The Political Determinants of Health: Elective and Chiefly Authority in South Africa Carol Mershon Professor Department of Politics Hugh S. and Winifred Cumming Chair in Politics University of Virginia PO Box 400787 Charlottesville VA 22904 USA [email protected]September 2016 Acknowledgements. For research assistance, I thank Rachel Okrent, Paromita Sen, and Jennifer Simons. For helpful comments, I thank Michelle Dion, Daniel Gingerich, Elizabeth Kaknes, Jon Kropko, Phil Potter, Olga Shvetsova, Denise Walsh, and seminar participants at the Institutions in Context Workshop, University of Tampere, Finland, Collegio Carlo Alberto, Turin, Italy, and University of Virginia. I gratefully acknowledge financial support from the Center for Global Health, Center for International Studies, Office of the Vice President for Research, Research Support in the Arts, Humanities, and Social Sciences, and Sesquicentennial Research Fund, all of the University of Virginia, along with financial support from the University of Tampere, Finland, and the Collegio Carlo Alberto, Turin. Abstract. Under what conditions do local political elites in new democracies promote citizen wellbeing? Pursuing this inquiry for South Africa, I assess hypotheses on party competition, popular participation, and chiefly authority as sources of local variation in public goods provision. Empirical analyses use an augmented version of a comprehensive census mortality sample. The evidence disconfirms the hypothesis that incumbents respond to competitive incentives to supply public goods. Partial, nuanced support appears for the hypothesis that citizen participation enhances public goods provision. Where chiefs are strong, restricted party competition under the ANC lowers the probability of infant and under-five death in majority Black African households, the largest set of households by far; where chiefs are strong, as voter turnout rises, the probability of infant and under-five death in majority Black African households diminishes. The article makes integrated theoretical and empirical contributions to scholarship on public goods, chieftaincy, and the workings of democracy.
44
Embed
The Political Determinants of Health: Elective and Chiefly ...carolmershon.weebly.com/uploads/1/2/8/3/12835113/political... · the recent literature on chiefly authority. Across sub-Saharan
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Political Determinants of Health: Elective and Chiefly Authority in South Africa
Carol Mershon
Professor Department of Politics
Hugh S. and Winifred Cumming Chair in Politics University of Virginia
PO Box 400787 Charlottesville VA 22904 USA
[email protected] September 2016 Acknowledgements. For research assistance, I thank Rachel Okrent, Paromita Sen, and Jennifer Simons. For helpful comments, I thank Michelle Dion, Daniel Gingerich, Elizabeth Kaknes, Jon Kropko, Phil Potter, Olga Shvetsova, Denise Walsh, and seminar participants at the Institutions in Context Workshop, University of Tampere, Finland, Collegio Carlo Alberto, Turin, Italy, and University of Virginia. I gratefully acknowledge financial support from the Center for Global Health, Center for International Studies, Office of the Vice President for Research, Research Support in the Arts, Humanities, and Social Sciences, and Sesquicentennial Research Fund, all of the University of Virginia, along with financial support from the University of Tampere, Finland, and the Collegio Carlo Alberto, Turin. Abstract. Under what conditions do local political elites in new democracies promote citizen wellbeing? Pursuing this inquiry for South Africa, I assess hypotheses on party competition, popular participation, and chiefly authority as sources of local variation in public goods provision. Empirical analyses use an augmented version of a comprehensive census mortality sample. The evidence disconfirms the hypothesis that incumbents respond to competitive incentives to supply public goods. Partial, nuanced support appears for the hypothesis that citizen participation enhances public goods provision. Where chiefs are strong, restricted party competition under the ANC lowers the probability of infant and under-five death in majority Black African households, the largest set of households by far; where chiefs are strong, as voter turnout rises, the probability of infant and under-five death in majority Black African households diminishes. The article makes integrated theoretical and empirical contributions to scholarship on public goods, chieftaincy, and the workings of democracy.
1
The Political Determinants of Health: Elective and Chiefly Authority in South Africa
Under what conditions do local political elites in new democracies promote the wellbeing
of the people they rule? This question is especially urgent in sub-Saharan Africa. The 1990s
witnessed the introduction of democratic regimes across Africa, the region with the highest rates
of infant and under-five mortality in the world (World Health Organization (WHO), 2016a,
2016b). In many African democracies, including post-apartheid South Africa, sub-national
governments shoulder a range of responsibilities devolved by national government, and thus are
in charge of delivering such public goods as water, sanitation, and health care (e.g., S. G.
The models weigh five classes of influences on the probability that the death of an infant
(child) occurs in household i within municipal health district j: a contextual dummy variable
tapping party competition in j, a contextual variable capturing citizen participation in j, the
gender of the infant (child) as the sole individual-level variable, a set of controls at the household
level, and another set of controls at the level of the municipal health district. At one stage of the
analysis, I conduct estimations for three sets of municipal health districts, according to presence
and strength of chiefly authority.15 All models use importance weights (on sampling for the
mortality ten percent datafile, DataFirst, 2015; personal communications with Statistics South
Africa personnel, July 10, 13, 14, 2015).
Empirical Analysis
The baseline model for infant mortality, identified as Model 1 in Table 1, looks at the
household’s majority racial group and the infant’s gender as influences on the occurrence of
infant death, along with the household’s location within a particular municipal health district. As
generate a theoretical prediction for the household-level likelihood of infant and child death.
Estimations including this variable are available from the author.
15 The online Tables SI-2A through SI-2C provide descriptive statistics on dependent and
independent variables for the three sets of municipal health districts.
18
the leftmost column of the table displays, babies in majority Black African and Coloured
households are relatively likely to die before their first birthday. Majority White households, on
the other hand, face a relatively low probability of experiencing the loss of an infant; the
coefficient on this dummy variable attains a marginal level of statistical significance. Female
babies in South Africa are less likely to die than are males, a finding aligning with extant public
health studies (e.g., Kahn, Garenne, Collinson, & Tollman, 2007).
[Table 1 about here.]
Model 2 incorporates a range of attributes that characterize municipal health districts. As
exhibited in the second column of Table 1, Model 2 leaves intact the impact of race and gender
on infant death. Moreover, the evidence from Model 2 overturns the expectation distilled in H1.
Specifically, the coefficient on the dummy variable tapping near-unipartism (as based on ENPS)
at the 2006 municipal election has a negative sign and is statistically significant: in municipal
health districts where one party dominates party competition, the likelihood that households
experience the loss of an infant is low relative to households in districts where multiple parties
engage in stiffer competition for council seats. In contrast, these data corroborate the
participatory hypothesis, H2: the greater the voter turnout in a municipal health district at the
2006 election, the lower the likelihood that a given household sustains the death of an infant.16
Looking at public goods, the greater a municipal health district’s share of adults lacking
16 As a robustness test, I conduct estimations in which deciles on the largest party’s margin of
victory replace the near-unipartism dummy as the proxy tapping inter-party competition (cf. note
12; details in Table SI-3). Similar results on H1 and H2 emerge. The numbers of observations
and groups decline in these estimations because there are missing data on 2006 electoral results
for one municipal health district, given boundary changes.
19
secondary education, the higher the probability that a household in the district suffers the death
of an infant. Two indicators of the municipal district’s delivery of goods typically presumed to
affect wellbeing—Vitamin A coverage and per capita primary health care spending—operate as
expected to significantly lower the probability of infant death in a given household. Remarkably,
the more widespread the immunization of infants in a municipal health district, the higher, not
the lower, the likelihood that a household in the district endures a baby’s death. Noteworthy, too,
is that overall district health care spending exerts little influence on infant mortality.
Model 3, the baseline model for the event of under-five death, reveals that children in
majority Black African and Coloured households are relatively likely to die before reaching the
age of five. Female children are relatively likely to survive beyond their fifth birthday. Majority
White households are relatively unlikely to undergo the tragedy of child death, but the
coefficient on this dummy lacks statistical significance.
The findings arrayed in the rightmost column of Table 1, on under-five death, resemble
those from Model 2, on infant death as determined by an array of attributes of municipal health
districts. One difference in Model 4 as compared to Model 2 is that the dummy for majority
White households lacks statistical significance. In Model 4, the result conflicting with H1, on
party competition, remains; support for H2, on citizen participation, persists. In line with Model
2, as scarcity of schooling rises, so too does the probability of child death. A surprise on infant
immunization recurs, now elevating the likelihood of under-five death. Vitamin A coverage and
per capita primary health care spending significantly reduce the probability of child mortality.17
17 As another robustness test, I replace the measures of deprivation in water and schooling with
the composite South African Index of Multiple Deprivation, which is based on 2011 Census data
and considers the material, employment, educational, and living environment domains (cf.
20
I use Models 2 and 4 as the basis for plots of the marginal effects of race and gender on
the probability of infant and child death, as differing across distinct municipal-level contexts of
partisan competition. As Figure 4 illustrates, the probability of infant death in majority Black
African households, as opposed to all other South African households, stands at 8.10 percent in
municipalities with relatively competitive party systems; it decreases to 7.07 percent where one
party dominates competition. Similarly, the probability of child death in Black African
households, as opposed to all other South African households, falls from 10.11 percent to 8.93
percent as the electoral arena shifts from competitive to one-party dominant. The marginal
effects of gender on the probability that households witness the death of babies and children, as
varying across distinct competitive contexts, are more limited; the advantage that female babies
and children have in survival lessens slightly where one party dominates the municipal party
system. Although the confidence intervals for these point estimates overlap, examining contrasts
of marginal effects reveals that these differences are statistically significant. (Details in Table SI-
4.)
[Figure 4 about here.]
I now evaluate hypotheses on chiefly authority by estimating variants of Models 2 and 4
while sorting the sample of households into three groups of municipal health districts: those
lacking chiefs altogether; those districts where chiefs are weak, that is, where the sampled
percentage of households in enumeration areas under chiefly jurisdiction lies above zero but
below the mean; and those districts where chiefs are strong, where the sampled percentage of
Noble, et al., 2013). The findings disconfirming H1 remain, while support for H2 disappears
(details in Table SI-3). Voter turnout and the SAIMD are relatively highly correlated (Pearson
correlation = -0.6327, p < 0.001).
21
households in enumeration areas under chiefly jurisdiction lies at or above the mean.
Table 2 reports the results of these estimations. The overarching message is that
influences on infant and under-five death vary across districts as distinguished by strength of
chiefly authority. Drilling deeper, Model 5 looks at the event of infant death among households
located in municipal health districts where chiefs are absent. As displayed at the far left of Table
2, the burden of infant mortality borne by majority Black African households reappears in this
class of districts. A significantly lower probability of infant death arises again in majority White
households located in districts without any chiefs. Here, for the first time, majority Coloured
households do not face a significantly greater likelihood of experiencing infant death than do
other households. Another first in the findings here is that female babies are not significantly less
likely to die than are male babies. Among households located in districts lacking chiefs, where
inter-party competition in the 2006 municipal election was relatively restricted (as tapped by the
near-unipartism dummy), the probability that a household suffers the loss of an infant is
relatively low; the coefficient attains marginal levels of statistical significance. Voter turnout in
the 2006 election no longer exerts a discernible impact on infant death. Thus, both H1 and H2
fall flat for households in this group of districts. As for public goods, the larger a district’s share
of people without piped water in or near their home, the greater the likelihood that a household in
the district suffers the death of an infant. Whereas infant immunization, Vitamin A coverage, and
per capita primary health care spending do not affect infant mortality, per capita total district
health spending reduces the probability that a household sustains the death of an infant.
[Table 2 about here.]
When we move to households located in districts where chiefs are weak, we discover that
only two factors exert a statistically significant impact on infant death: a baby’s birth in a
22
majority Black African household and in a municipal health district with a relatively high share
of adults lacking secondary schooling. As exhibited in the second column of Table 2, on Model
7, two other variables carry weight at marginal levels of significance: a baby’s birth in a majority
Coloured household and its gender. In these districts, as in those where chiefs are absent, the
evidence corroborates neither H1 nor H2.
Our attention shifts to households located in districts where chiefs are strong. As
registered in the third column of Table 2, devoted to Model 7, the probability of infant death is
significantly higher in majority Black African and majority Coloured households than in other
households in these districts. Alongside these factors, three other statistically significant
influences on infant death emerge in districts where chiefs are strong. Vitamin A coverage and
per capita primary health care spending diminish the likelihood that a given household will
experience the death of an infant. Educational deprivation raises that likelihood.
Look now at the right side of Table 2, on the event of under-five death, against the
backdrop of the findings arrayed on the table’s left side. Proceed step by step, from one set of
districts to another. Hence begin by comparing the results from Model 8 (fourth column), on
under-five death in households where chiefs are absent, and from Model 5, on infant death where
chiefs are absent. Major similarities emerge. Most important, the evidence runs counter to H1
and disconfirms H2. Next compare the findings from Model 9 (fifth column), on under-five
death where chiefs are weak, and from Model 6 on infant death in the same set of districts.
Despite some differences, similarities link the two groups of findings. In particular, support for
neither H1 nor H2 arises in these two estimations.
Now compare the findings from Model 10 (rightmost column), on under-five death where
chiefs are strong, and from Model 7, on infant death in such districts. For these settings, majority
23
Black African and majority Coloured households have a relatively high probability of facing the
tragedy of both under-five and infant death. Echoing the earlier findings, the interventions of
Vitamin A coverage and per capita primary health care spending lower the probability of child
death, whereas failures in the delivery of education increase the probability of child death. Most
pertinent to the argument, two differences separate the findings on infant and child death where
chiefs are strong: relatively restricted inter-party competition in the municipal heath district in
which a given household is located significantly reduces the probability of child death, which
flies in the face of the contestation hypothesis, H1; and relatively great voter turnout in a
household's health district reduces the probability of child death, supporting the participatory
hypothesis, H2. The coefficient on the latter variable attains marginal levels of statistical
significance.
It merits notice that in municipal health districts where chiefs are strong, compared to
other districts, relatively many people live without piped water and relatively many adults lack
secondary education. (See Tables SI-1, SI-2A through SI-2C). Even amid relative deprivation,
the proximal public goods of Vitamin A coverage and per capita primary health care spending
affect the distal public good of interest here: they work to lower the likelihood of infant and child
death. In this light, H4 is upheld. Where chiefs are strong, the contestation hypothesis, H1, fails
for both infant and under-five death, whereas the participatory hypothesis, H2, finds support only
for under-five death.
To better grasp the repeated failure of H1, I ask: does the identity of just which party
might or might not dominate municipal elections matter? I thus conduct estimations that
incorporate two additional dummy variables measured at the municipal health district level: one
marks those districts where the ANC came in first in the 2006 municipal elections; and the other
24
is an interactive dummy isolating those districts with both limited competition and the ANC as
the largest party in 2006. (For detailed results, see Table SI-5.)
I use such estimations as the basis for plots of the marginal effects of race and gender on
the probability of infant and child death, as differing across districts where the ANC does and
does not dominate partisan competition and looking only at districts where chiefs are strong.18 As
Figure 5 depicts, where chiefs are strong, the probability of infant death in majority Black
African households, as opposed to all other South African households, is 14.92 percent in
municipalities with relatively competitive local party systems or with the IFP as the leading party
(recall Figure 3). That probability declines to 7.71 percent where the ANC is dominant.
Likewise, where chiefs are strong, the likelihood that households suffer the loss of a child
decreases from 17.08 percent in districts with local party systems not dominated by the ANC, to
9.76 percent where the ANC sits astride the party system. The differences in survival just
discussed translate to seventy-two babies and seventy-three children in two similar
municipalities with strong chiefs for which the only difference would be whether or not the ANC
dominates the local party system. The marginal effects of gender on the probability that
households experience the death of babies and children, as varying by the fact of ANC
dominance where chiefs are strong, are more restricted; the advantage that female babies and
under-fives have in survival narrows slightly where strong chiefs and ANC dominance coincide.
Examining contrasts of marginal effects shows that the racial differences are statistically
significant, and those for gender are not. (For details, see Table SI-6.)
18 The plots focus only on districts where chiefs are strong because, due to collinearity, such
estimations for districts lacking chiefs and with weak chiefs omit one or both of the two
additional dummy variables used to isolate ANC dominance. See Table SI-5.
25
[Figure 5 about here.]
Still focusing on districts where chiefs are strong, I probe H2 further by calculating the
marginal effects of race and gender on the probability of infant and child death, as differing
across representative values of voter turnout. As Figure 6 illustrates, among districts with strong
chiefs, as voter turnout rises, the probability of infant and child death in Black African
households diminishes, whereas the probability that female babies and children die varies less.
Inspecting contrasts of marginal effects discloses that the racial differences are statistically
significant for both infant and under-five death. The differences for gender are not significant.
(See Table SI-7 for details.)
[Figure 6 about here.]
Conclusion
This article addresses a question of fundamental importance: what influence might local
political elites have on citizen wellbeing? It weighs answers emerging from scholarship on
contestation and participation as essential pillars of democracy and on the role of chiefly
institutions under democracy. The empirical analysis focuses on South Africa, a compelling site
for multiple reasons. Taking advantage of a comprehensive dataset on the distal public good of
infant and under-five mortality, I test hypotheses about the impact of interparty competition,
participatory pressures, and chieftaincy on health outcomes.
Since the end of apartheid, South African democracy has witnessed growth sufficient to
transform it into an upper middle-income economy. Yet one of the most striking findings here,
recurring throughout, is that majority Black African households are relatively likely to face the
tragedy of the death of an infant and a child under five. Racial inequities are a matter of life and
death for South African babies, children, and their families even now, after apartheid.
26
The ANC has dominated national-level South African politics since 1994. The ANC in
national government has championed and secured legislation that requires elected representatives
in local assemblies to pull chiefs into policy making. At no stage in this analysis do households
located in municipal health districts with relatively great party competition see a relatively low
likelihood of infant and child death. What is more, at multiple points the evidence directly
contradicts H1: relatively limited competition significantly reduces the probability of infant and
child death. Alongside the thoroughgoing failure of H1 stands partial, nuanced support for H2.
Analyzing all municipal health districts together, I find that households in districts with relatively
high voter turnout have a relatively low probability of experiencing the death of an infant or
child. Subdividing districts by strength of chiefly authority, it first appears that, only where
chiefs are strong, H2 holds for the event of child death; in all sets of districts the evidence on
infant death disconfirms that hypothesis. Scrutiny of marginal effects buttresses the finding that
limited competition enhances infant and child survival in Black African households and also
reveals that, where chiefs are strong, relatively restricted party competition under the ANC
lowers the probability of infant and under-five death in majority Black African households, the
most numerous category of households by far. Moreover, where ANC dominance and strong
chiefs coincide, as voter turnout rises, the likelihood of Black African infant and child death
dwindles. This evidence disconfirms H3 and supports H4 while thwarting H1 again and shoring
up H2.
This analysis of South African health outcomes is the first of its kind, given its theoretical
and empirical reach. Thus, along with novel evidence bearing on the research question, the paper
identifies avenues for further research on South Africa. First, it points to the interest of
investigating the determinants of such proximal public goods as safe water, along with the distal
27
public good of citizen wellbeing (e.g., on within-nation variation in Mexico, Díaz-Cayeros et al.
(2016)). Nonetheless, such a study lies outside the scope of this paper. Díaz-Cayeros et al. (2016)
examine both proximal and distal public goods, not by chance, in a book.19 Second, the findings
here guide enriched statistical analyses of the distal public goods of infant and child mortality in
South Africa. More finely grained data on electoral outcomes, chiefly authority, linguistic
fractionalization, and delivery of services relevant to health outcomes—not only water but also
sanitation, for instance—are available. Such measures could be used as more refined independent
variables, for the Census mortality datafile permits closer pinpointing of a given household’s
location.20
Third, the paper breaks new ground in its systematic empirical appraisal of the
hypotheses advanced, but does not uncover the mechanisms beneath its findings. Whether South
African chiefs co-produce with community members the distal public good of citizen health—or
other public goods—remains an open question. As noted, case study evidence from South Africa
suggests varying answers, with co-production present in some localities, absent in others, and
diverse patterns in localities without co-production (e.g., Oomen, 2005; Turner, 2014; Williams,
2010). Of particular interest, in some local cases chiefs cultivate networks with elected officials
to secure funds for local public goods, rather than mobilizing community cooperation to generate
19 Kramon and Posner (2013) show that relatively few studies look at both proximal and distal
public goods, engage in such an examination themselves, and do not assess hypotheses on the
impact of both elective and chiefly authority. On the last point, see below.
20 The publicly available data from the South African District Health Barometer on proximal
public goods do not allow for disaggregation: the DHB's measures of immunization, vitamin
coverage, and health spending pertain to the fifty-two health districts.
28
local public goods (Williams, 2010, esp. Chs. 6, 8). Additional research is needed to discern the
conditions accounting for any variation in co-production. For instance, public goods in two types
of chiefdoms could be compared: those experiencing the death of a chief and then, after an
interval, leadership under an inexperienced chief; and all other chiefdoms (cf. Baldwin, 2016, pp.
115-120). The data on South African chieftaincy permitting such a comparison are not publicly
available, however.
Another mechanism beneath the findings here might be that the ANC has used its
longstanding dominance at the national level to restrict resources available to, and public goods
provision within, municipalities controlled by opposition parties. Case-study evidence current as
of 2012 indicates that this has not occurred (Cameron, 2014; Resnick, 2014). Yet comparative
statics, based on a scenario that has only quite recently materialized, would offer the best
assessment of this possibility. The 2014 national elections handed the ANC its smallest seat
share in the National Assembly to date, 62.3 percent. Not only has national-level ANC
dominance eroded but also, crucially, local-level opposition parties have gained new strength:
the municipal elections of early August 2016 gave the ANC only 59.3 of the vote nationwide and
returned a number of municipal councils without a seat majority (de Vos, 2016; McMurry,
Martin, Lieberman, & de Kadt, 2016; Phillip, 2016). At the time of writing, the most variegated
set of municipal councils in the history of South African democracy has barely started work on
public goods provision. Future analysis of public goods under these governments is obviously
appealing.
This inquiry holds lessons for research beyond South Africa. Existing research highlights
the role of traditional leaders across sub-Saharan Africa, and pioneering studies examine how
chiefs in concert with both elective officials and community members co-produce proximal
29
public goods in Africa (e.g., Baldwin, 2016). Outside Africa, traditional leaders have local
political powers and at least the potential to affect public goods (e.g., Beath, Christia, &
Enikolopov, 2013; Cornell & Kalt, 2000; Díaz-Cayeros et al., 2014; Van Cott, 2010). To my
knowledge, however, no project other than this one appraises the impact of both elective and
traditional authority on the distal—literally vital—public goods of infant and under-five survival.
This article indicates how such work could proceed.
Most broadly, this study demonstrates that political scientists need to revisit the wisdom
on “elections as instruments of democracy” (Powell, 2000). Rich and diverse research schools in
the discipline analyze parties, party systems, and elections as essential links in the chain of
democratic representation and responsiveness. Yet where traditional leaders wield local power,
understanding the extent to which democracy serves citizen needs calls for investigation into
both elected and unelected authority. This article thus makes integrated theoretical and empirical
contributions to scholarship on public goods, chieftaincy, and the workings of democracy.
30
References
Acemoglu, D., Reed, T., & Robinson, J. A. (2014). Chiefs: Economic Development and Elite Control of Civil Society in Sierra Leone. Journal of Political Economy, 122(2), 319-368.
Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S., & Wacziarg, R. (2003). Fractionalization. Journal of Economic Growth, 8(2), 155-194.
Arvate, P. R. (2013). Electoral Competition and Local Government Responsiveness in Brazil. World Development, 43, 67-83.
Atkinson, D. (2007). Taking to the Streets: Has Developmental Local Government Failed in South Africa? In S. Buhlungu, J. Daniel, R. Southall, & J. Lutchman (Eds.), State of the Nation: South Africa 2007 (pp. 53-77). Cape Town: HSRC Press.
Baldwin, K. (2013). Why Vote with the Chief? Political Connections and Public Goods Provision in Zambia. American Journal of Political Science, 57(4), 794-809.
Baldwin, K. (2014). When Politicians Cede Control of Resources: Land, Chiefs, and Coalition-Building in Africa. Comparative Politics, 46(3), 253-271.
Baldwin, K. (2016). The Paradox of Traditional Chiefs in Democratic Africa. Cambridge: Cambridge University Press.
Banerjee, A., Kumar, S., Pande, R., & Su, F. (2011). Do Informed Voters Make Better Choices? Experimental Evidence from Urban India. Working Paper. Harvard University.
Banerjee, S. G., & Morella, E. (2011). Africa's Water and Sanitation Infrastructure: Access, Affordability, and Alternatives. Washington, D.C.: The World Bank.
Beall, J. (2005). Exit, Voice and Tradition: Loyalty to Chieftainship and Democracy in Metropolitan Durban, South Africa. Crisis States Research Centre working papers series 1, 59. Crisis States Research Centre, London School of Economics and Political Science.
Beall, J. (2006). Cultural Weapons: Traditions, Inventions and the Transition to Democratic Governance in Metropolitan Durban. Urban Studies, 43(2), 457-473.
Beall, J., Mkhize, S., & Vawda, S. (2005). Emergent Democracy and 'Resurgent' Tradition: Institutions, Chieftaincy and Transition in KwaZulu-Natal. Journal of Southern African Studies, 31(4), 755-771.
Beall, J., & Ngonyama, M. (2009). Indigenous Institutions, Traditional Leaders and Elite Coalitions for Development: the Case of Greater Durban, South Africa. Crisis States Research Centre working papers series 2, 55. Crisis States Research Centre, London School of Economics and Political Science.
Beall, J., Parnell, S., & Albertyn, C. (2015). Elite Compacts in Africa: The Role of Area-based Management in the New Governmentality of the Durban City-region. International Journal of Urban and Regional Research, 39(2), 390-406.
Beath, A., Christia, F., & Enikolopov, R. (2013). Do Elected Councils Improve Governance? Experimental Evidence on Local Institutions in Afghanistan. Working Paper. MIT.
Besley, T., Pande, R., & Rao, V. (2012). Just Rewards? Local Politics and Public Resource Allocation in South India. The World Bank Economic Review, 26(2), 191-216.
Bratton, M. (2012). Citizen Perceptions of Local Government Responsiveness in Sub-Saharan Africa. World Development, 40(3), 516-527.
Cameron, R. (2014). Vertical Decentralization and Urban Service Delivery in South Africa: Does Politics Matter? Development Policy Review, 32(s1), s81-s100.
Cheibub, J. A., Gandhi, J., & Vreeland, J. R. (2010). Democracy and Dictatorship Revisited. Public Choice, 143(1), 67-101.
31
Chhibber, P., & Nooruddin, I. (2004). Do Party Systems Count?: The Number of Parties and Government Performance in the Indian States. Comparative Political Studies, 37(2), 152-187.
Cleary, M. R. (2007). Electoral Competition, Participation, and Government Responsiveness in Mexico. American Journal of Political Science, 51(2), 283-299.
Cleary, M. R. (2010). The Sources of Democratic Responsiveness in Mexico. Notre Dame, IN: University of Notre Dame Press.
Coovadia, H., Jewkes, R., Barron, P., Sanders, D., & McIntyre, D. (2009). The Health and Health System of South Africa: Historical Roots of Current Public Health Challenges. The Lancet, 374(9692), 817-834.
Cornell, S., & Kalt, J. P. (2000). Where’s the Glue? Institutional and Cultural Foundations of American Indian Economic Development. The Journal of Socio-Economics, 29, 443-470.
Dahl, R. A. (1971). Polyarchy: Participation and Opposition. New Haven: Yale University Press.
DataFirst. (2015). “South African Census 2011, 10% Sample,” from Statistics South Africa. Retrieved from https://www.datafirst.uct.ac.za/dataportal/index.php/auth/login/?destination=catalog/485/get_microdata
de Kadt, D., & Larreguy, H. (2014). Agents of the Regime? Traditional Leaders and Electoral Behavior in South Africa. MIT Political Science Department Research Paper 2014-24.
de Kadt, D., & Lieberman, E. S. (2015). Do Citizens Reward Good Service? Voter Responses to Basic Service Provision in Southern Africa. MIT Political Science Department Research Paper 2015-25.
de Vos, P. (2016). What Happens After the Votes are Counted- Some Questions Answered. Retrieved from http://constitutionallyspeaking.co.za/what-happens-after-the-votes-are-counted-some-questions-answered/ - more-9460
Díaz-Cayeros, A., Estévez, F., & Magaloni, B. (2016). Strategies of Vote Buying: Democracy, Clientelism, and Poverty Relief in Mexico. New York: Cambridge University Press.
Díaz-Cayeros, A., Magaloni, B., & Ruiz-Euler, A. (2014). Traditional Governance, Citizen Engagement, and Local Public Goods: Evidence from Mexico. World Development, 53, 80-93.
Electoral Commission of South Africa (IEC). (2014). Municipal Election Results. Retrieved from: http://www.elections.org.za/content/Elections/Municipal-elections-results/
EISA (Electoral Institute for Sustainable Democracy in Africa). 2006a. Election Update 2006: South Africa. No. 1, February. Retrieved from http://www.content.eisa.org.za/pdf/eusa200601.pdf
EISA (Electoral Institute for Sustainable Democracy in Africa). 2006b. Election Update 2006: South Africa. No. 2, March. Retrieved from https://eisa.org.za/pdf/eusa200602.pdf
Ferree, K. E. (2006). Explaining South Africa's Racial Census. Journal of Politics, 68(4), 803-815.
Ferree, K. E. (2010). Framing the Race in South Africa: The Political Origins of Racial Census Elections. Cambridge: Cambridge University Press.
Filippov, M., Ordeshook, P. C., & Shvetsova, O. (2004). Designing Federalism: A Theory of Self-Sustainable Federal Institutions. Cambridge: Cambridge University Press.
32
Franck, R., & Rainer, I. (2012). Does the Leader's Ethnicity Matter? Ethnic Favoritism, Education, and Health in Sub-Saharan Africa. American Political Science Review, 106(2), 294-325.
Gottlieb, J. (2015). The Logic of Party Collusion in a Democracy: Evidence from Mali. World Politics, 67(01), 1-36.
Halleröd, B., Rothstein, B., Daoud, A., & Nandy, S. (2013). Bad Governance and Poor Children: A Comparative Analysis of Government Efficiency and Severe Child Deprivation in 68 Low- and Middle-income Countries. World Development, 48, 19-31.
Health Systems Trust. (2014). District Health Barometer 2013/2014. Retrieved from: http://www.hst.org.za/publications/district-health-barometer-201314
Jolobe, Z. (2007). 'Things Fall Apart: Can the Center Hold?' The State of Coalition Politics in the Cape Metropolitan Council. In S. Buhlungu, J. Daniel, R. Southall, & J. Lutchman (Eds.), State of the Nation: South Africa 2007 (pp. 78-94). Cape Town: HSRC Press.
Kahn, K., Garenne, M. L., Collinson, M. A., & Tollman, S. M. (2007). Mortality Trends in a New South Africa: Hard to Make a Fresh Start. Scandinavian Journal of Public Health, 35, 26-34.
Keefer, P., & Khemani, S. (2005). Democracy, Public Expenditures, and the Poor: Understanding Political Incentives for Providing Public Services. The World Bank Research Observer, 20(1), 1-27.
Koter, D. (2013). King Makers: Local Leaders and Ethnic Politics in Africa. World Politics, 65(2), 187-232.
Krämer, M. (2016). Neither despotic nor civil: the legitimacy of chieftaincy in its relationship with the ANC and the state in KwaZulu-Natal (South Africa). The Journal of Modern African Studies, 54(01), 117-143.
Kramon, E., & Posner, D. N. (2013). Who Benefits from Distributive Politics? How the Outcome One Studies Affects the Answer One Gets. Perspectives on Politics, 11(2), 461-474.
Krieger, N. (2008). Proximal, Distal, and the Politics of Causation: What's Level Cot to Do With It? American Journal of Public Health, 98(2), 221-230.
Lake, D. A., & Baum, M. A. (2001). The Invisible Hand of Democracy: Political Control and the Provision of Public Services. Comparative Political Studies, 34(6), 587-621.
Lieberman, E. S., & McClendon, G. H. (2012). The Ethnicity–Policy Preference Link in Sub-Saharan Africa. Comparative Political Studies, 46(5), 574-602.
Logan, C. (2009). Selected Chiefs, Elected Councillors and Hybrid Democrats: Popular Perspectives on the Co-existence of Democracy and Traditional Authority. The Journal of Modern African Studies, 47(1), 101-128.
Logan, C. (2013). The Roots of Resilience: Exploring Popular Support for African Traditional Authorities. African Affairs, 112(448), 353-376.
Lupu, N., & Riedl, R. B. (2012). Political Parties and Uncertainty in Developing Democracies. Comparative Political Studies, 46(11), 101-128.
Mares, I., & Young, L. (2016). Buying, Expropriating, and Stealing Votes. Annual Review of Political Science, 19(1), 267-288.
Mattes, R., & Bratton, M. (2007). Learning about Democracy in Africa: Awareness, Performance, and Experience. American Journal of Political Science, 51(1), 192-217.
33
Mayosi, B. M., Lawn, J. E., van Niekerk, A., Bradshaw, D., Abdool Karim, S. S., & Coovadia, H. M. (2012). Health in South Africa: Changes and Challenges Since 2009. The Lancet, 380(9858), 2029-2043.
McClendon, G. H. (2016). Race and Responsiveness: An Experiment with South African Politicians. The Journal of Experimental Political Science, 3(1), 60-74.
McDonnell, E. M. (2016). Conciliatory States: Elite Ethno-Demographics and the Puzzle of Public Goods Within Diverse African States. Comparative Political Studies, 49(11), 1513-1549.
McMurry, N., Martin, P., Lieberman, E., & de Kadt, D. (2016). Here are Four Reasons that South Africa's ANC Lost Ground in this Month's Elections. Monkey Cage, Washington Post. Retrieved from https://www.washingtonpost.com/news/monkey-cage/wp/2016/08/12/here-are-4-reasons-that-south-africas-african-national-congress-lost-ground-in-this-months-election/
Mershon, C. (2002). The Costs of Coalition. Stanford: Stanford University Press. Mershon, C., & Shvetsova, O. (2013). Party System Change in Legislatures Worldwide: Moving
Outside the Electoral Arena. Cambridge University Press. Muller, M. (2008). Free Basic Water — a Sustainable Instrument for a Sustainable Future in
South Africa. Environment and Urbanization, 20(1), 67-87. Municipal Systems Act, South Africa. (2000). Municipal Systems Act 32 of 2000. Retreived
http://faolex.fao.org/docs/pdf/saf93030.pdf National Health Act, South Africa. (2003). National Health Act 61 of 2003, as amended 2
September 2014. Retreived http://www.saflii.org/za/legis/consol_act/nha2003147.pdf Noble, M., Zembe, W., Wright, G. and Avenell, D. (2013). Multiple Deprivation and Income
Poverty at Small Area Level in South Africa in 2011. Cape Town: SASPRI. Ntsebeza, L. (2005). Democracy Compromised: Chiefs and the Politics of the Land in South
Africa. Leiden, Netherlands: Brill. Olson, M. (1993). Dictatorship, Democracy, and Development. The American Political Science
Review, 87(3), 567-576. Olson, M. (2000). Power and Prosperity: Outgrowing Communist and Capitalist Dictatorships.
New York: Basic Books. Oomen, B. (2005). Chiefs in South Africa: Law, Power & Culture in the Post-Apartheid Era.
Pietermaritzburg: University of KwaZulu-Natal Press. Ostrom, E. (1996). Crossing the Great Divide: Coproduction, Synergy, and Development. World
Development, 24(6), 1073-1087. Phillip, B. (2016). IFP to Receive Support from DA, EFF to Govern KZN districts. SABC News.
Retrieved from http://www.sabc.co.za/news/a/e5d386004de6805985bfc5ebaa746de2/IFPundefinedtoundefinedreceiveundefinedsupportundefinedfromundefinedDA,undefinedEFFundefinedtoundefinedgovernundefinedKZNundefineddistricts-20161708
Powell, G. B., Jr. . (2000). Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven: Yale Univeristy Press.
Putnam, R. D., Leonardi, R., & Nanetti, R. Y. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press.
Resnick, D. (2014). Urban Governance and Service Delivery in African Cities: The Role of Politics and Policies. Development Policy Review, 32(s1), s3-s17.
34
Riedl, R. B. (2014). Authoritarian Origins of Democratic Party Systems in Africa. Cambridge: Cambridge University Press.
Riedl, R. B., & Dickovick, J. T. (2014). Party Systems and Decentralization in Africa. Studies in Comparative International Development, 49(3), 321-342.
Schlozman, K. L., Verba, S., & Brady, H. E. (2012). The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy. Princeton: Princeton University Press.
Simeon, R., & Murray, C. (2001). Multi-Sphere Governance in South Africa: An Interim Assessment. Publius: The Journal of Federalism, 31(4), 65-92.
Statistics South Africa. (2011). CONTRALESA Partners with Census 2011 [Press release]. Retrieved from http://www.statssa.gov.za/?page_id=3974
Statistics South Africa. (2014). Statistics by Place. Retrieved from: http://beta2.statssa.gov.za/?page_id=964
Steytler, N. (2003). District Municipalities: Giving Effect to Shared Authority in Local Government. Law, Democracy, & Development, 7(2), 227-242.
Strøm, K., Müller, W. C., & Bergman, T. (2010). Cabinets and Coalition Bargaining: the Democratic Life Cycle in Western Europe. Oxford: Oxford University Press.
Taagepera, R., & Shugart, M. S. (1989). Seats and Votes: The Effects and Determinants of Electoral Systems. New Haven: Yale University Press.
Tafel, H. L. (2011). Regime Change and the Federal Gamble: Negotiating Federal Institutions in Brazil, Russia, South Africa, and Spain. Publius: The Journal of Federalism, 41(2), 257-285.
Touchton, M., & Wampler, B. (2014). Improving Social Well-Being Through New Democratic Institutions. Comparative Political Studies, 47(10), 1442-1469.
Traditional Leadership and Governance Framework Act, South Africa. (2004). Traditional Leadership and Governance Framework Act 41 of 2004, as Amended by Traditional Leadership and Governance Framework Amendment Act 23 of 2009. Retreived http://www.saflii.org/za/legis/consol_act/tlagfa2003431.pdf
Turner, R. L. (2014). Traditional, Democratic, Accountable? Navigating Citizen-Subjection in Rural South Africa. African Spectrum, 49(1), 27-54.
Van Cott, D. L. (2010). Indigenous Peoples' Politics in Latin America. Annual Review of Political Science, 13(1), 385-405.
Weghorst, K. R., & Lindberg, S. I. (2013). What Drives the Swing Voter in Africa? American Journal of Political Science, 57(3), 717-734.
Wilkinson, D., Gouws, E., Sach, M., & Karim, S. S. A. (2001). Effect of Removing User Fees on Attendance for Curative and Preventive Primary Health Care Services in Rural South Africa. Bulletin of the World Health Organization, 79, 665-671.
Williams, J. M. (2004). Leading from Behind: Democratic Consolidation and the Chieftaincy in South Africa. The Journal of Modern African Studies, 42(1), 113-136.
Williams, J. M. (2009). Legislating ‘Tradition’ in South Africa. Journal of Southern African Studies, 35(1), 191-209.
Williams, J. M. (2010). Chieftaincy, the State, and Democracy: Political Legitimacy in Post-Apartheid South Africa. Bloomington, IN: Indiana University Press.
World Bank. (2015a). Gini Index. Retrieved from: http://data.worldbank.org/indicator/SI.POV.GINI
35
World Bank. (2015b). Mortality Rate, Infant. Retrieved from: http://data.worldbank.org/indicator/SP.DYN.IMRT.IN
World Bank. (2015c). Mortality Rate, Under-5. Retrieved from: http://data.worldbank.org/indicator/SH.DYN.MORT
World Health Organization (WHO). (2016a). Global Health Observatory (GHO) Data: Infant Mortality. Retrieved from: http://www.who.int/gho/child_health/mortality/neonatal_infant_text/en/
World Health Organization (WHO). (2016b). Global Health Observatory (GHO) Data: Under-Five Mortality. Retrieved from: http://www.who.int/gho/child_health/mortality/mortality_under_five_text/en/
36
Figure 1. Proximal-Distal Continuum of Public Goods: Some Examples. PROXIMAL DISTAL | | | per capita primary health spdg immunization, Vitamin A coverage IMR, U5MR per capita district health spdg pct HH with piped water Note: See text for discussion of examples.
37
Figure 2. South African Municipal Party Systems, by Party Winning Most Seats in 2006 Municipal Elections.
0 .5 1 1.5 2 2.5 3
IFP (n=5)
DA (n=2)
ANC (n=44)
Largest Party's Margin Largest Party's Pct SeatsENPS Pct Voter Turnout
38
Figure 3. South African Municipal Party Systems, by Party Winning Most Seats in 2006 Municipal Elections and Chiefly Strength (No, Weak, and Strong Chiefs).
0 .5 1 1.5 2 2.5 3
stro
ngw
eak
none
IFP (n=5)
DA (0)
ANC (n=18)
IFP (0)
DA (0)
ANC (n=9)
IFP (0)
DA (n=2)
ANC (n=17)
Largest Party's Margin Largest Party's Pct SeatsENPS Pct Voter Turnout
Note: The horizontal axis shows values on the largest party’s margin of victory in seats, the largest party’s percentage of seats, the municipal councils’ effective number of parties in seats (ENPS), and percentage voter turnout in the 2006 municipal elections. Key to abbreviations: ANC=African National Congress; DA=Democratic Alliance; ENPS=Effective number of parties in seats; IFP=Inkatha Freedom Party.
39
Figure 4. Marginal Effects of Race and Gender on Probability of Infant and Under-Five Death, by Degree of Party Competition in Household's Municipal Health District.
-.05
0.0
5.1
.15
Effe
cts
on P
r(Inf
ant D
eath
) and
Pr(U
nder
-5 D
eath
)
0 1Limited Party Competition in 2006 Election in HH's District (0 = no, 1 = yes)
Majority Black African HH, Infant Death Female Infant DeathMajority Black African HH, Child Death Female Child Death
Average Marginal Effects with 95% CIs
40
Figure 5. Marginal Effects of Race and Gender on Probability of Infant and Under-Five Death, by Fact of ANC Dominance in Household's Municipality and Only Under Strong Chiefs.
0.0
5.1
.15
.2.2
5Ef
fect
s on
Pr(i
nfan
t Dea
th) a
nd P
r(Und
er-5
Dea
th)
0 1ANC Dominant in 2006 Municipal Election (0= no, 1 = yes)
Majority Black African HH, Infant Death Female Infant DeathMajority Black African HH, Child Death Female Child Death
Average Marginal Effects with 95% CIs
41
Figure 6. Marginal Effects of Race and Gender on Probability of Infant and Under-Five Death, at Representative Values of Voter Turnout in Household's Municipality and Only Under Strong Chiefs.
0.0
5.1
.15
.2Ef
fect
s on
Pr(I
nfan
t Dea
th) a
nd P
r(Und
er-5
Dea
th)
.40 .45 .50 .55 .60Voter Turnout in 2006 Election in HH's District
Majority Black African HH, Infant Death Female Infant DeathMajoity Black African HH, Child Death Female Child Death
Average Marginal Effects with 95% CIs
42
Table 1. Explaining the event of infant and under-five deaths in households; multilevel mixed effects logistic regression.
Infant Death in HH Under-Five Death in HH 1 2 3 4 Black African
1.438*** (0.265)
1.424*** (0.265)
1.330*** (0.216)
1.315*** (0.216)
Coloured 1.049*** (0.280)
1.086*** (0.281)
0.849*** (0.231)
0.879*** (0.232)
White -0.611† (0.333)
-0.600† (0.332)
-0.356 (0.260)
-0.348 (0.260)
Female
-0.089* (0.036)
-0.091* (0.036)
-0.065* (0.031)
-0.067* (0.031)
MHD near-unipartism 2006 election
-- -0.160* (0.070)
-- -0.153* (0.063)
MHD voter turnout 2006 election
-- -2.118* (0.972)
-- -2.413** (0.868)
MHD pct no piped water home/200m 2011
-- 0.021 (0.020)
-- 0.019 (0.018)
MHD pct adults no secondary school 2011
-- 0.067** (0.020)
-- 0.077*** (0.017)
MHD linguistic fractionalization 2011
-- -0.019 (0.016)
-- -0.018 (0.015)
MHD < 1 year fully immunized 2010
-- 0.022† (0.012)
-- 0.021† (0.011)
MHD 1-5 yrs Vitamin A coverage 2010
-- -0.031** (0.011)
-- -0.029** (0.010)
MHD per cap primary health care spending 2010
-- -0.033** (0.012)
-- -0.030** (0.011)
MHD per cap district health spending 2010
-- -0.012 (0.015)
-- -0.016 (0.013)
Municipal health district: identity
0.242 (0.034)
0.142 (0.027)
0.235 (0.031)
0.128 (0.023)
Wald χ2 148.23 201.95 194.72 269.22 Prob > χ2 0.0000 0.0000 0.0000 0.0000 Log likelihood -11649.00 -11461.498 -14213.22 -13982.263 N obs 39,122 38,631 39,122 38,631 N groups 52 51 52 51 † p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001. Notes: Reference category on HH majority racial group: Asian/Indian and other. Each model includes a constant, not reported. The numbers of observations and groups decline in Models 2 and 4 because there are missing data on 2006 electoral results for one municipal health district, given boundary changes.
43
Table 2. Explaining the event of infant and under-five deaths in households; multilevel mixed effects regression, by strength of chiefly authority. Infant death in HH Under-five death in HH Chiefly strength in HH’s district Chiefly strength in HH’s district Absent Weak Strong Absent Weak Strong
Model 5 6 7 8 9 10 Black African
1.133* (0.510)
1.452*** (0.327)
2.035* (1.009)
1.020* (0.420)
1.371*** (0.269)
1.693* (0.720)
Coloured 0.753 (0.519)
0.825† (0.465)
2.085* (1.055)
0.531 (0.429)
0.859* (0.384)
1.654* (0.772)
White -1.573* (0.620)
-0.101 (0.457)
0.806 (1.080)
-1.118* (0.478)
0.073 (0.361)
0.669 (0.779)
Female -0.105 (0.071)
-0.141† (0.077)
-0.064 (0.049)
-0.061 (0.062)
-0.153* (0.068)
-0.036 (0.043)
MHD near- unipartism 2006
-0.265† (0.149)
0.374 (1.180)
-0.133 (0.088)
-0.225† (0.128)
0.193 (0.993)
-0.199* (0.082)
MHD voter turnout 2006
-2.611 (2.398)
-18.001 (26.848)
-1.671 (1.519)
-1.092 (2.083)
-13.312 (22.604)
-2.243† (1.432)
MHD pct no piped water 2011
0.176** (0.062)
0.269 (0.306)
0.053 (0.036)
0.116* (0.054)
0.214 (0.259)
0.055 (0.034)
MHD pct adults no 2ndary school '11
-0.007 (0.048)
0.210** (0.062)
0.075* (0.029)
0.055 (0.040)
0.185** (0.054)
0.071* (0.028)
MHD linguistic fractional’n 2011
-0.085 (0.056)
0.043 (0.034)
-0.016 (0.022)
-0.045 (0.047)
0.034 (0.030)
-0.022 (0.021)
MHD < 1 year immunized 2010
-0.063 (0.042)
-0.032 (0.037)
0.002 (0.021)
-0.021 (0.035)
-0.016 (0.031)
-0.016 (0.020)
MHD age 1-5 Vitamin A 2010
-0.004 (0.019)
0.022 (0.047)
-0.108*** (0.018)
-0.000 (0.017)
0.023 (0.040)
-0.102*** (0.017)
MHD per cap PHC spdg 2010
0.033 (0.039)
0.084 (0.205)
-0.051*** (0.014)
0.010 (0.035)
0.042 (0.173)
-0.041** (0.013)
MHD per capita health spdg 2010
-0.085** (0.031)
0 (omitted)
0.020 (0.019)
-0.087** (0.028)
0 (omitted)
0.025 (0.018)
MHD: identity
1.33e-09 (0.072)
1.66e-00 (0.039)
0.086 (0.036)
4.36e-06 (0.045)
2.94e-07 (0.034)
0.092 (0.030)
Wald χ2 106.26 79.34 74.70 9.24 98.74 77.47 Prob > χ2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Log likelihood -3038.329 -2493.239 -5901.847 -3703.453 -3048.900 -7200.632 N obs 11,661 8,855 18,115 11,661 8,855 18,115 N groups 19 9 23 19 9 23 † p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001. Notes: Reference category on HH majority racial group: Asian/Indian and other. In districts where chiefs are weak, one variable is omitted due to collinearity.