African Elections As Vehicles For Change

African Elections as Vehicles for Change Nic Cheeseman Journal of Democracy, Volume 21, Number 4, October 2010, pp. 139-153 (Article) Published by The Johns Hopkins University Press DOI: 10.1353/jod.2010.0019 For additional information about this article http://muse.jhu.edu/journals/jod/summary/v021/21.4.cheeseman.html Access provided by Oxford University Library Services (17 Mar 2014 18:56 GMT) african฀elections as฀vehicles฀for฀change Nic฀Cheeseman Nic฀Cheeseman฀is฀university฀lecturer฀in฀African฀Politics฀at฀Oxford฀University.฀He฀has฀published฀widely฀on฀African฀democratization฀in฀journals฀ such฀as฀African฀Affairs฀and฀the฀Journal฀of฀Modern฀African฀Studies.฀His฀ monograph฀Democracy฀in฀Africa฀is฀forthcoming฀from฀Cambridge฀University฀Press. S ince฀the฀return฀of฀multipartism฀to฀sub-Saharan฀Africa฀in฀the฀early฀1990s,฀ transfers฀ of฀ power฀ between฀ rival฀ parties฀ have฀ been฀ rare—even฀ in฀ those฀ countries฀where฀an฀increase฀in฀political฀rights฀has฀gone฀hand฀in฀hand฀with฀ greater฀ respect฀ for฀ civil฀ liberties.฀ Instead,฀ presidents’฀ ability฀ to฀ control฀ state฀institutions฀and฀the฀flow฀of฀patronage฀has฀generated฀an฀incumbency฀ bias฀so฀strong฀that฀numerous฀countries฀in฀Africa฀have฀witnessed฀elections฀ without฀ change.฀ Since฀ 1989,฀ only฀ twelve฀ sub-Saharan฀ Africa฀ countries฀ have฀experienced฀opposition฀victories,฀and฀of฀those฀only฀Benin,฀Ghana,฀ Madagascar,฀and฀Mauritius,฀have฀met฀Samuel฀P.฀Huntington’s฀famous,฀if฀ crude,฀“two-turnover”฀test฀for฀a฀consolidated฀democracy.1฀ Explanations฀of฀why฀transfers฀of฀power฀have฀occurred฀in฀certain฀African฀ countries฀ but฀ not฀ in฀ others฀ have฀ typically฀ taken฀ the฀ form฀ of฀ finegrained฀ case฀ studies฀ that฀ describe฀ a฀ complex฀ mix฀ of฀ economic฀ decline,฀ falling฀support฀for฀incumbents,฀and฀the฀ability฀or฀inability฀of฀opposition฀ parties฀to฀form฀a฀broad฀alliance.2฀Comparative฀analysis฀has฀been฀rare,฀and฀ the฀ literature฀ has฀ said฀ little฀ about฀ a฀ general฀ pattern฀ that฀ merits฀ more฀ attention฀than฀it฀has฀so฀far฀received:฀In฀presidential฀systems,฀the฀prospects฀ for฀opposition฀success฀are฀conditioned฀by฀whether฀or฀not฀the฀incumbent฀ stands฀for฀reelection.฀In฀other฀words,฀when฀a฀sitting฀president฀cannot฀run฀ because฀of฀term฀limits,฀health,฀or฀internal฀party฀rules,฀the฀ruling฀party฀is฀ likely฀to฀perform฀significantly฀worse฀at฀the฀polls.฀ The฀ importance฀ of฀ incumbency฀ in฀ Africa฀ is฀ one฀ manifestation฀ of฀ a฀ more฀general฀trend฀noted฀by฀Gideon฀Maltz฀in฀these฀pages฀in฀2007.฀Comparing฀all฀elections฀in฀electoral฀authoritarian฀regimes฀between฀1992฀and฀ 2006,฀ Maltz฀ finds฀ that฀ while฀ incumbents฀ retained฀ power฀ in฀ 93฀ percent฀ Journal฀of฀Democracy฀฀Volume฀21,฀฀Number฀4฀฀October฀2010 ©฀2010฀National฀Endowment฀for฀Democracy฀and฀The฀Johns฀Hopkins฀University฀Press 140 Journal฀of฀Democracy Table฀1—PresidenTial฀Transfers฀of฀Power฀in฀MulTiParTy฀ africa,฀1990–20091 Transfers฀in฀ Incumbent฀Polls Transfers฀in฀ Open-Seat฀Polls Elections Benin฀(1991,฀1996),฀Cape฀Verde฀(1991),฀Madagascar฀ (1993,฀1996,฀2001),฀Malawi฀(1994),฀Senegal฀(2000),฀ Zambia฀(1991) Benin฀ (2006),฀ Cape฀ Verde฀ (2001),฀ Côte฀ d’Ivoire฀ (2000),฀ Ghana฀ (2000,฀ 2008),฀ Kenya฀ (2002),฀ Mali฀ (2002),฀Niger฀(1993),฀Sierra฀Leone฀(2007) Sample฀ includes฀ all฀ sub-Saharan฀ African฀ polities฀ holding฀ direct฀ multiparty฀ presidential฀ elections.฀Transfers฀are฀excluded฀if฀they฀were฀immediately฀preceded฀by฀a฀coup฀and฀therefore฀held฀in฀the฀absence฀of฀a฀ruling฀party. 1 of฀the฀elections฀that฀they฀contested,฀their฀successors฀won฀just฀52฀percent฀ of฀ the฀ time.3฀ The฀ pattern฀ that฀ Maltz฀ identifies฀ is฀ especially฀ noteworthy฀ in฀ the฀ African฀ context,฀ where฀ weak฀ institutions฀ and฀ personalized฀ power฀ tend฀to฀magnify฀the฀advantages฀of฀incumbency.฀Although฀open-seat฀polls฀ have฀been฀relatively฀rare,฀they฀account฀for฀half฀of฀all฀presidential฀transfers฀ of฀ power฀ from฀ one฀ party฀ to฀ another฀ between฀ 1990฀ and฀ 2009฀ (see฀ Table฀ 1฀above).฀Indeed,฀most฀of฀the฀elections฀cited฀as฀evidence฀of฀real฀democratic฀ gains฀ in฀ Africa—such฀ as฀ the฀ Ghanaian฀ election฀ of฀ 2000฀ and฀ the฀ Kenyan฀election฀of฀2002—occurred฀after฀term฀limits฀forced฀presidents฀to฀ step฀down.฀ Despite฀the฀clear฀significance฀of฀incumbency฀to฀political฀change฀around฀ the฀world,฀there฀have฀been฀few฀attempts฀to฀trace฀the฀mechanisms฀through฀ which฀ nonincumbent฀ elections฀ create฀ fresh฀ opportunities฀ for฀ opposition฀ parties฀ or฀ to฀ assess฀ whether฀ open-seat฀ polls฀ are฀ more฀ likely฀ to฀ lead฀ to฀ transfers฀of฀power฀in฀some฀regions฀rather฀than฀others.฀These฀lacunae฀represent฀significant฀limitations฀on฀our฀understanding฀of฀electoral฀competition,฀because฀incumbency฀matters฀for฀different฀reasons฀in฀different฀political฀contexts.฀ Consider฀the฀United฀States,฀where฀in฀the฀last฀thirty฀years฀only฀George฀ H.W.฀Bush฀lost฀as฀an฀incumbent฀and฀transfers฀of฀power฀between฀parties฀ have฀historically฀resulted฀from฀open-seat฀polls,฀the฀most฀recent฀example฀ being฀ the฀ election฀ of฀ Barack฀ Obama.4฀ Much฀ of฀ the฀ extant฀ work฀ on฀ the฀ significance฀of฀incumbency฀in฀the฀United฀States฀and฀other฀established฀democracies฀focuses฀on฀the฀claim฀that฀because฀the฀electorate฀is฀faced฀with฀ lesser-known฀candidates฀in฀open-seat฀elections,฀the฀significance฀of฀retrospective฀voting฀(casting฀one’s฀ballot฀based฀on฀a฀candidate’s฀past฀performance)฀is฀reduced,฀to฀the฀likely฀detriment฀of฀ruling฀parties.฀Although฀this฀ literature฀ offers฀ many฀ insights฀ into฀ the฀ advantages฀ conferred฀ by฀ incumbency฀in฀Europe฀and฀North฀America,฀it฀is฀not฀very฀helpful฀in฀accounting฀ for฀the฀character฀of฀nonincumbent฀elections฀in฀Africa.฀ First฀of฀all,฀voting฀behavior฀in฀Africa฀is฀typically฀less฀well-explained฀ by฀retrospective-voting฀models,฀and฀in฀any฀case฀evaluations฀of฀party฀and฀ individual฀performances฀are฀often฀more฀heavily฀shaped฀by฀patronage฀relations฀and฀communal฀identities฀than฀is฀the฀case฀in฀established฀democracies.฀ Nic฀Cheeseman 141 Second,฀the฀distinctive฀way฀in฀which฀political฀competition฀is฀conducted฀ in฀contemporary฀Africa฀exaggerates฀the฀significance฀of฀incumbency.฀The฀ weakness฀of฀political฀parties,฀the฀salience฀of฀ethnic฀identity,฀and฀the฀importance฀of฀neopatrimonial฀politics฀combine฀to฀create฀an฀electoral฀landscape฀in฀which฀institutions฀are฀typically฀weak฀and฀power฀is฀concentrated฀ in฀political฀leaders.฀Thus,฀while฀open-seat฀elections฀are฀likely฀to฀be฀even฀ more฀ strongly฀ correlated฀ with฀ transfers฀ of฀ power฀ in฀ Africa,฀ the฀ mechanisms฀through฀which฀incumbency฀works฀are฀likely฀to฀be฀distinctive.฀Finally,฀because฀the฀existing฀literature฀addresses฀democratic฀polities฀only,฀it฀ has฀little฀to฀say฀about฀the฀interrelationship฀of฀incumbency,฀turnover,฀and฀ democratization.฀Yet฀if฀democratic฀consolidation฀requires฀the฀removal฀of฀ old฀authoritarian฀powers฀and฀turnover฀is฀an฀indicator฀of฀democratic฀progress,฀open-seat฀elections฀may฀be฀a฀significant฀factor฀in฀the฀democratization฀process.฀ This฀essay฀draws฀on฀a฀comparative฀data฀set฀of฀African฀elections฀in฀the฀ multiparty฀era฀to฀demonstrate฀that฀open-seat฀polls฀are฀particularly฀likely฀ to฀ result฀ in฀ opposition฀ victories฀ in฀ sub-Saharan฀ Africa฀ for฀ two฀ reasons:฀ because฀ of฀ the฀ challenges฀ that฀ they฀ pose฀ for฀ ruling฀ parties฀ and฀ because฀ they฀are฀often฀more฀transparent฀and฀fair.฀While฀this฀combination฀does฀not฀ ensure฀ a฀ transfer฀ of฀ power,฀ it฀ does฀ create฀ a฀ rare฀ window฀ of฀ opportunity฀ for฀ political฀ change.฀ When฀ an฀ open-seat฀ poll฀ coincides฀ with฀ economic฀ hard฀times,฀government-related฀political฀scandals,฀and฀a฀strong฀opposition฀ campaign,฀turnover฀becomes฀not฀only฀possible฀but฀likely.฀Thus฀international฀and฀domestic฀actors฀seeking฀to฀enhance฀the฀prospects฀for฀political฀ change฀ would฀ do฀ well฀ to฀ prioritize฀ the฀ introduction฀ of,฀ and฀ respect฀ for,฀ term฀limits.฀Of฀course,฀turnover฀does฀not฀guarantee฀democratic฀gains,฀as฀ the฀recent฀experience฀of฀Kenya฀demonstrates,฀but฀it฀may฀help฀both฀to฀build฀ trust฀in฀the฀rules฀of฀the฀game฀and฀to฀boost฀the฀legitimacy฀of฀the฀political฀ process,฀and฀it฀is฀ultimately฀necessary฀for฀the฀institutionalization฀of฀competitive฀multiparty฀politics.฀ The Impact of Incumbency The฀vast฀majority฀of฀African฀political฀systems฀now฀feature฀constitutions฀ that฀place฀a฀two-term฀limit฀on฀the฀presidency.฀Of฀the฀thirty฀countries฀where฀ term฀limits฀have฀been฀reached,฀presidents฀refrained฀from฀seeking฀a฀third฀ term฀in฀Benin,฀Cape฀Verde,฀Ghana,฀Kenya,฀Mali,฀Mozambique,฀S~ao฀Tomé฀ and฀Príncipe,฀Seychelles฀(three-term฀limit),฀Sierra฀Leone,฀and฀Tanzania.5฀ In฀their฀2007฀article฀in฀these฀pages,฀Daniel฀Posner฀and฀Daniel฀Young฀cited฀ these฀cases,฀as฀well฀as฀the฀inability฀of฀the฀presidents฀of฀Malawi,฀Nigeria,฀ and฀Zambia฀to฀secure฀unconstitutional฀third฀terms,฀as฀evidence฀that฀African฀politics฀was฀“institutionalizing.”6฀Subsequent฀events,฀however,฀have฀ demonstrated฀that฀respect฀for฀term฀limits฀has฀not฀been฀ingrained฀across฀the฀ continent.฀To฀date,฀incumbent฀presidents฀have฀succeeded฀in฀amending฀the฀ constitution฀to฀remain฀in฀office฀in฀Angola,฀Burkina฀Faso,฀Chad,฀Djibouti,฀ 142 Journal฀of฀Democracy Table฀2—The฀iMPacT฀of฀incuMbency฀on฀ nonfounding-elecTion฀ouTcoMes,฀1990–20091 Incumbent฀(%) Open-Seat฀(%) Difference Ruling-party฀vote฀share 65 53 -12 Winning฀margin฀where฀ruling฀party฀won 41 31 -10 Voter฀turnout 65 62 -3 Rate฀of฀turnover 12 45 +33 1 Sample฀includes฀all฀sub-Saharan฀polities฀holding฀direct฀multiparty฀presidential฀elections. Gabon,฀ Namibia,฀ Niger,฀ Togo,฀ and฀ Uganda.7฀ Nonetheless,฀ over฀ the฀ last฀ two฀ decades฀ the฀ enforcement฀ of฀ term฀ limits฀ in฀ a฀ range฀ of฀ countries฀ has฀ resulted฀in฀a฀significant฀number฀of฀open-seat฀polls,฀making฀it฀possible฀for฀ the฀first฀time฀to฀examine฀the฀impact฀of฀incumbency฀on฀electoral฀outcomes฀ in฀Africa.฀ If฀one฀looks฀at฀all฀direct฀presidential฀elections฀in฀Africa฀between฀1990฀ and฀2009,฀it฀is฀clear฀that฀opposition฀parties฀perform฀far฀worse฀when฀running฀against฀a฀sitting฀president.฀Although฀there฀were฀nine฀cases฀in฀which฀ ruling฀parties฀lost฀despite฀fielding฀an฀incumbent,฀four฀of฀these฀defeats฀occurred฀during฀the฀rather฀exceptional฀circumstances฀of฀founding฀elections฀ (Benin฀ in฀ 1991,฀ Cape฀ Verde฀ in฀ 1991,฀ Malawi฀ in฀ 1994,฀ and฀ Zambia฀ in฀ 1991),฀and฀three฀are฀accounted฀for฀by฀just฀one฀country฀(Madagascar).฀By฀ contrast,฀ the฀ transfers฀ of฀ power฀ resulting฀ from฀ open-seat฀ elections฀ have฀ been฀ spread฀ more฀ evenly,฀ both฀ geographically฀ and฀ chronologically฀ (see฀ Table฀1).฀Moreover,฀a฀comparison฀of฀all฀nonfounding฀elections฀(see฀Table฀ 2)฀reveals฀that฀opposition฀parties฀were฀almost฀four฀times฀more฀likely฀to฀ win฀nonincumbent฀elections,฀while฀the฀vote฀share฀of฀ruling฀parties฀dropped฀ on฀average฀by฀12฀percent฀when฀they฀had฀to฀put฀up฀a฀new฀candidate.฀Even฀ when฀incumbent฀parties฀won,฀their฀margin฀of฀victory฀fell฀by฀10฀percent฀in฀ open-seat฀polls. Significantly,฀these฀results฀are฀not฀driven฀by฀the฀inclusion฀of฀electoral฀ authoritarian฀ systems,฀ in฀ which฀ landslide฀ victories฀ and฀ incumbent฀ elections฀inevitably฀go฀hand฀in฀hand.฀Even฀when฀excluding฀all฀countries฀rated฀ as฀Not฀Free฀by฀Freedom฀House,฀the฀share฀of฀elections฀won฀by฀the฀ruling฀ party฀ in฀ incumbent฀ and฀ open-seat฀ polls฀ remains฀ virtually฀ unchanged฀ at฀ 64฀percent฀and฀50฀percent,฀respectively.฀Nor฀is฀it฀the฀case฀that฀presidents฀ abide฀by฀term฀limits฀only฀in฀full฀democracies.฀Thus฀the฀pattern฀reported฀ here฀ does฀ not฀ emerge฀ simply฀ because฀ open-seat฀ elections฀ occur฀ only฀ in฀ countries฀with฀less฀repressive฀political฀environments฀that฀are฀more฀conducive฀ to฀ opposition฀ victories.฀ Even฀ when฀ the฀ sample฀ is฀ limited฀ to฀ Africa’s฀electoral฀democracies,฀ruling฀parties฀fare฀worse฀in฀open-seat฀polls.฀ In฀short,฀incumbency฀has฀an฀independent฀impact฀on฀the฀performance฀of฀ the฀opposition.฀ Three฀main฀factors฀explain฀why฀ruling฀parties฀struggle฀without฀an฀incumbent฀ candidate.฀ First฀ and฀ most฀ fundamentally,฀ weakly฀ institutionalized฀party฀structures฀and฀highly฀personalized฀politics฀mean฀that฀the฀battle฀ to฀select฀a฀new฀presidential฀candidate฀often฀results฀in฀a฀divided฀party฀un- Nic฀Cheeseman 143 able฀to฀roll฀out฀a฀united฀campaign.฀Second,฀nonincumbent฀candidates฀cannot฀point฀to฀a฀record฀of฀accomplishments,฀and฀so฀their฀promises฀to฀key฀client฀groups฀are฀less฀credible.฀At฀the฀same฀time,฀nonincumbent฀candidates฀ may฀face฀greater฀obstacles฀in฀funding฀their฀election฀campaigns.฀Finally,฀ unless฀the฀outgoing฀president฀has฀handpicked฀his฀successor,฀there฀may฀be฀ a฀ distance฀ between฀ the฀ incumbent฀ and฀ his฀ party’s฀ candidate.฀ The฀ president,฀who฀retains฀ultimate฀control฀over฀the฀machinery฀of฀government฀until฀ a฀new฀leader฀is฀inaugurated,฀may฀thus฀feel฀that฀he฀has฀less฀at฀stake฀in฀the฀ polls฀and฀prove฀more฀willing฀to฀preside฀over฀a฀free฀and฀fair฀contest.฀ Succession Struggles Maltz฀has฀argued฀that฀the฀major฀challenge฀to฀the฀continued฀hegemony฀ of฀electoral฀authoritarian฀regimes฀is฀effectively฀managing฀the฀succession฀ process.8฀Ruling฀parties฀in฀Africa฀are฀well฀aware฀of฀the฀danger฀posed฀by฀ leadership฀ battles.฀ Thus,฀ they฀ typically฀ either฀ avoid฀ asking฀ incumbents฀ to฀stand฀for฀renomination฀or฀they฀employ฀rubber-stamp฀processes฀structured฀to฀prevent฀intraparty฀competition.฀Where฀term฀limits฀exist฀and฀are฀ respected,฀however,฀ruling฀parties฀are฀forced฀to฀engage฀in฀the฀unsettling฀ act฀of฀choosing฀a฀new฀leader.฀This฀is฀a฀huge฀hurdle฀for฀parties฀in฀Africa,฀ because฀the฀centrality฀of฀the฀executive฀to฀political฀life฀and฀the฀low฀institutionalization฀of฀political฀parties฀tend฀to฀magnify฀the฀stakes฀of฀succession฀ politics.฀Rodger฀Govea฀and฀John฀Holm฀find฀that,฀as฀a฀result฀of฀the฀weakness฀of฀formal฀party฀structures,฀the฀majority฀of฀succession฀processes฀(61฀ percent)฀ between฀ 1963฀ and฀ 1988฀ were฀ “unregulated,”฀ meaning฀ that฀ the฀ outcome฀of฀the฀process฀was฀governed฀by฀the฀informal฀balance฀of฀personal฀ networks฀within฀the฀party฀rather฀than฀a฀straightforward฀provision฀in฀the฀ party฀ constitution.฀ Moreover,฀ many฀ unregulated฀ successions฀ deteriorate฀ into฀political฀crises฀and฀violence.9฀Less฀dramatically,฀the฀mismanagement฀ of฀ the฀ succession฀ process฀ can฀ entrench฀ internal฀ party฀ divisions,฀ leading฀ to฀greater฀infighting฀and,฀in฀extreme฀cases,฀the฀breakup฀of฀the฀party,฀all฀ of฀which฀makes฀it฀harder฀to฀put฀together฀a฀coherent฀and฀united฀election฀ campaign.฀The฀primary฀process฀is฀typically฀destabilizing฀for฀most฀political฀parties—even฀in฀well-established฀democracies.฀It฀is฀even฀more฀likely฀ to฀lead฀to฀lasting฀divisions฀where฀parties฀are฀poorly฀institutionalized,฀as฀in฀ sub-Saharan฀Africa.10฀ The฀story฀of฀Sierra฀Leone’s฀ruling฀party฀during฀the฀2007฀election฀cycle฀ provides฀a฀good฀example.฀In฀September฀2005,฀the฀Sierra฀Leone฀People’s฀ Party฀(SLPP)฀selected฀the฀sitting฀vice-president,฀Solomon฀Berewa,฀to฀replace฀outgoing฀president฀Ahmad฀Tejan฀Kabbah,฀thwarting฀the฀presidential฀ ambitions฀of฀Charles฀Margai.฀Margai,฀son฀of฀the฀country’s฀second฀prime฀ minister,฀ultimately฀decided฀to฀abandon฀the฀party฀and฀establish฀the฀rival฀ People’s฀Movement฀for฀Democratic฀Change฀(PMDC).฀The฀PMDC฀represented฀a฀major฀challenge฀to฀continued฀SLPP฀rule฀for฀two฀reasons.฀First,฀ the฀new฀party฀appealed฀to฀some฀of฀the฀same฀constituencies฀as฀the฀SLPP฀ 144 Journal฀of฀Democracy and฀threatened฀to฀siphon฀support฀away฀from฀the฀ruling฀party฀in฀its฀strongholds฀in฀the฀south฀and฀east฀of฀the฀country.฀Second,฀the฀SLPP฀already฀faced฀ a฀severe฀electoral฀challenge฀from฀Ernest฀Bai฀Koroma,฀the฀candidate฀of฀the฀ All฀People’s฀Congress฀(APC).11฀In฀the฀first฀round฀of฀the฀presidential฀election,฀Margai฀secured฀14฀percent฀of฀the฀vote,฀denying฀Berewa฀the฀55฀percent฀that฀he฀needed฀to฀avoid฀a฀second฀round.฀Then,฀in฀the฀runoff฀between฀ Berewa฀and฀Karoma,฀Margai฀turned฀his฀back฀on฀his฀former฀colleagues฀in฀ government฀and฀urged฀his฀supporters฀to฀back฀the฀opposition,฀enabling฀the฀ APC฀to฀make฀inroads฀into฀traditional฀SLPP฀areas฀and฀helping฀Karoma฀to฀ win฀the฀additional฀votes฀needed฀to฀secure฀a฀rare฀opposition฀victory.฀ In฀ Kenya,฀ the฀ ruling฀ Kenya฀ African฀ National฀ Union฀ (KANU)฀ paid฀ a฀ similarly฀high฀price฀for฀failing฀successfully฀to฀manage฀the฀battle฀to฀replace฀ Daniel฀arap฀Moi฀in฀the฀run-up฀to฀the฀2002฀election.฀Moi’s฀controversial฀ decision฀to฀impose฀his฀favored฀successor,฀Uhuru฀Kenyatta,฀resulted฀in฀a฀ number฀of฀prominent฀political฀leaders฀abandoning฀KANU฀to฀join฀the฀opposition—a฀turn฀of฀events฀that฀directly฀contributed฀to฀KANU’s฀defeat฀by฀ the฀National฀Alliance฀Rainbow฀Coalition฀(NARC).฀ In฀Ghana฀in฀2008,฀a฀divisive฀contest฀to฀replace฀President฀John฀Kufour฀ left฀the฀ruling฀New฀Patriotic฀Party฀(NPP)฀riven฀and฀marred฀its฀public฀image,฀ opening฀ the฀ door฀ to฀ an฀ opposition฀ win฀ at฀ the฀ polls.฀ The฀ seventeen฀ candidates฀who฀contested฀the฀NPP฀presidential฀primary฀spent฀vast฀sums฀of฀ money฀during฀a฀period฀of฀economic฀hardship,฀giving฀the฀impression฀that฀ the฀party฀was฀out฀of฀touch฀with฀the฀common฀man.฀Intraparty฀competition฀ intensified฀as฀Kufour฀meddled฀in฀the฀process฀in฀a฀bid฀to฀prevent฀longtime฀ rival฀Nana฀Addo฀Dankwa฀Akufo-Addo฀from฀securing฀the฀post.฀Although฀ Akufo-Addo฀ultimately฀won฀a฀relatively฀easy฀victory,฀the฀primaries฀left฀the฀ NPP฀deeply฀divided.฀The฀opposition฀National฀Democratic฀Congress฀(NDC)฀ made฀effective฀use฀in฀its฀campaign฀not฀only฀of฀presidential฀candidate฀John฀ Evans฀Atta฀Mills฀but฀also฀of฀former฀president฀Jerry฀John฀Rawlings.฀President฀Kufour,฀by฀contrast,฀refused฀to฀campaign฀for฀Akufo-Addo฀and฀would฀ not฀even฀return฀the฀phone฀calls฀of฀his฀own฀party’s฀campaign฀manager.12฀ The฀deep฀divisions฀that฀ran฀through฀the฀NPP฀extended฀to฀the฀selection฀ of฀constituency฀members฀of฀parliament฀(MPs).฀Losing฀primary฀candidates฀ reported฀that฀senior฀political฀leaders฀had฀“interfered”฀in฀their฀contests฀in฀ order฀to฀favor฀an฀ally฀from฀the฀“other฀side,”฀and฀consequently฀refused฀to฀ accept฀ the฀ legitimacy฀ of฀ their฀ defeat.13฀ Following฀ the฀ party฀ primaries,฀ a฀ significant฀number฀of฀losers฀felt฀sufficiently฀infuriated฀that฀they฀decided฀ to฀ run฀ as฀ independents—a฀ relatively฀ rare฀ occurrence฀ in฀ a฀ strongly฀ twoparty฀ system.฀ These฀ breakaway฀ candidacies฀ had฀ a฀ huge฀ impact฀ on฀ the฀ parliamentary฀elections฀and฀even฀the฀presidential฀election.฀Not฀only฀did฀ the฀ former-NPP฀ independents฀ not฀ campaign฀ on฀ behalf฀ of฀ Akufo-Addo,฀ they฀also฀diverted฀the฀attention฀of฀the฀official฀NPP฀candidates—who฀were฀ now฀embroiled฀in฀divisive฀local฀battles—away฀from฀campaigning฀for฀their฀ party’s฀presidential฀contender.฀In฀the฀end,฀Mills฀beat฀Akufo-Addo฀by฀just฀ 29,419฀ votes฀ in฀ a฀ runoff,฀ leaving฀ NPP฀ supporters฀ to฀ rue฀ these฀ internal฀ Nic฀Cheeseman 145 cleavages.฀The฀NPP฀also฀suffered฀in฀the฀parliamentary฀polls,฀where฀internal฀tensions฀most฀likely฀cost฀the฀party฀3฀seats฀where฀former฀NPP฀candidates฀won฀as฀independents,฀and฀5฀to฀10฀seats฀where฀official฀and฀dissident฀ NPP฀candidates฀split฀the฀vote,฀allowing฀for฀an฀NDC฀victory.฀Given฀that฀ the฀NDC฀won฀only฀7฀more฀seats฀than฀the฀NPP฀(114฀to฀107),฀it฀is฀easy฀to฀ see฀why฀senior฀NPP฀leaders฀believe฀that฀it฀was฀disunity฀that฀cost฀them฀the฀ election.14฀ Of฀ course,฀ the฀ challenges฀ posed฀ by฀ succession฀ battles฀ are฀ not฀ insurmountable.฀In฀the฀much฀discussed฀case฀of฀South฀Africa,฀the฀African฀National฀ Congress฀ (ANC)฀ went฀ through฀ an฀ extremely฀ divisive฀ succession฀ battle฀to฀choose฀Thabo฀Mbeki’s฀successor,฀who฀then฀went฀on฀to฀lead฀the฀ ANC฀ to฀ yet฀ another฀ landslide฀ victory฀ in฀ a฀ nonincumbent฀ election.฀ Yet฀ even฀in฀this฀case,฀where฀the฀ANC’s฀legitimacy฀as฀the฀party฀of฀liberation฀ continued฀to฀underpin฀its฀electoral฀dominance,฀the฀succession฀struggle฀left฀ its฀mark.฀The฀rivalry฀between฀Mbeki฀and฀his฀main฀adversary฀Jacob฀Zuma฀ intensified฀competition฀within฀the฀ANC,฀and฀a฀period฀of฀heated฀infighting฀ found฀its฀ultimate฀expression฀in฀the฀decision฀of฀a฀number฀of฀ANC฀leaders฀to฀leave฀the฀party฀to฀form฀the฀Congress฀of฀the฀People฀(COPE).฀COPE฀ presidential฀candidate฀Mosiuoa฀Lekota฀may฀have฀secured฀only฀1.3฀million฀ votes฀(7.4฀percent),฀but฀this฀was฀still฀a฀significant฀performance฀given฀that฀ COPE฀had฀been฀formed฀just฀months฀earlier.฀In฀short,฀succession฀battles฀ and฀open-seat฀elections฀tend฀to฀produce฀internal฀strains฀and฀challenges฀for฀ ruling฀parties,฀even฀though฀these฀are฀not฀always฀fatal. The Advantages of Incumbency Incumbents฀ enjoy฀ considerable฀ advantages฀ in฀ presidential฀ systems฀ where฀power฀and฀control฀over฀patronage฀are฀heavily฀centralized,฀but฀these฀ advantages฀may฀not฀extend฀to฀the฀successor฀in฀open-seat฀polls.฀Neopatrimonial฀political฀structures฀limit฀the฀constraints฀on฀executive฀action฀and฀ enable฀incumbents฀to฀treat฀public฀funds฀as฀personal฀largesse.฀Thus฀many฀ African฀presidents฀have฀been฀remarkably฀adept฀at฀building฀political฀machines฀while฀in฀office฀by฀diverting฀resources฀to฀their฀“homelands”฀in฀order฀ to฀reward฀supporters฀and฀punish฀enemies.฀The฀instrumental฀use฀of฀patronage฀to฀play฀divide-and-rule฀politics฀has฀characterized฀regimes฀as฀diverse฀ as฀ Moi’s฀ in฀ Kenya,฀ Hastings฀ Kamuzu฀ Banda’s฀ in฀ Malawi,฀ Gnassingbé฀ Eyadema’s฀in฀Togo,฀and฀Joseph฀Kabila’s฀in฀the฀Democratic฀Republic฀of฀ Congo.฀ Significantly,฀ building฀ an฀ effective฀ political฀ machine฀ is฀ a฀ longterm฀strategy฀that฀must฀begin฀anew฀as฀soon฀as฀each฀election฀is฀over;฀as฀ Mwai฀Kibaki฀learned฀recently,฀voters฀are฀likely฀to฀turn฀their฀backs฀on฀lastminute฀appeals฀if฀they฀feel฀that฀they฀have฀been฀abandoned฀for฀years.฀ Precisely฀because฀they฀are฀in฀power฀and฀can฀construct฀durable฀patronage฀networks,฀incumbents฀have฀a฀considerable฀advantage.฀They฀are฀also฀ better฀placed฀to฀make฀believable฀promises฀to฀prospective฀voters,฀who฀can฀ use฀a฀leader’s฀past฀performance฀to฀evaluate฀the฀reliability฀of฀his฀pledges฀to฀ 146 Journal฀of฀Democracy provide฀more฀goods฀in฀the฀future.฀This฀is฀significant,฀because฀in฀electoral฀ systems฀with฀few฀swing฀voters,฀electoral฀success฀is฀largely฀determined฀by฀ candidates’฀ ability฀ to฀ maximize฀ the฀ turnout฀ of฀ their฀ core฀ supporters.฀ As฀ recent฀research฀on฀Kenya฀has฀shown,฀the฀distribution฀of฀patronage฀and฀the฀ location฀ of฀ parties’฀ rallies฀ strongly฀ suggest฀ that฀ parties฀ consciously฀ and฀ strategically฀ work฀ to฀ generate฀ the฀ highest฀ turnout฀ possible฀ among฀ their฀ base฀and฀to฀attract฀some฀swing฀voters,฀while฀largely฀ignoring฀opposition฀ strongholds.15฀ Incumbents’฀ capacity฀ to฀ construct฀ political฀ machines฀ and฀ make฀credible฀clientelistic฀appeals฀gives฀them฀a฀comparative฀advantage฀in฀ generating฀turnout,฀so฀long฀as฀they฀have฀been฀sufficiently฀generous.16 The฀advantage฀of฀incumbency฀may฀not฀weigh฀so฀heavily฀in฀open-seat฀ polls,฀however.฀While฀nonincumbent฀candidates฀are฀likely฀to฀be฀wealthy,฀ they฀will฀not฀have฀had฀the฀same฀level฀of฀access฀to฀state฀resources.฀If฀the฀ sitting฀president฀will฀not฀bankroll฀the฀campaign฀of฀his฀successor,฀the฀resource฀gap฀between฀government฀and฀opposition฀candidates฀narrows.฀This฀ appears฀to฀have฀been฀the฀case฀in฀Kenya,฀where฀Moi’s฀backers฀refused฀to฀ fund฀Uhuru฀Kenyatta’s฀campaign,฀forcing฀him฀to฀rely฀instead฀on฀the฀Kenyatta฀family฀fortune;฀consequently,฀in฀the฀2002฀contest฀KANU฀spent฀just฀ a฀third฀of฀the฀$100฀million฀that฀it฀had฀lavished฀on฀the฀1992฀campaign.17฀ Nonincumbents฀may฀also฀find฀it฀difficult฀to฀convince฀voters฀of฀their฀credibility,฀and฀may฀struggle฀to฀win฀the฀loyalty฀of฀their฀predecessor’s฀supporters.฀In฀highly฀personalized฀systems,฀voters฀often฀identify฀with฀a฀particular฀ individual฀rather฀than฀a฀party,฀and฀the฀goodwill฀won฀by฀an฀incumbent฀may฀ be฀hard฀to฀transfer฀to฀his฀successor.฀ This฀problem฀can฀be฀exacerbated฀by฀ethnic฀rivalries,฀especially฀if฀the฀ incoming฀leader฀represents฀a฀different฀ethnic฀group฀or฀community฀from฀ that฀of฀his฀predecessor;฀supporters฀of฀the฀outgoing฀leader฀may฀be฀reluctant฀ to฀turn฀out฀en฀masse฀for฀the฀new฀man,฀while฀the฀new฀candidate’s฀supporters฀are฀less฀likely฀to฀be฀fully฀incorporated฀into฀the฀party’s฀political฀machine.฀The฀multiethnic฀nature฀of฀many฀African฀countries฀means฀that฀openseat฀ elections฀ will฀ often฀ result฀ in฀ a฀ change฀ in฀ the฀ ethnic฀ identity฀ of฀ the฀ ruling฀party’s฀presidential฀candidate.฀In฀fact,฀in฀highly฀diverse฀countries฀ such฀as฀Kenya฀and฀Mali,฀successful฀parties฀must฀be฀multiethnic฀coalitions฀ in฀ which฀ the฀ presidential฀ candidacy฀ rotates฀ among฀ the฀ different฀ groups.฀ (This฀ practice฀ has฀ been฀ institutionalized฀ in฀ Nigeria.)฀ Kenya฀ and฀ Ghana฀ well฀illustrate฀the฀challenges฀to฀ruling฀parties฀posed฀by฀the฀ethnic฀rotation฀ of฀presidential฀candidates.฀In฀Kenya฀in฀2002,฀it฀was฀partly฀in฀recognition฀ of฀the฀groundswell฀of฀frustration฀with฀24฀years฀of฀“Kalenjin฀rule”฀that฀Moi฀ chose฀Uhuru฀Kenyatta,฀a฀Kikuyu,฀as฀his฀successor.฀But฀Kenyatta฀had฀no฀ credibility฀among฀many฀of฀the฀smaller฀ethnic฀groups฀that฀Moi฀had฀carefully฀knitted฀into฀his฀ruling฀alliance.฀A฀huge฀number฀refused฀to฀transfer฀ their฀allegiance฀to฀Kenyatta,฀contributing฀to฀a฀landslide฀election฀victory฀ for฀Kibaki฀and฀the฀NARC.฀ The฀NPP฀faced฀a฀similar,฀if฀less฀dramatic,฀problem฀in฀Ghana฀when฀John฀ Kufour,฀ an฀ Ashanti,฀ was฀ replaced฀ by฀ Akufo-Addo,฀ an฀ Akyem.฀ Ashanti฀ Nic฀Cheeseman 147 voters฀were฀less฀enthusiastic฀about฀Akufo-Addo’s฀campaign฀than฀they฀had฀ been฀about฀Kufour’s฀in฀1996฀and฀2000.฀Moreover,฀given฀the฀tension฀between฀the฀Kufour฀and฀Akufo-Addo฀blocs,฀the฀NPP฀struggled฀to฀persuade฀ Ashanti฀voters฀that฀they฀would฀be฀looked฀after฀by฀an฀Akufo-Addo฀administration.18฀ Conversely,฀ although฀ both฀ the฀ Ashanti฀ and฀ the฀ Akyem฀ have฀ historically฀formed฀part฀of฀the฀NPP’s฀Akan฀voting฀bloc,฀many฀Akyem฀felt฀ that฀their฀support฀had฀been฀disregarded฀during฀Kufour’s฀tenure฀as฀party฀ leader.฀ As฀ a฀ result,฀ turnout฀ in฀ NPP฀ strongholds฀ was฀ far฀ lower฀ than฀ expected,฀leaving฀the฀party฀vulnerable.฀In฀the฀end,฀the฀NPP฀missed฀out฀on฀ an฀ absolute฀ first-round฀ majority฀ in฀ the฀ presidential฀ election฀ by฀ only฀ 0.9฀ percent,฀and฀ultimately฀lost฀the฀runoff฀by฀just฀40,586฀votes.฀Many฀NPP฀ leaders฀believe฀that฀the฀party’s฀complacent฀attitude฀toward฀its฀own฀supporters฀contributed฀to฀its฀downfall.19 To Stay or Go? Transfers฀ of฀ power฀ require฀ meaningful฀ elections฀ and฀ the฀ willingness฀ of฀the฀ruling฀party฀to฀concede฀defeat.฀The฀impact฀of฀nonincumbency฀on฀ these฀ two฀ processes฀ is฀ not฀ straightforward.฀ Open-seat฀ elections฀ create฀ anomalous฀ power฀ dynamics฀ because฀ the฀ new฀ presidential฀ candidate฀ is฀ nominally฀ in฀ charge฀ of฀ the฀ party฀ and฀ the฀ election฀ campaign,฀ while฀ the฀ sitting฀president฀retains฀control฀of฀the฀state’s฀coercive฀capacity฀until฀the฀ election฀results฀have฀been฀announced.฀Stealing฀an฀election฀in฀most฀cases฀ therefore฀requires฀the฀support฀of฀the฀outgoing฀leader.฀This฀bifurcated฀authority฀structure฀can฀play฀out฀in฀dramatically฀different฀ways.฀Where฀a฀sitting฀president฀handpicks฀his฀successor฀or฀the฀death฀of฀an฀incumbent฀means฀ that฀the฀presidential฀candidate฀becomes฀acting฀president,฀state฀power฀may฀ be฀ deployed฀ to฀ ensure฀ a฀ ruling-party฀ victory฀ in฀ highly฀ competitive฀ contests.฀Events฀in฀Togo฀following฀the฀death฀of฀longtime฀dictator฀Gnassingbé฀ Eyadema฀in฀2005฀followed฀this฀pattern,฀as฀members฀of฀the฀ancien฀régime฀ cracked฀down฀on฀the฀opposition฀in฀an฀attempt฀to฀ensure฀the฀succession฀of฀ the฀dictator’s฀son,฀Faure฀Essozimma฀Gnassingbé.฀ When฀a฀sitting฀president฀is฀succeeded฀by฀a฀candidate฀to฀whom฀he฀has฀ little฀attachment,฀however,฀he฀may฀be฀less฀willing฀to฀undermine฀the฀rules฀ of฀the฀game฀and฀more฀inclined฀to฀consider฀his฀own฀legacy.฀Should฀this฀be฀ the฀case,฀the฀opposition฀candidate฀is฀able฀to฀compete฀on฀a฀more฀level฀playing฀field,฀as฀in฀Ghana฀in฀2000฀and฀2008,฀Kenya฀in฀2002,฀and฀Sierra฀Leone฀ in฀2007.฀These฀processes฀are฀not฀easily฀measurable,฀as฀no฀comprehensive฀ data฀set฀of฀electoral฀fairness฀exists,฀and฀the฀two฀divergent฀outcomes฀produced฀by฀this฀bifurcated฀authority฀structure฀are฀likely฀to฀cancel฀each฀other฀ out฀in฀any฀“large฀N”฀analysis.฀ However,฀looking฀at฀a฀series฀of฀elections฀within฀a฀given฀country฀suggests฀that฀nonincumbent฀elections฀may฀be฀significantly฀more฀open.฀Kenya฀ is฀ a฀ case฀ in฀ point.฀ Desperate฀ to฀ remain฀ in฀ power,฀ Moi฀ prepared฀ for฀ the฀ elections฀of฀1992฀and฀1997฀by฀inspiring฀ethnic฀clashes,฀distorting฀the฀reg- 148 Journal฀of฀Democracy Table฀3—elecTion-year฀freedoM฀house฀scores฀ in฀Kenya฀and฀ghana1 Founding฀Election (incumbent) 2nd฀Election฀ (incumbent) 3rd฀Election฀ (nonincumbent) PR CL Total PR CL Total PR CL Total Kenya 4 5 9 6 6 13 4 4 8 Ghana 5 5 10 3 4 7 2 3 5 1 PR฀=฀political฀rights,฀CL฀=฀civil฀liberties.฀Freedom฀House฀scores฀range฀from฀1฀to฀7,฀with฀ lower฀scores฀signifying฀more฀freedom. istration฀ process,฀ and—just฀ to฀ be฀ sure—stuffing฀ ballot฀ boxes฀ on฀ polling฀ day.฀Clearly,฀the฀prospect฀of฀ceding฀power฀sat฀ill฀with฀Moi.฀In฀contrast฀to฀ this฀dismal฀picture,฀the฀2002฀open-seat฀election฀stands฀out฀as฀remarkably฀ free฀and฀fair.฀This฀is฀reflected฀in฀Table฀3฀above,฀which฀shows฀the฀Freedom฀ House฀scores฀for฀the฀year฀of฀the฀first฀three฀multiparty฀elections฀in฀Kenya฀ (1992,฀1997,฀and฀2002)฀and฀Ghana฀(1992,฀1996,฀and฀2000)฀in฀order฀to฀give฀ an฀indication฀of฀the฀political฀climate฀within฀which฀the฀polls฀were฀held.฀Not฀ only฀were฀the฀2002฀elections฀Kenya’s฀only฀polls฀to฀date฀with฀no฀civil฀unrest,฀but฀the฀counting฀process฀was฀widely฀seen฀to฀be฀accurate.฀Having฀been฀ forced฀to฀respect฀term฀limits,฀Moi฀proved฀to฀be฀far฀less฀invested฀in฀securing฀ a฀win฀for฀Kenyatta,฀even฀though฀Moi฀had฀hoped฀to฀wield฀power฀behind฀the฀ scenes฀under฀a฀Kenyatta฀presidency.฀ Ghanaian฀elections฀have฀followed฀a฀similar฀path.฀The฀opposition฀boycotted฀ the฀ 1992฀ polls,฀ citing฀ irregularities.฀ Although฀ the฀ 1996฀ election฀ represented฀ a฀ considerable฀ improvement฀ (in฀ part฀ due฀ to฀ the฀ building฀ of฀ interparty฀trust฀through฀the฀Inter-Party฀Advisory฀Committee฀established฀ by฀the฀electoral฀commission฀in฀1994),฀the฀political฀environment฀was฀far฀ from฀free฀and฀open,฀with฀few฀commentators฀confident฀that฀Rawlings฀was฀ prepared฀to฀lose.20฀By฀contrast,฀the฀nonincumbent฀election฀of฀2000฀was฀the฀ most฀transparent฀that฀Ghana฀had฀seen,฀as฀indicated฀by฀the฀significant฀improvement฀in฀Freedom฀House฀scores.฀Of฀course,฀the฀transfer฀of฀power฀also฀ owed฀much฀to฀Ghana’s฀gradual฀democratization฀under฀Rawlings฀and฀the฀ emergence฀of฀one฀of฀Africa’s฀most฀independent฀and฀professional฀electoral฀ commissions,฀but฀it฀was฀also฀significant฀that฀Rawlings’s฀personal฀power฀ and฀pride฀were฀not฀at฀stake.฀ The฀Ghanaian฀elections฀of฀2008฀offer฀an฀opportunity฀to฀trace฀the฀decision-making฀process฀of฀a฀defeated฀ruling฀party฀operating฀within฀a฀bifurcated฀power฀structure.฀When฀an฀NDC฀victory฀became฀inevitable,฀AkufoAddo฀came฀under฀enormous฀pressure฀from฀party฀activists฀and฀a฀significant฀ number฀ of฀ NPP฀ leaders฀ to฀ reject฀ the฀ results.21฀ Party฀ hard-liners฀ felt฀ that฀ the฀ reported฀ NDC฀ intimidation฀ of฀ NPP฀ agents฀ in฀ some฀ constituencies฀ justified฀ defying฀ the฀ electoral฀ commission฀ if฀ the฀ margin฀ of฀ victory฀ was฀ slim.฀A฀heated฀meeting฀of฀Akufo-Addo’s฀campaign฀team฀that฀was฀called฀ to฀decide฀the฀party’s฀response฀to฀its฀defeat฀was฀reportedly฀split฀down฀the฀ middle,฀ despite฀ the฀ self-professed฀ commitment฀ of฀ the฀ NPP฀ to฀ constitutionalism.฀Significantly,฀even฀though฀the฀election฀team฀ultimately฀agreed฀ to฀concede฀defeat,฀the฀decision฀was฀taken฀out฀of฀their฀hands฀by฀the฀sitting฀ Nic฀Cheeseman 149 president,฀ John฀ Kufour.฀ While฀ the฀ election฀ team฀ was฀ meeting฀ to฀ thrash฀ out฀its฀response,฀Kufour฀went฀on฀Ghanaian฀radio฀and,฀without฀consulting฀ those฀in฀the฀meeting,฀announced฀that฀the฀NPP฀would฀respect฀the฀decision฀ of฀the฀electoral฀commission.฀This฀declaration฀effectively฀ended฀the฀election;฀without฀Kufour’s฀backing,฀Akufo-Addo฀lacked฀the฀control฀over฀state฀ security฀forces฀that฀he฀would฀have฀needed฀in฀order฀to฀hold฀on฀to฀power฀ in฀the฀face฀of฀inevitable฀NDC฀protests.฀The฀president’s฀rivals฀within฀the฀ NPP฀see฀this฀as฀further฀evidence฀that฀Kufour,฀who฀had฀done฀little฀during฀ the฀campaign฀itself,฀was฀all฀too฀happy฀to฀frustrate฀Akufo-Addo’s฀personal฀ ambitions.22฀ The฀transfer฀of฀power฀in฀Sierra฀Leone฀was฀underpinned฀by฀a฀similar฀dynamic.฀ President฀ Kabbah฀ initially฀ promoted฀ Berewa’s฀ candidacy,฀ but฀ by฀ the฀time฀of฀the฀2007฀campaign฀relations฀between฀the฀two฀had฀soured.฀The฀ tension฀between฀the฀outgoing฀president฀and฀the฀new฀party฀standard-bearer฀ enabled฀the฀newly฀reconstituted฀National฀Electoral฀Commission฀(NEC)฀to฀ maintain฀its฀independence฀from฀the฀ruling฀SLPP฀during฀an฀extremely฀close฀ and฀ controversial฀ contest.฀ The฀ vote฀ count฀ for฀ the฀ presidential฀ runoff฀ revealed฀that฀in฀a฀number฀of฀polling฀stations฀more฀people฀had฀voted฀than฀were฀ registered.฀The฀NEC฀decided฀to฀invalidate฀all฀votes฀from฀these฀polling฀stations,฀thereby฀rejecting฀both฀fraudulent฀and฀legitimate฀ballots฀in฀these฀constituencies.฀Because฀the฀polling฀stations฀in฀question฀were฀largely฀in฀SLPP฀ strongholds,฀the฀ruling฀party฀planned฀to฀serve฀the฀head฀of฀the฀NEC฀with฀an฀ injunction฀to฀prevent฀her฀from฀declaring฀the฀election฀result,฀thus฀giving฀the฀ regime฀an฀opportunity฀to฀reassert฀control฀over฀the฀electoral฀process.฀ According฀ to฀ Berewa,฀ however,฀ Kabbah’s฀ stance฀ toward฀ his฀ successor฀ and฀ lack฀ of฀ support฀ during฀ the฀ election฀ completely฀ undermined฀ this฀ strategy.฀ Having฀ denied฀ his฀ party฀ the฀ usual฀ advantages฀ of฀ incumbency,฀ Kabbah’s฀refusal฀to฀intervene฀on฀Berewa’s฀behalf฀helped฀to฀weaken฀the฀ SLPP’s฀ momentum฀ and฀ to฀ embolden฀ the฀ NEC฀ to฀ declare฀ an฀ opposition฀ victory.฀At฀the฀SLPP฀party฀conference฀in฀2009,฀Berewa฀blamed฀the฀former฀ president฀for฀the฀party’s฀defeat,฀declaring฀that฀the฀“thing฀that฀Kabbah฀did฀ that฀hurt฀me฀a฀lot฀was฀when฀he฀broadcast฀that฀he฀was฀neutral.฀How฀can฀you฀ be฀neutral฀against฀your฀own฀party?”23 Of฀course,฀such฀rivalries฀are฀not฀inevitable.฀The฀sitting฀president฀and฀ the฀ party’s฀ candidate฀ may฀ pull฀ in฀ the฀ same฀ direction,฀ resulting฀ in฀ a฀ further฀compression฀of฀democratic฀space,฀as฀occurred฀in฀Nigeria฀in฀2007.฀In฀ that฀ case,฀ departing฀ president฀ Olusegun฀ Obasanjo฀ and฀ candidate฀ Umaru฀ Yar’Adua฀presided฀over฀the฀most฀farcical฀of฀polls,฀which฀were฀controlled฀ by฀the฀government฀from฀start฀to฀finish.฀Although฀nonincumbent฀elections฀ do฀not฀always฀produce฀a฀more฀level฀playing฀field,฀in฀cases฀where฀the฀relationship฀between฀the฀sitting฀president฀and฀the฀presidential฀candidate฀is฀not฀ intimate,฀the฀division฀of฀authority฀may฀hamper฀electoral฀manipulation฀and฀ make฀the฀retention฀of฀power฀through฀unconstitutional฀means฀less฀likely. The฀ combined฀ impact฀ of฀ succession฀ struggles,฀ of฀ the฀ difficulty฀ that฀ nonincumbent฀ candidates฀ may฀ have฀ in฀ mobilizing฀ patronage฀ networks,฀ 150 Journal฀of฀Democracy and฀ of฀ the฀ greater฀ willingness฀ of฀ retiring฀ presidents฀ to฀ oversee฀ free฀ and฀ fair฀elections฀means฀that฀open-seat฀polls฀represent฀real฀windows฀of฀opportunity฀for฀opposition฀parties฀in฀Africa.฀This฀finding฀is฀consistent฀with฀the฀ work฀of฀Marc฀Morjé฀Howard฀and฀Philip฀G.฀Roessler,฀whose฀analysis฀of฀a฀ global฀data฀set฀on฀competitive฀authoritarian฀regimes฀suggests฀that฀nonincumbent฀elections฀are฀more฀likely฀to฀have฀liberalizing฀outcomes,฀though฀ this฀is฀not฀inevitable.24฀ Of฀course,฀open-seat฀elections฀can฀only฀be฀part฀of฀the฀explanation฀for฀ any฀transfer฀of฀power.฀In฀Kenya฀in฀2002,฀Moi’s฀defeat฀was฀underpinned฀ by฀continued฀economic฀decline฀and฀a฀learning฀process฀among฀the฀opposition.฀In฀Ghana฀in฀2008,฀the฀global฀economic฀downturn,฀accusations฀of฀corruption,฀and฀a฀sense฀that฀it฀was฀time฀for฀a฀change฀contributed฀to฀the฀NPP’s฀ defeat.฀Furthermore,฀open-seat฀elections฀may฀not฀lead฀to฀turnover฀at฀all฀if฀ other฀conditions฀are฀unfavorable—as฀in฀South฀Africa฀in฀2009,฀where฀the฀ ruling฀ANC฀retained฀power฀despite฀a฀succession฀battle฀that฀exacerbated฀ intraparty฀rifts.฀Thus฀more฀work฀is฀needed฀to฀understand฀the฀conditions฀ under฀which฀opposition฀coordination฀and฀nonincumbency฀create฀the฀necessary฀conditions฀for฀political฀change. The Impact of Term Limits Another฀topic฀deserving฀of฀greater฀scholarly฀attention฀is฀whether฀term฀ limits,฀ by฀ increasing฀ the฀ likelihood฀ of฀ alternations฀ in฀ power,฀ promote฀ democratic฀consolidation.฀There฀are฀three฀good฀reasons฀for฀thinking฀that฀ turnover฀ can฀ play฀ such฀ a฀ positive฀ role฀ in฀ the฀ democratization฀ process.฀ First,฀ transfers฀ of฀ power฀ in฀ nonincumbent฀ elections฀ have฀ helped฀ to฀ remove฀entrenched,฀corrupt,฀and฀authoritarian฀parties฀from฀power,฀creating฀ opportunities฀for฀further฀political฀liberalization.฀The฀defeat฀of฀KANU฀in฀ Kenya฀ is฀ perhaps฀ the฀ most฀ striking฀ illustration฀ of฀ this.฀ Until฀ the฀ nonincumbent฀ election฀ of฀ 2002,฀ the฀ reintroduction฀ of฀ multipartism฀ in฀ Kenya฀ had฀not฀resulted฀in฀an฀improvement฀in฀governance฀or฀civil฀liberties.฀It฀was฀ not฀the฀introduction฀of฀multiparty฀elections฀but฀the฀removal฀of฀Moi฀from฀ power฀that฀opened฀up฀political฀space฀in฀Kenya.฀This฀is฀not฀an฀isolated฀example.฀Comparing฀the฀twenty฀countries฀in฀which฀opposition฀parties฀were฀ able฀to฀capture฀the฀presidency฀in฀nonfounding฀elections฀between฀1992฀and฀ 2006,฀Maltz฀finds฀that฀turnovers฀were฀typically฀followed฀by฀a฀significant฀ improvement฀in฀the฀quality฀of฀political฀rights฀and฀civil฀liberties.25฀ Second,฀turnover฀is฀important฀because฀it฀is฀perhaps฀the฀most฀powerful฀ sign฀ that฀ key฀ actors฀ have฀ a฀ genuine฀ commitment฀ to฀ democratic฀ values.฀ As฀ Huntington฀ has฀ argued,฀ we฀ can฀ only฀ say฀ that฀ democracy฀ is฀ consolidated฀when฀both฀the฀previous฀authoritarian฀ruling฀party฀and฀the฀party฀that฀ defeats฀it฀have฀demonstrated฀that฀they฀will฀willingly฀release฀the฀reins฀of฀ power฀at฀the฀appropriate฀time.26฀But฀an฀alternation฀in฀leadership฀does฀more฀ than฀simply฀prove฀a฀commitment฀to฀democratic฀values—it฀can฀also฀foster฀ them.฀The฀experience฀of฀witnessing฀a฀ruling฀party฀gracefully฀accept฀defeat฀ Nic฀Cheeseman 151 builds฀confidence฀among฀a฀range฀of฀actors฀that฀political฀leaders฀intend฀to฀ follow฀the฀rules฀of฀the฀game,฀and฀thus฀moves฀countries฀closer฀to฀a฀point฀ where฀democracy฀becomes฀the฀only฀game฀in฀town.27 Democratization฀is฀an฀iterative฀process฀composed฀of฀a฀number฀of฀tiny฀ steps.฀ It฀ is฀ advanced฀ when฀ democratic฀ rules฀ are฀ followed฀ and฀ set฀ back฀ when฀they฀are฀broken.฀In฀one฀of฀the฀only฀attempts฀to฀identify฀the฀microprocesses฀that฀may฀connect฀turnover฀and฀democratic฀consolidation฀in฀Africa,฀ Devra฀Moehler฀and฀Staffan฀I.฀Lindberg฀have฀recently฀shown฀that฀turnover฀ serves฀to฀reduce฀the฀gap฀in฀institutional฀trust฀between฀winners฀and฀losers.฀ Although฀this฀positive฀effect฀of฀turnovers฀erodes฀over฀time,฀Moehler฀and฀ Lindberg฀ conclude฀ that,฀ in฀ the฀ short฀ run,฀ alternations฀ in฀ power฀ help฀ “to฀ establish฀a฀self-reinforcing฀equilibrium฀by฀providing฀incentives฀for฀elites฀ on฀both฀sides฀to฀play฀by฀the฀democratic฀rules฀of฀the฀game.”28฀ Similarly,฀although฀Afrobarometer฀survey฀data฀have฀shown฀that฀the฀first฀ two฀decades฀of฀multipartism฀in฀Africa฀saw฀a฀gradual฀downward฀trend฀in฀ popular฀support฀for฀democracy,฀Michael฀Bratton฀finds฀that฀this฀was฀offset฀ in฀countries฀that฀had฀experienced฀an฀alternation฀in฀power,฀such฀as฀Ghana฀ and฀Mali.29฀In฀other฀words,฀transfers฀of฀power฀injected฀multiparty฀regimes฀ with฀a฀much฀needed฀dose฀of฀legitimacy.฀This฀is฀a฀significant฀finding฀because฀the฀stronger฀the฀popular฀support฀for฀democracy,฀the฀more฀costly฀it฀ is฀ for฀ leaders฀ to฀ abuse฀ democratic฀ institutions฀ and฀ indulge฀ authoritarian฀ tendencies;฀thus฀initial฀democratic฀gains฀are฀less฀likely฀to฀be฀eroded.฀Taken฀ together,฀all฀this฀suggests฀that—in฀addition฀to฀the฀obvious฀gain฀of฀simply฀ removing฀authoritarian฀leaders—nonincumbent฀elections฀can฀play฀an฀important฀role฀in฀the฀evolution฀of฀a฀national฀consensus฀in฀favor฀of฀democracy.฀ Clearly,฀there฀are฀good฀reasons฀to฀believe฀that฀the฀more฀turnover฀we฀see,฀ the฀more฀confidently฀we฀can฀speak฀of฀democratic฀consolidation. Yet฀it฀is฀also฀clear฀that฀the฀relationship฀between฀nonincumbency฀and฀democratization฀is฀complex฀and฀that฀the฀extent฀and฀lasting฀significance฀of฀the฀ democratic฀gains฀from฀turnover฀are฀questionable.฀Most฀obviously,฀there฀is฀ no฀guarantee฀that฀the฀opposition,฀when฀in฀power,฀will฀rule฀in฀a฀more฀democratic฀and฀transparent฀manner฀than฀their฀predecessors.฀In฀Benin,฀the฀victory฀ of฀Prime฀Minister฀Nicéphore฀Soglo฀in฀the฀1991฀presidential฀election฀led฀to฀ a฀steady฀process฀of฀democratic฀consolidation฀that฀has฀seen฀two฀turnovers,฀ but฀the฀process฀has฀been฀far฀from฀straightforward.฀Having฀lost฀the฀founding฀ election,฀authoritarian฀leader฀Mathieu฀Kérékou฀made฀one฀of฀Africa’s฀most฀ remarkable฀political฀comebacks฀to฀regain฀power฀at฀the฀1996฀polls.฀Turnover฀ may฀have฀aided฀the฀consolidation฀process฀in฀Benin฀then,฀but฀not฀because฀it฀ permanently฀removed฀the฀leader฀of฀the฀ancien฀régime฀from฀power.฀ Other฀countries฀have฀had฀far฀more฀bleak฀experiences฀with฀turnover.฀In฀ Zambia,฀it฀was฀opposition฀leader฀and฀trade-union฀hero฀Frederick฀Chiluba฀ who,฀having฀defeated฀Kenneth฀Kaunda฀in฀the฀founding฀election฀of฀1991,฀ established฀ the฀ most฀ corrupt฀ and฀ venal฀ regime฀ in฀ the฀ country’s฀ history.฀ Kenya฀has฀suffered฀a฀similarly฀mixed฀fate฀following฀the฀victory฀of฀opposition฀forces.฀Most฀obviously,฀the฀flawed฀polls฀of฀2007฀demonstrated฀that฀ 152 Journal฀of฀Democracy Kibaki฀was฀no฀more฀willing฀to฀give฀up฀power฀than฀Moi฀had฀been฀in฀1992฀ and฀1997.฀In฀part฀because฀of฀these฀events,฀Kenyans,฀who฀had฀been฀among฀ the฀world’s฀most฀optimistic฀people฀in฀2002,฀just฀years฀later฀had฀become฀ some฀of฀the฀most฀pessimistic.฀Yet฀despite฀the฀calamitous฀fallout฀from฀the฀ “Kenya฀crisis,”฀in฀which฀more฀than฀a฀thousand฀people฀died฀and฀more฀than฀ a฀hundred-thousand฀were฀displaced,฀Kenya฀remains฀a฀more฀open฀country฀ now฀than฀it฀was฀during฀the฀final฀fifteen฀years฀under฀Moi.฀ The฀ cases฀ of฀ Benin,฀ Kenya,฀ and฀ Zambia฀ suggest฀ that฀ the฀ impact฀ of฀ turnover฀ can฀ be฀ fully฀ understood฀ only฀ in฀ retrospect,฀ as฀ the฀ gains฀ of฀ an฀ isolated฀electoral฀turnover฀may฀prove฀to฀be฀illusory.฀Nonetheless,฀the฀potential฀for฀term฀limits฀to฀generate฀nonincumbent฀elections฀that฀contribute฀ to฀processes฀of฀democratic฀consolidation฀is฀clear.฀There฀are฀good฀reasons,฀ therefore,฀for฀international฀actors฀to฀push฀for฀term฀limits฀to฀be฀included฀ in฀constitutions฀and฀to฀insist฀that฀they฀be฀respected.฀At฀a฀time฀when฀international฀confidence฀in,฀and฀commitment฀to,฀African฀democracy฀appears฀ to฀be฀on฀the฀wane,฀this฀is฀good฀news:฀Term฀limits฀are฀remarkably฀easy฀to฀ monitor฀compared฀to฀other฀forms฀of฀political฀and฀economic฀conditionality,฀and฀they฀are฀a฀goal฀around฀which฀a฀wide฀range฀of฀international฀and฀ domestic฀actors฀are฀keen฀to฀coalesce,฀as฀was฀witnessed฀in฀anti–third-term฀ campaigns฀in฀Malawi,฀Nigeria,฀and฀Zambia.฀Supporting฀these฀movements฀ will฀not฀guarantee฀democratic฀change,฀but฀it฀will฀improve฀its฀prospects. NOTES 1.฀Samuel฀Huntington,฀The฀Third฀Wave:฀Democratization฀in฀the฀Late฀Twentieth฀Century฀(Norman:฀University฀of฀Oklahoma฀Press,฀1991). 2.฀ Lindsay฀ Whitfield,฀ “‘Change฀ for฀ a฀ Better฀ Ghana’:฀ Party฀ Competition,฀ Institutionalization฀ and฀ Alternation฀ in฀ Ghana’s฀ 2008฀ Elections,”฀ African฀ Affairs฀ 108฀ (August฀ 2009):฀ 621–41. 3.฀ Gideon฀ Maltz,฀ “The฀ Case฀ for฀ Presidential฀ Term฀ Limits,”฀ Journal฀ of฀ Democracy฀ 18฀ (January฀2007):฀134. 4.฀Thomas฀M.฀Holbrook,฀“Incumbency,฀National฀Conditions,฀and฀the฀2008฀Presidential฀ Election,”฀PS:฀Political฀Science฀&฀Politics฀(October฀2008):฀709–12. 5.฀Term฀limits฀are฀not฀employed฀in฀Côte฀d’Ivoire,฀Equatorial฀Guinea,฀Gambia,฀GuineaBissau,฀Mauritania,฀Sudan,฀and฀Zimbabwe.฀ 6.฀Daniel฀N.฀Posner฀and฀Daniel฀J.฀Young,฀“The฀Institutionalization฀of฀Political฀Power฀in฀ Africa,”฀Journal฀of฀Democracy฀18฀(July฀2007):฀126–40. 7.฀Posner฀and฀Young,฀“Institutionalization฀of฀Political฀Power฀in฀Africa,”฀132.฀ 8.฀Maltz,฀“The฀Case฀for฀Presidential฀Term฀Limits.” 9.฀Rodger฀Govea฀and฀John฀Holm,฀“Crisis,฀Violence฀and฀Political฀Succession฀in฀Africa,”฀ Third฀World฀Quarterly฀18฀(March฀1998):฀129–48. 10.฀Marty฀Cohen฀et฀al.,฀The฀Party฀Decides:฀Presidential฀Nominations฀Before฀and฀After฀ Reform฀(London:฀University฀of฀Chicago฀Press,฀2008). Nic฀Cheeseman 153 11.฀Magnus฀Ohman,฀“The฀2007฀Presidential฀and฀Parliamentary฀Elections฀in฀Sierra฀Leone,”฀Electoral฀Studies฀(December฀2008):฀764–68. 12.฀Author’s฀interview฀with฀Konadu฀Apraku,฀NPP฀election-campaign฀manager,฀Accra,฀ Ghana,฀8฀July฀2009. 13.฀Author’s฀interview฀with฀Nana฀Yaw฀Ofori-Kuragu,฀independent฀NPP-affiliated฀MP,฀ Accra,฀Ghana,฀11฀July฀2009. 14.฀Author’s฀interview฀with฀Ambrose฀Dery,฀NPP฀MP฀and฀deputy฀minority฀leader฀in฀Parliament,฀Accra,฀Ghana,฀7฀July฀2009. 15.฀Simeon฀Nichter,฀“Vote฀Buying฀or฀Turnout฀Buying?฀Machine฀Politics฀and฀the฀Secret฀ Ballot,”฀American฀Political฀Science฀Review฀102฀(February฀2008):฀19–31;฀Jeremy฀Horowitz,฀ “Ethnic฀Groups฀and฀Campaign฀Strategy฀in฀Kenya’s฀2007฀Election,”฀Working฀Group฀in฀African฀Political฀Economy฀Working฀Paper฀17,฀December฀2009. 16.฀Pedro฀C.฀Vicente฀and฀Leonard฀Wantchekon,฀“Clientelism฀and฀Vote฀Buying:฀Lessons฀ from฀Field฀Experiments฀in฀African฀Elections,”฀Oxford฀Review฀of฀Economic฀Policy฀25฀(Summer฀2009):฀292–305. 17.฀ David฀ Throup,฀ “The฀ Kenya฀ General฀ Election:฀ December฀ 27,฀ 2002,”฀ CSIS:฀ Africa฀ Notes฀14฀(January฀2003). 18.฀Author’s฀interview฀with฀Ben฀Ephson,฀political฀editor฀of฀the฀Daily฀Dispatch,฀Accra,฀ Ghana,฀10฀July฀2009. 19.฀Arthur฀Kennedy,฀Chasing฀the฀Elephant฀into฀the฀Bush:฀The฀Politics฀of฀Complacency฀ (Bloomington,฀Ind.:฀AuthorHouse,฀2009). 20.฀Richard฀ Jeffries,฀ “The฀ Ghanaian฀ Elections฀ of฀ 1996:฀ Towards฀ the฀ Consolidation฀ of฀ Democracy?”฀African฀Affairs฀97฀(April฀1998):฀189–208. 21.฀Author฀interview฀with฀Elizabeth฀Ohene,฀former฀NPP฀minister฀of฀education,฀Accra,฀ Ghana฀(16฀July฀2009). 22.฀Author฀interview฀with฀Yaw฀Osafo-Maafo,฀former฀NPP฀minister฀of฀finance,฀Accra,฀ Ghana฀(18฀July฀2009). 23.฀Silas฀Gbandia,฀“Ex-President฀Kabba฀Weeps฀as฀Berewa฀Says฀He฀Is฀a฀Sell-Out,”฀Sierre฀Leone฀News,฀6฀March฀2009,฀www.thesierraleonenews.com/national/122-ex-presidentkabba-weeps-as-berewa-says-he-is-a-sell-out. 24.฀ Marc฀ Morjé฀ Howard฀ and฀ Philip฀ G.฀ Roessler,฀ “Liberalizing฀ Electoral฀ Outcomes฀ in฀ Competitive฀ Authoritarian฀ Regimes,”฀ American฀ Journal฀ of฀ Political฀ Science฀ 50฀ (April฀ 2006):฀365–81. 25.฀Maltz,฀“Case฀for฀Presidential฀Term฀Limits,”฀135. 26.฀Huntington,฀Third฀Wave. 27.฀Adam฀Przeworski,฀Democracy฀and฀the฀Market:฀Political฀and฀Economic฀Reforms฀in฀ Eastern฀Europe฀and฀Latin฀America฀(Cambridge:฀Cambridge฀University฀Press,฀1991). 28.฀Devra฀Moehler฀and฀Staffan฀I.฀Lindberg,฀“Narrowing฀the฀Legitimacy฀Gap:฀Turnovers฀ as฀a฀Cause฀of฀Democratic฀Consolidation,”฀Journal฀of฀Politics฀71฀(October฀2009):฀1463. 29.฀Michael฀Bratton,฀“The฀Alternation฀Effect฀in฀Africa,”฀Journal฀of฀Democracy฀15฀(October฀2004):฀147–58.