition has passed. Thus, it seems likely that many of the factors that characterise a transitional regime, and which – according to Field and Siavelis24 – are likely to preclude the adoption of more inclusive candidate selection mech- anisms may have faded over time. For example, both political uncertainty and the pressure to maintain elite alliances (lest the transition to democracy unravel) are likely to have be- come less pressing as new ‘rules of the game’ have become established. This may have eroded the constraints that make the adoption of inclusive candidate selection less com- mon in newer democracies, though it may not have erased them entirely. Other constraints may not fade at the same rate. In particular, many of the political parties (and political party systems) that operate within sub-Saharan Africa’s more democratic regimes remain weakly institutionalized.25 A significant proportion of African parties lack the organizational strength of their counterparts in other regions, including basic capabilities such as the abil- ity to identify their own members. Alone, this may not always deter African political parties from adopting ambitious reforms to candidate selection mechanisms – but it may make 21 Shane Mac Giollabhuí, “How Things Fall Apart: Candidate Selection and the Cohesion of Dominant Parties in South Africa and Namibia,” Party Politics 19, no. 4 (2013): 577–600, https://doi.org/10.1177/1354068811407599. 22 Ichino and Nathan, “Primary Elections in New Democracies.” 23 Thomas Carothers, “The End of the Transition Paradigm,” Journal of Democracy 13, no. 1 (2002): 5–21. 24 Field and Siavelis, “Candidate Selection Procedures in Transitional Polities.” 25 Matthias Basedau and Alexander Stroh, “Measuring Party Institutionalization in Developing Coun- tries: A New Research Instrument Applied to 28 African Political Parties,” Working Paper (German Institute of Global and Area Studies, February 1, 2008); Edalina Rodrigues Sanches, Party Systems in Young Democracies: Varieties of Institutionalization in Sub-Saharan Africa (Abingdon: Routledge, 2018). MIASA Working Paper 2021(1) 8 divergence between a party’s intention in adopting reforms and the impact they have in practice particularly likely in the African context. Might divergent rates of change among these different factors help to explain why political parties in sub-Saharan Africa appear to be moving towards more inclusive candidate selection mechanisms, but in a slow and une- ven fashion? Our analysis suggests that they do. 2 Research design Analysing the case of the NDC in Ghana offers a chance to address shortcomings in our un- derstanding of why political parties reform their candidate selection mechanisms, and the challenges they encounter when they do so. Ghana is one of Africa’s most stable emerging democracies. Since the re-introduction of multi-party democracy in 1992, there have been several alternations in power between the NDC and its main competitor, the NPP. These alternations have been both peaceful and regular, with power changing hands every eight years. This pattern evidences a two-party system that is both quite stable and relatively well institutionalized – characteristics that set Ghana apart from many other African countries. While both the NDC and the NPP have a nationwide presence – something mandated by Ghanaian party laws – and eschew appeals to an ethnic base, both have clear regional strongholds.26 While the NPP draws most of its support from the Ashanti and Eastern Re- gions, the NDC typically wins the bulk of the vote in the Volta Region and the northern parts of the country. This, combined with a first-past-the-post electoral system, means that the NDC and NPP are the gate-keepers of Ghana’s political world. Previous research has identi- fied Ghana’s political parties, and their primary processes, as one of the most important bar- riers to entry into politics, particularly for women.27 The NDC shares some important similarities with other political parties in the region, suggesting its experience has something to teach us more broadly. It was founded in 1992 by Jerry John Rawlings, a military leader who initially came to power via a coup before ruling as Ghana’s elected President (1993-2001) after the return to multiparty politics. Like most parties associated with authoritarian pasts, the NDC has a particular interest in demonstrat- ing its democratic credentials. It has also enjoyed success in several elections (1992, 1992, 2008 and 2012), growing beyond its founder to govern the country under Presidents John Atta Mills (January 2009 – July 2012) and John Dramani Mahama (July 2012 – January 2016). 26 Anja Osei, “Formal Party Organisation and Informal Relations in African Parties: Evidence from Ghana,” The Journal of Modern African Studies 54, no. 1 (2016): 37–66, https://doi.org/10.1017/S0022278X15000981. 27 Gretchen Bauer, “Ghana: Stalled Patterns of Women’s Political Representation,” in The Palgrave Handbook of Women’s Political Rights, ed. Susan Franceschet, Mona Lena Krook, and Netina Tan, Gender and Politics (London: Palgrave Macmillan, 2019), 607–25; Gretchen Bauer and Akosua K. Darkwah, “‘Some Money Has to Be Going...’ Discounted Filing Fees to Bring More Women into Par- liament in Ghana,” in Gendered Electoral Financing: Money, Power and Representation in Compar- ative Perspective, ed. Ragnhild L. Muriaas, Vibeke Wang, and Rainbow Murray, Gender and Compar- ative Politics (New York, NY: Routledge, 2019), 133–54. MIASA Working Paper 2021(1) 9 The NDC has formally adopted an ideological position, identifying as a social democratic party and joining Socialist International, the worldwide organization of social democratic, socialist and labour parties. Yet, as in the case of most African parties, ideology forms a rel- atively minor part of its appeals to voters, which tend to be based on broader promises to deliver development and fight corruption, as well as appeals to group identity (which in Ghana manifests primarily on a regional, rather than an ethnic, basis). All of this makes the NDC – and the changes it made to its parliamentary primaries between 2015 and 2019 – a particularly important case when it comes to expanding our understanding of what drives political parties to reform their candidate selection mecha- nisms, the challenges that parties confront in implementing reforms, and the impact that reforms have in practice. From a theoretical point of view, studying Ghana allows us to build on the work of Field and Siavelis on candidate selection in transitional polities.28 The case of the NDC can shed light on how parties choose to reform their candidate selection proce- dures as the constraints initially imposed by democratic transition fade over time. The NDC’s candidate selection mechanisms are also substantively significant, since political parties across Africa often view Ghana’s parties as worthy of emulation. For example, in 2017, Kenya’s opposition alliance, the National Super Alliance (NASA) planned to replicate the NPP’s method of compiling a parallel vote tally, a tactic seen (by the NPP and, presumably, NASA) as having played a crucial role in preventing electoral manipulation in Ghana’s 2016 election.29 These plans were scuppered when the NPP experts, who formed part of a team of election observers from the Democratic Union of Africa, were denied entry on their arrival into Kenya. Nevertheless, this example shows that the NDC’s motivations for reforming its candidate selection mechanisms, and its experience in implementing those reforms, have significance beyond Ghana’s borders. A qualitative case study is well-suited to addressing our central research questions. Though relying on a single case may limit our ability to generalize, it allows us to untangle complex causal relationships that play out over time.30 Here we confront not only the (very high) likelihood that multiple factors contributed to the outcomes of interest, but also com- plexity in terms of timing: we need to be able to examine both underlying (long-term) causes, as well as more temporally proximate triggers of the decisions made by the NDC. As with many complex social processes, the connection between triggers and long-term causes is such that ‘seemingly rival explanations may often be complementary.’31 28 “Candidate Selection Procedures in Transitional Polities.” 29 Nic Cheeseman, Gabrielle Lynch, and Justin Willis, “Digital Dilemmas: The Unintended Conse- quences of Election Technology,” Democratization 25, no. 8 (2018): 1397–1418, https://doi.org/10.1080/13510347.2018.1470165. 30 Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (Cambridge, Massachusetts: MIT Press, 2005). 31 Paul Pierson, Politics in Time (Princeton: Princeton University Press, 2004), 102. MIASA Working Paper 2021(1) 10 We draw on a wide range of empirical evidence. We use Ghanaian (and to a lesser extent, international) media reports to document both the publicly stated reasons for the NDC’s adoption of reforms in 2015 – and their abandonment in 2019 - as well as the party’s public assessment of the impact of those reforms. To probe whether these public accounts align with narratives from within the party, we draw on 28 semi-structured interviews with key informants. The majority of these interviews were conducted in Ghana during the last four months of 2019 with different authors responsible for interviewing different categories of respondents depending on their personal expertise and networks.32 Respondents included senior NDC officials (past and present) who were closely involved in the reforms, current NDC MPs as well as (unsuccessful) NDC parliamentary aspirants or candidates with first-per- son experience of the 2015 primaries, and leading political experts within Ghana. We also rely on some descriptive analysis of the constituency-level election results from 2016 to test the plausibility of the effects attributed to the reforms by those within the NDC. 3 The case of the National Democratic Congress in Ghana The mechanisms used by the NDC to select its parliamentary candidates have not been static over time. On the whole, the trend has been one of gradual democratization, decen- tralization and formalization of candidate selection procedures, with the NDC historically lagging slightly behind the NPP in terms of the inclusiveness of its selection procedures.33 Since at least the 2004 election both the NDC and NPP have had fairly standard rules for candidate selection set out in their party constitutions.34 By 2015, the NDC and the NPP had very similar mechanisms for selecting parliamentary candidates. In each constituency, elec- toral colleges of around 500 delegates, comprised of four (in the case of the NDC) or five (in the case of the NPP) party executives from each polling station – of which there might be about 90 in a constituency, as well as a small number of additional delegates (typically around 15) selected the parliamentary candidates for the relevant party.35 For the 2015 pri- maries, however, the NDC opted to expand its primary selectorate significantly, opening up primary voting to all party members: thousands, rather than hundreds, of people were now able to participate in each constituency. These changes made the NDC one of the most progressive parties in Africa as far as the inclusiveness of the selectorate for parliamentary primaries is concerned. Prior to the reforms on which we focus in this article, the NDC’s candidate selection procedures for 32 A small number of interviews, conducted by G. Bauer between May 2016 and June 2018 and part of a different project, provide supplementary evidence. 33 Cyril K Daddieh and George M Bob-Milliar, “In Search of ‘Honorable’ Membership: Parliamentary Primaries and Candidate Selection in Ghana,” Journal of Asian and African Studies 47, no. 2 (2012): 204–20, https://doi.org/10.1177/0021909611421905. 34 Daddieh and Bob-Milliar; Osei, “Political Parties in Ghana.” 35 Bauer and Darkwah, “Some Money Has to Be Going...”; Ichino and Nathan, “Primary Elections in New Democracies.” MIASA Working Paper 2021(1) 11 parliamentary elections reflected the median, or most-common, approach adopted by po- litical parties in sub-Saharan Africa. Examining 64 parties from 25 countries in the region, Merete Seeberg and her co-authors found that roughly half of these parties (55%), including the NDC, involved party delegates in the selection of parliamentary candidates.36 In contrast, 29% of the parties employed more exclusive approaches (party leaders only) while just 16% took the most inclusive approach (allowing ordinary party members and/or all voters to be involved in the choice of candidates). Consequently, the changes generated very positive media coverage for the NDC, with one reporter stating that the reforms were ‘a democratic first and a marked improvement in the system employed by its rival the New Patriotic Party.’37 Significantly, the NDC’s decision to expand the selectorate became very closely tied to the introduction of a new biometric party register. Plans for this were first mooted in the wake of the 2012 election, and in 2013 the NDC solicited bids for the construction of the new register and production of biometric membership cards for members. Balking at the cost of this – the bids received came in at more than 7 million euro – the NDC ultimately attempted to roll-out a “home grown” solution. Relying on data included in the national voter register, the NDC issued party members with new membership cards featuring a photo, but it did not itself collect biometric data, nor use biometric data (other than photos) to verify the identity of those included in the new register or check for multiple registrations. The two initiatives – the construction of a new party register and the expansion of the se- lectorate – became almost inseparable both in practice and in terms of how the reforms were discussed by NDC insiders. In media reports and in our interviews, party leaders, MPs and aspirants rarely talked about one without immediately mentioning the other, fre- quently presenting the two initiatives as part of a single package of reform. As they put it, the decision to allow all party members to vote in the 2015 primary elections was premised on the idea that members would be identified by the new register, while the expense of the register was justified (in part) on the basis that it would facilitate the expansion of the selec- torate. As we discuss in more detail later, this conflation of the two initiatives had important ramifications for how the NDC evaluated the impact of its reforms in the wake of the 2016 election. Motivations for the NDC reforms in 2015 Our interviews with NDC party leaders, aspirants, candidates and sitting MPs (who won the 2016 elections), together with an analysis of media reports documenting public statements made by those actors, reveal four sets of reasons that motivated the opening up of the NDC 36 Seeberg, Wahman, and Skaaning, “Candidate Nomination, Intra-Party Democracy, and Election Violence in Africa.” 37 GhanaWeb, “NDC Presidential, Parliamentary Primaries Slated for Nov. 7,” GhanaWeb, August 14, 2015, https://www.ghanaweb.com/GhanaHomePage/NewsArchive/NDC-presidential-parliamen- tary-primaries-slated-for-Nov-7-375040. MIASA Working Paper 2021(1) 12 primaries in 2015. Untangling the relative importance of these is empirically challenging. Different types of actors within the party typically emphasized different reasons for the re- forms, while certain reasons were accorded greater prominence in public narratives about the reforms and less in more private discussions. The reason that was typically first cited by those within the NDC – both in media reports and interviews – was a desire to democratize the candidate selection process, in the sense of bringing more party members into the pro- cess as primary voters, bolstering grass-roots participation in this crucial party activity. Yet a second reason tended to follow this, and was typically given much greater emphasis by those we interviewed – especially aspirants and MPs: a desire to reduce the role of money in the primary elections, often referred to as the monetization of politics. A third reason, building the party’s organizational capacity, was cited much less frequently in interviews. However, this factor was particularly important to senior leaders in the party who drove the reforms forward, suggesting that it played a central role. Finally, a fourth reason, only occa- sionally cited (and almost exclusively by senior party leaders) was a concern that the NDC needed to keep up with (and preferably outpace) reforms adopted by the rival NPP. As Cyril Daddieh and George Bob-Milliar have observed, the NDC’s ‘revolutionary ideology and populist rhetoric predisposed it to a preference for bottom–up processes and for consensus building in its operations.’38 This preference appears to have played a role in the decision to expand the primary selectorate in 2015. Almost all of those we interviewed about the 2015 reforms referred to a desire to democratize the candidate selection process, and for many this was the first reason they identified. While our respondents only rarely described the reforms explicitly as an attempt to strengthen internal democracy, most im- plied as much, frequently referencing a desire to increase the participation of party mem- bers in decision-making. For example, one relatively senior MP (first elected in 2012), stated that the primary motivation for the reforms was a desire ‘to get everybody involved in the decision’39 while another explained that ‘it just made sense that they wanted more people involved.’40 In 2015, party leaders had made similar statements to the media, sometimes linking this desire to increase participation to a need to select candidates who had genuine grass-roots support. Thus, party leaders expected that expanding participation in the pri- maries would generate electoral benefits down the track, making it harder for “big men” without genuine local support to gain nominations. For example, in 2015, the NDC’s Deputy General Secretary, Koku Anyidoho explained that delegates sometimes selected candidates who did not resonate with voters, driving down the NDC’s vote in parliamentary races com- pared to the presidential poll.41 Other party leaders repeated these sentiments in interviews, with some explaining that broadening participation and making the primaries more inclu- sive would force candidates to become more familiar with the needs of their constituents. 38 “In Search of ‘Honorable’ Membership,” 211. 39 NDC MP, interviewed by S. Dodsworth, 12 November 2019, Accra. 40 NDC MP, interviewed by S. Dodsworth, 14 November 2019, Accra. 41 GhanaWeb, “NDC Presidential, Parliamentary Primaries Slated for Nov. 7.” MIASA Working Paper 2021(1) 13 This would, in their view, not only empower the grassroots members of the party but ensure NDC candidates were better positioned to win during the national election campaign. While a desire to strengthen democracy by expanding participation was often the first response to questions about what drove the NDC to reform its primary process in 2015, it was typically not what respondents emphasized most strongly, nor even the reason that was most prominent in contemporaneous media reports. This suggest some degree of so- cial desirability bias: respondents may have explained the reforms by reference to democ- racy and participation due to a perception that this was the “correct” response. In interviews – and media reports – a different motivation tended to be given greater emphasis, though it was one that also had serious implications for the quality of democracy: vote-buying. Spe- cifically, a desire to eliminate the “cocoa season” described by Staffan Lindberg,42 the costs of which many felt had escalated to a point that was unacceptable and unsustainable. In media interviews at the time the reforms were announced, party leaders asserted that by increasing the number of people who voted in the primary elections, the new system would make vote-buying impossible, unless one had – in the words of Anyidoho – ‘all the money in the Bank of Ghana.’43 Similarly, in interviews, aspirants stressed the need to move away from a primary system in which they were expected to provide “envelopes” (money) and material goods to delegates. One sitting MP (first elected in 2012 and again in 2016) stated, ‘the system of delegates, getting delegates is prone to – how do you call it – the monetiza- tion of the process. The highest bidder wins.’44 Aspirants explained the logic behind the re- forms: with the new procedures expanding the selectorate, candidates could not possibly pay all voters in the way delegates had been paid in the past. One aspirant who (unsuccess- fully) challenged an incumbent MP in the Upper West region said that in the past one could “buy” all of the executives (delegates) for GHC 500 each but that would not be possible when the voter register was expanded 10-fold. Many others made similar statements about the (anticipated) “impossibility” of buying victory when all party members were able to vote. All of the senior NDC party leaders we interviewed indicated that reducing the role of money in the primaries as well as their overall cost was one of the main reasons for the 2015 reforms. Some emphasized that the practice of vote-buying not only made the pri- mary elections expensive, but led to corruption and dishonesty. As a very senior NDC party official put it, the reforms were ‘the most radical way’ to ‘eliminate corruption in the [pri- mary] elections.’45 At the same time, senior NDC party leaders also spoke frequently of the third motivation, namely, building the party’s organizational capacity – something not raised by those who had actually competed in the primaries. According to senior NDC party officials, the reforms to the 2015 party primaries were also intended to build the 42 Staffan I. Lindberg, “‘It’s Our Time to “Chop”’: Do Elections in Africa Feed Neo-Patrimonialism Ra- ther than Counter-Act It?,” Democratization 10, no. 2 (2003): 121. 43 GhanaWeb, “NDC Presidential, Parliamentary Primaries Slated for Nov. 7.” 44 NDC MP, interviewed by S. Dodsworth, 3 October 2019, Accra. 45 Interviewed by S. Alidu and G.A. Bukari, 2 October 2019, Accra. MIASA Working Paper 2021(1) 14 infrastructure of the party, creating an accurate and reliable database of party members while raising revenue, as members would need to pay their membership dues to obtain their new party card. While these kinds of motivations received less prominence in media coverage, they were emphasized by party leaders interviewed by the press. As early as 2013, the NDC’s General Secretary called on party members to support the new party register and planned reforms to the primary process because they would help transform the party, en- suring it was active on continuous basis – not only during elections.46 Later, in August 2015, he explained how the new party register made the impending expansion of the selectorate possible and would help to reduce disputes relating to candidate selection47 since so many conflicts ‘hover around who a real member is.’48 He went on to highlight the importance of the new party register – which he presented as almost inseparable from the expansion of the selectorate – in terms of organizational capacity: You cannot run an organization without knowing the number of members you have in that organization, so it is a basic requirement of any management situation that, if you are in charge of any organization, you have to know how big or how small the organization is to be able to monitor the growth or decline in membership of that organization.49 According to senior party leaders the twin reforms – expanded selectorate and new party register – would allow the party to make more accurate predictions about election out- comes (since changes in party membership would signal increases or decreases in support), identify where it should invest campaign resources, and increase its ability to generate rev- enue (since party members would be more motivated to pay their membership dues). Finally, senior party leaders sometimes mentioned a fourth motivation for the 2015 reforms: keeping up with – and preferably overtaking – the NPP. While the NPP continues to use a delegate-based system for its parliamentary primaries, it had expanded the selec- torate for the selection of its presidential candidate prior to the 2012 election. In interviews, one expert informant as well as several party leaders explained that this made the NDC look as if it was lagging behind the NPP in terms of democracy. In media reports, senior party figures admitted that the NDC was emulating the NPP, with one Ghanaian newspaper de- scribing the reforms in its headline as an order from former President Jerry John Rawlings ‘to copy NPP.’50 In an interview, one very senior NDC figure noted how the NDC’s origins – 46 Jasmine Arku, “NDC Reforms Distinct from NPP — Asiedu Nketia,” Graphic Online, October 9, 2013, https://www.graphic.com.gh/news/politics/ndc-reforms-distinct-from-npp-asiedu-nketia.html. 47 Under the delegate-based system, disputes often arose at the vetting stage, as formal rules state that aspirants are only eligible to contest if they have been a party member for a certain length of time.. 48 Quoted in Gifty Arthur, “NDC Holds Presidential & Parliamentary Primaries On November 7,” The Herald, August 14, 2015. 49 quoted in Arthur. 50 Daniel Nonor, “JJ Orders NDC to Copy NPP,” The Chronicle, October 7, 2013, https://al- lafrica.com/stories/201310072122.html. MIASA Working Paper 2021(1) 15 created out of Rawlings’ PNDC – made the NDC appear less democratic than the NPP in the minds of some – and this was something the party had long sought to change: When we got to 2002, NPP had emerged strongly and people were comparing how they were doing their things versus how we were doing ours… So, people began increasingly looking at our party as less democratic than NPP that had also begun selecting their candidates through open elections and so on. So, you cannot isolate yourself from an environment like this.51 Such statements suggest that both competition and a degree of contagion were underlying drivers of the reform, though factors that had a less immediate or proximate effect on the decision made in 2015. The impact of the 2015 reforms: Intended vs unintended effects The reforms introduced by the NDC in 2015 had a variety of effects. Some of these aligned with the party’s intentions. Most notably, expanding the selectorate to include all party members did increase participation in the candidate selection process. Where in the past, candidates had been selected by several hundred people, in 2015 the NDC’s candidates were (for the most part) selected by thousands.52 This roughly 10-fold expansion of partici- pation represented a significant achievement: according to Ghana’s Electoral Commission, more than 1.2 million people voted in the NDC’s primary elections.53 In comparison, around 140,000 people voted in the NPP primaries. Unfortunately for the NDC, this expansion of participation did not generate many of the benefits that party leaders had anticipated – and had a number of unanticipated negative effects. There is some disagreement about whether the reforms achieved the goal of reduc- ing the monetization of the primary process. In interviews, a number of aspirants and can- didates in the primaries, as well as some sitting MPs, asserted that the reform had increased rather than decreased the financial (and time) cost of the primaries, something that they attributed directly to the expansion of participation. They complained that they had been expected to compensate not just a group of delegates but all voters – up to 10,000 in a single constituency. A common refrain was that it was as if candidates were financing a gen- eral election two times over. One incumbent NDC MP lamented: You now [in 2015] have to bribe the 50,000. It's very expensive and time consuming ... Now the entire community was involved. In fact, you spend so much money, and then your energy. It was so difficult.54 51 Interviewed by S. Alidu and G.A. Bukari, 2 October 2019, Accra. 52 In a small number of constituencies, parliamentary candidates were unopposed in the primary election. 53 Ebenezer Afanyi Dadzie, “95% of NDC Members Affirm Mahama’s Candidature for 2016 Polls,” Cit- iFM Online, November 22, 2015, http://citifmonline.com/2015/11/95-of-ndc-members-affirm-ma- hamas-candidature-for-2016-polls/. 54 Interviewed by S. Dodsworth. 3 October 2019, Accra. MIASA Working Paper 2021(1) 16 Yet there were also some aspirants who suggested that, although the overall cost of run- ning a primary campaign was increased by the need to engage with a larger number of people (pushing up costs for things like transportation and refreshments), direct vote-buy- ing (in the sense of monetary payments explicitly being made in exchange for votes) did become more difficult. One (unsuccessful) aspirant who contested in the Volta region ex- plained: Whereas [with] 9 branch members you can always bribe them easily – it happened in 2012 where the night before our primaries people were distributing 500 Ghana cedis when some of us had only 50 Ghana cedis to give. So, in 2015 the good thing about it is that you couldn’t bribe people on election day or a day before the election day, it was difficult.55 There were also some candidates who stated that the goal of reducing the monetization of the process was accomplished for certain people; the candidates who had good track-rec- ords in their communities and genuine grassroots support (among which they typically identified themselves). Such candidates could – according to these respondents – count on the votes of ordinary party members and so were liberated from the financial expectations of delegates.56 The most serious unintended consequences to flow from the reformed primary pro- cess in 2015 related to the new party register, with which the expansion of the selectorate had become so closely tied. Figures at the top of the NDC party structures identified ‘the development of the database’ as ‘the biggest source of challenges.’57 The construction of the new party register would have constituted a logistical challenge under any circum- stances, but for the NDC this was exacerbated, in some constituencies, by the fact that those who were opposed to the expansion of the selectorate were responsible for building the new party register. Moreover, the fact that the new register would determine who could – and could not – vote in the primary elections meant that there was a clear incentive to ma- nipulate it. According to one very senior NDC official, this resulted in ‘many people finding ways of compromising the database,’ aided by the fact that the party ‘did not have the ca- pacity to introduce the necessary checks and security features’58 It is therefore unsurprising that problems with the register were cited by nearly all of our informants whether they were party leaders, aspirants, or candidates. Some stated that certain NDC members had been prevented from registering or had their names removed from the new register (causing them to be turned away on the day of the primaries), while others suggested that NPP 55 NDC parliamentary aspirant from 2015, interviewed by S. Alidu, October 2019, Accra. 56 There may be a degree of social desirability at play in some responses, though we found that as- pirants and MPs were typically quite open about the fact that they spent a lot of money on their campaigns. The role of money in Ghana’s primary elections is very widely acknowledged and well documented: see Westminster Foundation for Democracy, “Cost of Politics in Ghana” (London: Westminster Foundation for Democracy, 2017). 57 Interview with senior NDC party official, by S. Alidu and G.A. Bukari, 2 October 2019, Accra. 58 Interviewed by S. Alidu and G.A. Bukari, 2 October 2019, Accra. MIASA Working Paper 2021(1) 17 members might have “infiltrated” the register in order to help elect weaker NDC candidates (benefiting the NPP in the national election). Media reports document similar allegations, most notably in Atebubu-Amantin Constituency in the Brong-Ahafo Region, where NDC party members set fire to a copy of the new register in the wake of allegations that it had been manipulated by aspirants.59 Although the true extent of any manipulation of the new party register is difficult (perhaps impossible) to verify, from the vantage point of party leaders, the expanded regis- ter engendered significant acrimony and division within the NDC. While many ordinary party members were able to vote in the 2015 primaries, it is clear that (at least) a small mi- nority were disenfranchised by the new register – their names were omitted either due to poor implementation, or deliberate manipulation. In the wake of the NDC’s election loss in 2016, the reformed primary process and new party register quickly became targets of blame. Senior NDC figures made allegations– both in media reports, and in our interviews – that the reforms had caused disunity and disillusionment, supressing turnout among the NDC’s supporters, splitting its parliamentary vote in constituencies where dissatisfied aspir- ants had run as independents, and leading many NDC supporters to vote “skirt and blouse” (voting for the NDC in the presidential race, but for the NPP’s parliamentary candidate). These allegations were purportedly borne out by the findings of the “Kwesi Botchwey Re- port,” the 455-page document produced by a committee (led by former Finance Minister Professor Kwesi Botchwey) that the NDC had tasked with investigating the causes of its de- feat in 2016.60 However, close analysis of constituency level election data strongly suggests the negative impact of the reforms on the NDC’s electoral fortunes have been overestimated by many within the NDC. Although gaps between the NDC’s presidential and parliamentary votes show that a significant number of people did vote “skirt and blouse” in some constit- uencies, there are very few constituencies in which this ultimately led the NDC to lose the parliamentary seat, and even fewer where this can plausibly be attributed to the expansion of the primary selectorate and new party register. Specifically, there were 17 constituencies where the NDC’s presidential vote outstripped its parliamentary vote by a number exceed- ing the NPP’s (parliamentary) victory margin.61 For many of these, there is little evidence of controversy relating to the primaries in contemporaneous media reports. In others, there are reports of disputes within the NDC triggered by the primary process – some of which led to court cases or failed aspirants standing as independents. However, many of these 59 Daniel Yao Dayee, “NDC Burns Bloated Register At Atebubu,” Daily Guide Ghana, October 24, 2015, https://www.modernghana.com/news/651305/ndc-burns-bloated-register-at-atebubu.html. 60 The report has not been publicly released, though some media outlets claimed to have seen leaked versions of it. 61 Authors’ calculations based on number of ballots cast in 2012 and 2016, using data from Election Passport (available at http://www.electionpassport.com). These constituencies were Atebubu- Amantin, Dormaa West, Tain, Ekumfi, Dade Kotopon, Bunkpurugu, Kpandai, Salaga South, Savelugu, Tempane, Zebilla, Lawra, Nandom, Sissala East, Krachi East, Jomoro, and Sefwi Akontombra. MIASA Working Paper 2021(1) 18 disputes centred on the vetting process rather than the new register, having been triggered by disagreements about the eligibility of aspirants to compete. In only four constituencies – Atebubu-Amatin, Dade Kotopon, Jomoro and Lawra – do media reports provide fairly clear evidence of disputes centred on attempts to manipulate the new party register. More- over, in one of these constituencies – Jomoro – the real drain on the NDC’s parliamentary vote was Samia Yaaba Nkrumah, the candidate for the Convention People's Party, who had served as the MP for Jomoro from 2008 to 2012 and is daughter of Ghana’s first President, Kwame Nkrumah. Similarly, voter turnout was down by almost 10% (compared to 2012) in the Volta Region62 – the NDC’s most important stronghold, where lower turnout clearly benefited the NPP in the presidential race. This drop in voter turnout had less impact on the parliamentary results. While it almost certainly contributed to the NDC’s shock loss in Krachi East, where the NPP’s candidate won with a margin of roughly 50 votes, the NDC won all the other par- liamentary seats in the Volta region. Party leaders, aspirants and MPs frequently attributed the drop in voter turnout to disillusionment and resentment generated by the manipulation of the party register during the primaries, but at best, dissatisfaction with the changes made to the primary process in 2015 explain this only partially. The largest drops in turnout tended to coincide with disputes centred on the NDC’s vetting process, as was the case in Ketu South. Moreover, the drop in turnout was not uniform: in the Brong Ahafo Region (which is generally considered a swing region) there had been reports of NDC members being disenfranchised in the primaries, but voter turnout was essentially unchanged com- pared to 2012. Previous research has also highlighted a variety of factors – quite independent of the primary reforms – that led to the NDC to lose the 2016 election. These include general disil- lusionment with eight years of NDC rule, combined with the expectation – now quite firmly entrenched in the minds of many Ghanaian voters – that political power changes hands every eight years.63 Poor economic performance and a strong campaign by the NPP also cost the NDC.64 In the eyes of many voters, the NDC’s mismanagement of the country was amply demonstrated by frequent and increasingly long-lasting rolling blackouts in the pe- riod from 2012 to 2016. These blackouts became known as “dumsor” (roughly translated as “off and on” in Twi) and inflicted significant economic damage: one report calculated that 62 Authors’ calculations using data from Election Passport (available at http://www.electionpass- port.com). 63 Ransford Edward Van Gyampo, Emmanuel Graham, and Eric Yobo, “Ghana’s 2016 General Elec- tion : Accounting of the Monumental Defeat of the National Democratic Congress (NDC),” Journal of African Elections 16, no. 1 (2017): 24–45, https://doi.org/10.20940/JAE/2017/v16i1a2. 64 Joseph R. A. Ayee, “Ghana’s Elections of 7 December 2016: A Post-Mortem,” South African Journal of International Affairs 24, no. 3 (2017): 311–30, https://doi.org/10.1080/10220461.2017.1378124; Nic Cheeseman, Gabrielle Lynch, and Justin Willis, “Ghana: The Ebbing Power of Incumbency,” Jour- nal of Democracy 28, no. 2 (2017): 92–104. MIASA Working Paper 2021(1) 19 dumsor cost Ghana’s economy almost $1 billion in 2014 alone.65 Together, these factors meant that the NDC faced an uphill battle in the 2016 election. Although the negative electoral impacts of the 2015 reforms were overestimated by many within the NDC, it is clear that they did generate considerable suspicion and disunity within the party. This was not due to the formal design of the reforms, but rather the manner in which they were implemented. Regardless, the (mis)perception that the changes to the primary process had played a role in the NDC’s defeat, ultimately led to a reversal of the reforms for the 2019 primaries. The decision to abandon the 2015 reforms In 2017, the NDC’s National Executive Council (NEC) voted unanimously to scrap the bio- metric party register because it lacked credibility and to “manually” construct another new party register.66 Later, in 2019, the NEC announced that the NDC would put the selection of parliamentary candidates for the 2020 election back in the hands of delegates – albeit a slightly higher number of them than in the past: the electoral colleges for each constituency would comprise nine delegates from each polling station (rather than four, as in 2012) and the number of additional constituency-level delegates would increase. The party also intro- duced several new rules designed to avoid internal disputes relating to the primaries and minimise the impact of those that did occur. For example, aspirants in the 2019 primaries in Volta region were required to sign declarations stating that they would respect the results, abstain from taking court action and refrain from running as an independent candidate should their primary campaign be unsuccessful.67 To a person, all of the NDC party leaders interviewed for our study seemed to express some regret at the reversion to the delegate system and to indicate that the expanded se- lectorate is ‘the future of the party.’68 Indeed, in their views, the original reasons for seeking the reform still prevail. Moreover, they all seemed in agreement that the problems were not with the overall goal of democratizing the party primary process – by both expanding par- ticipation and reducing monetization – but with the implementation of the reform. At heart, the problem was that the party leadership had not fully appreciated the enormity of both the technical and political task that it was taking on. Several of the NDC MPs we interviewed expressed regret that the party had decided to return to the previous primary process, indi- cating a personal preference for the more inclusive process used in 2015. That more 65 Charles Ackah, “Electricity Insecurity and Its Impact on Micro and Small Businesses in Ghana,” ISSER Report (University of Ghana, Legon: Institute of Statistical, Social and Economic Research, 2015). 66 GhanaWeb, “NDC Biometric Register Not Credible – Portuphy,” GhanaWeb, October 19, 2017, https://www.ghanaweb.com/GhanaHomePage/NewsArchive/NDC-biometric-register-not-credi- ble-Portuphy-592118. 67 Ghanaian Times, “Volta NDC to Vet 63 Aspirants for Parliamentary Primary …Rawlings Aide Joins Race,” Ghanaian Times, July 23, 2019, http://www.ghanaiantimes.com.gh/volta-ndc-to-vet-63-aspir- ants-for-parliamentary-primary-rawlings-aide-joins-race/. 68 Interview with senior NDC party official, by S. Alidu and G.A. Bukari, 2 October 2019, Accra. MIASA Working Paper 2021(1) 20 inclusive process was, they felt, safer and more reliable, since the risk of being “outbid” for the votes of the delegates was reduced – providing you had demonstrated your value to the community. Notably, while some women MPs indicated that the reversion to the dele- gate system had little impact on them personally, a small number felt it had made their campaigns for re-selection significantly harder because it increased the risk they would be outspent by rival candidates. Such statements, together with the fact that the number of female candidates se- lected by the NDC in 2019 dropped to 27, compared to 40 female candidates fielded in 2016, suggests that the reversion of the primary reforms may have had a gendered impact. This possibility is something that we have been unable to fully explore in this paper and which warrants further attention in the future. Our evidence suggests that the more inclusive pri- mary process employed in 2015 did not necessarily reduce the cost of primary elections on the whole, but there are signs that it did sometimes render voter-buying less effective in some circumstances because it allowed aspirants – particularly female aspirants –with good track-records in their communities to make greater use of this non-financial political capital. This is an effect that has previously been suggested, but not directly tested by prior re- search.69 4 Implications and conclusion Ultimately, our evidence shows a substantial divergence between what the NDC hoped to achieve through its reforms to parliamentary candidate selection in 2015, and the actual effects of those reforms. The party did succeed in expanding participation in the primary process. However, this did not deliver the electoral benefits that NDC leaders had antici- pated, in part because the construction of the new party register – a necessary step in the implementation of the reforms – was poorly managed, triggering internal disputes and dis- trust. The reforms also had no immediate success in reducing the monetization of candidate selection. In fact, many aspirants found that the costs of contesting the primaries increased. Yet, in certain cases, the expansion of the selectorate did reduce the efficacy of vote-buying, allowing some aspirants to better leverage other sources of political capital. Had the re- forms been retained, this might have shifted the behaviour of aspirants, reducing the role of money over the long term. The NDC was also unable to realise anticipated gains in terms of the party’s infrastructure and organizational capacity. The biggest gain anticipated here was a comprehensive and reliable register of party members, but the new register con- structed in 2015 was flawed and ultimately abandoned. It is also clear that the process of implementing the reforms did trigger disputes, disunity and distrust within the NDC, though our evidence suggests that the electoral cost of this has tended to be overstated by 69 Nahomi Ichino and Noah L Nathan, “Democratizing the Party: The Effects of Primary Election Re- forms in Ghana,” Working Paper, 2016, http://sticerd.lse.ac.uk/seminarpapers/pspe21032017.pdf. MIASA Working Paper 2021(1) 21 NDC insiders. It is highly likely that divisions and disputes did contribute to the NDC’s poor electoral performance in 2016 in certain constituencies, but only a small portion of those disputes can be credibly laid at the door of the reforms to the primary process. As noted earlier, Field and Siavelis70 identify several constraints that deter political parties in transitional democracies from adopting inclusive candidate selection mecha- nisms. In such contexts, they argue, high political uncertainty, weak and limited party or- ganizations, strategic complexity, and pressure to maintain elite alliances encourage party leaders to adopt more exclusive procedures. This article helps to fill in the next part of that story. As the example of the NDC shows, the constraints that Field and Siavelis identify fade but do not disappear entirely over time. If a country – like Ghana – successfully moves past the period of transition and begins the slower, more gradual (but still contingent) process of democratic consolidation, political parties are likely to see the democratization of their internal procedures as something they ought to do. This may be felt especially strongly in the case of the successors to former single or military parties – such as the NDC – that are likely to be particularly keen to distance themselves from their past. Moreover, as con- straints like political uncertainty, strategic complexity and pressure to maintain elite alli- ances (so as to sustain a democratic transition) reduce, the adoption of inclusive candidate selection procedures may become more possible. Yet these shifts in the structural determi- nants of candidate selection mechanism do not – on their own – cause specific reforms. More proximate causes, such as those that motivated the decision of the NDC, must trigger a concrete decision to reform. When parties do decide to implement reforms, however, they are rarely entirely free of the constraints of the past. Thus, weak party organization may still act as a significant barrier to the introduction of inclusive selection mechanisms, even as it forms part of the motivation for adopting them. This helps to explain why we are seeing an increasing number of political parties in sub-Saharan Africa’s more democratic countries experimenting with more inclusive candidate selection mechanisms, as the NDC has done in Ghana and the APC in Nigeria. It also explains why this shift towards inclusive mechanisms is slow, uneven, and sometimes subject to reversal. Our findings clearly show the importance of studying the process of reform as a dy- namic, contentious and highly contingent process. Explaining why political parties employ certain mechanisms to select candidates requires us to understand more than just the ef- fects that a given procedure has once it is put in place: we must also understand the effects of the changes that must be made to move from one system to another. This is not just a question of institutional design, but of practical implementation. Examining the practical challenges of reforming primaries is valuable not only to academics, but also to political parties seeking to make their own candidate selection procedures more inclusive. Our case study suggests, for example, that the sequencing of reforms is very important. Any party considering a shift from a primary process based on delegates, to one in which all party 70 “Candidate Selection Procedures in Transitional Polities.” MIASA Working Paper 2021(1) 22 members can participate, would be well advised to invest in building a credible register of its members first. Notably, this does seem to be the lesson that the NPP has drawn from the NDC’s experience. In an interview, one senior NPP leader stated that their party agreed with the general principle of expanding participation in the primary process, but observed that in light of what happened to the NDC, changes would need to be made slowly: the involve- ment of all party members in the selection of parliamentary candidates was thus a long term goal. Detailing some of the difficulties that the NDC encountered in implementing its more inclusive system – and noting the NDC’s own assessment that this had contributed to their electoral loss – he nevertheless stated: I think maybe in the near future we all have to think through that process again where every card bearing member of the party is given the opportunity to vote… ultimately that is where we may all have to go when we are having primaries.71 This statement reveals that the negative experience of the NDC has not doomed the pro- spects for a more inclusive candidate selection process to one day be introduced by the NPP – and perhaps by other parties across Africa that look to Ghana for examples. Yet, the mech- anisms used to select parliamentary candidates have a critical influence on who has access to political power. Attempts to change these mechanisms will inevitably create winners and losers, no matter how well-informed they are by past experience. Reforms to candidate se- lection mechanisms will therefore remain a serious practical and political challenge for po- litical parties everywhere. 71 Interview with senior NPP official, by S. Alidu, 15 November 2019, Accra. 23 MIASA Working Paper 2021(1) Bibliography Ackah, Charles. “Electricity Insecurity and Its Impact on Micro and Small Businesses in Ghana.” ISSER Report. University of Ghana, Legon: Institute of Statistical, Social and Economic Research, 2015. Arku, Jasmine. “NDC Reforms Distinct from NPP — Asiedu Nketia.” Graphic Online, October 9, 2013. https://www.graphic.com.gh/news/politics/ndc-reforms-distinct-from- npp-asiedu-nketia.html. Arthur, Gifty. “NDC Holds Presidential & Parliamentary Primaries On November 7.” The Herald. August 14, 2015. Ayee, Joseph R. A. “Ghana’s Elections of 7 December 2016: A Post-Mortem.” South African Journal of International Affairs 24, no. 3 (2017): 311–30. https://doi.org/10.1080/10220461.2017.1378124. Barnea, Shlomit, and Gideon Rahat. “Reforming Candidate Selection Methods: A Three- Level Approach.” Party Politics 13, no. 3 (2007): 375–94. https://doi.org/10.1177/1354068807075942. Basedau, Matthias, and Alexander Stroh. “Measuring Party Institutionalization in Developing Countries: A New Research Instrument Applied to 28 African Political Parties.” Working Paper. German Institute of Global and Area Studies, February 1, 2008. Bauer, Gretchen. “Ghana: Stalled Patterns of Women’s Political Representation.” In The Palgrave Handbook of Women’s Political Rights, edited by Susan Franceschet, Mona Lena Krook, and Netina Tan, 607–25. Gender and Politics. London: Palgrave Macmillan, 2019. Bauer, Gretchen, and Akosua K. Darkwah. “‘Some Money Has to Be Going...’ Discounted Filing Fees to Bring More Women into Parliament in Ghana.” In Gendered Electoral Financing: Money, Power and Representation in Comparative Perspective, edited by Ragnhild L. Muriaas, Vibeke Wang, and Rainbow Murray, 133–54. Gender and Comparative Politics. New York, NY: Routledge, 2019. Carothers, Thomas. “The End of the Transition Paradigm.” Journal of Democracy 13, no. 1 (2002): 5–21. Cheeseman, Nic, and Miles Larmer. “Ethnopopulism in Africa: Opposition Mobilization in Diverse and Unequal Societies.” Democratization 22, no. 1 (2015): 22–50. https://doi.org/10.1080/13510347.2013.809065. Cheeseman, Nic, Gabrielle Lynch, and Justin Willis. “Digital Dilemmas: The Unintended Consequences of Election Technology.” Democratization 25, no. 8 (2018): 1397– 1418. https://doi.org/10.1080/13510347.2018.1470165. ———. “Ghana: The Ebbing Power of Incumbency.” Journal of Democracy 28, no. 2 (2017): 92–104. Daddieh, Cyril K, and George M Bob-Milliar. “In Search of ‘Honorable’ Membership: Parliamentary Primaries and Candidate Selection in Ghana.” Journal of Asian and African Studies 47, no. 2 (2012): 204–20. https://doi.org/10.1177/0021909611421905. Dadzie, Ebenezer Afanyi. “95% of NDC Members Affirm Mahama’s Candidature for 2016 Polls.” CitiFM Online, November 22, 2015. http://citifmonline.com/2015/11/95-of- ndc-members-affirm-mahamas-candidature-for-2016-polls/. MIASA Working Paper 2021(1) 24 Dayee, Daniel Yao. “NDC Burns Bloated Register At Atebubu.” Daily Guide Ghana. October 24, 2015. https://www.modernghana.com/news/651305/ndc-burns-bloated- register-at-atebubu.html. Elischer, Sebastian. Political Parties in Africa: Ethnicity and Party Formation. New York: Cambridge University Press, 2013. Field, Bonnie N., and Peter M. Siavelis. “Candidate Selection Procedures in Transitional Polities: A Research Note.” Party Politics 14, no. 5 (2008): 620–39. https://doi.org/10.1177/1354068808093393. George, Alexander L., and Andrew Bennett. Case Studies and Theory Development in the Social Sciences. Cambridge, Massachusetts: MIT Press, 2005. Ghanaian Times. “Volta NDC to Vet 63 Aspirants for Parliamentary Primary …Rawlings Aide Joins Race.” Ghanaian Times, July 23, 2019. http://www.ghanaiantimes.com.gh/volta-ndc-to-vet-63-aspirants-for- parliamentary-primary-rawlings-aide-joins-race/. GhanaWeb. “NDC Biometric Register Not Credible – Portuphy.” GhanaWeb, October 19, 2017. https://www.ghanaweb.com/GhanaHomePage/NewsArchive/NDC- biometric-register-not-credible-Portuphy-592118. ———. “NDC Presidential, Parliamentary Primaries Slated for Nov. 7.” GhanaWeb, August 14, 2015. https://www.ghanaweb.com/GhanaHomePage/NewsArchive/NDC- presidential-parliamentary-primaries-slated-for-Nov-7-375040. Giollabhuí, Shane Mac. “How Things Fall Apart: Candidate Selection and the Cohesion of Dominant Parties in South Africa and Namibia.” Party Politics 19, no. 4 (2013): 577– 600. https://doi.org/10.1177/1354068811407599. Gyampo, Ransford Edward Van, Emmanuel Graham, and Eric Yobo. “Ghana’s 2016 General Election : Accounting of the Monumental Defeat of the National Democratic Congress (NDC).” Journal of African Elections 16, no. 1 (2017): 24–45. https://doi.org/10.20940/JAE/2017/v16i1a2. Ichino, Nahomi, and Noah L Nathan. “Democratizing the Party: The Effects of Primary Election Reforms in Ghana.” Working Paper, 2016. http://sticerd.lse.ac.uk/seminarpapers/pspe21032017.pdf. ———. “Primary Elections in New Democracies: The Evolution of Candidate Selection Methods in Ghana.” In Routledge Handbook of Primary Elections, edited by Robert G. Boatright, 1st ed., 369–83. New York: Routledge, 2018. International Republican Institute, and National Democratic Institute. “IRI/NDI Nigeria International Election Observation Mission Final Report.” Washington D.C.: IRI and NDI, 2019. https://www.iri.org/sites/default/files/nigeria_election_report_updated.pdf. Kebonang, Zein, and Wankie Rodrick Wankie. “Enhancing Intra-Party Democracy: The Case of the Botswana Democratic Party.” Journal of African Elections 5, no. 2 (2006): 141– 50. Koter, Dominika. “Costly Electoral Campaigns and the Changing Composition and Quality of Parliament: Evidence from Benin.” African Affairs 116, no. 465 (2017): 573–96. https://doi.org/10.1093/afraf/adx022. Lindberg, Staffan I. “‘It’s Our Time to “Chop”’: Do Elections in Africa Feed Neo-Patrimonialism Rather than Counter-Act It?” Democratization 10, no. 2 (2003): 121. Mlomo, Mpho. “The Roles and Responsibilities of Members of Parliament in Facilitating MIASA Working Paper 2021(1) 25 Good Governance and Democracy.” In Botswana, Politics and Society, edited by W. A. Edge and M. H. Lekorwe. Pretoria: J.L. van Schaik, 1998. Nonor, Daniel. “JJ Orders NDC to Copy NPP.” The Chronicle. October 7, 2013. https://allafrica.com/stories/201310072122.html. O’Gorman, Melanie. “Why the CCM Won’t Lose: The Roots of Single-Party Dominance in Tanzania.” Journal of Contemporary African Studies 30, no. 2 (2012): 313–33. https://doi.org/10.1080/02589001.2012.669566. Omobowale, A. O. “Flawed Political Party Primaries In Nigeria’s Fourth Republic: The Case Of The Peoples’ Democratic Party (PDP).” African Journal for the Psychological Study of Social Issues 11, no. 1 (2008): 282–91. Osei, Anja. “Formal Party Organisation and Informal Relations in African Parties: Evidence from Ghana.” The Journal of Modern African Studies 54, no. 1 (2016): 37–66. https://doi.org/10.1017/S0022278X15000981. ———. “Political Parties in Ghana: Agents of Democracy?” Journal of Contemporary African Studies 31, no. 4 (2013): 543–63. https://doi.org/10.1080/02589001.2013.839227. Pierson, Paul. Politics in Time. Princeton: Princeton University Press, 2004. Rahat, Gideon. “Candidate Selection: The Choice Before the Choice.” Journal of Democracy 18, no. 1 (2007): 157–70. https://doi.org/10.1353/jod.2007.0014. Rahat, Gideon, and William P. Cross. “Political Parties and Candidate Selection.” In Oxford Research Encyclopedia of Politics. Oxford: Oxford University Press, 2018. https://oxfordre.com/view/10.1093/acrefore/9780190228637.001.0001/acrefore- 9780190228637-e-625. Sanches, Edalina Rodrigues. Party Systems in Young Democracies: Varieties of Institutionalization in Sub-Saharan Africa. Abingdon: Routledge, 2018. Seeberg, Merete Bech, Michael Wahman, and Svend-Erik Skaaning. “Candidate Nomination, Intra-Party Democracy, and Election Violence in Africa.” Democratization 25, no. 6 (2018): 959–77. https://doi.org/10.1080/13510347.2017.1420057. Tenuche, Marietu. “The Albatross Called Primary Elections and Political Succession in Nigeria.” Canadian Social Science 7, no. 4 (2011): 121–30. https://doi.org/10.3968/j.css.1923669720110704.063. Wang, Vibeke, and Ragnhild Louise Muriaas. “Candidate Selection and Informal Soft Quotas for Women: Gender Imbalance in Political Recruitment in Zambia.” Politics, Groups, and Identities 7, no. 2 (2019): 401–11. https://doi.org/10.1080/21565503.2018.1564056. Westminster Foundation for Democracy. “Cost of Politics in Ghana.” London: Westminster Foundation for Democracy, 2017. MIASA Working Paper 2021(1) 26 Biographical notes Susan Dodsworth is a Lecturer at the University of Queensland’s School of Political Science and International Studies. Prior to this, she worked as a Research Fellow at the University of Birmingham’s International Development Department. Her research examines the role of political institutions – including parliaments – in development and democratization, and the challenges faced by election observers across sub-Saharan Africa. Full details of her re- search are available at her website: www.susandodsworth.com Dr. Seidu Alidu is currently a Senior Lecturer at the Department of Political Science and a Research Fellow at the Centre for Asian Studies, at the University of Ghana. He was a Do- nahue Institute Scholar on United States’ Political Thought at the University of Massachu- setts in 2014. He is a visiting lecturer to twelve Master of Research and Public Policy (MRPP) Universities in Africa including the University of Lagos, University of Ibadan, University of Dar-es-Salam, University of Nairobi and the University of Botswana. He was a Maria Sibylla Merian Institute for Advanced Studies in Africa (MIASA) fellow on Parliaments and Democ- racy in Africa. He has worked closely with and for numerous policy think tanks including the Partnership for African Social and Governance Research (PASGR) based in Nairobi Kenya; Centre for International Development Issues Nijmegen based in Netherlands; the Institute for Development Studies at the University of Sussex, the Institute for Democratic Govern- ance; the Friedrich-Ebert Stiftung; the United Nations Development Programme, the Varie- ties of Democracy Project based in the University of Gothenburg in Sweden, the Ghana Armed Forces Command and Staff College, the German-African Governance Partnership Or- ganization and the Social Science Research Council based in New York. He has published widely on his subject area with his papers appearing in Peace Review, Criminal Justice Stud- ies, Transitional Justice Review, Journal of African Political Economy and Development, Con- temporary Journal of African Development and Ghana Studies. Gretchen Bauer holds a PhD in political science from the University of Wisconsin-Madison, and is professor of political science and international relations at the University of Delaware, in the USA. She conducts research on women's political leadership in Africa, with a current focus on women in parliament and cabinet in Ghana, and on which she has published many books and journal articles. She has been a Visiting Researcher at the Institute for Public Pol- icy Research in Windhoek, Namibia (2002) and at the University of Botswana in Gaborone (2009). During 2016 she was a Fulbright Scholar at the University of Ghana in Accra. At the University of Delaware, she directed the Young African Leaders Initiative (YALI) Mandela Washington Fellowship from 2014 to 2018. MIASA Working Paper 2021(1) 27 Gbensuglo Alidu Bukari is Political Economist and a Lecturer in Politics and Governance at the Department of Politics and Governance, University for Development Studies, Tamale- Ghana. His areas of research span the fields of political economy, democratic studies and development studies. His research interests include; political economy, political economy of elections, political economy of development, political economy of governance, politics, economics and public policy. He has published in some leading journals including Journal of Applied Economics and Finance, Africa Journal of Public Sector Development and Gov- ernance, Research in World Economy Journal, Legon Journal of the Humanities and Social Science Journal.