Political participation in Africa: The role of individual resources

ilable at ScienceDirect

Electoral Studies xxx (2013) 1–17

Contents lists ava

Electoral Studies

journal homepage: www.elsevier .com/locate/e lectstud

Political participation in Africa: The role of individualresources

Ann-Sofie Isaksson*

Department of Economics, University of Gothenburg, Vasagatan 1, Box 640, 405 30 Gothenburg, Sweden

a r t i c l e i n f o

Article history:Received 21 January 2013Received in revised form 1 July 2013Accepted 17 September 2013

JEL classification:D01D72O12O55

Keywords:Political participationResourcesAfricaAfrobarometer

* Tel.: þ46 (0) 31 7861249; fax: þ46 (0) 31 78613E-mail address: [email protected]

0261-3794/$ – see front matter � 2013 Elsevier Ltdhttp://dx.doi.org/10.1016/j.electstud.2013.09.008

Please cite this article in press as: IsakssoStudies (2013), http://dx.doi.org/10.1016/j.

a b s t r a c t

This paper aims to examine the role of individual resources in explaining African politicalparticipation. If political participation is costly and requires inputs in terms of individualresources, and citizens in young developing country democracies face comparatively highparticipation costs and have more limited individual resource endowments than citizens inmore established democracies, a resource approach to political participation should beparticularly relevant in the African setting. On the contrary, however, empirical findingsdrawing on recent data for more than 27,000 respondents in 20 emerging African de-mocracies suggest weak explanatory power of the resource perspective. Often, the rela-tively resource poor actually participate to a greater extent than the more resource rich.The results are encouraging in that they suggest fairly broad-based political participation,but also call attention to the need to evaluate the motivational forces behind the decisionto take part.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Widespread political participation, defined as citizenacts to influence the selection of and/or the actions takenby political representatives, has an intrinsic democraticvalue. In fact, it makes sense to argue that democracy re-quires political participation to be legitimate (Bratton et al.,2005). Striving for political equality in the sense that thepreferences of each citizen should count equally, one needsto assess who participates in the political process andwhose preferences are represented in politics.

It is widely agreed that the propensity to participatepolitically is not evenly distributed across citizens; studiesof Western democracies suggest that those who participateconstitute an unrepresentative set of citizens, dispropor-tionally coming from more advantaged groups in society

26.u.se.

. All rights reserved.

n, A.-S., Political particelectstud.2013.09.008

(Brady et al., 1995; Verba et al., 1995; Lijphart, 1997; Bartels,2005; Griffin and Newman, 2005). If policy preferences alsovary across socio-economic groups (see e.g. Verba and Nie,1972; Verba et al., 1978), and elected officials are moreresponsive to the preferences of those who participatepolitically than to those who do not (see e.g. Bartels, 2005;Boulding and Wampler, 2010; Gilens, 2005; Griffin andNewman, 2005), skewed participation risks translatinginto skewed government policy. This is troubling, since itsuggests that inequality of influence and resources is cu-mulative (Dahl, 1961); economic inequality may causeinequality in terms of political participation, which in turnmay imply that policies increasingly address the prefer-ences of more well-off citizens, thus adding to economicinequality (Bartels, 2005).

The resource perspective, pioneered by the U.S.-basedwork of Brady, Verba and Schlozman (Brady et al., 1995;Verba et al., 1995), stresses that political participation iscostly and requires inputs in terms of individual resourceslike skills and time. Focussing on how resource differences

ipation in Africa: The role of individual resources, Electoral

mailto:[email protected]

www.sciencedirect.com/science/journal/02613794

http://www.elsevier.com/locate/electstud

http://dx.doi.org/10.1016/j.electstud.2013.09.008

http://dx.doi.org/10.1016/j.electstud.2013.09.008

1 For an overview, see Feddersen (2004).

A.-S. Isaksson / Electoral Studies xxx (2013) 1–172

among people differentially constrain their ability to meetthe costs of participating, it can potentially help explain astratified pattern of political activity. While based on ob-servations from an established Western democracy, itshould arguably be especially relevant in a developingcountry setting, where citizens are likely to have a weakerresource base and where poorly developed infrastructureshould lead to high participation costs. However, judgingfrom existing empirical evidence on the patterns of politi-cal participation in developing countries, this is far fromclear (see Section 2).

The present paper explores political participation inAfrica. Drawing on recent data on over 27,000 respondentsin 20 emerging African democracies, the aim is to examinethe role of individual resources in explaining African po-litical participation. The empirical findings suggest that theresource perspective does a poor job at explaining politicalparticipation in Africa; it is not uncommon that we see therelatively resource poor participating to a larger extentthan the more resource rich. The results are encouraging inthat they suggest fairly broad-based political participation,but also call attention to the need to further evaluate themotivational forces behind the decision to take part.

Understanding the patterns of political participation inAfrica – where poverty is widespread and democratic in-stitutions are still emerging – seems particularly important.For poverty reduction, it should be central that the demo-cratic process represents the many and not the few. And, ifpolitical participation is required to legitimise democracy,then studying its determinants in the African context, wherethedemocratic states areyounger andmore fragile, shouldbecritical (Kuenzi and Lambright, 2007). However, consideringthat the African democracies are young, and that untilrecently there have not been any reliable and comparabledata on democratic attitudes and behaviour in Africa, it is notsurprising that the work on African political participation isrelatively scarce. While a number of studies – most notablythose of Michael Bratton and his colleagues (see Bratton,1999, 2008; Bratton et al., 2005, 2010) and that of Kuenziand Lambright (2010) – are relevant to consider for our pur-poses (see Section 2), evaluating the role of individual re-sources for political participation is not their primary focus.

The present study differs from existing work on Africanpolitical participation in several important ways. To myknowledge, it is the first study to focus exclusively on the roleof individual resources for explaining African political partic-ipation, taking care to control for individual and contextualfactors affecting the costs and benefits of participation aswellas to explore the sensitivity of the resource estimates to theuse of different measures and sub-samples. Moreover, andvery important in terms of external validity considering thatthe African democracies are young and evolving, compared toprevious work it uses more recent and more comprehensivedata. As such, the results should add significantly to our un-derstanding of the prerequisites for broad-based citizenengagement in the emerging African democracies.

2. Resources and participation

The resource perspective, stressing the role of individualresources for meeting the costs of participating, was

Please cite this article in press as: Isaksson, A.-S., Political particStudies (2013), http://dx.doi.org/10.1016/j.electstud.2013.09.008

developed by Brady, Verba and Schlozman in themid 1990s(Brady et al., 1995; Verba et al., 1995). Earlier studies ofpolitical participation linked socio-economic status toparticipation – finding the better educated and those withhigher incomes to be more likely to participate (Verba andNie, 1972; Wolfinger and Rosenstone, 1980). However, intheir influential work on American political participation,Brady, Verba and Schlozman developed this thinking, dis-cussing the causal mechanisms that link socio-economicstatus to participation. Their findings highlight the differ-ential resource requirements for different forms of partic-ipation, for instance indicating that in the U.S., resources interms of time, money and civic skills matter less for votingthan for other political acts.

Rational theories of participation, dating back to Downs(1957),1 assume that individuals evaluate the costs andbenefits of participating politically, and decide to partici-pate when the expected net benefit of doing so is positive.This assumption need not be restrictive. Interpreting thecosts and benefits of participation in a broad sense, thebenefits of political activity refer to the motivational forcesbehind the decision to take part, such as conflicting in-terests stimulating engagement (see e.g. Solt, 2008), a willto conform to participatory norms (see e.g. La Due Lake andHuckfeldt, 1998; Knack and Kropf, 1998; Gerber et al.,2008), or material incentives based in clientelist offersfrom rulers who exchange personal favours for politicalsupport (see e.g. Wantchekon, 2003; Christensen and Utas,2008; Lindberg and Morrison, 2008; Vicente, 2008). Thecosts of political participation refer to its demands in termsof e.g. time, money, knowledge and information.

The resource perspective focuses on the cost side of thisequation, evaluating how resource differences amongpeople differentially constrain their ability to meet thecosts of participating (Verba et al., 1995). If participation iscostly, the individual’s decision on whether or not to takepart is, just as the decision to consume any good, con-strained by a budget restriction determined by the in-dividual’s resource base (Solt, 2008). By considering theeffects of resources on political participation, one canassess the impact of relaxing the budget constraint relevantfor participation. Conditional on variation in the costs andbenefits of participating (see Section 3.2), having more ofthe relevant resources should imply that one is better ableto meet the costs of participation and thus participatemore.

The resource perspective seems particularly importantwhen studying political participation in developing coun-tries with young democratic systems. Compared to citizensin more established democracies, citizens in these coun-tries may face higher participation costs as a result ofpoorly developed infrastructure (e.g. political infrastruc-ture in terms of polling stations, community meeting hallsetc.; physical infrastructure enabling citizens to reach thenearest political infrastructure; and infrastructure for in-formation transmission), or theymay have a less developedindividual resource base. Both would result in the resourceconstraint relevant for political participation more often


2 Time and money are clearly related; with money you can hiresomeone to free up time, and using time for political participation has anopportunity cost in terms of foregone income. However, they are not thesame. In contrast to money, there is a fixed upper-bound on time,meaning that money can be more unevenly distributed (Verba et al.,1995).

3 For a discussion of the effects of asymmetric information on electoralparticipation, see Feddersen and Pesendorfer (1999).

4 For more information on Afrobarometer sampling procedures andsurvey methods see Bratton et al. (2005) and the Afrobarometer Network(2007).

5 Note, however, that the Afrobarometer is not meant to be generalisedto all of Sub-Saharan Africa. The selection of countries is intentionallybiased towards liberalising regimes, meaning that a uthoritarian regimesand countries in conflict are under-represented (Afrobarometer Network,2007).

A.-S. Isaksson / Electoral Studies xxx (2013) 1–17 3

being binding, meaning that the impact of resources onparticipation should be especially important.

As noted, the conventional finding – often based onstudies from the U.S. – is that citizens with low incomesand little education participate less than their richer andmore educated counterparts. Comparing across otherWestern democracies, however, the results are more mixed(Verba et al., 1978; Norris, 2002), and considering theempirical evidence available for developing countries thepicture becomes even less clear-cut. Studies of democraticattitudes and behaviours in Asia and Latin America suggestthat poor people participate politically no less than morewell-off citizens (Yadav, 2000; Krishna, 2002, 2008, onIndia; Booth and Seligson, 2008, on Latin America; andBratton et al., 2010, on East Asia). If anything, the empiricalfindings actually seem to indicate that poor citizensparticipate more. Education, on the other hand, is found tobe positively associated with participation in India(Krishna, 2002, 2008) and Latin America (Booth andSeligson, 2008), but not in East Asia where those withoutformal schooling turn out to be the ones most likely to vote(Bratton et al., 2010).

Turning to empirical evidence on participatory patternsin Africa, the findings of Bratton (1999), who study de-terminants of political participation in Zambia, and ofBratton et al. (2010) and Kuenzi and Lambright (2010), whoboth study determinants of electoral participation usingdata from 1999 to 2000 (Afrobarometer round 1) for 12 and10 African countries respectively, suggest no effect of eco-nomic standing and mixed effects of education. Consid-ering democratic attitudes and behaviours in a sample withrespondents from 15 African countries, Bratton (2008), onthe other hand, finds comparatively high participation ratesamong poorer citizens. As noted, however, while thesestudies routinely include some resource variables, evalu-ating the role of individual resources for political partici-pation is not their primary focus. Rather, these studiesstress the relevance of the availability of institutions thatlink citizen to state, or ‘agencies of mobilization’ such aspolitical parties (Bratton, 1999; Kuenzi and Lambright,2010). I return to this when discussing appropriate con-trol variables in Section 3.2.

To evaluate the role of individual resources forexplaining political participation in the young and evolvingAfrican democracies, it is important to have recent datawith wide coverage, to carefully control for contextual andindividual variation in the costs and benefits of participa-tion and to explore the robustness of the resource estimatesto alternative resource measures and sub-samples.

The present study focuses on resources in terms of time,money, human capital and information, all of which appearimportant for political participation in a developing coun-try context. Political participation will always involve in-vestments of time. With little time at hand, you will berestricted in terms of political activity, and arguablyparticularly so in a developing country with poorly devel-oped infrastructure. Compared to time, the connectionbetween money and participatory acts might appear lessdirect. However, in a developing country with widespreadpoverty, lack of moneymaywell restrict an individual from,say, travelling to the polling station or the community


meeting hall or from being able to devote time to politicalparticipation.2 Human capital, next, helps the individualunderstand the political process and build civic skills suchas communication and organisational abilities, and hencefacilitates political participation (Verba et al., 1995). In adeveloping country context, where illiteracy is sometimeswidespread, this issue should be particularly pressing.Illiterate citizens have trouble making sense of informationabout the political process and are constrained in terms ofcommunicating their views. Information, finally, is oftenput forth as an important cost of political participation (LaDue Lake and Huckfeldt, 1998). How do you vote? Forwhom do you vote? In what other ways, and for whatpurpose, should you participate politically?3 Processinginformation of this type requires resources in terms of timeand human capital. However, considering that we also needthe information to be available, it appears suitable toconsider information access as a resource in its own right.Again, this issue should be particularly pertinent in adeveloping country context where access to informationsources like TV, newspapers and the Internet cannot betaken for granted.

3. Data and empirical setup

To examine the role of individual resources in explainingAfrican political participation I employ new data from theAfrobarometer survey. The Afrobarometer is a comprehen-sivemulti-country surveyproject collectingdata onpoliticaland economic attitudes andbehaviour of African citizens. Assuch, it provides a unique opportunity to study mass polit-ical participation in a large African multi-country sample.The fourth and most recent wave of the survey, which isused here, was conducted in 2008–2009 and covers over27,000 respondents from 20 African countries – Benin,Botswana, Burkina Faso, Cape Verde, Ghana, Kenya, Lesotho,Liberia, Madagascar, Malawi, Mali, Mozambique, Namibia,Nigeria, Senegal, South Africa, Tanzania, Uganda, Zambiaand Zimbabwe. The survey covers a representative sampleof each country’s voting age population4 (with a standardsample size of 1200 observations per country, except inNigeria, South Africa and Uganda where sample sizes arearound twice this size) and asks a standard set of questionsin all countries, thus allowing for cross-national compari-sons.5 I estimate the following benchmark probit for thepolitical participation PPir of individual i in region r:



prob½PPir ¼ 1� ¼ FðRiaþ Gibþ Xidþ grÞ
That is, the probability that individual i in region r partic-ipates is taken to depend on vectors of resources Ri, socio-demographic group affiliations Gi, and additional individ-ual controls Xi, allowing for region fixed effects gr . Fð,Þdenotes the standard normal cumulative distributionfunction. Conditional on Gi, Xi and gr , which are meant tocapture individual and contextual variation in the costs andbenefits of participating (see Section 3.2), we expect thathaving more of the concerned resources should ease theresource constraint on participating, and thus enable moreparticipation.
While focus is on micro (individual) variation inparticipation, I take care to control for macro (national) andmeso (regional) variation in the same. Our 20 Africansample countries have in common that they are relativelyyoung democracies and that they are poor by internationalstandards. As discussed above, these conditions are rele-vant when assessing the resource perspective, since theymay imply that the resource constraints relevant for po-litical participation more often are binding. At the sametime, however, the countries considered are by no meanshomogeneous. Focussing on 20 countries I am unable toclosely examine individual country experiences (for a briefoverview of the post-independence democratic develop-ment of our sample countries, see Table A1; for in-depthaccounts of the same see e.g. Bratton and Van de Walle,1997; Lindberg, 2006a). However, considering that macrolevel determinants of participation – such as countries’historical experiences, institutional arrangements andeconomic and political conditions – are likely to affect notonly the average level of political participation but also theassociation between our focus micro level factors andparticipation, pooled sample estimations accounting forregion fixed effects will be complemented by individualcountry estimations, allowing us to consider country vari-ation in parameter estimates.

6 The other questions on inter-electoral participation focus on joiningwith others to raise an issue, taking part in demonstrations or protestmarches, and contacting local government officials.

7 Observations are weighted using within country weights adjustingthe sample to be nationally representative with respect to region, urban-rural distribution etc.

8 Out of the observations that do not have missing values on either‘voting’ or ‘meeting’, 54 percent engages in both activities, 23 percentonly votes, 13 percent only attends community meetings, and 11 percenttakes part in neither of the activities.

3.1. Dependent variable

The outcome variable of interest is political participa-tion. As noted in Section 1, one can think of politicalparticipation as citizen acts to influence the selection ofand/or the actions taken by political representatives. Assuch, it can take many forms. On top of voting, which is themost common, and in a sense, the most basic form of po-litical participation, citizens can work in election cam-paigns, engage in the local community, contact politicalleaders, attend demonstrations etc. Important for ourpurposes, political acts like these can vary in what indi-vidual resources they require. Moreover, they presumablyvary in what information they display, in the extent towhich they are mainstream or unconventional, in whetherthey are undertaken alone or in groups, and in the extent towhich they are unequally distributed across citizens (forfurther discussion see e.g. Verba et al., 1995; Lijphart, 1997).Acknowledging that political participation is a multidi-mensional concept that encompasses a wide and hetero-geneous set of activities, I cannot claim to capture it in full.What I can do, however, is to make sure to consider both


electoral and inter-electoral participation, i.e. voting as wellas political activity taking place between elections. Study-ing participation in the emerging African democracies,where important aspects of political activity take placeinformally (Bratton et al., 2005), this should be particularlyimportant.

Hence, I consider two alternative dependent variables:voting (electoral participation) and attending communitymeetings (inter-electoral participation). For voting, I createa dummy variable taking the value one if the respondentreports to have voted in the most recent [year 200X] na-tional election, and zero otherwise. Those who were tooyoung to vote at the time of the election are excluded fromthe estimation. The data contains information on severalforms of inter-electoral participation. However, consideringhow diverse these activities are – presumably varying on alldimensions described above – using a composite inter-electoral participation index would hide substantial het-erogeneity. Instead, I choose to focus on the most commonform of inter-electoral participation in the data, namelyattending community meetings.6 I create a dummy variabletaking the value one if the respondent reports to haveattended a community meeting during the past year, andzero otherwise (for variable descriptions and summarystatistics see Tables A2 and A3). In Section 4.2 I evaluate towhat extent the results can be generalised to other forms ofinter-electoral participation.

Looking at Figs. 1 and 2, we can note that there is a greatdeal of country variation in political participation.7 Theshare of respondents who report to have voted in the lastelection ranges from 64 percent in Zambia to 92 percent inBenin, and the share of respondents who report to haveattended a community meeting during the past year rangesfrom 32 percent in Cape Verde to 92 percent inMadagascar.In Botswana, Lesotho, Madagascar and Zimbabweattending community meetings is actually more commonthan voting, highlighting the importance of not focussingsolely on electoral participation when studying Africanpolitical participation. In the remaining countries, however,voting is the more common political act. Generallyspeaking, although the correlation coefficients are modest,being engaged in one of the activities means that you aremore likely to also engage in the other.8

With respect to the high share of respondents reportingto vote, a few notes are in order. Importantly, our self-reported voting shares are not strictly comparable to offi-cial country turnout figures, which tend to be lower (seeTable A4). First of all, the voting survey question simplyasks the respondent whether he/she voted in the ‘last [year200X] national election’. Hence, in the many cases where


Zam

bia

Nig

eria

Leso

tho

Zim

babw

e

Bots

wan

a

Mad

agas

car

Sout

h A

frica

Uga

nda

Burk

ina

Faso

Nam

ibia

Mal

i

Moz

ambi

que

Sene

gal

Keny

a

Mal

awi

Libe

ria

Cap

e Ve

rde

Tanz

ania

Gha

na

Beni

n

0.2

.4.6

.81

Fig. 1. Share of respondents reporting to have voted in the last national election.

Cap

e Ve

rde

Nig

eria

Nam

ibia

Sout

h A

frica

Gha

na

Zam

bia

Moz

ambi

que

Beni

n

Mal

i

Zim

babw

e

Sene

gal

Keny

a

Burk

ina

Faso

Uga

nda

Libe

ria

Bots

wan

a

Tanz

ania

Mal

awi

Leso

tho M

adag

asca

r

0.2

.4.6

.81

Fig. 2. Share of respondents reporting to have attended a community meeting during the past year.

9 Instead using a measure of political efficacy (the Afrobarometer e.g.checks whether the respondent knows the name of the local MP) wouldbe problematic in this context due to reverse causality from politicalparticipation to knowledge of the political process.


parliamentary and presidential elections are held concur-rently we do not know which of the two the respondentrefers to. Moreover, even if the two elections were not heldconcurrently, if the respondent only took part in one of thetwo elections, it seems likely that he/she would rememberand report the one election he/she in fact took part in,meaning that self-reported voting shares would be inflatedcompared to the official turnout rates. Second, differencescould arise due to sampling. Although the Afrobarometer ismeant to be nationally representative with respect to eachcountry’s voting age population, it is not unreasonable toassume that there might be some over-sampling of in-dividuals, say those with a steady address, who are alsomore likely to vote. Still, considering that casting a ballot isoften viewed as a civic duty, to some extent the discrepancybetween self-reported voting shares and official turnoutrates is most likely due to survey respondents over-reporting voting. The official and self-reported turnoutfigures are significantly correlated however (see Table A4),and provided that the degree of over-reporting does notvary systematically with individual resources, it should notbias our estimates. Nevertheless, in Section 4.2 I evaluatethe sensitivity of results to respondents over-reportingvoting.


3.2. Explanatory variables

The explanatory variables considered can be dividedinto resource indicators, regional controls, socio-demographic group affiliations and additional individualcontrols. The resource indicators capture individualresource endowments in terms of human capital, money,information and time. To measure human capital I usedummies indicating whether the respondent’s highestlevel of education is at primary, secondary or post-secondary level.9 To capture economic standing, I followBratton et al. (2005) and create a ‘lived poverty index’ basedon the responses to the question, ‘Over the past year, howoften, if ever, have you or anyone in your family gonewithout: (a) enough food to eat, (b) enough clean water forhome use, (c) medicines or medical treatment, (d) enoughfuel to cook your food?’, with response categories rangingfrom0 for ‘never’ to 4 for ‘always’ for each item. Similarly, to



proxy for resources in terms of information, I create anindex based on responses to the question, ‘How often doyou get news from the following sources: a) radio, b)television, and c) newspapers?’, with response categoriesranging from 0 for ‘never’ to 4 for ‘every day’. To proxy fortime availability, finally, I include a dummy variable indi-cating whether the respondent has full-time employment.While individuals in full-time employment tend to be moreresource rich in terms of money and human capital, theyarguably have less time on their hands. In Section 4.2 Ievaluate the sensitivity of results to using alternativeresource variables.

To what extent can the resource estimates be inter-preted causally? Whereas reverse causality from partici-pation to our resource variables should not be a majorconcern – childhood education precedes political involve-ment, and it seems a fair assumption that for the absolutemajority of adults, work- and family-related decisions areprior to political participation10 – we need to take accountof potential omitted variable bias. This involves consideringour theoretical priors and carefully controlling for con-founding factors.

The theoretical predictions are clear. Thinking of re-sources as means to meet the costs of participation, more isbetter – having more of the relevant resources should easethe resource constraint on participating, and thus enablemore participation (see the discussion in Section 2). How-ever, individual resource endowments are presumablycorrelated with variation in the costs and benefits ofparticipation, meaning that being interested in the role ofresources for meeting the costs of participation we need tocontrol for contextual and individual variation in the costsand benefits of participating.

With respect to contextual variation, studies of macro-level determinants of participation suggest that participa-tion costs and benefits vary across countries with factorslike democratic tradition, economic conditions, and politi-cal institutions (see e.g. Jackman, 1987; Lijphart, 1997;Posner and Simon, 2002; Kostadinova, 2003; Fornos et al.,2004; Lindberg, 2006b). Similarly, studies of meso-leveldeterminants of participation suggest that participationcosts and benefits could vary across regions within coun-tries depending on access to political and physical infra-structure, e.g. distance to the nearest polling station andthe quality of the road to get there, the salience of localpolicy concerns and community variation in participatorynorms (see e.g. La Due Lake and Huckfeldt, 1998; Knack andKropf, 1998; Krishna, 2002; Norris, 2002). For Africa inparticular, the results of Bratton (1999) and Kuenzi andLambright (2010) suggest that an important determinantof political participation is the availability of institutionsthat link citizen to state – a factor which will clearly varyacross regions. Likewise, the prevalence of clientelist ac-tivity – another often suggested determinant of African

10 Although we cannot rule out that someone can choose, say, a line ofwork as a result of political engagement (Verba et al., 1995) this ought tobe quite rare. Moreover, whereas you might seek information more oftenbefore an election if you plan to vote, the information variable focuses oninformation exposure on a more regular basis.


voting patterns and political engagement (see e.g.Wantchekon, 2003; Christensen and Utas, 2008; Lindbergand Morrison, 2008; Vicente, 2008) – is likely to vary notonly across countries, but also across regions withincountries. If the concerned individual resource endow-ments also vary systematically across regions, this couldbias our estimates. Being interested in isolating the effect ofindividual resources on political participation it is thus notenough to control for country level variation in the costsand benefits of participation – we need more detailedcontextual controls. 246 sub-national region dummiesincluded in all estimations should help pick up the influ-ence of contextual factors affecting the costs and benefits ofpolitical participation.

Second, we need to control for individual level factorsthat could affect the costs and benefits of participation andpotentially contaminate the resource estimates. I dividethese factors into socio-demographic group affiliations andadditional individual controls. To begin with, we know thatpolitical participation tends to vary across socio-demographic groups. In particular, earlier studies basedon smaller African samples suggest a gender-gap inparticipation (Bratton, 1999; Bratton and Logan, 2006;Bratton et al., 2010), greater turnout among older citizens(Bratton et al., 2005, 2010; Kuenzi and Lambright, 2010)and among citizens living in rural areas (Bratton, 1999;Kuenzi and Lambright, 2010; Bratton et al., 2010). Whileit seems plausible that individual resources, such as skillsand time, are differentially available to these socio-demographic groups, we cannot assume that they are theonly factors generating the observed socio-demographicvariation in participation. If there are indeed other factorsat play (say, differences in participatory norms), not ac-counting for the socio-demographic group affiliations willbias the resource estimates. Hence, I control for gender,urban/rural residence and age.

The additional individual controls refer to individualattributes that are arguably closely related to both politicalparticipation and resource endowments. In particular,there is reason to believe that people with differentresource endowments also vary in terms of civic engage-ment, clientelist motivations, networks, and policy prefer-ences – factors that may also affect participation.

With respect to civic engagement, the information var-iable will presumably not only capture information avail-ability; it is also likely to pick up a tendency to seek outinformation, meaning that both participation and infor-mation exposure could be influenced by omitted variablesrelated to civic engagement. To control for civic-mindedness, I include a control for political interest.

Turning to clientelism, it is not unreasonable to assumethat the poor, due to having great needs, may be moresusceptible to clientelist appeals. If clientelist offers stim-ulate participation (Christensen and Utas, 2008; Vicente,2008), this would counteract the hypothesised tendencyof the poor participating less due to not being able to meetthe costs of participation. As noted, the sub-national regiondummies should help pick up regional variation in theprevalence of clientelism. However, to proxy for individualvariation in the influence of the same, I include a variableon the respondent’s attitudes towards clientelist activity



(assuming that people who are more favourable to cli-entelism also are more likely to accept/seek clientelistoffers).

Regarding network effects, a person’s education andemployment status will influence what people he/shecomes in contact with, and certain socio-economic groupsmay be more inclined to discuss politics and may holdstronger norms of democratic participation. Consider thecase of education. It should help the individual develop thehuman capital needed to meet the costs of participationand to build politically relevant social capital (La Due Lakeand Huckfeldt,1998). Being interested in isolating the effectof the former, one would have to control for the latter. Toproxy for networks, or politically relevant social capital, Iinclude a variable indicating whether the respondent dis-cusses politics with friends.

With respect to policy preferences, it is not unreason-able to assume that resource endowments affect whatpolicy issues lie close at heart, and that policy preferencescould motivate political participation. In particular, youreconomic standing will not only determine whether youcan, say, afford to take the bus to the polling station, it willalso help define your pecuniary interest in distributionalconflict – potentially an important motivation behindparticipation (see the discussion in Solt, 2008). To controlfor distributional policy preferences, I use a question askingthe respondent to rate how the government deals withnarrowing the gap between rich and poor.

Importantly, due to endogeneity concerns the additionalindividual controls are not interpreted causally,11 but areincluded in separate estimations as proxies for omittedfactors that could otherwise bias our resource estimates.

4. Results

This section evaluates the role of individual resourceendowments in explaining African political participation.After considering the results of the benchmark estimationswe move on to assess to what extent the findings aresensitive to alternative samples and specifications.

4.1. Resources and participation – main findings

Regressing our political participation measures – votingand meeting – on the resource variables and regionalcontrols (Table 1, Regressions 1 and 4), the results are un-expected in light of our theoretical priors. Whereas povertyis significantly related to political participation, the asso-ciation is in the unexpected direction if viewing money as aresource constraining participation – the poorer you are,the more likely you are to participate. On average, a onestandard deviation higher poverty index score implies an

11 Not only are these factors likely to affect participation, it is alsoreasonable to assume that participating politically stimulates politicalinterest, helps build politically relevant social capital, makes a personmore exposed to clientelist appeals, as well as possibly contributes tostronger views on certain policy issues. Also, political interest and tosome extent politically relevant social capital are very proximate to ouroutcome measure political participation, and thus presumably driven by asimilar set of explanatory factors.


approximately 1 percentage point higher probability tovote, and a 2 percentage point higher probability to attendcommunity meetings. Similarly, the indicator included tocapture restricted time availability – if the respondent isemployed full time – is positively related to both voting andattending community meetings. Compared to people whoare not in full-time employment – and conditional on ed-ucation, information and economic standing – citizensworking full-time are around 4 percentage points morelikely to vote and 5 percentage points more likely to attendcommunity meetings. Viewing time as a resource relevantfor political participation, and believing that people in full-time employment are comparatively restricted in terms ofthe time they have to spend on political activity, this issurprising. Likewise, education is not significantly relatedto attending community meetings, and actually negativelyrelated to voting; compared to people with no schooling, aperson with secondary or post-secondary education is 6–7percentage points less likely to vote. Believing that humancapital is required for citizens to understand the electionprocess –who the candidates are, what they stand for etc. –and that it facilitates active participation in communitymeetings, this is surprising.

Only information is significantly related to participationin the expected direction. On average, a one standard de-viation higher score in the information index implies aroughly 1 percentage point higher probability to vote and a3 percentage point higher probability to attend communitymeetings. Hence, at this stage, only information comesacross as a resource potentially relevant for meeting thecosts of political participation.12

These results should not necessarily be taken at facevalue, however; to evaluate the explanatory power of re-sources as a means of meeting the costs of participating weneed to control for systematic variation in the costs andbenefits of participating. As discussed in Section 3.2, thesub-national region dummies included in all estimationsshould help control for contextual factors that could createregional variation in the costs and benefits of politicalparticipation. However, there is still the possibility that ourresource variables pick up omitted individual level factorsaffecting the decision to participate politically.

We know that political participation tends to vary acrosssocio-demographic groups. A quick look at the groupmeans in Table 2 reveals that in line with previous studiesbased on smaller African samples, in our 20 sample coun-tries women tend to be less politically active than men,rural citizens participate to a greater extent than theirurban counterparts, and older people participatemore thanyounger individuals. Moreover, the groups vary in terms ofresource endowments; with the exception of our proxy fortime availability, women, older citizens and people living inrural areas tend to be more resource poor than theirrespective comparison groups. Given our priors that theconcerned resources are relevant for meeting the costs ofparticipating politically, one would thus expect that these

12 Excluding the information index from the estimations does notchange the unexpected poverty or education effects (the results areavailable upon request).


Table 1Political participation in Africa: the role of resources (probit marginal effects).

Dependent variable (1) Voting (2) Voting (3) Voting (4) Meeting (5) Meeting (6) Meeting

ResourcesPoverty 0.009**

(0.004)0.004(0.004)

0.004(0.004)

0.024***(0.005)

0.017***(0.005)

0.017***(0.005)

Full-time 0.036***(0.008)

0.014(0.009)

0.013(0.008)

0.049***(0.011)

0.006(0.012)

0.004(0.012)

Education_Primary �0.014(0.012)

0.013(0.012)

0.007(0.012)

0.007(0.011)

0.030***(0.012)

0.021*(0.012)

Education_Secondary �0.059***(0.016)

0.012(0.014)

0.001(0.014)

�0.026(0.016)

0.050***(0.016)

0.036**(0.016)

Education_Post-sec. �0.070***(0.020)

�0.001(0.017)

�0.021(0.017)

0.014(0.019)

0.068***(0.018)

0.043**(0.018)

Information 0.012***(0.005)

0.016***(0.005)

0.008*(0.005)

0.026***(0.007)

0.033***(0.007)

0.020***(0.007)

Socio-demographic groupsRural 0.045***

(0.010)0.042***(0.010)

0.108***(0.011)

0.104***(0.010)

Female �0.022***(0.007)

�0.013*(0.007)

�0.075***(0.008)

�0.060***(0.008)

Age 0.020***(0.001)

0.019***(0.001)

0.021***(0.001)

0.021***(0.001)

Age squared �0.0002***(0.0000)

�0.0002***(0.0000)

�0.0002***(0.0000)

�0.0002***(0.0000)

Add. individual controls No No Yes No No YesRegion dummies Yes Yes Yes Yes Yes Yes

Observations 23,070 23,070 23,070 25,893 25,893 25,893

Notes: Standard errors (clustered by the 246 regions) in parentheses; *significant at 10%, **significant at 5%, ***significant at 1%. Observations are weightedusing combined within � across country weights. The within country weights adjust the samples to be nationally representative with respect to gender,region, urban-rural distribution etc. The across country weights adjust all country samples to the same size (N ¼ 1200). For a description of the additionalindividual controls see Table A2.


groups participate comparatively little. For women this istrue. For older people and citizens living in rural areas, onthe other hand, we see the opposite – i.e. relatively highparticipation rates. Hence, while accounting for resourcedifferentials could presumably help explain the lowerparticipation among women compared to men (seeIsaksson et al., 2013), the key to explaining the relativelyhigh participation rates among older people and citizensliving in rural areas seemingly lies outside the resourceperspective. Introducing the socio-demographic groupaffiliation variables into the regressions (Table 1, Re-gressions 2 and 5), the participatory inequalities remain.Apparently, factors other than resource differentialscontribute to the observed socio-demographic variation inparticipation.

Returning to the resource estimates, looking at Re-gressions 2 and 5 it appears that some of the unexpectedresults from the previous estimations were driven byrelatively high participation rates among comparativelyresource poor socio-demographic groups (older citizensand citizens living in rural areas). Education, poverty andworking full-time are no longer significantly related – inthe unexpected direction – to voting. Similarly, workingfull-time is no longer significantly related – again in theunexpected direction – to attending community meetings.On the other hand, we can still observe the expected pos-itive association between information exposure and polit-ical participation, and in addition we now actually see theexpected positive association between education and


community meeting attendance. Compared to people withno schooling, a person with primary school education is 3percentage points more likely to attend community meet-ings. For individuals with secondary or post-secondaryeducation the difference is about twice that (the differ-ence in magnitude being statistically significant). Yet again,the result that community meeting attendance is higheramong the poor remains. And in the cases where the sta-tistically significant relationships in the unexpected direc-tion are no longer present, there are still no signs of theexpected associations that would suggest that the con-cerned resources were relevant for meeting the costs ofparticipation.

To sum up the results so far, it seems the resourceperspective does a poor job at explaining African politicalparticipation. If a resource is relevant for meeting the costsof participating, more of that resource should mean moreparticipation. If anything, however, the estimations suggestthat having little time (i.e. working full-time) and littlemoney (i.e. being poorer) is associated with more partici-pation. The latter result is in line with the findings ofBratton (2008) on Africa, and with some studies of politicalparticipation in Asia and Latin America (Yadav, 2000; Boothand Seligson, 2008; Bratton et al., 2010). Hence, rather thanconstraining participation, it seems working full-time andbeing poor is related to motivational factors that stimulateparticipation. Education and information, on the otherhand, come out as potentially relevant for meeting thecosts of participation. However, education seems to matter


Table 2Group means in terms of political participation and resource endowments.

Rural Urban Male Female Age < 30 Age 30–49 Age > 49 Full sample

Voting 0.792 0.736 0.792 0.752 0.656 0.816 0.852 0.772Meeting 0.721 0.547 0.703 0.610 0.556 0.709 0.751 0.657Educationa 0.345 0.643 0.502 0.407 0.599 0.423 0.225 0.454Information �0.346 0.535 0.107 �0.154 0.082 �0.024 �0.239 �0.023Poverty 0.082 �0.153 �0.020 0.012 �0.101 0.033 0.115 �0.004Full-time 0.141 0.234 0.214 0.135 0.136 0.229 0.138 0.175

Notes: Observations are weighted using combined within � across weights (see the description in Table 1).a Refers to having some secondary school or more education. The group differences in means are statistically significant at the 1% level, except for male vs.

female poverty where the difference is significant at the 5% level (for the age categories, the significance test is based on an F-test of all parameters beingequal to zero).


only for taking part in community meetings, and whereasinformation appears to matter for both voting andattending community meetings it has relatively modesteffects.

Are the resource estimates driven by omitted factorsrelated to the individual resource base, as opposed to whatwe are trying to measure, i.e. the importance of therespective resources for meeting the costs of participating?When (in line with the discussion in Section 3.2) includingcontrols for politically relevant social capital, political in-terest, clientelism and distributional policy preferences13

(Regressions 3 and 6), the resource estimates remain qual-itatively the same. Time andmoney still donot comeout as afactors constraining political participation; poverty is stilluncorrelatedwith voting and positively related to attendingcommunity meetings, and there are still no signs of anegative association between political participation andbeing full-time employed. Hence, the positive relation be-tween poverty and community meeting attendance re-mains stable in the face of controls for distributional policypreferences and attitudes towards clientelism. And simi-larly, controlling for people in full-time employment havingaccess to more politically relevant social capital or beingmore civic-minded – factors which could counteract thesupposednegative effect of having little time –working full-time still does not stand out as a factor constraining politicalparticipation. The positive effects of education (onattending community meetings) and information (on bothvoting and attending community meetings) remain, butdrop in size and arenowsomewhat less precisely estimated.Hence, accounting for higher levels of politically relevantsocial capital among the well-educated and a tendency ofcivic minded individuals to seek information, resources interms of human capital and information still seem to bearsome relevance for meeting the costs of participation.

The fact that the results remain qualitatively the same inthe face of controls so closely related to political partici-pation (e.g. political interest), seems to indicate that theyare quite stable. Still, however, our explanatory frameworksuggests that if a resource is relevant for meeting the costsof participation, more of that resource should mean moreparticipation. Finding the opposite – as in the case of theobserved greater community meeting attendance among

13 Due to the endogeneity concerns discussed in Section 3.2, I viewthese indicators merely as proxies for omitted variables and do notinterpret their estimates.


the poor – indicates that some other factor, not consideredhere, is at play. What we can say, is that when controllingfor potentially confounding factors, the results still suggestthat the concerned resources are not important enough todominate these unknown factors.

4.2. Further testing

The results so far indicate that the resource perspectivehas weak explanatory power in our African sample. Thissection explores to what extent these findings are robustacross different sub-samples and to using alternativeresource and participation measures. The results, brieflysummarized below, are presented in the Appendix.

First of all, and as discussed in Section 3, macro leveldeterminants of participation – e.g. countries’ historicalexperiences and institutional arrangements – are likely toaffect not only the average level of political participation ina country, but also the association between our focus microlevel factors and participation. To evaluate the extent ofcountry variation in the resource estimates, I run separateregressions (equivalent to Regressions 2 and 5 in Table 1)for the 20 country sub-samples (see Table A5). Just like thepooled sample estimations, the individual country re-gressions offer next to none support for the view that timeand money are resources relevant for meeting the costs ofparticipating. If anything, they tend to suggest that the poorand those with little time on their hands are more likely toparticipate politically. Furthermore, one can note that theassociation between education and information on the onehand and political participation on the other is far fromuniform across countries.

Next, the results do not appear contingent on the choiceof resource indicators (see Table A6). If instead of using thepoverty index, which imposes linearity on the associationbetween economic standing and participation, we usedummy variables distinguishing between the non-poor, thesomewhat poor, and the very poor, and dummy variablescapturing self-reported living conditions (‘very bad’, ‘fairlybad’, ‘neither good nor bad’, ‘fairly good’ and ‘very good’),the basic conclusion that the poor, if anything, participatemore than the better off remains the same. Similarly, ifinstead of using the information index (which imposeslinearity) we use a dummy variable indicating if therespondent owns a radio, or dummy variables indicatinghow often the respondent gets news from a radio, the basicconclusion remains the same. Finally, if using an alternativetime indicator (available only in the second round of the



Afrobarometer, results are available upon request), whichunlike our benchmarkmeasure captures time spent on self-employment and household work, time still does not standout as restricting participation.14

Turning to the dependent variables, the comparativelygreat number of missing observations on the voting indi-cator (see Table A3) is due to excluding (the 2916) re-spondents whowere not of voting age at the time of the lastnational election. Could the difference in effective samplebe what generates variation in resource estimates betweenthe voting and community meeting estimations? As it ap-pears, no; restricting the sample to observations that havenomissing values on either indicator (Table A7, Regressions1–2) the results remain the same.

Next, we should evaluate the sensitivity of results torespondents possibly over-reporting voting. First of all,coding those who claim not to remember whether theyvoted as non-voters rather than missing observations(Table A7, Regression 3), seeing that this response couldserve as an escape from having to admit that you did notvote, does not change the results. Similarly, restricting thesample to include only respondents from the five countrieswith the smallest discrepancy between self-reportedvoting share and official turnout (Table A7, Regression 4),and to only include observations where the interviewerjudges the respondent as honest15 (Table A7, Regression 5),the results stand.

Another concern would be if people’s voting behaviour(or tendency to over-report voting) is affected by restrictedcivil liberties or democratic practices in their country ofresidence (see e.g. Lindberg, 2004; Collier and Vicente,2009). Reasonably, an individual could have plenty of re-sources in terms of time, money, information and humancapital, but still abstain from voting due to voter intimi-dation or as a result of perceiving the election as unfair. Tocheck if this is why we find that the resource perspectivehas relatively weak explanatory power, in two alternativevoting regressions (Table A7, Regressions 6–7) I restrict thesample to include only countries judged as ‘free’ byFreedom House, and countries with Polity IV democracyscores higher than five (see Table A1). The basic resultsstand.

The second dependent variable – community meetingattendance – is meant to shed light on political participa-tion taking place between elections. To check if the findingsare relevant for different forms of inter-electoral politicalparticipation, and not just for attending communitymeetings, I construct a composite variable based on thefirst principal component of four binary indicators

14 Moreover, the non-significant resource parameters are not the resultof collinearity among our resource variables. Inspecting the varianceinflation factors in the benchmark setup suggests that we do not have aproblem of multi-collinearity.15 Based on the question: ‘What was the respondent’s attitude towardsyou during the interview? Was he/she: honest, in between, ormisleading?’with 79 percent of the respondents being judged as ‘honest’,19 as ‘in between’ and 2 as ‘misleading’. Being a subjective judgement onpart of the interviewer we cannot assume that this assessment is true andfair. Nevertheless, the question is useful as a rough check of datareliability.


revealing if during the past year the respondent has 1)attended a community meeting, 2) joined others to raise anissue, 3) taken part in a demonstration or protest march,and 4) contacted a local government representative. Usingthis indicator as dependent variable (Table A7, Regression8) the results remain qualitatively the same. As it seems,the findings obtained when focussing on communitymeeting attendance could be relevant for other forms ofinter-electoral participation as well.

5. Conclusions

This study explored the role of individual resources inexplaining African political participation. If citizens indeveloping countries with young democratic systems facecomparatively high participation costs and have a lessdeveloped individual resource base compared to citizens inmore established democracies, this should mean that theresource constraints relevant for political participationmore often are binding. Against this background, ourstarting point was that the influential resource perspective,which stresses that political participation is costly and re-quires inputs in terms of individual resources, should beparticularly relevant in the African setting. The empiricalfindings indicate otherwise, however.

Empirical analysis of a recent and comprehensive datamaterial, covering political and economic attitudes andbehaviour of over 27,000 respondents across 20 Africancountries, suggest weak explanatory power of the resourceperspective. The estimations offer no support for the viewthat time andmoney are resources relevant for meeting thecosts of participating. If anything, they suggest that poorercitizens and people with little time on their hands are morelikely to participate. And while education and informationseem to bear some relevance for meeting the costs ofparticipation, education matters only for attending com-munity meetings, the information effects are relativelymodest, and the individual country estimations revealconsiderable country heterogeneity in their estimates.

The main results are robust over a wide range of alter-native specifications. They remain intact to regional con-trols included to account for contextual variation in thecosts and benefits of political participation, and to indi-vidual controls included as proxies for omitted variablesrelated to the person’s resource base as well as to the de-cision to take part. Moreover, they withstand the use ofalternative resource indicators, using an alternative mea-sure for inter-electoral participation, and restricting thesample to only include respondents from countries with asmall discrepancy between self-reported and officialturnout, to respondents judged as honest, and to re-spondents from countries with relatively well-functioningdemocracies. Breaking down the pooled sample into theindividual country sub-samples adds to the picture that theresource perspective has weak explanatory power in theemerging African democracies.

In line with findings from other developing countriessuggesting that poor people value democracy and partici-pate politically no less than more well-off citizens, the re-sults of the present study seem to indicate broad-basedpolitical participation, which should be a prerequisite for a



well-functioning democracy. Finding the opposite – i.e.strong participatory inequalities originating in unequalresource endowments – would have suggested a scenariowhere higher participation among the resource rich mightreinforce existing economic inequalities.

At the same time, however, if the relatively resourcepoor participate to a greater extent than the compara-tively resource rich, this not only constitutes a partici-patory inequality in itself but also raises the questionwhy. Focussing on the relevance of resources for meetingthe costs of participating, the present paper exploresfactors enabling participation rather than the motivationsbehind the choice to participate. Finding comparativelyhigh participation rates among the relatively resourcepoor – who should supposedly be particularly con-strained in their decision of whether to take part – in-dicates that some other factor, not considered here, is atplay. If high participation among resource poor citizens isa sign of the often suggested importance of personalisedrelationships and clientelist appeals in African politics,this would not come across as good news. Neither woulda situation where the poor participate politically simplybecause their opportunity cost of doing so is very low, nora scenario where the resource rich do not participate to

Appendix

Table A1Post-independence democratic development in sample countrie

Country Coloniser(main)

Post-independence democratic developmenta

Benin France 1960–91 Military rule, one party rule, and restrictpractices. 1991 – Democracy

Botswana UK 1966 – DemocracyBurkina Faso France 1960–78 Military rule, one party rule, and restrict

practices. 1978–80 Democracy. 1980–2002 Militardemocratic practice. 2002 – Emerging democracy.

Cape Verde Portugal 1975–90 One party rule. 1991 – DemocracyGhana UK 1957–92 Periods of democracy, military rule, one

and restricted democratic practices. 1992 – DemocKenya UK 1963–2002 Emerging democracy, one party rule, r

democratic practices. 2002 – DemocracyLesotho UK 1966–70 Democracy, 1970–93 Military rule and re

democratic practices, 1993–2002 Democracy/Emedemocracy. 2002 – democracy

Liberia US 1847–1984 Emerging democracy, one party, milita1984–97 Restricted democratic practice/transition1997–2001 Emerging democracy. 2001–06 Restricpractice/transitional governments. 2006 – Democr

Madagascar France 1960–89 Periods of military rule, one party rule, ademocratic practices. 1989–93 Multiparty transitio

Malawi UK 1964–93 One party rule, 1994 – democracyMali France 1960–91 Military and one party rule. 1992 – DemMozambique Portugal 1975–90 One party rule, 1990–94 Multiparty tran

1994 – DemocracyNamibia S. Africa 1990 – DemocracyNigeria UK 1960–99 Democracy, military rule, restricted dem

1999 – Dem./Emerging democracySenegal France 1960–2000 Periods of emerging democracy, one p

democratic practices. 2000 – Democracy.South Africa UK 1910–94 Restricted Democratic Practice (white ru

Transition period; 1994 – Democracy


the same extent because they are able to influence out-comes via alternative – corrupt – means. To be able toevaluate and tackle systematic participatory inequalities,we need to understand the basis of existing disparities.While the provision of information and education mightstimulate political engagement and help citizens makemore informed choices, the results of the present paperseem to suggest that the key to understanding variationin African political participation lies outside the resourceperspective. We need further knowledge about the na-ture of, and the motivations behind, political participa-tion in Africa.

Acknowledgements

I am thankful to Arne Bigsten, Michael Bratton, NiklasHarring, Göran Holmqvist, Andreas Kotsadam, StaffanLindberg, Halvor Mehlum, Måns Nerman, Bo Rothstein,Måns Söderbom, Jo Thori Lind and seminar participants atthe 2010 NCDE conference in Helsinki and at the Devel-opment Workshop at the University of Gothenburg forvaluable suggestions (any remaining mistakes are of coursemy own). I gratefully acknowledge financial support fromSida.

s.

Firstmulti-partyelectionb

PolityIV scorec

Freedomhouseratingd

ed democratic 1991 7 Free (2)

1965 8 Free (2)ed democraticy regime/restricted

1978 2 Partly free (4.5)

1991 n.a. Free (1)party rule,racy/Emerging Democracy

1956 8 Free (2)

estricted 1992 7 Partly free (3)

strictedrging

1965 8 Free (2.5)

ry rule.al governments.ted democraticacy.

2005 7 Partly free (4.5)

nd restrictedn, 1993 – Democracy

1989 7 Partly free (3)

1994 6 Partly free (4)ocracy 1992 7 Free (2)sition, 1994 6 Partly free (3.5)

1989 6 Partly free (2.5)ocratic practices. 1979 4 Partly free (4)

arty rule, and restricted 1978 8 Free (2.5)

le), 1990–94 1994 9 Free (1.5)

(continued on next page)


Table A1 (continued )

Country Coloniser(main)

Post-independence democratic developmenta Firstmulti-partyelectionb

PolityIV scorec

Freedomhouseratingd

Tanzania UK 1964–92 One Party rule, 1992–1995 Multiparty transition,1995 – Emerging Democracy

2000 2 Partly Free (3.5)

Uganda UK 1962–66 Democracy, 1966–96 Periods of military rule, one partyrule, and restricted democratic practices, 1996 – Restricteddemocratic practice

1962 1 Partly Free (4.5)

Zambia UK 1964–90 Emerging Democracy/one party rule, 1991–2006Democ./Emerging democ., 2006 – Democ.

1991 5 Partly free (4)

Zimbabwe UK 1980–87 Emerging Democracy, 1987 – Restricted Democratic Practice 1979 1 Not free (6.5)

a From the African Elections Database (2010).b First post-independence multi-party parliamentary election judged as ‘free’ or ‘partly free’ by the International Institute for Democracy and Electoral

Assistance (2010) (or first equivalent election held in a period in which the country is judged as a democracy by the African Elections Database).c Polity IV (Polity IV Project, 2010) democracy score for 2005 (for 2006 in Liberia due to democratic transition in 2005), 0–10 with higher values meaning

better democracy (see Marshall and Jaggers, 2002).d Freedom house combined political rights and civil liberties rating from 2005, 1–7 with 1–2.5 judged as ‘free’, 3–5 as ‘partly free’, and 5.5–7 as ‘not free’

(see Freedom House, 2010).

Table A2Variable descriptions.

Dependent variablesVoting: Dummy variable equal to one if the respondent reports to have voted in the ‘most recent [20XX] national

elections’; zero otherwise. ‘Don’t know’/’Can’t remember’ responses, as well as those who were too young tovote at the time of the election (including those turning 18 during the year of the election), are coded as missing values.

Meeting: Dummy variable equal to one if the respondent reports to have attended a community meeting duringthe past year; zero otherwise (‘don’t know’/‘can’t remember’ responses coded as missing values).

Resource variablesEducation (based on question of what is the respondent’s highest level of education):No-school: Dummy variable equal to one if the respondent has no formal schooling; zero otherwise (used as reference category in estimation).Education_Primary: Dummy variable equal to one if the respondent’s highest level of education is at primary school level (including those withincomplete primary); zero otherwise. Education_Secondary: Dummy variable equal to one if the respondent’s highestlevel of education is at secondary school level (including those with incomplete secondary); zero otherwise.Education_Post-secondary: Dummy variable equal to one if the respondent’s highest level of education is atpost-secondary school level (including those with incomplete post-secondary); zero otherwise.

Full-time: Dummy variable equal to one if the respondent has full-time paid employment; zero otherwise(if no employment or part-time employment).

Information: An index with mean zero and standard deviation one, higher values meaning that the person hasgreater access to information. Constructed as the first principal component of the responses to, ‘How oftendo you get news from the following sources: (a) radio, (b) television and (c) newspapers?’ with response categoriesranging from 0 for ‘never’ to 4 for ‘every day’.

Poverty: A poverty index with mean zero and standard deviation one, higher values meaning that you are poorer. Constructed as the first principalcomponent of the answers to, ‘Over the past year, how often, if ever, have you or anyone in your family gone without:(a) enough food to eat, (b) enough clean water for home use, (c) medicines or medical treatment, (d) enough fuel tocook your food?’, with response categories ranging from 0 for ’never’ to 4 for ’always’ for each item.

Socio-demographic group affiliationsFemale: Dummy variable equal to one if the respondent is female; zero otherwise.Rural: Dummy variable equal to one if the respondent lives in a rural area; zero otherwise.Age variables: Age in years and age squared.

Additional individual controlsPolitically relevant social capital: Dummy variable equal to one if the respondent reports to occasionally or frequently discuss politics

with friends/family; zero if reporting to never do so.Political interest: Dummy variable equal to one if the respondent claims to be somewhat or very interested in public affairs;

zero if not at all or not very interested.Clientelism: Dummy variable equal to one if in the choice between statement (a) ‘since leaders represent everyone,

they should not favour their own family or group’, and (b) ‘once in office, leaders are obliged to help their homecommunity’, the respondents agrees/strongly agrees with statement (b). The dummy variable takes the value zeroif instead the respondent agrees/strongly agrees with (a), agrees with neither statement or chooses the ‘don’t know’ response category.

Distributional policy preferences: Three dummies based on the question ‘how well or badly would you say the currentgovernment is handling narrowing gaps between rich and poor?’. Bad inc. gap: Dummy equal to one if the respondentthinks the government handles narrowing gap between rich and poor very badly or fairly badly. Good inc. gap: Dummyequal to one if the respondents think the government handles narrowing gap between rich and poor very well or fairly well.Undecided: Dummy equal to one if the respondent is undecided with respect to the above question (used as reference category in estimation).

Regional controlsRegion dummies: 246 sub-national regions referring to the first-order administrative division in a country, in the survey manual

denoted ‘region/province’


Please cite this article in press as: Isaksson, A.-S., Political participation in Africa: The role of individual resources, ElectoralStudies (2013), http://dx.doi.org/10.1016/j.electstud.2013.09.008

Table A3Summary statistics.

Variable Observations Mean Std. dev. Min Max

Political participationVoting 24,621 0.77 0.42 0 1Meeting 27,521 0.64 0.48 0 1

ResourcesPoverty 27,417 0.00 1.00 �1.99 4.24Full-time 27,613 0.18 0.39 0 1No-school 27,669 0.16 0.36 0 1Education_Primary 27,669 0.37 0.48 0 1Education_Secondary 27,669 0.37 0.48 0 1Education_Post-secondary 27,669 0.11 0.31 0 1Information 27,562 0.00 1.00 �1.57 1.79

Socio-demographic group affiliationsRural 27,713 0.62 0.49 0 1Female 27,713 0.50 0.50 0 1Age 27,379 36.33 14.50 18 100Age squared 27,379 1529.91 1282.53 324 10,000

Additional individual controlsPol. relevant social capital 27,492 0.69 0.46 0 1Political interest 27,446 0.63 0.48 0 1Clientelism 27,696 0.36 0.48 0 1Bad inc. gap 27,689 0.73 0.45 0 1Good inc. gap 27,689 0.21 0.41 0 1Undecided inc. gap 27,689 0.06 0.23 0 1

Table A4Official turnout versus self-reported voting.

Country Official turnout (% of voting age population)in last national election prior to the surveya

% share of respondents reporting tohave voted in last national electionb

Benin 2007 Parliamentary: 62 92Botswana 2004 Parliamentary: 44 67Burkina Faso 2007 Parliamentary: 40 75Cape Verde 2006 Presidential: 79; 2006 Parliamentary: 80 85Ghana 2004 Presidential: 80; 2004 Parliamentary: 80 90Kenya 2007 Presidential: 55; 2007 Parliamentary: 55 83Lesotho 2007 Parliamentary: 39 65Liberia 2005 Presidential: 59; 2005 Parliamentary: 71 82Madagascar 2007 Parliamentary: n.a. 69Malawi 2004 Presidential: 58; 2004 Parliamentary: 55 82Mali 2007 Presidential: 48; 2007 Parliamentary: 39 79Mozambique 2004 Presidential: 36; 2004 Parliamentary: 36 80Namibia 2004 Presidential: 81; 2004 Parliamentary: 80 79Nigeria 2007 Presidential: n.a.; 2007 Parliamentary: n.a. 65Senegal 2007 Presidential: 55; 2007 Parliamentary: 28 80South Africa 2004 Parliamentary: 57 74Tanzania 2005 Presidential: 68; 2005 Parliamentary: 65 90Uganda 2006 Presidential: 61; 2006 Parliamentary: 60 73Zambia 2006 Presidential: 56; 2006 Parliamentary: 56 64Zimbabwe 2008 Presidential: 47; 2008 Parliamentary: 45 65

a Source of official turnout figures: International Institute for Democracy and Electoral Assistance (2010).b Refers to those of voting age at the year of the election. The correlation between official and self-reported turnout (18 observations) is 0.57 and sta-

tistically significant (where presidential and parliamentary elections are held concurrently, the higher official turnout rate of the two is used).



Table A5Estimations by country (probit marginal effects).

Benin Botsw. Burk.F. CapeV. Ghana Kenya Lesotho Liberia Madag. Malawi Mali Mozam. Namibia Nigeria Senegal S.Afr. Tanzania Uganda Zambia Zimbab.

Panel A: Dependent variable is votingResourcesPoverty 0.01

(0.01)�0.03(0.02)

0.01(0.02)

0.02*(0.01)

�0.01(0.01)

0.04**(0.02)

0.01(0.02)

�0.00(0.01)

�0.05**(0.02)

�0.01(0.02)

0.01(0.01)

0.02(0.02)

�0.03**(0.01)

0.00(0.01)

0.02(0.01)

�0.01(0.01)

�0.01(0.01)

�0.01(0.01)

0.03(0.02)

0.03*(0.02)

Full-time �0.03(0.04)

0.08**(0.04)

0.01(0.05)

0.01(0.03)

0.02(0.02)

0.02(0.03)

0.00(0.06)

�0.02(0.04)

0.01(0.04)

�0.00(0.05)

�0.06(0.07)

0.07*(0.04)

0.04(0.03)

0.02(0.03)

0.10***(0.03)

�0.01(0.03)

�0.00(0.03)

0.07**(0.03)

�0.12**(0.06)

0.00(0.05)

Ed_Prim. �0.00(0.02)

0.06(0.05)

�0.06*(0.04)

0.04(0.04)

�0.08**(0.03)

0.02(0.05)

0.04(0.06)

0.03(0.03)

0.06(0.06)

�0.06(0.05)

�0.01(0.03)

�0.06(0.07)

0.18***(0.05)

0.04(0.06)

0.05(0.03)

�0.04(0.08)

0.06(0.05)

0.04(0.04)

0.03(0.07)

0.09(0.08)

Ed_Sec. �0.01(0.02)

0.05(0.07)

�0.05(0.05)

0.04(0.04)

�0.09**(0.04)

0.05(0.05)

�0.06(0.07)

0.09**(0.03)

0.03(0.07)

�0.14**(0.07)

0.07(0.06)

�0.04(0.08)

0.20***(0.07)

0.05(0.06)

0.08**(0.04)

�0.04(0.07)

0.00(0.05)

0.04(0.04)

0.08(0.07)

0.16*(0.09)

Ed_Post. 0.01(0.03)

�0.00(0.08)

0.05(0.07)

0.03(0.05)

�0.07(0.07)

0.11***(0.04)

0.09(0.09)

0.01(0.05)

�0.15(0.11)

�0.13(0.11)

0.09**(0.04)

�0.21(0.15)

0.12**(0.05)

0.01(0.06)

0.05(0.05)

�0.00(0.08)

�0.12(0.13)

0.05(0.05)

0.01(0.09)

0.13(0.09)

Info. 0.01(0.01)

�0.05**(0.02)

0.01(0.02)

0.03**(0.01)

0.02(0.01)

0.02(0.02)

0.03(0.02)

0.02(0.02)

�0.01(0.03)

�0.00(0.02)

0.06***(0.02)

�0.02(0.02)

0.00(0.02)

0.02(0.02)

0.01(0.02)

0.03*(0.02)

0.03**(0.02)

�0.02(0.02)

0.04*(0.02)

0.06***(0.02)

Soc.dem. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Reg.dum. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Obs. 1082 1000 1002 1057 919 987 1079 1009 1227 933 1120 731 927 1901 1065 1950 859 2146 1009 1067

Panel B: Dependent variable is meetingResourcesPoverty 0.07***

(0.02)0.06***(0.02)

0.01(0.02)

0.02(0.02)

0.03(0.02)

�0.01(0.02)

�0.02*(0.01)

0.00(0.01)

0.01(0.01)

�0.02(0.01)

0.04**(0.02)

0.03(0.02)

�0.02(0.02)

0.03**(0.01)

0.07***(0.02)

0.05***(0.01)

�0.00(0.01)

�0.01(0.01)

0.02(0.02)

0.03*(0.02)

Full-time 0.11**(0.05)

�0.05(0.04)

�0.02(0.06)

�0.06*(0.04)

0.05(0.04)

0.04(0.04)

�0.15***(0.06)

�0.00(0.05)

0.03*(0.02)

�0.05(0.05)

�0.12(0.08)

0.10**(0.05)

�0.03(0.04)

0.11***(0.03)

0.04(0.05)

�0.04(0.03)

0.06*(0.03)

�0.00(0.04)

�0.00(0.05)

�0.02(0.05)

Ed_Prim. �0.01(0.04)

0.07(0.05)

0.08**(0.04)

0.08(0.06)

�0.08(0.05)

0.16**(0.07)

0.01(0.05)

0.05(0.04)

0.04(0.03)

0.04(0.04)

0.01(0.04)

0.03(0.08)

0.04(0.09)

0.12**(0.06)

�0.02(0.04)

�0.13(0.08)

0.14**(0.06)

0.06(0.04)

�0.12(0.08)

0.04(0.07)

Ed_Sec. �0.05(0.05)

0.12**(0.06)

�0.02(0.05)

0.23***(0.07)

�0.11*(0.06)

0.20***(0.08)

�0.02(0.06)

0.07*(0.04)

0.01(0.03)

0.05(0.04)

0.10(0.07)

0.04(0.08)

0.08(0.09)

0.14**(0.06)

0.01(0.05)

�0.13(0.09)

0.17***(0.04)

0.09**(0.04)

�0.09(0.08)

0.22***(0.08)

Ed_Post. �0.05(0.10)

0.13**(0.05)

0.06(0.08)

0.30***(0.09)

�0.12(0.09)

0.22***(0.06)

�0.04(0.08)

0.14***(0.04)

0.03(0.03)

0.08(0.06)

0.00(0.08)

�0.04(0.12)

0.14(0.10)

0.12*(0.06)

0.09(0.07)

�0.07(0.09)

�0.05(0.13)

0.07(0.05)

�0.08(0.10)

0.16**(0.07)

Info. 0.04*(0.02)

0.05***(0.02)

0.03(0.02)

0.03*(0.02)

0.05**(0.02)

0.05**(0.02)

0.01(0.02)

0.09***(0.02)

0.01(0.01)

0.00(0.02)

0.08***(0.02)

0.01(0.02)

�0.01(0.02)

0.02(0.02)

0.05*(0.02)

0.04**(0.02)

0.00(0.02)

0.06***(0.02)

0.06***(0.02)

�0.00(0.02)

Soc.dem. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Reg.dum. Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Obs. 1147 1156 1068 1184 1092 1046 1152 1148 1283 1108 1184 1041 1188 2069 1126 2242 1037 2355 1126 1141

Notes: Robust standard errors in parentheses; *significant at 10%, **at 5%, ***at 1%; Observations are weighted using within country weights adjusting the sample to be nationally representative with respect toregion, urban–rural distribution etc.; ‘Soc. dem.’ refers to the socio-demographic group affiliations and ‘Reg. dum.’ to the 246 region dummies (see Table A2).

A.-S.Isaksson

/ElectoralStudies

xxx(2013)

1–17

14Pleasecite

thisarticle

inpress

as:Isaksson,A

.-S.,Politicalparticipation

inAfrica:

Therole

ofindividual

resources,ElectoralStudies

(2013),http://dx.doi.org/10.1016/j.electstud.2013.09.008

Table A6Using alternative poverty and information measures (probit marginal effects).

Dependent var. (1) Vote (2) Vote (3) Vote (4) Vote (5) Meet (6) Meet (7) Meet (8) Meet

ResourcesEducation_Primary 0.014

(0.012)0.013(0.012)

0.012(0.012)

0.014(0.012)

0.029**(0.012)

0.030**(0.012)

0.030***(0.012)

0.030***(0.012)

Education_Secondary 0.014(0.014)

0.011(0.014)

0.015(0.013)

0.017(0.013)

0.049***(0.016)

0.047***(0.016)

0.059***(0.016)

0.057***(0.016)

Education_Post-sec. 0.001(0.017)

�0.005(0.017)

0.006(0.017)

0.009(0.017)

0.067***(0.018)

0.063***(0.018)

0.083***(0.017)

0.084***(0.017)

Full-time 0.015*(0.009)

0.012(0.009)

0.014(0.009)

0.016*(0.009)

0.005(0.012)

0.006(0.012)

0.006(0.012)

0.010(0.012)


0.014***(0.005)

0.032***(0.007)

0.031***(0.007)

Poverty 0.004(0.004)

0.004(0.004)

0.016***(0.005)

0.017***(0.005)

Alternative poverty measuresa

Somewhat poor 0.014*(0.008)

0.028***(0.009)

Very poor 0.025**(0.011)

0.024*(0.013)

Very bad �0.015(0.015)

0.019(0.016)

Fairly bad �0.012(0.011)

0.019*(0.011)

Fairly good 0.003(0.011)

0.018(0.011)

Very good �0.006(0.020)

�0.009(0.020)

Alternative information measuresb

Own radio 0.040***(0.008)

0.063***(0.010)

Medium radio 0.008(0.014)

0.064***(0.017)

Often radio 0.034***(0.011)

0.090***(0.013)

Socio-dem. controls Yes Yes Yes Yes Yes Yes Yes Yes

Region dummies Yes Yes Yes Yes Yes Yes Yes Yes

Observations 23,070 22,974 23,039 23,070 25,893 25,772 25,861 25,893

Notes: Standard errors (clustered by the 246 regions) in parentheses; *significant at 10%, **significant at 5%, ***significant at 1%. Observations are weightedusing combined within � across country weights (see description in Table 1).

a Refer to dummy variables distinguishing between the non-poor (reference category), the somewhat poor, and the very poor (based on the question ‘Overthe past year, how often, if ever, have you or anyone in your family gone without enough food to eat?’), and dummy variables capturing self-reported livingconditions (‘very bad’, ‘fairly bad’, ‘neither good nor bad’, which is used as reference category, ‘fairly good’ and ‘very good’).

b Refer to dummy variables indicating if the respondent owns a radio, and dummy variables indicating how often the respondent gets news from a radio –

seldom (‘never’ or ‘less than once a month’, used as reference category), medium (‘a few times a month’) or often (‘a few times a week’ or ‘every day’).

Table A7Altering the sample and dependent variables.

Columns 1–7 Probit marginal effects, Column 8 OLS regression coefficients

(1) (2) (3) (4) (5) (6) (7) (8)Dependent var. Vote Meet Alt. vote Vote Vote Vote Vote Alt. meet

ResourcesPoverty 0.004

(0.004)0.016***(0.005)

0.005(0.004)

0.002(0.006)

0.006(0.004)

0.005(0.005)

0.002(0.004)

0.056***(0.010)

Full-time 0.014(0.009)

�0.000(0.012)

0.014(0.009)

0.001(0.016)

0.014(0.009)

0.017(0.012)

0.020**(0.009)

0.043*(0.023)

Education_Primary 0.015(0.012)

0.023**(0.012)

0.015(0.012)

0.024(0.023)

0.013(0.013)

0.015(0.015)

0.010(0.012)

0.095***(0.022)

Education_Secondary 0.014(0.014)

0.050***(0.016)

0.014(0.014)

0.040(0.026)

0.012(0.015)

�0.008(0.017)

0.002(0.015)

0.174***(0.031)

Education_Post-sec. 0.001(0.017)

0.065***(0.018)

�0.000(0.017)

�0.008(0.033)

0.008 0.004(0.019)

�0.002(0.019)

0.292***(0.038)(0.018)


0.033***(0.007)

0.018***(0.005)

0.027***(0.007)

0.012**(0.006)

0.020***(0.007)

0.014***(0.005)

0.119***(0.013)

(continued on next page)



Table A7 (continued )

Columns 1–7 Probit marginal effects, Column 8 OLS regression coefficients

(1) (2) (3) (4) (5) (6) (7) (8)Dependent var. Vote Meet Alt. vote Vote Vote Vote Vote Alt. meet

Socio-demographic Yes Yes Yes Yes Yes Yes Yes Yes

Region dummies Yes Yes Yes Yes Yes Yes Yes Yes

Observations 22949 23019 23156 4921 18166 9272 16088 24964

Notes: Standard errors (clustered by region) in parentheses; *significant at 10%, **at 5%, ***at 1%; Observations are weighted using combined within� acrosscountry weights; Regressions 1–2 restrict the sample to observations that have missing values on neither the voting nor the meeting indicator (the samplesnevertheless differ slightly due to having to drop observations in regions where all respondents have the same value on the outcome variable); Regression 3uses an alternative voting dummy coding ‘don’t know’ responses as zeros instead of missing values; Regression 4 restricts the sample to include only re-spondents from the five countries with the smallest discrepancy between self-reported voting share and official turnout, namely Cape Verde, Ghana,Liberia, Namibia and Zambia; Regression 5 restricts the sample to only include observations where the interviewer judges the respondent as honest; Re-gressions 6–7 restrict the sample to countries judged as ‘free’ by FreedomHouse and countries with Polity IV democracy scores higher than five, respectively;Regression 8 uses an inter-electoral participation index as dependent variable.


References

African Elections Database, 2010. Available at: http://africanelections.tripod.com/.

Afrobarometer Network, February 2007. Round 4 Survey Manual. Afro-barometer Network.

Bartels, L.M., 2005. Economic Inequality and Political Representation.mimeo. Princeton University.

Booth, J.A., Seligson, M.A., 2008. Inequality and democracy in LatinAmerica: Individual and contextual effects of wealth on politicalparticipation. In: Krishna, A. (Ed.), Poverty, participation and de-mocracy: A global perspective. Cambridge University Press, New York.

Boulding, C., Wampler, B., 2010. Voice, votes, and resources: evaluatingthe effect of participatory democracy on well-being. World Dev. 38(1), 125–135.

Bratton, M., 1999. Political participation in a new democracy: institutionalconsiderations from Zambia. Comp. Polit. Stud. 32 (5), 549–588.

Bratton, M., 2008. Poor people and democratic citizenship in Africa. In:Krishna, A. (Ed.), Poverty, Participation and Democracy: a GlobalPerspective. Cambridge University Press, New York.

Bratton, M., Chu, Y., Lagos, M., 2010. Who votes? Implications for newdemocracies. Taiwan J. Democr. 6 (1), 1–30.

Bratton, M., Logan, C., 2006. The Political Gender Gap in Africa: SimilarAttitudes, Different Behaviours. Afrobarometer Working Paper No. 58.

Bratton, M., Mattes, R., Gyimah-Boadi, E., 2005. Public Opinion, De-mocracy, and Market Reform in Africa. Cambridge University Press,Cambridge.

Bratton, M., Van de Walle, N., 1997. Democratic Experiments in Africa:Regime Transitions in Comparative Perspective. Cambridge UniversityPress, Cambridge.

Brady, H.E., Verba, S., Lehman Schlozman, K., 1995. Beyond Ses: a resourcemodel of political participation. Am. Polit. Sci. Rev. 89 (2), 271–294.

Christensen, M.M., Utas, M., 2008. Mercenaries of democracy: the ‘poli-tricks’ of remobilized combatants in the 2007 general elections, SierraLeone. Afr. Affairs 107 (429), 515–539.

Collier, Vicente, 2009. Votes and Violence: Evidence from a Field Exper-iment in Nigeria. CSAE WPS/2008-16.

Dahl, R., 1961. Who Governs? Yale University Press, New Haven andLondon.

Downs, A., 1957. An Economic Theory of Democracy. Harper and Row,New York.

Feddersen, T.J., 2004. Rational choice theory and the paradox of notvoting. J. Econ. Perspect. 18 (1), 99–112.

Feddersen, T.J., Pesendorfer, W., 1999. Abstention in elections withasymmetric information and diverse preferences. Am. Polit. Sci. Rev.93 (2), 381–398.

Fornos, C.A., Power, T.J., Garand, J.C., 2004. Explaining voter turnout inLatin America, 1980–2000. Comp. Polit. Stud. 37 (8), 909–940.

Freedom House, 2010. Freedom in the World Data 2005. Available at:http://www.freedomhouse.org/template.cfm?page¼351&ana_page¼292&year¼2005 (accessed 09.07.10).

Gerber, A.S., Green, D.P., Larimer, C.W., 2008. Social pressure and voterturnout: evidence from a large-scale field experiment. Am. Polit. Sci.Rev. 102 (1), 33–48.

Gilens, M., 2005. Inequality and democratic responsiveness. Public Opin.Quart. 69 (5), 778–796.


Griffin, J.D., Newman, B., 2005. Are voters better represented? J. Polit. 67(4), 1206–1227.

International Institute for Democracy and Electoral Assistance, 2010.Available at: http://www.idea.int/vt/ (accessed 21.04.10).

Isaksson, A.-S., Kotsadam, A., Nerman, M., 2013. The gender gap in Africanpolitical participation: testing theories of individual and contextualdeterminants. J. Dev. Stud. (in press).

Jackman, R.W., 1987. Political institutions and voter turnout in theindustrialised democracies. Am. Polit. Sci. Rev. 81 (2), 405–424.

Knack, S., Kropf, M.E., 1998. For shame! The effect of community coop-erative context on the probability of voting. Polit. Psychol. 19 (3),585–599.

Kostadinova, T., 2003. Voter turnout dynamics in post-communistEurope. Eur. J. Polit. Res. 42 (6), 741–759.

Krishna, A., 2002. Enhancing political participation in democracies: whatis the role of social capital? Comp. Polit. Stud. 35 (4), 437–460.

Krishna, A., 2008. Do poor people care less for democracy? Testing in-dividual-level assumptions with individual-level data. In: Krishna, A.(Ed.), Poverty, participation and democracy: A global perspective.Cambridge University Press, New York.

Kuenzi, M., Lambright, G., 2007. Voter turnout in Africa’s multiparty re-gimes. Comp. Polit. Stud. 40 (6), 665–690.

Kuenzi, M., Lambright, G., 2010. Who votes in Africa? An examination ofelectoral participation in 10 African countries. Party Polit.. http://dx.doi.org/10.1177/1354068810376779.

La Due Lake, R., Huckfeldt, R., 1998. Social capital, social networks, andpolitical participation. Polit. Psychol. 19 (3), 567–584.

Lijphart, A., 1997. Unequal participation: democracy’s unresolveddilemma. Am. Polit. Sci. Rev. 91 (1), 1–14.

Lindberg, S.I., 2004. The democratic qualities of competitive elections:participation, competition and legitimacy in Africa. CommonwealthComp. Polit. 42 (1), 61–105.

Lindberg, S.I., 2006a. Democracy and elections in Africa. The Johns Hop-kins University Press, Baltimore.

Lindberg, S.I., 2006b. The surprising significance of African elections. J.Democr. 17 (1), 139–151.

Lindberg, S.I., Morrison, M.K.C., 2008. Are African voters really ethnic orclientelistic? Survey evidence from Ghana. Polit. Sci. Quart. 123 (1),95–122.

Marshall, M.G., Jaggers, K., 2002. Polity IV Project: Political RegimeCharacteristics and Transitions 1800–2002. Dataset Users’ Manual.Integrated Network for Societal Conflict Research (INSCR) and Pro-gram Center for International Development and Conflict Manage-ment (CIDCM), University of Maryland.

Norris, P., 2002. Democratic Phoenix: Reinventing Political Activism.Cambridge University Press, Cambridge, UK.

Polity IV project, 2010. Polity IV Project: Political Regime Characteristicsand Transitions 1800–2002. Available at: http://www.systemicpeace.org/inscr/inscr.htm (accessed 09.07.10).

Posner, D.N., Simon, D.J., 2002. Economic conditions and incumbentsupport in Africa’s new democracies: evidence from Zambia. Comp.Polit. Stud. 35 (3), 313–338.

Solt, F., 2008. Economic inequality and democratic political engagement.Am. J. Polit. Sci. 52 (1), 48–60.

Verba, S., Nie, N.H.,1972. Participation inAmerica. Harper &Row,NewYork.Verba, S., Nie, N.H., Kim, J.-O., 1978. Participation and Political Equality: a

Seven Nation Comparison. Cambridge University Press, Cambridge.


http://africanelections.tripod.com/

http://africanelections.tripod.com/

http://refhub.elsevier.com/S0261-3794(13)00151-0/sref2












































http://www.freedomhouse.org/template.cfm?page%3D351%26ana_page%3D292%26year%3D2005












http://www.idea.int/vt/



















http://dx.doi.org/10.1177/1354068810376779

http://dx.doi.org/10.1177/1354068810376779






















http://www.systemicpeace.org/inscr/inscr.htm

http://www.systemicpeace.org/inscr/inscr.htm










Verba, S., Schlozman, K.L., Brady, 1995. Voice and Equality: CivicVoluntarism in American Politics. Harvard University Press, Cam-bridge, MA.

Vicente, P., 2008. Is Vote Buying Effective? Evidence from a Field Exper-iment in West Africa.

Wantchekon, L., 2003. Clientelism and Voting Behaviour: Evidence from aField Experiment in Benin. World Polit. 55 (3), 399–422.


Wolfinger, R.E., Rosenstone, S.J., 1980. Who Votes? Yale University Press,New Haven.

Yadav, Y.S., 2000. Understanding the Second Democratic Upsurge: Trendsof Bahujan Participation in Electoral Politics in the 1990s. In:Frankel, F.R., Hasan, Z., Bhargava, R., Arora, B. (Eds.), TransformingIndia: Social and Political Dynamics of Democracy. Oxford UniversityPress, Delhi.
















Political participation in Africa: The role of individual resources

Documents

Transcript of Political participation in Africa: The role of individual resources