Democracy and child health in developing countries

Abstract

The rise of democracy across the world brought with it expectations that governments would be more attentive and responsive to the welfare of the people, creating better services and better health. Indeed, a considerable body of scholarship finds that democracy has significant, direct effects on multiple measures of well-being, particularly life expectancy and infant mortality. Despite several recent critiques, the paramount theme is that democracy is good for health. This study contributes to this literature by assessing the relationship between democracy and child diarrhea and malnutrition across 52 developing countries. Using a multilevel modeling strategy, the analysis examines the country-level effects of democracy and development on child health, while simultaneously taking into account wealth, education, and other household characteristics at the individual level. Contrary to much previous scholarship, democracy does not exhibit a significant association with diarrhea or malnutrition. Instead, gross domestic product (GDP) per capita and improved sanitation and water have substantial effects on child health at the country level. At the individual level, household wealth and maternal education have the largest health-enhancing impact on child diarrhea and malnutrition. Furthermore, the size and strength of the relationship between wealth and health does not vary by political regime. These results demonstrate the enduring importance of socioeconomic status regardless of political context, and they support a small but growing literature that calls the democracy–health link into question.

Keywords

Child health cross-national democracy health inequality multilevel

The past hundred years are not lacking in important events. Nevertheless, among the great variety of developments that have occurred in the twentieth century, I did not, ultimately, have any difficulty in choosing one as the preeminent development of the period: the rise of democracy. This is not to deny that other occurrences have also been important, but I would argue that in the distant future, when people look back at what happened in this century, they will find it difficult not to accord primacy to the emergence of democracy as the preeminently acceptable form of governance.

(Amartya Sen, 1999a: 3–4)

Historically, there has never been a famine in a functioning multi-party democracy, and Nobel Prize winner Amartya Sen asserts that this relationship is causal, not simply coincidental. Without elections, opposition parties, and uncensored public criticism, ruling groups do not suffer political consequences from their failure to prevent famines. However, ruling groups in a democracy face penalties from such failures, and the threat of those penalties gives leaders the necessary incentives to prevent them (Sen, 1999b). Thus, for Sen, the global expansion of democracy is important because of its potential to enhance well-being. The rise of democracy across the world is indeed a remarkable feature of the twentieth century. Among the world’s 195 polities, 125 are classified as electoral democracies (Puddington, 2015). Many politicians and policy makers share Sen’s optimism that this growth of democracy will foster peace and security, economic development, human rights, and general well-being and prosperity (US State Department, 2011).

Scholars share this optimism also, as evidenced by the proliferation of research documenting the positive effects of democracy in the last several decades across multiple academic disciplines (e.g. Besley and Kudamatsu, 2006; Brown, 1999; Ghobarah et al., 2004; Klomp and De Haan, 2009; Lake and Baum, 2001). Despite recent critiques of the literature (Ross, 2006), the paramount theme throughout much of this scholarship is that democracy improves health and well-being, both directly and indirectly through a variety of intervening mechanisms. Franco et al. (2004) go so far as to claim that

The way societies organise themselves through their political regimes and their egalitarian policies could have a more important role in health than structural variables such as wealth … Increasing democratisation may be a way to counteract the deleterious effect on health of the unequal distribution of economic resources on a global scale. (p. 1421)

This study examines the effects of democracy on child diarrhea and malnutrition across 52 developing countries. It contributes to extant cross-national scholarship on the democracy–health link in several key ways. First, using a multilevel modeling strategy, it examines the country-level effects of democracy on individual-level child health, while also controlling for socioeconomic status (SES) and other household characteristics that are important but typically omitted in cross-national democracy research. This analytic strategy also allows for the possibility that the effect of SES on health varies as a function of democracy. Second, previous scholars call for a more rigorous test of the democracy–health link using multi-country health surveys and more specific measures of physical morbidity (Safaei, 2006). Using data from a collection of standardized, population-based surveys in developing countries, this study answers that call. I assess two indicators of child health: diarrhea and malnutrition. These variables are derived from height and weight measurements and parental reports collected by carefully trained survey teams in the field (Sommerfelt and Stewart, 1994). As such, they are more valid and reliable than the more commonly used country-level indicators of infant mortality and life expectancy (Scanlan, 2004). Third, much previous literature includes large samples of countries at various stages of economic development. The sample for this analysis is limited to developing countries only so that results are not biased by rich countries that also tend to be more democratic. This approach is more appropriate since developed and developing countries have different pasts and trajectories. Therefore, traditionally accepted explanatory variables might be more fitting for rich versus poor countries (Viterna et al., 2008).

How does democracy improve well-being?

The essential feature of democracy, and the one that links most clearly to health, is contestation (Dahl, 1971). Democratic governments permit regular competition among divergent values and interests, and contestation occurs when an opposition party has some chance of winning political office as a result of elections (Alvarez et al., 1996). Indeed, the core principle of democracy is that the current in-group could be the out-group at the next election (Przeworski et al., 2000). Because of this, politicians must develop platforms that attract adherents in order to maximize their chances of election (or re-election; Lake and Baum, 2001). Elected officials listen to what their constituents want and need because they must face their criticism and seek their support (Sen, 1999b). Widespread misery is unpopular, so democratically elected political leaders concern themselves with human development more so than autocrats (Gerring et al., 2012). Citizens who live in a democracy can penalize their political leaders if they fail to adequately protect the population from preventable illness and premature mortality (Wigley and Akkoyunlu-Wigley, 2011).

Democracy enhances health, then, by creating an avenue for people to press their government to make public investments in social services (Goldsmith, 1986; Sorensen, 1991). Politicians gather information about the preferences of the general population through voting (Frey and Al-Roumi, 1999). Policy preferences then impact the size and scope of public policies, particularly in regard to social spending (Brooks and Manza, 2006). In a democracy, the costs of taking action to bring about a change in leadership are low, which means that politicians are tightly constrained by the preferences of the majority electorate (Lake and Baum, 2001). Thus, democratic governments devote more resources toward basic needs and social services such as education and health care (Nelson, 2007). This is because competitive elections keep politicians accountable to a very broad populace. In contrast, autocrats serve the narrow interests of a much smaller group, such as the military or the wealthy elite (Besley and Kudamatsu, 2006). Because democratically elected leaders must satisfy such a broad populace, they ensure that health-enhancing public goods will reach a wide segment of the population (Wigley and Akkoyunlu-Wigley, 2011).

From ancient Greece until today, scholars equate democracy with a political space that is favorable to the needs of the disadvantaged. But the logic of this claim rests on the axiomatic assumption that democracies are a mechanism for redistribution (Gerring et al., 2012). The electoral incentives of a democracy work not only to raise spending on public goods that will benefit the health of the population but also to increase efficiency and reallocate resources toward areas that politicians believe will win them more votes (McGuire, 2006). Political elites engage in competitive struggles for votes, which gives them an incentive to seek out previously ignored voters in order to be elected. As candidates broaden their support base, they provide more social services in order to garner the support of the poor and interest groups who advocate pro-poor policies (Gauri and Khaleghian, 2002; Wigley and Akkoyunlu-Wigley, 2011). Early modernization theorists posit that over time, the more marginalized members of society organize into interest groups, develop an electoral base, and gain power in the legislature. As political power shifts from the elite minority to the middle and lower class majority, inequality declines in developing countries (Lenski, 1966; Lipset, 1959). In turn, democracies foster a culture of equality that empowers every citizen, including the very poor, to demand that the government satisfies their rights and needs (Gerring et al., 2012).

Evidence for the democracy–health link

Most cross-national studies regarding this topic in the past several decades propose that democracy improves health and overall welfare. The typical approach is to demonstrate the statistical significance of democracy in regression models that control for economic development and other relevant variables. For example, several studies based on analytic samples that include both developed and developing countries show that democracy is a strong predictor of improvements in physical quality of life and the provision of basic needs (e.g. Frey and Al-Roumi, 1999; London and Williams, 1990; Moon, 1991; Moon and Dixon, 1985; Young, 1990). It is also more commonly associated with improved infant mortality and life expectancy (e.g. Besley and Kudamatsu, 2006; Klomp and De Haan, 2009; Lake and Baum, 2001; Moore et al., 2006; Przeworski et al., 2000; Safaei, 2006; Wigley and Akkoyunlu-Wigley, 2011; Zweifel and Navia, 2000). Some even claim that the effects of democracy on life expectancy and infant mortality surpass those of economic development (Franco et al., 2004). Although less frequent, other scholars focus their analyses solely on samples of developing countries. The results are similar: democracy is consistently associated with longer life expectancy and lower infant mortality rates (e.g. Lena and London, 1993; Tsai, 2006; Wickrama and Mulford, 1996).

There are comparatively few studies on the relationship between democracy and other measures of child health in developing countries. Jenkins and Scanlan (2001) find that the expansion of political and civil rights reduces child malnutrition. Smith and Haddad (2000) similarly claim that if all countries were able to reach high levels of democracy, the worldwide prevalence of underweight children would decline by 29.4 million. Klomp and De Haan (2009) find beneficial effects of democracy on a composite-dependent variable that includes prevalence rates for HIV, tuberculosis (TB), respiratory infection, child malnutrition, diarrhea, and immunizations. Other than these, extant research on the relationship between democracy and health most commonly operationalizes health as infant mortality or life expectancy.

Of course, not every scholar finds a strong relationship between democracy and health. Indeed, this relationship is called into question by several more recent cross-national studies that do not find direct effects of democracy on infant mortality (Ross, 2006; Shandra et al., 2012; Swiss et al., 2012) or other health outcomes (Burroway, 2012; Scanlan, 2010; Wejnert, 2008). This may be due to sample selection biases (Ross, 2006) or lack of longitudinal analyses (Swiss et al., 2012) in previous research. Disparate findings could also be due to the tendency to analyze developed and developing countries together in one sample. Wejnert (2008) finds that democracy improves women’s health in developed but not developing countries. Others similarly echo the call to separate developed from developing countries in cross-national analyses (Viterna et al., 2008). Despite these critiques, the dominant theme of the literature is that democracy is good for health.

Evidence for the redistributive hypothesis is inconclusive. Many cross-national empirical studies find support for an inverse relationship between democracy and income inequality as Lipset and Lenski predict (e.g. Huber et al., 2006; Lee, 2005; Moon, 1991; Muller, 1988; Reuveny and Li, 2003). However, many others disconfirm the association (for reviews, see Gradstein and Milanovic, 2004; Sirowy and Inkeles, 1990). At the same time, another body of research examines whether or not democracies increase social spending and provide more equitable social services than autocracies and whether or not this fosters better health. Again, evidence is mixed (e.g. Besley and Kudamatsu, 2006; Gauri and Khaleghian, 2002; Ghobarah et al., 2004; Lake and Baum, 2001). Some scholars suggest that democracies do favor expansion of social spending but not necessarily broader allocation of resources (see Nelson, 2007, for a review). The analysis that follows contributes to these literatures by determining whether democracy moderates the effect of household wealth. If democratic governments empower the disadvantaged, foster equality, and provide more health resources for everyone, then personal wealth should be less consequential for purchasing good health under a democratic regime.

Measuring democracy

The attempt to accurately measure democracy has a long history. Scholarship in the late 1950s and1960s used objective measures such as the composition of legislative bodies or voter turnout statistics. However, recent research relies much more heavily on subjective measures that are determined by expert judges who assign ratings to countries (Bollen and Paxton, 2000). Certain widely used datasets such as Freedom House and Polity are almost canonical (Herrera and Kapur, 2007). But several scholars also write extensively on the distinct challenges of measuring democracy, how such measures are conceptualized, and the inherent measurement errors and weaknesses that arise from subjective indices (e.g. Bollen, 1980, 1990, 1993; Munck and Verkuilen, 2002a, 2002b). My goal is not to reiterate the debate surrounding the technical measurement of subjective democracy indices. Instead, I review several of the ways in which these indicators are typically used and then empirically evaluate their relationship with child diarrhea and malnutrition.

Perhaps, the most common method of operationalizing democracy is to use a continuous scale that represents a country’s level of democracy at a certain point in time. This considers democracy as a measure of degree and addresses questions about whether different increments are associated with improved health (Brown, 1999). Others find it more useful to treat regime type dichotomously. As Zweifel and Navia (2000) put it, ‘As in the proverbial case of pregnancy, countries are either democratic or they are not’ (p. 103). In this case, democracy is not considered as a matter of degree. Rather, the condition of being democratic (vs authoritarian) is what matters. In addition, because countries are often concentrated in the upper or lower extremes of continuous democracy indices, some researchers dichotomize such measures due to the distribution of the data (Brown, 1999).

The use of cross-sectional levels of democracy makes an implicit assumption that democracy is related to health, regardless of the length of time that democratic regimes exist in a particular country. Democracy scores at one point in time ignore the fact that democracy could incrementally affect health in the long term (Muller, 1988). Moreover, newer democracies may not perform as well as those that are long established (McGuire, 2006). Measures that reflect the stock of democracy over time reflect the possibility that a cumulative history of democracy might be more important for influencing health (Ross, 2006). Democracies and autocracies create deep legacies that extend back for decades, perhaps longer. The effects of these legacies are cumulative and unfold over time (Gerring et al., 2012). Thus, an average democracy score over time or a count of the years a regime has been democratic may be a better way to capture the democracy’s influence. Additionally, the effect of democracy could be nonlinear over time. It is possible that over many years, the benefits of democracy could eventually reach a plateau and then diminish (Ross, 2006). Such nonlinearities could be captured by the quadratic of the average democracy score over time (Lake and Baum, 2001).

In addition to its direct effects, democracy may also moderate the relationship between household wealth and child health. Redistributive theories of democracy claim that democratically elected officials allocate resources in ways that benefit a broad segment of the population, including (or perhaps especially) the poor. If this is the case, then household wealth should be less consequential for health in a democratic context because everyone gains from the expansion of health-enhancing social welfare services. In the absence of democracy, private wealth would be more important for purchasing health. A cross-level interaction between democracy (country-level) and wealth (individual-level) would test the proposition that the effect of household wealth on child health varies according to political regime. If wealth has a negative affect on diarrhea and malnutrition, then the interaction term should be positive. That is, the interaction term will have a dampening effect on the slope of household wealth.

Methods

Data and sample

I draw individual-level data (level 1) from the Demographic and Health Surveys (DHS), a collection of nationally representative, population-based surveys conducted in developing countries (Macro, 2009). Following convention, I collect data on democracy from the Polity IV Project (Marshall and Jaggers, 2010). I use this dataset for several reasons. First, the indicators of democracy created for Polity are most ideally suited for measuring contestation (Lake and Baum, 2001). The Polity IV Project conceptualizes democracy as the presence of institutions and procedures through which citizens are able to effectively convey their preferences about alternative leaders and policies (Marshall and Jaggers, 2010). Second, Polity stands out among the available democracy datasets because of its transparent and detailed coding rules, wide range of cases, and long historical time frame (McGuire, 2006; Munck and Verkuilen, 2002a). Third, Polity is the ‘industry standard’ and is highly correlated with other democracy measures (Gerring et al., 2012; Wigley and Akkoyunlu-Wigley, 2011). I collect data on other country-level variables from the World Bank (2010).

The sample includes countries with available data on the dependent variables collected during DHS phases 3 through 5 (1995–2008) for comparability in time and content. Thus, the analysis pools data on 292,204 children in 52 developing countries. (See Appendix 1 for a list of the countries included.)¹ One limitation of the DHS is that data are only available for developing countries. However, by limiting the sample to developing countries, this study answers the call of previous scholars to separate developed from developing countries in cross-national analyses of the effects of democracy (Viterna et al., 2008; Wejnert, 2008). Much of the research on the democracy–health link uses analytic samples of both developed and developing countries combined. Yet, the positive association between democracy and economic development is well established (Przeworski et al., 2000). Consequently, developed countries that are also highly democratic could be biasing the democracy coefficient and driving the regression line. Even controlling for level of economic development would not eliminate that bias, since the two are conflated. Regression lines are sensitive to influential observations that, when removed, can change the line considerably (Moore and McCabe, 2006). Thus, the removal of highly developed, highly democratic countries from the sample could substantially affect the slope of the regression line.

Furthermore, because developed and developing countries have different pasts and future trajectories, different sets of variables may be more appropriate for explaining well-being in the two sets of countries. To determine whether the factors that matter in developing countries are different than those that matter to developed countries, scholars must test theories of democracy in separate samples (Viterna et al., 2008). Democracy does not work in the same fashion in advanced industrialized countries as it does in developing countries (Matland, 1998; Wejnert, 2008). Economically affluent regions devise incentives and protective strategies that sustain democratic regimes, whereas democracy is less institutionalized and resilient in less developed regions. For these reasons, cross-national analyses that incorporate the entire world create distorted views about patterns of democracy (Wejnert, 2005). Although scholars typically give universal explanations of the patterns between democracy and health, theoretically, the results of a study on the entire world could be markedly different than the results of a study on a subset of the world (Viterna et al., 2008).

Estimation technique

While previous research finds important effects of democracy on health, it is limited to country-level associations. Such ecological-level studies cannot conclude that the findings would still hold after taking into account the compositional characteristics of countries (McTavish et al., 2010). Scholars have yet to examine whether democracy is associated with child health after adjusting for individual-level characteristics. I address this methodological issue with a series of hierarchical generalized linear logit models (HGLMs) estimated with the HLM 6.08 software developed by Raudenbush et al. (2004). These models predict the odds that a child has had a recent episode of diarrhea or is underweight based on a set of both individual- and country-level explanatory variables. The advantage of this technique is that the net effects of one level can be estimated while controlling for variation in the other level. That is, this estimation technique assesses the relationship between democracy and child health, while controlling for variation in maternal and household characteristics within countries. Ignoring the nesting of children within countries and including country-level variables violates the assumption of independent standard errors and inflates the risk of a type 1 error. However, hierarchical analysis provides unbiased and efficient estimates of the coefficients, as well as proper standard error estimates (Guo and Zhao, 2000; Raudenbush and Bryk, 2002).

Multi-level models can be explained in two steps. First, at level 1, the log-odds of being stunted or having a recent episode of diarrhea [log (p_ij/1 − p_ij)] for the ith child in the jth country is expressed as a function of country intercepts (β_0j), a set of fixed individual-level characteristics (βX_ij), and an error term (r_ij)

\log (\frac{p_{i j}}{1 - p_{i j}}) = β_{o j} + β X_{i j} + r_{i j}

Second, at level 2, the parameters from the first step become the dependent variables and are regressed on a set of country-level predictors. Each country intercept (β_0j) and the slopes of the individual-level characteristics (β_ij) are expressed as a function of a general intercept term (γ_0j), a set of country-level characteristics (γC_j), and an error term (ε_0j)

β_{o j} = γ_{0 j} + γ C_{j} + ε_{0 j}

(1)

β_{i j} = γ_{i j} + γ C_{j} + ε_{i j}

(2)

Equation (1) represents the random intercept component of the model that estimates the effects of democracy on child health, while holding the individual-level characteristics constant. Equation (2) represents the random coefficient component of the model that estimates the effect of democracy on the slope of household wealth (cross-level interactive effect). A positive interaction term would indicate that democracy dampens the negative slope of wealth on health. This would suggest that the effect of wealth is weaker in a democracy. If democratic regimes are truly distributional and provide public services that benefit the health of all, then private wealth would not be as consequential for purchasing better well-being.

One limitation of multilevel models is that they do not control for unobserved country-level factors in the same manner as a traditional fixed effects model. This is problematic because omitted group-level variables can potentially bias the individual-level slope estimates (Chaplin, 2003). I deal with this problem using group-mean centering: individual-level variables are differenced from their country means at level 1. This produces the most accurate slope estimates for the individual-level variables, even in the presence of unobserved country-level variables that are correlated with the individual-level variables (Chaplin, 2003; Raudenbush and Bryk, 2002). Group-mean centering also has a theoretical advantage. It implies that an individual’s relative position within a country influences the outcome. This is appropriate for cross-country comparisons in which levels of wealth or education, for example, may have different value depending on context (Enders and Tofighi, 2007).

I examine correlation matrices in order to assess possible multicollinearity among variables (see Appendices 3 and 4). Because some of the country-level development variables are moderately correlated, I further test for multicollinearity at level 2 by estimating variance inflation factors (VIFs). None of the VIFs are higher than 5, indicating that multicollinearity is not problematic in this analysis (Studenmund, 2001). I assess the potential influence of outlier countries using Cook’s distance statistics, and results remain consistent (Fox, 1997). I employ histograms to examine the distribution of variables, and use the natural log to correct for skewness when a variable violates the assumption of normality (details below). Finally, as is conventional in cross-national research (e.g. Burroway, 2015; Shandra et al., 2012), I estimate the models with robust standard errors to correct for heteroskedasticity (Raudenbush et al., 2004).

Measures

Dependent variables

The analysis includes two binary measures of child health. Diarrhea is reported by the child’s mother. Following Hatt and Waters (2006), a child is coded 1 for having diarrhea if she/he is ill with diarrhea at any time during the 2 weeks preceding the interview. Of course, the accuracy of the diarrhea variable is affected by the reliability of the mother’s recall of the illness episode, but a 2-week period ensures that recall errors will not be serious (Macro, 2009). Diarrhea is commonly caused by various bacteria, viruses, and parasites. This could be due to lack of access to safe water for drinking/cooking/cleaning, sanitary living conditions, proper sewage, hygienic food storage and preparation, and adequate nutritional resources (World Health Organization (WHO), 2013).

Malnutrition is calculated using anthropometric measures of height and weight, which provides a good indication of health status that is independent of maternal perceptions (Hill and Upchurch, 1995). A child is malnourished if she/he is underweight, defined as low weight-for-age. A child is coded 1 for being underweight if she/he is more than two standard deviations below the median of the WHO (2008) child growth standards for his/her height, age, and gender. The prevalence of underweight children is the most common indicator of malnutrition.² Many factors contribute to malnutrition, including food deprivation, lack of micronutrients, and infections and illness (World Bank, 2010).

One advantage of analyzing diarrhea and malnutrition is that these measures are more valid and reliable than the more commonly used country-level measures of infant mortality and life expectancy, which are based on estimates from a combination of vital registration systems, sample surveys, and censuses. Complete vital registration systems are unavailable in most developing countries. Data collected by countries vary by source and method for any given time and place, which makes cross-country comparison extremely difficult. For this reason, infant mortality and life expectancy are derived from indirect estimation techniques that use all available information and attempt to reconcile differences among sources. Infant mortality estimates are sometimes based on extrapolations from outdated surveys that refer to an earlier reference period. Annual life expectancy estimates are interpolated from 5-year period data. In both cases, this means that they do not reflect true events as much as observed data (UNICEF, 2014; World Bank, 2010).

Because incidence data on disease is often unavailable, mortality rates are used as a proxy to identify vulnerable populations (World Bank, 2010). But ideally, scholars would use data on specific physical illnesses derived from standard population surveys that are internationally comparable (Safaei, 2006). Indeed, this is the most significant contribution of the DHS. DHS questionnaires and data collection procedures are standardized to ensure they are comparable across countries. There are a number of other features of the DHS that ensure the estimates accurately represent the true health situation in developing countries, including coverage that is national in scope, high participation rates (typically over 90%), and high-quality interviewer training (Corsi et al., 2012). Carefully trained survey teams follow standardized guidelines in physically weighing and measuring children in the field. Training includes classroom instruction, as well as practice field experience and quality control tests to ensure proficiency (Sommerfelt and Stewart, 1994). Thus, in using DHS indicators for child diarrhea and malnutrition, this study answers the call for more rigorous tests of the democracy–health link using specific measures of physical morbidity and standardized multi-country population surveys (Safaei, 2006).

Aside from the validity and reliability of anthropometric surveys, diarrhea and malnutrition are theoretically meaningful because they are at the core of a nation’s quality of life (Scanlan, 2004). Together, diarrhea and malnutrition are the two leading causes of death among children under the age of five worldwide. Given that both are preventable and treatable, these childhood illnesses indicate the extent of fundamental deprivation in a population (UNICEF, 2014; WHO, 2013). The prevalence of such childhood illnesses suggests greater food security and development concerns in the population, since a society whose children are in jeopardy is one that is most likely facing many other challenges of poverty and underdevelopment as well (Scanlan, 2004). Understanding the patterns of inequality in child diarrhea and malnutrition not only affects children in the present, but it also has long-term consequences for the future. Adults who have suffered from poor health and malnutrition in childhood are less physically and intellectually productive and experience more chronic morbidity (Blackwell et al., 2001; Smith and Haddad, 2000). Healthy children, however, are generally more likely to develop cognitive, emotional, and social skills and, subsequently, succeed in school and society. This makes child health a fundamental underpinning of development in developing countries because it shapes how nations progress (UNICEF, 2007).

Democracy

The analysis that follows uses four indicators of democracy to reflect different ways of measuring contestability. First, I measure level of democracy using a democracy score from the Polity IV Project.³ This variable combines coding on the competitiveness and regulation of political participation, openness and competitiveness of executive recruitment, and constraints on the chief executive or lack thereof. The democracy score is an interval variable that ranges from −10 (high autocracy) to 10 (high democracy).⁴ Second, I include a dichotomous measure of democracy in order to consider the possibility that the condition of being democratic may matter more for child health than the degree. A country is coded as democratic if it has a democracy score of 1 or above and coded as 0 otherwise.⁵ This separates the sample into countries that are either democratic or autocratic.⁶

Finally, I use the Polity dataset to create two indicators of a country’s long-term stock of democracy. I measure history of democracy as the proportion of years a country has been democratic since 1960. ‘Democratic’ is defined dichotomously as above. These dichotomous scores are summed and then divided by the number of years for which data are available since 1960 (including the year of observation). Other scholars (e.g. Ross, 2006) simply sum the years a country has been democratic since 1900. I use a proportion both because the year of observation varies slightly in the sample and because the number of years for which data are available also varies. Furthermore, I use 1960 because that is the earliest year for which data are available for many sub-Saharan African countries in the sample that were newly gaining independence around that time.⁷ History² is the quadratic form of this variable and is meant to capture any nonlinear effects of democracy over time.

Development

As is standard practice, I control for level of economic development in all models. Prominent development scholars stress that economic development is the primary concern and the most effective way to improve well-being in developing countries (i.e. Collier, 2007; Sachs, 2005). Economic development improves health by increasing the food supply, the availability of doctors, and educational expansion (Buchmann and Hannum, 2003; Cooper et al., 2003; Jenkins and Scanlan, 2001). Increased wealth also tends to correspond to higher standards of living and more advanced medical technology (Shandra et al., 2012). I measure economic development as gross domestic product (GDP) per capita in purchasing power parity dollars (World Bank, 2010) and use the natural log to correct for its highly skewed distribution.

Other aspects of development may be equally, if not more, important than GDP per capita for child diarrhea and malnutrition. Inadequate access to clean water and sanitation facilities accounts for a substantial amount of the burden of illness and death in developing countries. About half of the developing world lacks even a simple latrine and about one-sixth lacks clean water (WHO and UNICEF, 2004). Contaminated water sources provide habitats for mosquitoes and other organisms that cause malaria and parasitic diseases, and consuming contaminated water results in a number of water-borne illnesses that cause diarrhea. Moreover, without clean water and appropriate sanitation systems, diseases spread rapidly. Most endemic diarrhea is transmitted from person to person because of inadequate hygiene (WHO, 2003). Thus, from an epidemiological perspective, lower rates of access to sanitation and water put everyone at risk of illness. Recent cross-national analyses demonstrate that improving access to water and sanitation substantially reduces infant and child mortality, even adjusting for known confounding variables (Burroway, 2015; Cheng et al., 2012). Data on the percentage of the population with access to an improved sanitation facility and the percentage of the population with access to an improved water source are drawn from the World Bank (2010). Due to their multicollinearity, I follow recent cross-national research and measure improved sanitation and water as the average of these two indicators (Burroway, 2015; Shandra et al., 2012).⁸

SES and other household characteristics

In addition to the country-level political and development context, each model includes a number of individual-level variables that are important for child health: household wealth, mother’s education, mother’s employment status, urban residence, mother’s age, household head’s sex, child’s age, child’s sex, and household size. SES is commonly operationalized as household wealth and mother’s education – the two most well-studied correlates of child health in developing countries at the individual level (Bollen et al., 2001). Household wealth is measured as a composite index that represents the cumulative living standard of a household. Following previous research (Heaton et al., 2005a), this index is calculated as the percentage of household items (including clean water, flush toilet, radio, television, electricity, refrigerator, bicycle, motorcycle, car, telephone, and finished floor) present in the home.⁹ Higher SES allows families to prevent illness or minimize the consequences once it occurs (Link and Phelan, 1995). For example, wealthier families are better able to provide health care, medication, appropriate clothing, bed nets, food, and a clean, safe place to live (Heaton et al., 2005a). On one hand, diseases and treatments may change over time, but the association between wealth and health remains because people with more resources will always use them to garner a health advantage (Link and Phelan, 2002). On the other hand, poverty limits the distribution and effectiveness of health-enhancing technologies, thus shaping vulnerability to disease (Stratton et al., 2008).

Mother’s education is measured as a series of categorical variables including primary, secondary, and higher, with no education as the reference group. On a basic level, maternal education improves child health through general access to information and greater health knowledge (Thomas et al., 1990). As a result, educated women are more likely to interact effectively with health-care providers, comply with treatment regimens, and break from tradition in adopting newer innovations in nutrition (Smith and Haddad, 2000). But beyond this, education brings women more respect in the household and in the community, strengthening their bargaining power and putting them in a better position to make decisions regarding childcare (Nussbaum, 2003). Education also gives women more opportunities to work outside the home and earn an income, which gives them greater control over resources in the household (Sen, 1999b).

Mother employed is a dichotomous variable, coded as 1 for currently employed and 0 for unemployed. Gainful employment outside the home brings women an independent wage and more influence in the allocation of household resources (Heaton et al., 2005b; Sen, 1999b). Urban residence is included as a binary variable, coded 1 for urban and 0 for rural. Infant mortality tends to be lower in urban areas, urban women tend to have greater access to health care, and urban families tend to have higher SES (Heaton et al., 2005a). Mother’s age is measured in years. Increasing maternal age is associated with improved child health and well-being because older mothers are healthier themselves and tend to have higher levels of educational attainment (Barclay and Myrskyla, 2016; Sutcliffe et al., 2012). A dichotomous variable indicates the presence of a female household head (coded as 1 for yes and 0 for no). Across multiple cultural contexts, female-headed households are much more likely to experience poverty than other households (Buvinic and Gupta, 1997). This suggests that female-headed households are at greater risk for child illness as well.

Child’s age (measured in months) and child’s sex (coded as 1 for male, 0 for female) are also included. Children are particularly susceptible to being underweight during the weaning and post-weaning period (6–24 months; WHO, 2011). Thus, malnutrition should decline with age. There is some evidence that the incidence of diarrhea declines with age as well (Walker et al., 2012). Some research claims that gender-biased allocation of food and health services results in higher illness and mortality rates among females, especially where there is a strong preference or need for sons (Hossain et al., 2007; Thomas, 1994). Household size is measured as total number of persons living in the household. A child’s health can be compromised in large families where resources are shared among many, and competition for food is greater. Plus, the risk of unsanitary conditions and disease spread is heightened when many people live in close proximity (Heaton et al., 2005a; LeGrand and Phillips, 1996). Descriptive statistics for all of the variables used in the analysis are provided in Appendix 2. Country-level and individual-level correlation matrices are displayed in Appendices 3 and 4, respectively.

Results

Table 1 assesses the relationship between democracy and child diarrhea in a series of HGLMs. Each model displays one measure of democracy at a time, while controlling for economic development, sanitation/water, and characteristics of the child’s household. Models 1 through 4 show the direct effects of democracy. Model 5 displays the cross-level interaction between democratic and household wealth. Table 2 replicates the analysis for child malnutrition. Each cell in these tables displays odds ratios (ORs) and standardized ORs (in bold and italics) in order to compare effect sizes across variables.¹⁰

Table 1.

HGLMs of child diarrhea.

	Model 1	Model 2	Model 3	Model 4	Model 5
Country level (N = 52)
Democracy score	1.005
	1.071
Democratic (vs autocratic)		1.109			1.016
		1.053			1.046
History of democracy			0.828
			−1.170
History²				0.769
				−1.277
Improved sanitation/water	0.991*	0.991*	0.991*	−0.992	0.991*
	−1.655	−1.661	−1.622	− 1.593	−1.659
GDP per capita	1.031	1.029	1.039	1.048	1.028
	1.075	1.070	1.096	1.118	1.069
Individual level (N = 292,204)
Household wealth	0.995***	0.995***	0.995***	0.995***	0.996***
	−1.443	−1.443	−1.443	−1.443	−1.329
Mother’s education
Primary	1.006	1.006	1.006	1.006	1.006
	1.017	1.017	1.017	1.017	1.016
Secondary	0.918	0.918	0.918	0.918	0.919
	−1.269	−1.269	−1.269	−1.269	−1.263
Higher	0.690***	0.690***	0.690***	0.690***	0.692***
	−2.801	−2.801	−2.802	−2.802	−2.783
Mother employed	1.086***	1.086***	1.086***	1.086***	1.087***
	1.259	1.259	1.259	1.259	1.262
Urban residence	1.034	1.034	1.034	1.034	1.033
	1.097	1.097	1.096	1.096	1.095
Mother’s age	0.989***	0.989***	0.989***	0.989***	0.989***
	−1.230	−1.230	−1.230	−1.230	−1.230
Female household head	1.066***	1.066***	1.066***	1.066***	1.066***
	1.193	1.193	1.193	1.193	1.194
Child’s age	0.977***	0.977***	0.977***	0.977***	0.977***
	−2.970	−2.970	−2.970	−2.970	−2.970
Male	1.109***	1.109***	1.109***	1.109***	1.110***
	1.334	1.334	1.334	1.334	1.335
Household size	1.005	1.005	1.005	1.005	1.005
	1.047	1.047	1.047	1.047	1.046
Cross-level interaction
Democratic × wealth					0.998
					−1.004

GDP: gross domestic product.

Each cell contains odds ratios and standardized factor changes in bold and italics.

p < 0.05, **p < 0.01, ***p < 0.001 (one-tailed tests).

Table 2.

HGLMs of child malnutrition.

	Model 1	Model 2	Model 3	Model 4	Model 5
Country level (N = 52)
Democracy score	0.991
	−1.137
Democratic (vs autocratic)		0.988			0.995
		−1.035			−1.014
History of democracy			1.347
			1.282
History²				1.344
				1.259
Improved sanitation/water	0.978***	0.978***	0.977***	0.977***	0.978***
	−3.447	−3.424	−3.543	−3.579	−3.427
GDP per capita	0.679***	0.683***	0.670***	0.668***	0.683***
	−2.522	−2.487	−2.601	−2.623	−2.486
Individual level (N = 292,204)
Household wealth	0.984***	0.984***	0.984***	0.984***	0.982***
	−3.063	−3.063	−3.063	−3.063	−3.325
Mother’s education
Primary	0.778***	0.778***	0.778***	0.778***	0.778***
	−2.006	−2.006	−2.006	−2.006	−2.009
Secondary	0.585***	0.585***	0.585***	0.585***	0.584***
	−4.431	−4.431	−4.431	−4.431	−4.455
Higher	0.406***	0.406***	0.406***	0.406***	0.405***
	−12.263	−12.263	−12.258	−12.257	−12.344
Mother employed	1.001	1.001	1.001	1.001	1.001
	1.003	1.003	1.003	1.003	1.002
Urban residence	0.914*	0.914*	0.914*	0.914*	0.915*
	−1.283	−1.283	−1.283	−1.283	−1.281
Mother’s age	0.998	0.998	0.998	0.998	0.998
	−1.039	−1.039	−1.039	−1.039	−1.039
Female household head	0.943***	0.943***	0.943***	0.943***	0.943***
	−1.178	−1.178	−1.178	−1.178	−1.178
Child’s age	1.016***	1.016***	1.016***	1.016***	1.016***
	2.156	2.156	2.156	2.156	2.156
Male	1.132***	1.132***	1.132***	1.132***	1.132***
	1.410	1.410	1.410	1.410	1.410
Household size	1.014***	1.014***	1.014***	1.014***	1.015***
	1.152	1.152	1.152	1.152	1.153
Cross-level interaction
Democratic × wealth					1.001
					1.004

GDP: gross domestic product.

Each cell contains odds ratios and standardized factor changes in bold and italics.

p < 0.05, **p < 0.01, ***p < 0.001 (one-tailed tests).

Table 1 shows that the democracy score (model 1) and the dummy variable for being democratic (model 2) are slightly positive. However, neither of these variables reaches significance. History of democracy (model 3) and the quadratic term (model 4) are both negative, but they do not reach significance either. Many scholars posit that contestation allows citizens to express their policy preferences and keep politicians accountable to their will, which ultimately translates into better population health and well-being. However, neither current level of democracy nor a long-term history of democracy exhibits the robust relationship with child health that others report. Furthermore, model 5 does not provide any evidence that the slope of household wealth varies as a function of political regime. Democracy does not seem to moderate the effect of household wealth on health.

Nevertheless, Table 1 suggests that a number of other factors improve child health. At the country level, sanitation and water have a substantial effect on diarrhea. For every standard deviation increase in sanitation/water, the odds of diarrhea decline by 1.6–1.7 standard deviations. At the individual level, the largest contributors to improved child health are child’s age, household wealth, and highly educated mothers. For every standard deviation increase in child’s age, the odds of diarrhea decrease by almost 3 standard deviations. For mothers with higher education, the odds of child diarrhea decline by about 2.8 standard deviations. A standard deviation increase in household wealth translates into a decrease in the odds of diarrhea by 1.4 standard deviations. As expected, older mothers have healthier children, but female-headed households experience a health disadvantage. Contrary to expectation, the odds of diarrhea are higher for male children and children of employed mothers. Employment outside of the home may produce contradictory effects on child health because it reduces the time available for childcare (Bhattacharya, 2006). It is also possible that female employment in developing countries reflects the worse socioeconomic position of particularly impoverished women who are forced to work (Drovandi and Salvini, 2004).

The main findings from Table 2 are similar to those of Table 1, with some slight fluctuations. All four measures of democracy remain insignificant. Sen claims that competitive elections ensure food security because political leaders are accountable to the populace and, therefore, provide for their basic needs. Therefore, democracy should reduce the incidence of malnutrition. However, the results here show that democracy does not have a statistically significant association with child malnutrition. The cross-level interaction term is also nonsignificant. This means that whether a regime is democratic or autocratic, the direction and strength of the relationship between household wealth and child health remains the same. However, other country-level variables are highly consequential for child health. Economic development and the provision of sanitation and water are associated with reducing the odds of a child being underweight.

At the individual level, the effects of mother’s education and household wealth are the largest in magnitude. For every standard deviation increase in household wealth, the odds of being underweight decline by about 3 standard deviations. The beneficial impact of mother’s education rises with every level of attainment. Primary education reduces the odds of being underweight by about 2 standard deviations, secondary education by 4.4 standard deviations, and higher education by 12.3 standard deviations (compared to mothers with no education).

Table 2 also shows that larger households and female-headed households incur some health disadvantages, but living in an urban area seems to bring health benefits. Comparing Table 1 to Table 2, the effects of child’s age are mixed. As children age, their odds of diarrhea decline but their odds of being underweight increase. This could partially reflect the difference between chronic and acute health conditions (Cameron and Williams, 2009). Older children require more nutrition to support growth, thus older children who do not receive appropriate nutrients could be more susceptible to long-term, chronic malnutrition. At the same time, because the immune system develops over time, older children might have better immunities to protect against bacteria and viruses that cause diarrhea. Male children have higher odds of a recent episode of diarrhea and higher odds of being underweight. Some scholars using the DHS similarly find that girls are less likely to be malnourished than boys (Hill and Upchurch, 1995), and others find no difference (Shin, 2007). Given equal care and feeding, female children are more likely to survive than male children (Bhattacharya, 2006). Girls may also be less prone to disease episodes (Hill and Upchurch, 1995).

Summarizing the main patterns of Tables 1 and 2, the contestability of a political regime does not seem to generate the health impacts that previous literature predicts. Neither the level of democracy, the dichotomous condition of being democratic, nor a history of democracy is statistically significant. Economic development, and perhaps even more importantly, the provision of sanitation and water are more consequential for improvements in child health at the country level. At the household level, wealth and maternal education have the largest effects on child health, even net of a variety of other household characteristics. This is consistent with a substantial body of research that emphasizes household SES as the most important factor in maintaining good health (Bollen et al., 2001). Furthermore, the redistribution hypothesis suggests that democracies allocate resources in ways that improve population health. If this is the case, household wealth should be less essential for purchasing good health in a democracy because everyone benefits from the wide distribution of resources and expansion of public services. However, the cross-level interaction between democracy and wealth is not significant. Thus, the size and strength of the relationship between wealth and diarrhea/malnutrition is consistent whether the regime is democratic or autocratic. These results are robust to a number of alternative model specifications and sensitivity analyses, as described in the endnotes.

Discussion and conclusion

The worldwide trend toward rising democracy over the last several decades contributes to growing optimism that democracy might literally change the world by securing peace and deterring aggression, opening markets, promoting development, upholding human rights, protecting the environment, and improving health. Spreading democracy brought with it expectations that competitive elections would make governments more attentive and responsive to the welfare of their people, creating better services and better health. Scholarly attention to the topic grew in tandem, as researchers and policy makers alike seek to understand the consequences of this rising trend. As a result, a rather considerable body of literature claims that democracy directly affects multiple measures of well-being, particularly life expectancy and infant mortality.

However, democracy is not a guaranteed solution for solving today’s pressing health inequalities (Scanlan, 2010). The results reported here support a small but growing contingent of scholars who call the democracy–health link into question (e.g. Ross, 2006; Swiss et al., 2012; Wejnert, 2008). The analysis differs from prior literature on democracy and health in several key methodological ways. First, it provides an even more rigorous test of the effects of democracy by taking into account within-country variation in maternal and household characteristics. This is especially important given the significance of variables such as household wealth and maternal education for health improvement. Second, the analysis concentrates on developing countries only. Although some earlier studies find democracy effects on health in analyses of developing countries (e.g. Lena and London, 1993; Wickrama and Mulford, 1996), more recent studies do not find this robust direct effect (e.g. Swiss et al., 2012; Wejnert, 2008). There is less evidence for the democracy–health link specifically in developing countries. Most researchers group developed and developing countries together in their statistical analyses. But the factors that shape variation in health and democracy in the developed world are different from those in the developing world. For instance, emerging democracies in poor countries are still unstable and not as effective in redistributing resources as the more established democracies of the developed world (Huber et al., 2006; Swiss et al., 2012). Previous studies assuming that the relationship between democracy and health holds across the board should perhaps be reconsidered (Viterna et al., 2008).

Third, the findings of this study may differ from previous research because of its specific focus on child diarrhea and malnutrition. Safaei (2006) notes that the use of physical morbidity measures from standardized multi-country surveys provides a more rigorous test of the democracy–health link than the traditional ecological approach. This analysis answers that call. I draw individual-level data from population-based surveys administered by trained personnel who weigh and measure children in the field (Sommerfelt and Stewart, 1994). This makes the health indicators used here more valid and reliable than the traditional measures of infant mortality and life expectancy that are based on indirect estimates derived from multiple sources and years. Diarrhea and malnutrition are just two of the proximate causes of mortality. Although they account for a large share of infant and child mortality, there are still many other causes. It is possible that other proximate causes could be more responsive to democracy than those studied here. Democracy works, in theory, because people press the government for health services and policies that they want and need. It is possible that diarrhea and malnutrition are perceived as so commonplace that they are overlooked in favor of fighting other health crises that are perceived as more severe.

For example, in recent years, international organizations brought HIV/AIDS, TB, and malaria into the spotlight, promoting health interventions such as antiretroviral drug treatments for HIV, vaccines for TB, and bed nets for malaria. The everyday suffering from diarrhea and malnutrition may, in contrast, seem more routine and, therefore, receive less attention. In addition, individuals tend to demand curative services over preventive services, and these demands influence policymaking (McGuire, 2006). But interventions like clean water, education, or food security that will reduce diarrhea and malnutrition are structural and preventive. As Sen (1999b) points out, ‘In a democracy, people tend to get what they demand, and more crucially, do not typically get what they do not demand’ (p. 156). The excessive preoccupation with curative services at the expense of preventive may help explain the reason that day-to-day hunger endures even if governments increase spending on certain health-enhancing services (McGuire, 2006). The results of this study point to the need to incorporate a wider range of health outcomes in studies of democracy, rather than limit the analysis to the conventional country-level measures of life expectancy or infant mortality. More direct indicators of health that assess the ability to avoid preventable illness provide a more meaningful basis for gauging the effects of democratic political institutions (Wigley and Wigley-Akkoyunlu, 2011).

In addition to the methodological contributions, this study engenders several theoretical implications as well. Contestation is the key mechanism that motivates democratic regimes to provide better health, but non-democratic regimes sometimes produce healthy populations. Some of the most remarkable health and development improvements of the twentieth century occurred under authoritarian rule, even while democracies in developing countries were plagued by persistent poverty and health inequality (Gerring et al., 2012). For example, Vietnam, China, and Cuba achieved great health improvements without political contestation. Instead, disciplined, quasi-military party organizations and high levels of government expenditure on health-enhancing social infrastructure improved population health (Gauri and Khaleghian, 2002). Some countries might not allow political freedom, but they promote economic and social rights that enhance the physical quality of life. At the same time, some democracies might support political rights but not guarantee the kind of economic and social environment that truly improves well-being (Owens, 1987; Young, 1990).

Theories of democracy also rest on the presupposition that contestable governments are responsive to the needs of the electorate. But this assumes that states have the resources and the power to act in ways that improve population well-being. World systems theorists argue that as economic transactions cross-national boundaries at an ever-increasing rate, the growing wealth and power of transnational corporations (TNCs) undermine the power of the state and marginalize the state as an actor (Evans, 1997). The democratization of periphery countries is a welcome trend; however, almost as soon as democratic social movements take hold in periphery countries, they are beleaguered by external forces that assert their own policy agendas (Chase-Dunn, 1999). Fledgling democracies are caught between meeting constituent demands for welfare improvement and brokering compliance with foreign investors who push states to relax their policies on taxation, profit repatriation, and labor market protections (Li and Resnick, 2003). As TNCs gain control in economic, political, and social domains, states are limited in their capacity to implement policies in their own long-term interests (Kentor and Boswell, 2003). Foreign investors hamper health and well-being by discouraging social welfare policies that are beneficial to the local population but that do not serve TNC interests (Beer and Boswell, 2002; Shen and Williamson, 2001). Through tax evasion and disguising taxable profits, TNCs also reduce the government resources that would otherwise fund health and social services (Wimberley, 1990). Thus, in the face of declining state power, countries in the periphery of the world-economy are less able to assert their own governance priorities than those in the core, and this, in turn, affects population health.

The need to attract foreign investment and provide a favorable climate for TNCs challenges the capacity for democratic governance (Jensen, 2003; Li and Reuveny, 2003). Several researchers explore the political conditions for attracting foreign investment, and the results are contested (e.g. Jensen, 2003; Li and Resnick, 2003). However, much less is known about the ability of democratic governments to shield their citizens from the harmful pressures of TNCs (see Shandra et al., 2004, for an exception). In an era of increasing globalization, are democratically elected leaders still accountable to their constituents, or do they answer to foreign investors who bring the potential for economic growth? When democratic governments negotiate with foreign investors, how do they promote the development needs of the population and still provide an appealing environment for TNCs? These are important questions since attracting foreign investment is an integral part of development strategies in many countries (Jensen, 2003).

Theories of democracy also claim that democratic governments improve health in part because they empower the disadvantaged and give them the capacity to press elected leaders for what they need. But the results of this study may call this assumption into question. Policy preferences alone do not govern distributional outcomes. Cross-national differences in government social policies are also determined by working-class capacity for collective action. Distributive outcomes are often a product of power, not preferences (Bradley et al., 2003). It is possible that working-class parents do not have the collective power necessary to translate their policy preferences into social policy. In fact, public policy does not seem to be tracking the preferences of poor voters in most newly democratic developing countries (McGuire, 2006).

Instead, democracies are more responsive to the needs of the middle and upper classes. Public health-care programs that are directed at basic services for the disadvantaged are of little value to wealthier segments of the population who rely on private health care. Thus, political pressure from the economic elite often diverts state funding away from public services (Ghobarah et al., 2004). Or, even if democratic governments do channel more money into public services, those services subsidize middle and upper classes who can already afford to buy the services privately (Carlton-Ford, 2010; Ross, 2006). In this case, the benefits of democracy might not trickle down to the children who need them the most, that is, poor children afflicted by diarrhea and malnutrition.

For these reasons, it is important to look at the distribution of income within states (Ghobarah et al., 2004), which is overlooked in population studies that model democracy and health at the macro level. By accounting for the distribution of income and education within countries, this study underscores the enduring salience of household SES for child well-being. Spreading democracy does not necessarily secure the poor and does not necessarily improve population health. Democracies supposedly ease the struggle against poverty and disadvantaged living conditions, but they do not eliminate the struggles altogether (Wejnert, 2011). Indeed, the results demonstrate that low-income, low-educated families struggle to meet the basic needs of their children whether they live in a democratic country or not. Regardless of political regime, the effect of wealth on child well-being remains the same.

In addition to household SES, GDP per capita and population access to sanitation/water are of central importance to child well-being. Although some scholars claim that democracy contributes to development, others claim the reverse. This relationship is strongly contested, and decades of studies in this area yield mixed and inconclusive empirical evidence (see Gasiorowski, 2000; Przeworski and Limongi, 1993; Sirowy and Inkeles, 1990, for reviews). It is possible that democracy and development are mutually reinforcing and that it is the interlock of the two that helps both take root more firmly (Boutros-Ghali et al., 2002). Untangling this causality is beyond the scope of this article, but deserves future research. The results presented here demonstrate that holding political regime constant, GDP per capita and sanitation/water are robust predictors of cross-national variation in child health.¹¹ Future research may help determine whether democracies have an indirect effect on health by improving the economy and basic societal living conditions. Of course, cross-sectional data are limited, and longitudinal models may be more appropriate for answering these kinds of questions.

One limitation of this study, as well as many others like it, is the disconnect between procedure and quality. In theory, contestability is the defining feature of democracy, and I follow convention in measuring contestability with Polity IV data. Studies in this field are replete with similar analyses. But competitive elections are procedures that are emblematic of democracy. Child diarrhea and malnutrition are outcomes that may call the quality of democratic institutions into question. Leading scholars successfully standardized the term ‘democracy’ on the basis of procedural definitions. However, as democratization continues across a wide spectrum of cases and circumstances, early efforts to standardize the term should be supplemented with more careful attention to the substance of democratic governments (Collier and Levitzky, 1997).¹² Despite several important strides in this direction, data are limited in time and scope (e.g. Altman and Pérez-Liñán, 2002; Bühlmann et al., 2012; Diamond and Morlino, 2005). The measurement of democratic quality remains an area ripe for future research.

But beyond these measurement issues, the results of this study may call into question the theoretical significance of contestability and accountability. The effectiveness of these mechanisms in improving population health and development rests on the assumption that the interests of politicians coincide with the interests of voters. Yet, politicians have goals and values of their own. Once elected, they may pursue policy objectives that differ from the priorities of citizens. Even if politicians want to do nothing but serve the public, they may have different visions of development and how to best achieve it (Manin et al., 1999). After all, development priorities take many forms, such as foster economic growth, improve food distribution, increase educational opportunity, expand access to clean water, improve health, or reduce social exclusion (Nussbaum, 2004; Sen, 1985). Furthermore, many stakeholders compete for their priorities to be preeminent: the impoverished, the middle class, the powerful elite, and foreign investors. Who are democratically elected officials accountable to? And whose vision for development wins?

Despite his optimism about the potential for democracy to improve well-being worldwide (Sen, 1999a), the results of this study are not antithetical to Sen’s theories of development. In fact, he realizes that there is danger in overselling the importance and advantages of democracies. For example, India may have eradicated famines, but the country is still plagued by chronic undernutrition and illiteracy. Furthermore, Sen defines development as a process of expanding freedom. Yet, freedom refers not only to the guarantee of political rights but also to the enrichment of people’s choices, accomplishments, and capabilities. In Sen’s (1999b) words, freedom is ‘our ability to live as we would like’ (p. 13). Endemic poverty, chronic ill health, lack of education, and the absence of gainful employment are major sources of ‘unfreedom’ that prevent people from thriving and contributing fully to society (Sen, 1999b: 3). The results of this study demonstrate that many individuals are not ‘free’ even in otherwise free countries. Democracy is not an automatic remedy to all social ailments, and democratic institutions ‘cannot be viewed as mechanical devices for development’ (Sen, 1999b: 158).

Footnotes

Appendix

Appendix 4.

Individual-level correlation matrix.

	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Diarrhea	1.00
2. Underweight	0.03	1.00
3. Primary education	0.04	−0.06	1.00
4. Secondary education	−0.03	−0.11	−0.41	1.00
5. Higher education	−0.03	−0.07	−0.16	−0.13	1.00
6. Mother employed	0.02	0.03	0.03	−0.16	0.004	1.00
7. Household wealth	−0.06	−0.19	−0.16	0.39	0.34	−0.18	1.00
8. Urban residence	−0.02	−0.12	−0.07	0.26	0.20	−0.09	0.51	1.00
9. Mother’s Age	−0.04	0.01	−0.04	−0.12	0.05	0.15	−0.01	−0.02	1.00
10. Fem House Head	0.01	−0.03	0.03	0.05	0.01	0.04	−0.04	0.05	−0.01	1.00
11. Child’s age	−0.14	0.09	0.004	−0.01	0.002	0.05	0.03	0.01	0.20	−0.001	1.00
12. Male	0.02	0.02	−0.01	0.01	0.004	−0.005	0.01	0.002	−0.001	−0.001	0.003	1.00
13. Household size	0.01	0.04	−0.04	−0.09	−0.08	0.04	−0.002	−0.07	0.16	−0.07	−0.01	−0.001	1.00

Acknowledgements

The author is grateful to David Brady for detailed and constructive comments. The author also wishes to thank Kathleen Fallon, Linda George, Linda Burton, Kim Blankenship, and S. Philip Morgan for helpful suggestions. An earlier version of this article was presented at the annual Political Economy of the World Systems meetings at Clark University in Worcester, Massachusetts.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Altman

Pérez-Liñán

(2002) Assessing the quality of democracy: Freedom, competitiveness and participation in eighteen Latin American countries. Democratization 9(2): 85–100.

Alvarez

Cheibub

Limongi

. (1996) Classifying political regimes. Studies in Comparative International Development 31: 3–36.

Barclay

Myrskyla

(2016) Advanced maternal age and offspring outcomes: Reproductive aging and counterbalancing period trends. Population and Development Review 42(1): 69–94.

Beer

Boswell

(2002) The resilience of dependency effects in explaining income inequality in the global economy: A cross-national analysis, 1975-1995. Journal of World-Systems Research 8: 30–59.

Besley

Kudamatsu

(2006) Health and democracy. American Economic Review 96(2): 313–318.

Bhattacharya

(2006) Economic development, gender inequality, and demographic outcomes: Evidence from India. Population and Development Review 32: 263–291.

Blackwell

Hayward

Crimmins

. (2001) Does childhood health affect chronic morbidity later in life? Social Science & Medicine 52: 1269–1284.

Bollen

(1980) Issues in the comparative measurement of political democracy. American Sociological Review 45: 370–390.

Bollen

(1990) Political democracy: Conceptual and measurement traps. Studies in Comparative International Development 25: 6–24.

10.

Bollen

(1993) Liberal democracy: Validity and method factors in cross-national measures. American Journal of Political Science 37: 1207–1230.

11.

Bollen

(2001) Cross-National Indicators of Liberal Democracy 1950–1990 (2nd ICPSR version). Ann Arbor, MI: Inter-University Consortium for Political and Social Research.

12.

Bollen

Paxton

(2000) Subjective measures of liberal democracy. Comparative Political Studies 33: 58–86.

13.

Bollen

Glanville

Stecklov

(2001) Socioeconomic status and class in studies of fertility and health in developing countries. Annual Review of Sociology (27): 153–185.

14.

Boutros-Ghali

. (2003) The Interaction between Democracy and Development. Paris: United Nations Educational, Scientific and Cultural Organization.

15.

Bradley

Huber

Moller

. (2003) Distribution and redistribution in postindustrial democracies. World Politics 55: 193–228.

16.

Brooks

Manza

(2006) Social policy responsiveness in developed democracies. American Sociological Review 71: 474–494.

17.

Brown

(1999) Reading, writing, and regime type: Democracy’s impact on primary school enrollment. Political Research Quarterly 52: 681–707.

18.

Buchmann

Hannum

(2003) The Consequences of Global Educational Expansion: Social Science Perspectives. Cambridge, MA: American Academy of Arts and Sciences.

19.

Bühlmann

Merkel

Müller

. (2012) The democracy barometer: A new instrument to measure the quality of democracy and its potential for comparative research. European Political Science 11(4): 519–536.

20.

Burroway

(2012) A cross-national analysis of sex-specific HIV prevalence rates and women’s access to property, land, and loans in developing countries. International Journal of Sociology 42: 47–67.

21.

Burroway

(2015) Women’s rights save lives: A cross-national analysis of infant and child mortality and women’s access to land, property, and loans in developing countries. Sociology of Development 1(4): 418–441.

22.

Buvinic

Gupta

(1997) Female-headed households and female maintained developing countries? Economic Development and Cultural Change 45(2): 259–280.

23.

Cameron

Williams

(2009) Is the relationship between socioeconomic status and health stronger for older children in developing countries? Demography 46: 303–324.

24.

Carlton-Ford

(2010) Major armed conflicts, militarization, and life chances: A pooled time-series analysis. Armed Forces & Society 36(5): 864–889.

25.

Chaplin

(2003) Hierarchical linear models: Strengths and weaknesses. Paper presented at the association for public policy analysis and management annual meeting, Washington, DC, 26 October.

26.

Chase-Dunn

(1999) Globalization: A world-systems perspective. Journal of World-Systems Research 5(2): 187–215.

27.

Cheng

Schuster-Wallace

Watt

. (2012) An ecological quantification of the relationships between water, sanitation and infant, child, and maternal mortality. Environmental Health 11: 4.

28.

Cingranelli

Richards

(2010) The Cingranelli-Richards (CIRI) Human Rights Dataset (version 2010.08.15). Available at: http://www.humanrightsdata.com/p/data-documentation.html

29.

Collier

Levitsky

(1997) Democracy with adjectives: Conceptual innovation in comparative research. World Politics 49(3): 430–451.

30.

Collier

(2007) The Bottom Billion: Why the Poorest Countries Are Failing and What Can Be Done about It. Oxford: Oxford University Press.

31.

Cooper

Getzen

Laud

(2003) Economic expansion is a major determinant of physician supply and utilization. Health Services Research 38: 675–696.

32.

Corsi

Neuman

Finlay

. (2012) Demographic and health surveys: A profile. International Journal of Epidemiology 41(6): 1602–1613.

33.

Dahl

(1971) Polyarchy: Participation and Opposition. New Haven, CT: Yale University Press.

34.

Diamond

Morlino

(eds) (2005) Assessing the Quality of Democracy. Baltimore, MD: Johns Hopkins University Press.

35.

Drovandi

Salvini

(2004) Women’s autonomy and demographic behaviour. Population Review 43: 1–31.

36.

Enders

Tofighi

(2007) Centering predictor variables in cross-sectional multi-level models: A new look at an old issue. Psychological Methods 12: 121–138.

37.

Evans

(1997) The eclipse of the state? Reflections on stateness in an era of globalization. World Politics 50(1): 62–87.

38.

Fox

(1997) Applied Regression Analysis, Linear Models, and Related Models. Thousand Oaks, CA: SAGE.

39.

Franco

Alvarez-Dardet

Ruiz

(2004) Effect of democracy on health: Ecological study. British Medical Journal 329: 1421–1423.

40.

Freedom House (2010) Freedom in the world. Available at: https://freedomhouse.org/report-types/freedom-world

41.

Frey

Al-Roumi

(1999) Political democracy and the physical quality of life: The cross-national evidence. Social Indicators Research 47: 73–97.

42.

Gasiorowski

(2000) Democracy and macroeconomic performance in underdeveloped countries: An empirical analysis. Comparative Political Studies 33: 319–349.

43.

Gauri

Khaleghian

(2002) Immunization in developing countries: Its political and organizational determinants. World Development 30(12): 2109–2132.

44.

Gerring

Thacker

Alfaro

(2012) Democracy and human development. The Journal of Politics 74(1): 1–17.

45.

Ghobarah

Huth

Russett

(2004) Comparative public health: The political economy of human misery and well-being. International Studies Quarterly 48: 73–94.

46.

Goldsmith

(1986) Democracy, political stability, and economic growth in developing countries. Comparative Political Studies 18: 517–531.

47.

Gradstein

Milanovic

(2004) Does Liberté = Egalité? A survey of the empircal links between democracy and inequalty with some evidence on the transition economies. Journal of Economic Surveys 18(4): 515–537.

48.

Guo

Zhao

(2000) Multilevel modeling for binary data. Annual Review of Sociology 26: 441–462.

49.

Hatt

Waters

(2006) Determinants of child morbidity in Latin America: A pooled analysis of interactions between parental education and economic status. Social Science & Medicine 62: 375–386.

50.

Heaton

Forste

Hoffmann

. (2005a) Cross-national variation in family influences on child health. Social Science & Medicine 60: 97–108.

51.

Heaton

Huntsman

Flake

(2005b) The effects of status on women’s autonomy in Bolivia, Peru, and Nicaragua. Population Research and Policy Review 24: 283–300.

52.

Herrera

Kapur

(2007) Improving data quality: Actors, incentives, and capabilities. Political Analysis 15: 365–386.

53.

Hill

Upchurch

(1995) Differences in child health: Evidence from the demographic and health surveys. Population and Development Review 21: 127–151.

54.

Hossain

Phillips

Pence

(2007) The effect of women’s status on infant and child mortality in four rural areas of Bangladesh. Journal of Biosocial Science 39: 355–366.

55.

Huber

Nielsen

Pribble

. (2006) Politics and inequality in Latin America and the Caribbean. American Sociological Review 71: 943–963.

56.

Jenkins

Scanlan

(2001) Food security in less developed countries, 1970 to 1990. American Sociological Review 66: 718–744.

57.

Jensen

(2003) Democratic governance and multinational corporations: Political regimes and inflows of foreign direct investment. International Organization 57: 587–616.

58.

Kaufman

Kraay

Mastruzzi

(2010) Worldwide governance indicators. Available at: http://info.worldbank.org/governance/wgi/index.aspx#home

59.

Kentor

Boswell

(2003) Foreign capital dependence and development: A new direction. American Sociological Review 68: 301–313.

60.

Klomp

De Haan

(2009) Is the political system really related to health? Social Science & Medicine 69: 36–46.

61.

Lake

Baum

(2001) The invisible hand of democracy: Political control and the provision of public services. Comparative Political Studies 34: 587–621.

62.

Lee

C-S

(2005) Income inequality, democracy, and public sector size. American Sociological Review 70: 158–181.

63.

LeGrand

Phillips

(1996) The effect of fertility reductions on infant and child mortality: Evidence from Matlab in rural Bangladesh. Population Studies 50: 51–68.

64.

Lena

London

(1993) The political and economic determinants of health outcomes: A cross-national analysis. International Journal of Health Services 23: 585–602.

65.

Lenski

(1966) Power and Privelege: A Theory of Social Stratification. New York: McGraw-Hill.

66.

Resnick

(2003) Reversal of fortunes: Democratic institutions and foreign direct investment inflows to developing countries. International Organization 57: 175–211.

67.

Reuveny

(2003) Economic globalization and democracy: An empirical analysis. British Journal of Political Science 33: 29–54.

68.

Link

Phelan

(1995) Social conditions as fundamental causes of disease. Journal of Health and Social Behavior 80–94.

69.

Link

Phelan

(2002) McKeown and the idea that social conditions are fundamental causes of disease. American Journal of Public Health 95: 730–732.

70.

Lipset

(1959) Some social requisites of democracy: Economic development and political development. American Political Science Review 53: 69–105.

71.

London

Williams

(1990) National politics, international dependency, and basic needs provision. Social Forces 69: 565–584.

72.

Long

Freese

(2014) Regression Models for Categorical Dependent Variables Using Stata, 3rd edn. College Station, TX: Stata Press.

73.

McGuire

(2002) Democracy, social provisioning, and under-5 mortality: A cross-national analysis. Paper presented at the annual meeting of the American Political Science Association, Boston, MA, 1 September.

74.

McGuire

(2006) Democracy, basic service utilization, and under-5 mortality: A cross-national study of developing states. World Development 34: 405–425.

75.

Macro (2009) Demographic and health surveys: ICF Macro. Available at: http://www.measuredhs.com/

76.

McTavish

Moore

Harper

. (2010) National female literacy, individual socio-economic status, and maternal health care use in Sub-Saharan Africa. Social Science & Medicine 71: 1958–1963.

77.

Manin

Przeworski

Stokes

(eds) (1999) Democracy, Accountability, and Representation. Cambridge: Cambridge University Press.

78.

Marshall

Jaggers

(2010) Polity IV project: Political regime characteristics and transitions, 1800–2009 (version p4v2008). Available at: http://www.systemicpeace.org/polity/polity4.htm

79.

Matland

(1998) Women’s representation in National Legislatures: Developed and developing countries. Legislative Studies Quarterly 23(1): 109–125.

80.

Montgomery

Gragnolati

Burke

. (2000) Measuring living standards with proxy variables. Demography 37: 155–174.

81.

Moon

(1991) The Political Economy of Basic Needs. Ithaca, NY: Cornell University Press.

82.

Moon

William

(1985) Politics, the state, and basic human needs. American Journal of Political Science 29: 661–694.

83.

Moore

McCabe

(2006) Introduction to the Practice of Statistics. New York: W.H. Freeman.

84.

Moore

Teixeira

Sheill

(2006) The health of nations in a global context: Trade, global stratification, and infant mortality rates. Social Science & Medicine 63: 165–178.

85.

Muller

(1988) Democracy, economic development, and income inequality. American Sociological Review 53: 50–68.

86.

Munck

Verkuilen

(2002a) Conceptualizing and measuring democracy: Evaluating alternative indices. Comparative Political Studies 35: 5–34.

87.

Munck

Verkuilen

(2002b) Generating better data: A response to discussants. Comparative Political Studies 35: 52–57.

88.

Nelson

(2007) Elections, democracy, and social services. Studies in Comparative International Development 41: 79–97.

89.

Nussbaum

(2003) Women’s education: A global challenge. Signs: Journal of Women in Culture and Society 29: 325–355.

90.

Nussbaum

(2004) Beyond the social contract: Capabilities and global justice. Oxford Development Studies 32: 3–18.

91.

Owens

(1987) The Future of Freedom in the Developing World: Economic Development as Political Reform. New York: Pergamon Press.

92.

Przeworski

Limongi

(1993) Political regimes and economic growth. Journal of Economic Perspectives 7: 51–69.

93.

Przeworski

Alvarez

Cheibub

. (2000) Democracy and Development: Political Institutions and Well-Being in the World, 1950-1990. New York: Cambridge University Press.

94.

Puddington

(2015) Freedom in the World 2015: Discarding Democracy: A Return to the Iron Fist. Washington, DC: Freedom House.

95.

Raudenbush

Bryk

(2002) Hierarchical Linear Models: Applications and Data Analysis Methods, 2nd edn. Thousand Oaks, CA: SAGE.

96.

Raudenbush

Bryk

Cheong

. (2004) HLM 6: Hierarchical Linear and Nonlinear Modeling. Lincolnwood, IL: Scientific Software International.

97.

Reuveny

(2003) Economic openness, democracy, and income inequality. Comparative Political Studies 36(5): 575–601.

98.

Ross

(2006) Is democracy good for the poor? American Journal of Political Science 50: 860–874.

99.

Sachs

(2005) The End of Poverty: Economic Possibilites for Our Time. New York: Penguin Press.

100.

Safaei

(2006) Is democracy good for health? International Journal of Health Services 36: 767–786.

101.

Scanlan

(2004) Women, food security, and development in less-industrialized societies: Contributions and challenges for the new century. World Development 32(11): 1807–1829.

102.

Scanlan

(2010) Gender, development, and HIV/AIDS: Implications for child mortality in less industrialized countries. International Journal of Comparative Sociology 51(3): 211–232.

103.

Sen

(1985) Commodities and Capabilities. Amsterdam: Elsevier.

104.

Sen

(1999a) Democracy as a universal value. Journal of Democracy 10: 3–17.

105.

Sen

(1999b) Development as Freedom. New York: Anchor Books.

106.

Shandra

London

(2012) The international monetary fund, structural adjustment, and infant mortality: A cross-national analysis of sub-Saharan Africa. Journal of Poverty 16: 194–219.

107.

Shandra

Nobles

London

. (2004) Democracy, dependency, and infant mortality: A quantitative, cross-national analysis of less developed countries. Social Science & Medicine 59: 321–333.

108.

Shen

Williamson

(2001) Accounting for cross-national differences in infant mortality decline (1965–1991) among less developed countries: Effects of women’s status, economic dependency, and state strength. Social Indicators Research 53: 257–288.

109.

Shin

(2007) Child health in Peru: Importance of regional variation and community effects on children’s height and weight. Journal of Health and Social Behavior 48: 418–433.

110.

Sirowy

Inkeles

(1990) The effects of democracy on economic growth and inequality: A review. Studies in Comparative International Development 25(1): 126–157.

111.

Smith

Haddad

(2000) Explaining Child Malnutrition in Developing Countries: A Cross-Country Analysis. Washington, DC: International Food Policy Research Institute.

112.

Sommerfelt

Stewart

(1994) Children’s Nutritional Status: DHS Comparative Studies No. 12. Calverton, MD: Macro International.

113.

Sorensen

(1991) Democracy, Dictatorship and Development. London: Palgrave Macmillan.

114.

Stratton

O’Neill

Kruk

. (2008) The persistent problem of malaria: Addressing the fundamental causes of a killer. Social Science & Medicine 67: 854–862.

115.

Studenmund

(2001) Using Econometrics: A Practical Guide. Boston, MA: Addison Wesley Longman.

116.

Sutcliffe

Barnes

Belsky

. (2012) The health and development of children born to older mothers in the United Kingdom: Observational study using longitudinal cohort data. British Medical Journal 345: e5116. Available at: https://dx-doi-org.web.bisu.edu.cn/10.1136/bmj.e5116

117.

Swiss

Fallon

Burgos

(2012) Reaching a critical mass: Women’s political representation and child health in developing countries. Social Forces 91: 531–558.

118.

Thomas

(1994) Like father, like son; like mother, like daughter: Parental resources and child height. The Journal of Human Resources 29: 950–988.

119.

Thomas

Strauss

Maria-Helena

(1990) How does mother’s education affect child height. The Journal of Human Resources 26: 183–211.

120.

Tsai

M-C

(2006) Does political democracy enhance human development in developing countries? A cross-national analysis. American Journal of Economics and Sociology 65: 233–268.

121.

UNICEF (2007) State of the World’s Children 2008. New York: United Nations Children’s Fund.

122.

UNICEF (2014) State of the World’s Children 2015: Reimagine the Future: Innovation for Every Child. New York: United Nations Children’s Fund.

123.

US State Department ( 2011) Democracy. Available at: http://www.state.gov/g/drl/democ/ (accessed 1 March 2011).

124.

Vanhanen

(2000) A new dataset for measuring democracy: 1810–1998. Journal of Peace Research 37: 251–265.

125.

Viterna

Fallon

Beckfield

(2008) How development matters: A research note on the relationship between development, democracy and women’s legislative representation. International Journal of Comparative Sociology 49: 455–476.

126.

Walker

CLF

Perin

Aryee

. (2012) Diarrhea incidence in low- and middle-income countries in 1990 and 2010: A systematic review. BMC Public Health 12: 220.

127.

Wejnert

(2005) Diffusion, development, and democracy, 1800–1999. American Sociological Review 70: 53–81.

128.

Wejnert

(2008) Effects of global democracy on women’s reproductive health: 1970–2005, cross-world analysis. Marriage and Family Review 44: 154–172.

129.

Wejnert

(2011) Issues in health, democracy, and development. Research in Political Sociology, Democracies: Challenges to Societal Health 19: 1–10.

130.

Wickrama

KAS

Mulford

(1996) Political democracy, economic development, disarticulation, and social well-being in developing countries. Sociological Quarterly 37: 375–390.

131.

Wigley

Akkoyunlu-Wigley

(2011) The impact of regime-type on health: Does redistribution explain everything? World Politics 63(4): 647–677.

132.

Wimberley

(1990) Investment dependence and alternative explanations of third world mortality: A cross-national study. American Sociological Review 55: 75–91.

133.

World Bank (2010) World Development Indicators. Available at: http://databank.worldbank.org (accessed 15 January 2010).

134.

World Health Organization (WHO) (2008) Training Course on Child Growth Assessment. Geneva: WHO.

135.

World Health Organization (WHO) ( 2011) Children’s environmental health > environmental risks > malnutrition. Available at: http://www.who.int/ceh/risks/otherisks/en/index1.html (accessed 4 February 2011).

136.

World Health Organization (WHO) (2013) Diarrhoeal disease: Fact sheet N330. Available at: http://www.who.int/mediacentre/factsheets/fs330/en/

137.

World Health Organization and United Nations International Children’s Emergency Fund (WHO and UNICEF) (2004) Meeting the MDG Drinking Water and Sanitation Target: A Mid-term Assessment of Progress. Geneva: WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation.

138.

Young

(1990) Do some authoritarian governments foster physical quality of life? Social Indicators Research 22: 367–384.

139.

Zweifel

Navia

(2000) Democracy, dictatorship, and infant Mortality. Journal of Democracy 11: 99–114.