International Human Rights Law, Global Economic Reforms, and Child Survival and Development Rights Outcomes

Abstract

Are recent trends in international law supporting child rights and promoting neoliberal economic reforms complementary or contradictory? To answer this question, we identify the component parts of child rights mobilization, recent global economic reforms, and child rights outcomes to theorize the particular relationships among them. Focusing on child survival and development rights in 99 poor and middle-income countries from 1983 to 2001, we find that countries' acquiescence to established international law concerning economic rights influences the successful implementation of most of these rights, while the ratification of child rights treaties does not show an effect during the period studied. National links to child rights nongovernmental organizations are also associated with improved child rights outcomes, as is being selected to receive a loan from the World Bank (for reducing child labor and increasing immunizations). We find weak support for the hypothesis that the implementation of loan conditionalities is more deleterious for rights that are costlier to implement. We also find that achieving the goal of neoliberal economic reforms—trade openness—results in less successful implementation of most child rights outcomes considered. Finally, in a related analysis, we find that the ratification of child rights treaties, as well as the adoption and implementation of structural adjustment agreements, enhances the presence of child-related organizations within countries.

Child rights represent a key site for learning more about the interaction of human rights and global economic discourses. In recent decades, an emerging worldwide consensus about the importance of child rights has coincided with the spread of neoliberal economic policies. To what extent do these twin forces work together, and to what extent do they operate in opposition to one another? In this analysis, we draw on theories of the global system to suggest ways to integrate analyses of human rights and economics.

On the one hand, there are some clear commonalities between recent human rights and economic trends. Both human rights and economic neoliberalism aspire to universality (see, e.g., Soysal 1994; compare Landgren 2005) and celebrate the autonomous individual (Boyle & Carbone 2006; Boyle 2002; Frank & Meyer 2002; Meyer & Jepperson 2000). International governmental and international nongovernmental organizations “IGO” and INGOs) are the global carriers of both these discourses (see Boli & Thomas 1997, 1999). Human rights and neoliberalism have both been critiqued as imperialistic (Anghie 2004; Rajagopal 2003), but both claim to be in the best interests of the poor and oppressed. As these similarities suggest, human rights and neoliberalism are symmetrical processes. Together, the two regimes create world contexts within which human rights violations and free markets become the business of the whole world rather than just particular nations. On the other hand, these two separate but intertwined discourses can place contradictory requirements on nation-states. To understand these potential contradictions, it is useful to trace the recent evolution of child rights and neoliberal economic policies.

Several transformations in thinking have placed children at the center of recent human rights discourse. First, personhood and the human rights of persons, in contrast to collective interests, became central to the world human rights regime (see, e.g., Boyle 2002; Berkovitch 1999; Ramirez 2002; Ramirez & Meyer 1998). For example, labeling female genital cutting as a rights violation required a weakening of the collective power of first nations and then families (Boyle 2002). Individualism made it possible to think of children as individual agents rather than simply as members of families or the future of a nation-state.¹ These transformations also allowed the idea that individuals had a right to education to gain precedence over the idea that education was necessary to contribute to overall national development (Chabbott 1999). A focus on the empowerment of individuals through the assignment of individual rights was an important precursor to child rights.

Second, child rights emerged at the intersection of a perceived need to protect children and an expansion of rights beyond civil and political rights to include certain economic entitlements. Child rights were initially invoked at international conferences bringing together many professionals and representatives from IGOs and INGOs (compare Merry 2006; Canan & Reichman 2002; Goldman 2005). The global mobilization eventually led to international lawmaking. The 1959 United Nations Declaration on the Rights of the Child set forth obligations with respect to children but was not binding law. The International Labour Organization's (ILO) 1973 Minimum Age Convention, which set restrictions on child labor, had the primary goal of protecting children (Boyle et al. 2006). Meanwhile, the United Nation's 1976 International Covenant on Economic, Social, and Cultural Rights (ICESCR) officially recognized the positive obligation of states to provide their citizens with basic services, such as education and health care. This represented a shift from prior rights treaties because, for a state to comply with it, it was insufficient for the state to stop doing something bad, such as torture (Rubenstein 2004). The treaty went further, requiring states to undertake “concerted, rational, well-planned steps forward to finance and build housing, health clinics, and schools” (Rubenstein 2004:851). Although most countries in the world have ratified the ICESCR, it continues to be controversial, and there is disagreement about how to best implement it (Roth 2004; Rubenstein 2004).

The trends toward protecting children and recognizing economic rights came together in the United Nation's 1989 Convention on the Rights of the Child (the CRC). The CRC is a sweeping document. It includes nearly 40 substantive sections, each identifying a category of child rights. For example, the Convention obligates states to insure the survival and development of children (Section 6[2]). It requires states to diminish infant and child mortality and to insure that all children have access to health care (Section 24). It requires states to make primary education compulsory and free, to make secondary and higher education accessible to all, and to encourage school attendance and reduce drop-out rates (Section 28). In terms of child labor, the CRC recognizes “the right of the child to be protected from economic exploitation and from performing any work that is likely to be hazardous or to interfere with the child's education” (Section 32[1]). The Convention gives children the right to a name from birth (Section 7[1]), the right to free expression (Section 13[1]) and association (Section 15[1]), and the right to rest and leisure (Section 31[1]). The CRC, signed by all but two countries in the world, means that child rights are human rights, at least in terms of international law.²

Nevertheless, child rights are in some ways different from other rights, and this is likely to affect their implementation. The CRC is unique because it includes both civil/political rights and social/economic/cultural rights. With respect to civil and political rights, these rights have traditionally been available to children only indirectly through their parents. For example, it is widely agreed that states should allow adults religious freedom. There is undoubtedly less consensus that states should support minor children in the free exercise of a religion (given that this may be at odds with the beliefs of parents), despite this provision in the widely ratified CRC. While both free-religion principles are enshrined in international law, the child rights principle is more open to challenge. Social, economic, and cultural rights are somewhat controversial in and of themselves, as noted above, although there appears to be more of a consensus to provide them for children than to provide them for adults. (Consider the number of countries that have ratified the CRC [192] versus the ICESCR [149].) Even those who support economic rights acknowledge that they put an immense burden on governments, especially in poorer states (Antrobus 2001; Matz 2002; Rudra 2002). Finally, some rights in the CRC, such as the right to put education ahead of labor and the right to have a name at birth, are viewed as culturally biased (Nieuwenhuys 1998; Panter-Brick 2000; Pupavac 2001; Burr 2002; White 2002; Jones 2005; Myers 2001). The question of why so many countries ratified the CRC despite these controversies is an interesting one but is a topic for another analysis. In terms of the analysis here, the important point is that the unique character of child rights may prompt countries to ignore or downplay their CRC commitments.

Marta Santos Pais (1999), director of Evaluation, Policy, and Planning at UNICEF, groups the myriad rights in the CRC into four categories (survival and development rights; rights against discrimination; rights to have best interests considered; and rights to self-expression and participation). In the short term, the United Nations and its affiliated organizations assigned the highest priority to the first of these, child survival and development (United Nations 1990). If successful implementation of the CRC is occurring, child survival and development rights are therefore the area where progress should be most evident. For this reason, we choose to focus on this category of child rights in the present analysis.

Turning to another aspect of contemporary globalization, neoliberal economic policies developed independently of child rights discourses. Beginning in the late 1970s, with the trend accelerating in the early 1980s, the World Bank began making loans conditional on the implementation of certain economic policies (Kapur et al. 1997; Babb 2005). By agreeing to implement policy reforms, third world countries were able to acquire much-needed funds from international financial institutions (Babb 2005). Loan conditionalities were laid out in structural adjustment agreements (SAAs) that emphasized the free market principles of restricted state spending, balanced state budgets, and the privatization of industries (Eckaus 1986; Weisbrot et al. 2006). Conditionalities included the deregulation of labor markets, financial markets, and agricultural prices, and the removal of trade barriers (Babb 2005; Sadasivam 1997). Although particular conditionalities were somewhat varied across countries, all represented a repudiation of the more state-centered Keynesian economic policies typical in developing countries in the 1970s (Pender 2001).

By 1995, the World Bank faced increasing criticism for its conditionality requirements from both activists and academics (Kapur et al. 1997; Donnelly 2002). As a response, the organization made a commitment to focus more on alleviating poverty and less on economic growth per se. It began to include political reforms designed to foster “good governance” in its conditionalities (Babb & Carruthers 2008). It changed the title of conditionalities from “structural adjustment agreements” to “comprehensive development frameworks.” As time passed, it became clear that SAAs and comprehensive development frameworks were not that different from one another. For example, the World Bank continued using the term structural adjustment as both a category and a title of loans after 1995. More important, the organization remained committed to its free market ideology and continued to make loans conditional on the implementation of liberal economic reforms. In its World Development Report 2000/2001, the World Bank expanded the goals of economic growth to include poverty reduction but maintained its view that neoliberal strategies were the best mechanisms to achieve those goals (World Bank 2000).

When countries attempt to comply with both human rights treaties and SAAs, they can encounter contradictions. Specifically, to fully implement the CRC requires substantial state outlays of funds, while SAAs emphasize fiscal austerity. Surprisingly, there has been little analysis of the integrated effects of these two global trends for child rights outcomes.

The Effects of Global Human Rights and Neoliberal Economic Discourses

Law and economic theories are rarely integrated in the law and society field (Edelman 2004). This is true at the international law level. One global theory—World Polity Theory—emphasizes law, policy, and culture. Scholars in this tradition have studied human rights extensively (see, e.g., Hafner-Burton & Tsutsui 2005; Frank & McEneaney 1999 on laws concerning same-sex relations; Boyle 2002 on anti–female genital cutting policies; Cole 2005 on state ratifications of human rights treaties). They emphasize the embeddedness of nation-states in a wider world cultural context of meaning and models. In the world polity, ideas and structures spread throughout many countries quickly despite the vast differences across those countries. IGOs and INGOs are key players in this global civil society; they are important in spreading valued principles among nation-states (Boli & Thomas 1999). In terms of explaining the high levels of support for the CRC, World Polity Theorists would point to how such ratification marks countries as legitimate participants in global civil society. At a more pragmatic level, ratification can also increase the amount of funding countries receive from international sources (see Barrett & Tsui 1999). In terms of whether child rights reforms are actually implemented, World Polity Theorists would look to the role played by nonstate transnational actors as well as straightforward treaty adoption.

In this tradition, Cole (2005) examines the circumstances under which states are willing to ratify international human rights treaties. He finds support for both World Polity Theory and more rationalist arguments. Supporting World Polity Theory, he finds that many countries ratified treaties around the time of international conferences promoting the content of those treaties. He also finds a contagion effect, suggesting normative pressure to ratify. On the rationalist side, the strength of enforcement mechanisms has been a better predictor of ratification than how closely a country's practices actually corresponded to the content of the treaty. Countries have generally avoided ratifying treaties with strong enforcement mechanisms. Cole's (2005) study illustrates the varied meaning of treaty ratification for different countries and suggests that ratification may not necessarily closely correspond to implementation.

Hafner-Burton and Tsutsui (2005) address the issue of implementation in their study of countries' ratification of human rights treaties and human rights violation records from 1978 to 1999. Their analysis uncovers two key findings. First, they find that initially countries ratify human rights treaties even when they are violating human rights and have no apparent commitment to insure human rights in the future. Hafner-Burton and Tsutsui (2005) explain that these countries have signed on to treaties to enhance their legitimacy. Second, these scholars find that even wholly symbolic institutional commitments have moved governments toward compliance with human rights in the long run. Hafner-Burton and Tsutsui (2005) call this the “paradox of empty promises” and explain that it occurs because treaty commitments provide nonstate human rights advocates with the tools to force reforms (see Keck & Sikkink 1998). The emphasis of their study on nonstate actors is consistent with other work that notes how associations and grassroots activists have worked to translate women's rights principles into a local vernacular (compare Merry 2006) and use networks to share advocacy strategies (compare Smith 1995).

What is important in determining whether rights are enforced, according to Hafner-Burton and Tsutsui (2005), is whether a support system emerges after treaties are ratified, providing resources and vigilantly calling attention to abuses. Their study suggests that laws may serve as magnets for human rights organizations to enter nations and force reforms. World Polity Theorists have studied the impact of treaty ratification and the presence of INGOs on human rights outcomes, but they have not yet integrated countries' links to international financial institutions into their analyses.

While World Polity Theorists have focused on human rights adoption and implementation, other scholars of globalization focus more on economic outcomes. In this literature, one key question is whether loan conditionalities in the form of SAAs have a positive effect on economic growth. Research to date has suggested that they do not in the short term (e.g., Gylfason 1987; Przeworski & Vreeland 2000). In the long term, the results have been even more pessimistic (Harrigan & Mosley 1991). A study by the World Bank itself in 1994 indicated that SAAs had not led to the economic growth rates anticipated for countries in Africa (World Bank 1994). The report attributed the disappointing results to nations' failure to implement and enforce SAAs fully and wholeheartedly (World Bank 1994:15; see also Stone 2004; Bird et al. 2004). This explanation was undercut when Mexico, a model country in terms of implementing SAA requirements, also experienced an economic crisis in the mid-1990s (Pender 2001). In addition, East Asian countries confounded World Bank experts by showing the greatest growth rates in the early 1990s despite relatively strong state interventions in national markets (Pender 2001). These cases cast doubt on the “improper implementation” explanation for sluggish growth.

In addition to economic growth, studies have also focused on the impact of SAAs on inequality within countries. In general, these studies find an increase in income inequality associated with SAAs (e.g., Portes & Hoffman 2003; Vreeland 2003). Cumulatively, the data suggest that structural adjustment has not been as beneficial to national development as was initially anticipated.

A smaller body of research has considered the impact of international loans and loan conditionalities on non-economic outcomes. For example, Mutangadura and colleagues (2002) conducted a pooled time-series analysis to determine the impact of levels of debt on primary-to-secondary transition rates and secondary school enrollments in countries in Africa from 1980 to 1997. The researchers categorize debt into three categories: (1) debt constituting 0 to 50 percent of GDP; (2) debt constituting 50.5 to 100 percent of GDP; and (3) debt constituting more than 100 percent of GDP. Using these categories and controlling for education expenditures, Mutangadura et al. (2002) find that higher levels of debt are associated with lower transition rates and lower levels of secondary enrollment. This is consistent with the findings that SAAs lead to increases in income inequality within countries.

Turning to ethnographic data, Makene's (2007) case study of education in Tanzania helps explain why high national debt levels can result in less schooling. One path to the fiscal austerity required by SAAs was implementing user fees for public services, such as education. However, the CRC requires that primary education be free for children. (The education user fee was ultimately rejected by the World Bank in 2002.)³Makene (2007) focuses on this contradiction in Tanzania. She finds that many government officials were unaware that they had committed their country to inconsistent requirements by ratifying the CRC while still charging user fees for education. They were only aware of the reforms promoted by the International Monetary Fund (IMF) and the World Bank. This suggests that, when contradictions do exist, the economic requirement is likely to take precedence at the national level. At the individual level, Makene finds because of the fees, some parents will enroll only some of their children in school.

Education is one important area of child survival and development rights; access to health care is another. Relevant to this, Bradshaw and others (1993) consider the effect of national debt on under-five mortality rates in third world countries. Using mortality measures from 1987 and debt obligations and other economic measures from the decade prior to that year, these researchers learned that debt levels have indirectly increased under-five mortality rates. Higher levels of debt are not directly associated with higher levels of infant mortality, but they slow economic growth, and this in turn has increased infant mortality.

Another important study of the non-economic consequences of SAAs was conducted by Abouharb and Cingranelli (2006). These authors conducted a global comparative analysis of countries from 1981 to 2000 to assess the impact of SAAs on physical integrity rights. They found that receiving and implementing an SAA from the World Bank had a deleterious effect on physical integrity rights, including freedom from torture, political imprisonment, extrajudicial killings, and disappearances. They suggest that the negative impact of SAAs was not intended by the international financial institutions but rather was the result of states' needing to settle domestic unrest that arose when fiscal austerity measures were imposed. Child rights are distinct from traditional human rights in some ways, so it is unclear the extent to which the same processes would apply to them. Taken together, previous studies on the impact of SAAs on rights-related outcomes suggest that SAAs are deleterious to rights, at least indirectly.

Only a handful of studies integrate analyses of global economic pressures with other types of globalization. Our analysis adds to this small number by integrating the literature on human rights reforms and the literature on neoliberal economic reforms. This is important in terms of both empirics and theory development. We believe that the most useful approach is to delineate theoretically how particular aspects of human rights mobilization and neoliberal economic reform impact particular children's rights. To this end, we consider three measures of human rights reform: (A1) ICESCR ratification, (A2) child rights treaty ratification, and (A3) child rights INGO mobilization, and three measures of neoliberal economic reforms: (B1) agreeing to an SAA, (B2) implementing an SAA, and (B3) trade openness. In terms of outcomes, we consider four measures that should be associated with the implementation of child rights to survival and development: (C1) reducing child labor, (C2) increasing infant immunizations, (C3) increasing primary school enrollments, and (C4) increasing secondary school enrollments. In the next section, we theorize the relationships among the component parts of child rights and neoliberal economic expansion.

Theory Development and Hypotheses

In this section we discuss how human rights and neoliberal economic trends may interact to affect outcomes for children. Throughout the section, we provide a series of hypotheses that we later test using an over-time analysis of 99 developing countries from 1983 to 2001, the final year in which education user fees were endorsed by the World Bank. Our dependent variables, as noted, center around child labor, health, and education. In terms of international law, we first consider whether and when countries ratify the ICESCR. Countries will have ratified this treaty if they are supportive of economic rights, an important component of child rights to survival and development. Countries that have ratified the ICESCR are more likely to have the cultural and structural support necessary to implement those child rights. We therefore hypothesize:

Hypothesis 1A: States that ratified the International Covenant on Economic, Social, and Cultural Rights will show greater progress toward child rights goals (reducing child labor, and increasing immunizations and educational enrollments).

We also expect that countries that ratify more child-oriented treaties will have better outcomes for children:

Hypothesis 1B: States that ratify child rights treaties will show greater progress toward child rights goals (reducing child labor, and increasing immunizations and educational enrollments) than those that do not.

This hypothesis will be supported if there is in fact a tight correspondence between the new international legal framework for child rights and child rights outcomes. There is reason to believe this is true. For example, research has shown that the ratification of the CRC is associated with the appointment of children's ombudspersons—a tool that can be used by activists within countries (Gran & Aliberti 2003; compare Finnemore 1996; Frank et al. 2000; McNeely 1995). On the other hand, the relationship between policy and outcomes may be more complex. Although global actors are good at creating the appearance of consensus (Goldman 2005), closer inspection sometimes reveals suppressed disagreements (Silliman 1999). As a consequence, nations that voice a formal commitment to young people may still fail to take action on the ground. If this is true, Hypothesis 1B will not be supported.

Our next hypothesis is derived from earlier World Polity Theory research findings that the presence of INGOs has an important impact on country outcomes. INGOs signal links to global civil society and can put pressure on countries to honor their commitments:

Hypothesis 2: The greater the number of child rights INGOs (CRINGOs) within a country, the greater the progress toward child rights goals (reducing child labor, and increasing immunizations and educational enrollments).

Child rights treaty ratification may draw CRINGOs into countries, leading to greater reforms on the ground. At the same time, global support for child rights may energize CRINGOs to mobilize, even in countries that have not ratified child rights treaties. The important point is that CRINGOs will advocate for children, which should improve child rights outcomes.

In theorizing the effect of global economic trends on child rights outcomes, it is important to separate the component parts of both the trends and the outcomes and to theorize the relationships separately. Failing to do so would conflate many disparate forces. First, on the economic side, we distinguish the entry into a loan agreement from the actual implementation of loan conditionalities (compare Abouharb & Cingranelli 2006). In order to qualify for loans, countries must demonstrate some level of commitment to world polity principles. The World Bank is an organization whose power depends not only on its economic capacity but also on the creation and maintenance of its legitimacy (Goldman 2005). It is thus in the interest of the World Bank to work with the United Nations and other human rights organizations. Because it does not want to be associated with child rights violations, the World Bank will avoid loaning money to countries that have poor records for implementing child rights.

Hypothesis 3: The adoption of SAAs will have a positive association with progress toward child rights goals (reducing child labor, and increasing immunizations and educational enrollments).

Hypothesis 3 is consistent with Abouharb and Cingranelli's (2006) finding that SAA adoption is associated with better physical integrity rights records. The World Bank is sympathetic to human rights discourses and so are likely to reward nations that observe human rights with loans.

The effect of the implementation of SAAs should be somewhat different from the determination of who receives them. Here the difficult task of reforming a national economy comes to the fore. For this analysis, we consider the general characteristics that most loan conditionalities shared during the period studied: fiscal austerity and cutting government spending. Since child rights reforms compete with other public projects for scarce government resources, we hypothesize that the costlier the reform, the slower its implementation. In terms of the relative cost of child survival and development rights, it is likely more costly to keep an additional child in primary and secondary school than it is to provide an immunization. Countries spend more on education overall than they do on health. (We conducted analyses of government spending on health and education as a percentage of GDP and found that, in nearly every country, the percentage of GDP spent on education was much larger than total expenditures on health.) To counteract the high cost of education, the World Bank recommended user fees for a time, and these can reduce the motivation of parents to enroll all of their children in school, at least in the short term (see Makene 2007). Thus, we hypothesize:

Hypothesis 4: The implementation of SAAs will have a more negative effect on levels of educational enrollment than on levels of immunizations.

It is difficult to assess the cost to the state of reducing the level of child labor relative to increasing enrollments and immunizations, so we cannot test our theoretical proposition concerning cost for this variable.

Beyond the entry into SAAs and their implementation, if SAAs are ultimately successful in allowing states to achieve certain economic goals, those achievements may also have an impact on child rights outcomes. One of the goals of structural adjustment is to open a country's market to the global economy. Proponents of neoliberal economic policy emphasize the creation of jobs in developing countries through trade liberalization and argue that this can reduce the need for children to work to supplement family income. For example, Drezner (2004) has found that as the Vietnamese gained access to the global market, there was a reduction in children working in rice paddies. Furthermore, wealthy countries' consumers may reject developing country exports that are made with child labor. On a less positive note, others emphasize how detrimental working conditions and reduced social protection associated with trade openness could reduce children's access to health care services and education (Richards & Gelleny 2007). We predict, during the period we consider, that open markets will have a deleterious effect on local citizens because increased global competition can put local firms and farms out of business. The effect on local citizens may spill over to children, displacing them from the labor market, while decreasing families' ability to pay for education and primary health care.

Hypothesis 5A: Trade openness will reduce child labor.

Hypothesis 5B: Trade openness will be associated with lower levels of infant immunizations and educational enrollments within countries.

In sum, we expect national acquiescence to international law to be associated with greater improvements in child survival and development rights. We also expect more CRINGOs to be associated with better implementation of child rights, as they provide both structures to carry out those rights and pressure on governments to make those rights a priority. Finally, we hypothesize separate associations of SAA adoption, implementation, and goal achievement with the different types of child survival and development rights.

Research Design

The goals of our analysis were to determine whether child rights mobilization and recent global economic reform have affected outcomes for children in developing countries over time. To evaluate our hypotheses about the evolution and impact of the child rights and neoliberal economic regimes, we utilized cross-national longitudinal data at yearly intervals for more than 99 poor and middle-income countries (the N varies slightly across the dependent variables) between 1983 and 2001. We focused on poor and middle-income countries because they were the countries at risk of experiencing loan conditionalities during this period. The time period ran from the early use of SAAs to the year in which the World Bank officially moved away from the user fee requirement for education.

Dependent Variables

Child survival and development rights outcomes were the core dependent variables in this analysis. To represent these child rights goals, we used measures of national-level rates of child labor, the percentage of infants immunized, and rates of primary and secondary education enrollment. We do not claim that these four indicators represent the full range of child rights. They are, however, a reasonable representation of the child survival and development category of child rights.

Child Labor

For our historical analysis, we measured child labor as the percentage of children between ages 10 and 14 who were economically active. For this indicator, child labor did not include household labor or apprenticeships.⁴ Household surveys conducted in the 1990s and early 2000s found percentages of child labor similar to the ILO figures in 65 countries, suggesting that country estimates of this indicator are reliable (Fares & Raju 2007). For the child labor outcome model, we excluded countries that claimed to have no child labor in the beginning year of the analysis.⁵ In addition to the questionability of this claim, floor effects made an analysis of these countries' progress impossible. None of these countries recorded changes in child labor in any subsequent years. The CRC Committee hopes that nations will reduce harmful child labor and increase immunizations and educational enrollments. We therefore reverse-coded the child labor variable to make it easier to compare the effects of the independent variables across models.

Immunization

The world has seen a shift in views toward health over time—from a charitable calling to a fundamental human right (Inoue & Drori 2006; see CRC Article 24). Immunization against diphtheria, pertussis (whooping cough), and tetanus (the DPT immunization) is one of the basic interventions for child survival. Data came from the World Bank World Development Indicators (WDI; 2006). The variable is measured as the percentage of 12- to 23-month-old children who are immunized. WDI data on immunization are derived from WHO and UNICEF estimates (http://www.who.int/immunization_monitoring/en/). These estimates are based on official reports from countries and adjusted by WHO and UNICEF officials, who consider potential biases as well as contributions from local experts.

Primary and Secondary Gross Enrollment Rates

To measure improvements in educational access, we used primary and secondary gross enrollment rates, which are comparable across states (Benavot & Riddle 1988). Among all types of ratio measures, gross enrollment rates are most widely used in cross-national studies on global expansion of mass education and by UNESCO and the World Bank (Benavot & Riddle 1988). Gross enrollment rates are calculated by taking the number of students at the primary, secondary, or tertiary educational level and standardizing each by the relevant school-age population, which can vary across countries. Thus, gross enrollment rates take national variations in the duration and structure of schooling into account, allowing a fair degree of cross-national comparability (Benavot & Riddle 1988).⁶

Independent Variables

Treaties Ratified by Each Country

Our analysis utilized measures of the strength of child rights mobilization since 1983. We considered, first, the ratification of the ICESCR and, second, the ratification of two major child rights treaties, that is, the CRC (1989) and the ILO Minimum Age Convention (1973). Countries that have ratified the ICESCR recognize the positive obligation of states to provide certain services to citizens, paving the way for the implementation of child survival and development. Countries that have ratified the child rights treaties have committed themselves even more specifically to the implementation of child survival and development. Figure 1 shows the ratification trend for the two child rights treaties. Since all but two countries have ratified the CRC, the primary variation for this treaty relates to the timing of ratification, which ranges from 1989 to 1996. The ILO Minimum Age Convention had a steady stream of ratifications when it was first proposed. There was then a lull in ratifications until its ratification became a precondition for the ratification of the more popular ILO Worst Forms of Child Labor Convention (which was available for ratification only in 1999, at the very end of our time period) (Boyle et al. 2006).⁷ Thus, in some sense, ratifications of the Minimum Age Convention capture the varying commitment to two child-rights oriented treaties. Early adopters have the strongest commitment to child rights, late adopters are part of the more recent “norm cascade,” and those who never adopted the treaty show the lowest level of commitment to child rights. We hypothesized that treaty ratification should facilitate progress toward child rights goals.

Figure 1. Cumulative Number of Ratifications of the Two Child Rights Conventions by Non-OECD Countries, 1973ndash;2001

CRINGO-Domestic Civil Society Linkage

The second measure of child rights mobilization is the number of INGOs to which citizens or organizations of a country belong. INGO linkage captures the extent of a state's participation in world society, in this case, in the sphere of child rights (Boli & Thomas 1999). Membership data came from the Yearbook of International Organizations (Union of International Associations, various years), which reports country memberships annually from 1982. Using the 2007 version, the most current version available, we identified 93 CRINGOs. An INGO was regarded as a CRINGO if its purposes included the promotion of child rights, the promotion of education, eliminating child labor, eliminating child trafficking, eliminating corporal punishment, or similar goals or activities. Figure 2 indicates that the average number of country affiliations with CRINGOs increased from just over 5 in 1980 to around 13 in 2001.

Figure 2. Mean of Country Child Rights INGOs, Non-OECD Countries, 1980–2001

SAAs and Implementation

A number of analyses have suggested that the conditionalities placed on nation-states wishing to borrow money from the World Bank or the IMF have hurt the most vulnerable members of societies, including children (Tomasevski 2005; R. Evans 2002). Our study assessed these claims across countries and over time. We focused on loans from the World Bank because, unlike IMF loans, these are not awarded to alleviate pre-existing financial crises within countries (which would create selection effect problems). Figure 3 illustrates the number of World Bank structural adjustment loans given to OECD and non-OECD countries (Organisation for Economic Co-operation and Development) during the 1983 to 2001 period. From Figure 3, it is clear that almost all recipients of loans were developing countries and that the number of loans began to increase in the early 1980s. To measure SAA adoption, each year a country entered an SAA counted as a 1; all other years counted as 0. We assumed that the effect of SAA implementation would not occur immediately, but rather would take some time. Consistent with the literature (Abouharb & Cingranelli 2006), our data indicated that on average SAAs last for three years. Thus, it is reasonable to assume that changes associated with the implementation of SAAs should appear in the three years following the adoption of an SAA. To capture implementation, therefore, we coded the three years following a SAA adoption as 1; we coded all other years as 0. As noted in our introduction, for a period in the late 1990s, the World Bank may have been somewhat less focused on fiscal austerity in the awarding and implementation of SAAs. To test this, we included a time-period dummy variable for pre- and post-1995 in analysis that is not shown, but it did not affect the direction or statistical significance of coefficients.

Figure 3. World Bank Structural Adjustment Loans, 1980–2001

Trade

Trade as percentage of GDP was included as a more general measure of the economic openness of a country. This came from the World Bank World Development Indicators (2006). Our measure of trade was the sum of exports and imports as a share of GDP. Assuming it takes some time for trade openness to affect child rights implementation, we included the trade level three years prior to the year in which a child rights outcome was measured.

Control Variables

Lagged Dependent Variables

Autocorrelation is common in most time-series models as many variables exhibit a high level of serial autocorrelation (Ostrom 1990; Baltagi 1995; Kennedy 1998). The four dependent variables used in this analysis are changing only incrementally over time. We addressed this by including a three-year lagged term for each dependent variable. As our analysis covered the period 1983 to 2001, lagged terms included values from 1980 to 1998. Using lagged dependent variables was appropriate in our analysis for two reasons. First, while scholars caution that using lagged dependent variables as a solution for autocorrelation can lead researchers to mistakenly discount the importance of other independent variables, particularly if they do not vary dramatically over time (Achen 2000), the results will still be unbiased but more conservative. Second, substantively, this research focuses on factors influencing changes in the levels of child rights outcomes over time rather than the determinant of raw levels of child rights outcomes. Thus, controlling for previous levels of the four child rights outcome measures was crucial for the analysis.

Democracy

The Polity IV score variable was included to evaluate different effects across autocratic and democratic governments with respect to child rights outcomes. This indicator was derived from the coding of the institutionalized polity score (Marshall & Jagger 2002), where a variable rages from −10 (full autocracy) to +10 (full democracy). The polity score is updated on a yearly basis and we used the 2004 version of it, which covers the period 1880 to 2004.

Foreign Direct Investment

Although the majority of foreign direct investment (FDI) flows between developed countries, there has been a pronounced growth of FDI in developing countries over the last few decades (UNCTD 2006), and this has been shown to affect at least some aspects of child well-being (Neumayer & Soyasa 2005). The World Bank provides information on inflows of FDI as a percentage of GDP. We included in our equations the FDI level three years prior to the year in which a child rights outcome was measured.

GDP per Capita

In order to account for a country's general level of economic development, we included GDP per capita and annual change in GDP per capita. GDP per capita is a standard control variable in cross-national research. The GDP per capita measure came from the World Bank World Development Indicators (2006) and is logged to reduce skewness. GDP per capita change measures increases or decreases in GDP per capita from the previous year (t−1).

Total Population, Percentage of Population Under 14 Years Old, and Percentage of Urban Population

An important influence on both child survival and child development may be the relative size of the child population. For example, once a country has committed to universal schooling, we might expect public and private investments in schooling to rise or fall depending on the sheer size of the school-aged cohort. Likewise, migration within a country, particularly urbanization, might have an influence on both the supply of and demand for child labor, and on the resources required for schooling and immunization. Data for these variables came from the World Bank World Development Indicators (2006).

Statistical Analysis

We investigated the effects of economic and cultural globalization on child rights outcomes with a pooled time-series analysis of poor and middle-income countries from 1983 to 2001.⁸ The unit of analysis was country-year. The data file included 19 years for most of the countries. Pooled time-series analysis is an appropriate and accessible method for measuring dynamic historical trends such as the increase in globalization and its effects on child rights outcomes across many different states. The period under analysis covered those years in which loan conditionalities and child rights requirements promoted inconsistent national policies with respect to user fees for education. In addition, the time period incorporated considerable variation in national economic and political circumstances.

Utilizing Hausman's Chi-Square Test, we determined that fixed-effects models are preferable to random-effects models for the models presented in this article.⁹ We used panel-corrected standard errors to take into account heteroskedastic standard errors. The fixed-effects model took the general form:

${\text{Y}}_{\text{it}}=\sum \left({\alpha }_{\text{i}}{\text{D}}_{\text{i}}\right)+{\beta }_{\text{it}}{\text{X}}_{\text{it}}+{\varepsilon }_{\text{it}}$

where Y _it represents the dependent variables in state i in year t, and D _i represents a vector of state-specific indicator variables with individual-specific constants α _i. X _it represents a vector of independent variables with coefficients β _it. ɛ_i represents the error term. This fixed-effects strategy allowed us to model general effects of economic and cultural globalization while respecting cross-national variation (Sassen 1998; Brady & Wallace 2000). Using this general structure we specifed our models as follows:

$\begin{array}{c}{\text{Child Rights Outcome}}_{\text{it}}={\alpha }_{+\beta \text{1}}\hspace{0.17em}\text{Past}\hspace{0.17em}{\text{level}}_{\text{t}-3}+{\beta }_{\text{2}}\text{ICESCR}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}{+}_{\beta \text{3}}\text{CR}\hspace{0.17em}{\text{Treaties}}_{\text{t}}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}+{\beta }_{\text{4}}{\text{CRINGO}}_{{\text{S}}_{\text{t}}}+{\beta }_{\text{5}}\text{SAA}\hspace{0.17em}{\text{approval}}_{\text{t}}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}+{\beta }_{\text{6}}{\text{SAA implementation}}_{\text{t}}+{\beta }_{\text{7}}\hspace{0.17em}{\text{trade}}_{\text{t}-3}\\ \hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}+\sum {\left({\beta }_{\text{k}}\hspace{0.17em}\text{other}\hspace{0.17em}\text{control}\right)}_{\text{t}}+{\varepsilon }_{\text{it}}\end{array}$

where Child Rights Outcome _it represents, respectively, the percentage of children not engaged in child labor, primary education enrollment rates, secondary education enrollment rates, or the percentage of 12- to 23-month-old children who are DPT-immunized in state i in year t. Past level is the level of the respective child rights outcome three years earlier. ICESCR is a dummy variable for whether the country has ratified the ICESCR by year t. Treaties is the number of child rights treaty ratified (0, 1, or 2) and CRINGOs is the number of memberships in child-rights INGOs, both in year t. Structural adjustment variables are World Bank loan approval and implementation, and trade openness (t−3). Controls include country-specific political (democracy versus autocracy), economic (GDP per capita, GDP growth per capita, and FDI [t−3]), and demographic (population, proportion of under 14 in population, and urban population) variables. For primary and secondary enrollments, we added the child labor level in addition to other controls to examine if working hinders primary and secondary school enrollment. While they improved the reliability of our results, a drawback of fixed-effects models is that they do not allow one to test for regional effects. We hope to take up possible differences across regions in future studies.

For indicators that were not measured in every year—child labor, primary and secondary education, and CRINGOs—we interpolated missing values. We tested the interpolation process of the three dependent variables by interpolating the entire time series based only on one observation per five-year period. Regressing these interpolations on the true values yielded R-squares of greater than 0.99 for child labor, 0.98 for primary enrollment, and 0.99 for secondary enrollment. Data availability improved over the time period studied. We coded country membership information in CRINGOs using data from 1981 and then every five years beginning in 1985, interpolating the values between those years. Previous studies using this approach for missing INGO data include Meyer et al. (1997), Boli and Thomas (1999), Frank et al. (2000), Hironaka (2002), and Belkin and Schofer (2003). They find INGO data highly serially-autocorrelated and defend interpolation of INGO data as appropriate.

To test the robustness of our results, we considered outliers and influential cases. We used a number of diagnostic procedures including partial plots and Cook's D to identify these cases.¹⁰ The removal of these cases individually or as a group had no substantial effect on the models.¹¹ We also performed various sensitivity tests on the models, using scatterplots of the residuals. We did not find substantial problems of heteroskedacity or non-normality. Although there was a modest level of correlations among some variables, raising some concern of multicollinearity, the overall stability of the sign and significance across models with different specifications supported the robustness of the results presented here.

Results

The results of our analysis are shown in Table 1. For each of the four dependent variables, we included two models, one excluding and one including the ratification of the ICESCR. Columns 1A and 1B model the effects for child-labor (reverse-coded); Columns 2A and 2B model the effects for our measure of children's health—levels of DPT immunization; and Columns 3A–3B and 4A–4B model the effects for primary and secondary school enrollment, respectively. We hypothesized the effects of the first six variables listed in Table 1 and included the remaining variables as controls.

Table 1. Unstandardized Coefficients From the Fixed-Effects Regression of Various Indicators on Child Rights Outcomes, Non-OECD Countries, 1983–2001

Note: Standard errors in parentheses; two-tailed tests for control variables, one-tailed tests for all others.

^* p<0.05; ^**p<0.01; ^**p<0.001.

Turning first to the effects of international law, we found that the ratification of child rights treaties had no effect on the implementation of the child survival and development variables when other factors were controlled. States that ratified two child rights treaties were no more likely to produce better child rights outcomes than states that ratified none or one in any of the eight models.¹² On the other hand, the older and more foundational ICESCR did have a positive effect on three of the four indicators of child rights (all except child labor). Thus, Hypothesis 1A received considerable support while Hypothesis 1B did not.

These results are somewhat different from Hafner-Burton and Tsutsui's (2005) findings. In contrast to their results for human rights, we did not find a negative relationship between treaty ratification and respect for child survival and development rights. The number of child rights treaties ratified showed a positive, although statistically insignificant, effect across the four measures of child rights outcomes. States were not ratifying child rights treaties to cover up egregious child rights violations. These differences in findings across the two studies confirm that the initial impetus to ratify treaties and the pressure to comply once treaties are in place is different in the sphere of child rights than for more traditional human rights. The findings here suggest that child rights principles are less accepted in the world polity than other forms of human rights. Countries that are amenable to the concept of economic rights (as evidenced by their ratification of the ICESCR) represent a segment of the world polity that does appear to accept the legitimacy of child rights.

Next we turn to the impact of child rights mobilization within countries, as measured by the number of child rights INGOs with affiliate members within each country. Supporting Hypothesis 2, states whose citizens affiliated with a greater number of CRINGOs were more likely to show a positive trend in child rights outcomes. The presence of CRINGOs was linked to reductions in child labor (Models 1A and 1B), increases in immunizations (Models 2A and 2B), and increases in secondary enrollment rate (Model 4A and 4B). It also had a statistically significant positive effect on primary enrollment rates when the ratification of the ICESCR was not considered (Model 3A). On average, for each 10 linkages to CRINGOs, the countries in our sample experienced a modest increase of a 0.17 percentage point in the rate of children not engaged in child labor (Model 1A), an increase of 2.70 percentage points in DPT immunization (Model 2A), and larger increases of 1.76 percentage points in primary school enrollment (Model 3A) and 1.34 percentage points in secondary enrollment (Model 4A). This is consistent with Hafner-Burton and Tsutsui's (2005) finding that the support system that develops following treaty ratification is especially important for implementation.

Model 3B indicated that CRINGOs did not have a significant impact on primary educational enrollments over and above the effect of a country having ratified the ICESCR. This suggests that the ICESCR laid the foundation for high levels of participation in national educational systems. The ICESCR may have encouraged the participation of CRINGOs in this process so that we saw overlapping effects for ICESCR ratification and CRINGO levels. The two variables explained the same variance in enrollment outcomes. This makes sense because Article 13 of the ICESCR requires states to provide free compulsory primary education and to progressively introduce free secondary education. Article 14 of the Covenant requires countries to devise a detailed implementation plan for fulfilling the obligations of Article 13.¹³ Overall, the results in Table 1 provided support for Hypothesis 2 that a greater presence of CRINGOs in a country leads to improvements in child rights outcomes.

The combined findings for ICESCR ratification, child treaty ratification, and CRINGO linkages raise questions about the relationship between international law and CRINGOs. For example, perhaps child rights treaty ratification has an indirect effect on child rights outcomes by fostering the growth of CRINGO linkages. International law may improve child rights outcomes by legitimizing child rights discourse in global civil society (Hafner-Burton & Tsutsui 2005). This, in turn, may facilitate the diffusion of CRINGOs into national civil society and empower domestic mobilization around child rights (compare Liu & Boyle 2001). After completing our analysis of Table 1, we explicitly examine this possibility.

Our remaining hypotheses related to neoliberal economic reforms and goals. We first hypothesized (Hypothesis 3) that the World Bank would select countries with good child rights records for loans, resulting in a positive association between loan approval and child rights outcomes. This hypothesis was supported with respect to child labor and immunizations. On average, children not in the labor force were 0.66 percentage points higher during the loan negotiation process (Model 1A). SAA adoption also had a significant positive effect on levels of immunization. During the year of negotiation of a loan from the World Bank, levels of DPT immunizations increased by 1.42 percentage points on average (Model 2A). The adoption of SAAs did not affect primary and second education enrollment levels (Models 3A–3B and 4A–4B). Overall, Hypothesis 3 received partial support.

In terms of implementation, our findings showed little direct effect of SAAs on either education or immunizations. The effect on immunizations was positive but statistically insignificant. The effect on both primary and secondary enrollments was negative, but again, statistically insignificant. This provides only very weak support for our hypothesis that SAA implementation would have a more deleterious effect on child rights reforms that are costlier to implement (Hypothesis 4). Future research with more precise measures of costliness is necessary.

Turning to child labor, we found that the level of child labor went down during SAA implementation. The effect of reducing child labor was very small but statistically significant in both Models 1A and 1B. Since the World Bank is a key player in global economics, and it has explicitly endorsed international labor standards that call for abolishing child labor (Nelson 2000), activists may have been watching closely to see if its loans had a deleterious effect on child labor practices within countries. To avoid controversy, the World Bank may have been particularly observant of child labor practices during the implementation of loan conditionalities. Another possibility is that children are displaced from work as the cost of adult labor gets cheaper, a possible side effect of SAA implementation.

While trade openness had no effect on child labor, it slightly lowered the level of DPT immunization (Models 2A and 2B) and primary (Models 3A and 3B) and secondary enrollment rates (Models 4A and 4B). Our Hypothesis 5B concerning these effects was supported. This is consistent with the literature arguing that trade-induced volatility undermines social protections for the most vulnerable in developing countries (Rudra 2002). Overall, our findings on economic factors suggested that loan conditionalities were not directly impeding child survival and development rights outcomes, but the goals sought through those conditionalities were in fact affecting in some areas.

Because of the history of SAAs, we thought that child rights trends might be different pre- and post-1995. We therefore included a post-1995 dummy in each model. This did not affect the direction or significance of any coefficient in any model, including coefficients for the SAA-related variables, so we did not include the dummy in Table 1.

Finally, Table 1 also shows the effects of other key variables. Democracy was significantly associated with primary education enrollments. GDP was positively associated with primary and secondary enrollments. In contrast, GDP had a negative association with child labor. This is probably a floor effect—the wealthiest countries' child labor rates cannot get much lower. Total population was also associated with increases in child labor, as well as secondary enrollments (for the latter variable, when the overall rate of child labor was controlled). Child labor rates had a direct adverse effect on primary education. The larger the percentage of the population under 14, the less the progress on levels of immunization and secondary enrollments. High levels of urbanization appeared to facilitate access to immunizations and secondary education. FDI had no significant effects on child rights outcomes. These relationships provide additional context for our main findings.

For those interested in international law, the most surprising finding in Table 1 was the lack of effect of child rights treaties on child survival and development outcomes. To further investigate this relationship, we conducted a pooled, fixed-effects time-series analysis of the total number of CRINGOs within countries for the same 1983–2001 period, looking for possible effects of the treaty ratification variable. Table 2, Model 1, shows the baseline relationships involved. Child rights treaty ratification had a strong, positive effect on CRINGO memberships. This is consistent with Barrett and Tsui's (1999) finding that global elites pay more attention to norm-compliant nations. As one would expect, the more democratic and economically developed a country was, the greater the number of CRINGO memberships.¹⁴ In Models 2 through 5, we added additional variables signifying general economic globalization, SAAs, ICESCR ratification, and the number of CRINGOs in the world, respectively. Even after controlling for these factors, child rights treaty ratification continued to show a positive, significant effect on the number of country memberships in CRINGOs. This means that child rights treaties have an indirect positive effect on child survival and development rights outcomes. Ratifying these treaties early in the period was associated with greater increases in CRINGOs, and from the results in Table 1, we knew that the presence of CRINGOs was associated with improved implementation of all the child rights considered here.

Table 2. Unstandardized Coefficients From the Fixed-Effects Regression of CRINGO Membership, Non-OECD Countries, 1983–2001

Note: Standard errors in parentheses; two-tailed tests for control variables, one-tailed tests for all others.

^* p<0.05; ^**p<0.01; ^***p<0.001

The ratification of the ICESCR was not associated with increasing numbers of CRINGOs during the period considered. Quite possibly, the ICESCR prompted mobilization in an earlier period, and its direct effect on CRINGOs had subsided by 1983. Interestingly, results from Table 2 suggest that SAA adoption and implementation have significant positive effects on the number of CRINGOs within a country as well. This relationship may indicate counter-hegemonic globalization (P. Evans 2000) in which transnational networks of people and organizations mobilize against neoliberal principles and practices (Keck & Sikkink 1998; Donnelly 2002). This poses an intriguing question about the paradoxes of structural adjustment for future scholars to consider.

Conclusion

Two things are clear from our findings. First, international law, mobilization efforts, and global economic reforms are interrelated spheres of globalization that should all be considered when it comes to the implementation of child survival and development rights. For example, while the ratification of international child rights treaties and the implementation of SAAs had little direct effect on child rights outcomes, both increased nation-states' linkages to global civil society through the presence of CRINGOs, and CRINGOs did have a direct positive effect on child rights outcomes. Second, within these spheres, different factors have different effects on different types of child right outcomes. For example, the effect of negotiating an SAA had a positive effect on some child rights outcomes, but trade openness had a negative effect on most of them. Aspects of global economic reforms and aspects of international laws had disparate effects, and this further varied depending on the child rights outcome considered. The implication is that researchers must break down spheres of globalization into various component parts when theorizing connections across them.

Delving more specifically into our findings, in the area of international law, we found that countries that have acquiesced to the global principle of economic, social, and cultural rights, as evidenced by their ratification of the ICESCR, had better records of child rights improvement than other countries. It appears that more established international law had the greater effect because national ratification of child rights treaties, in contrast to the ICESCR, had no direct effect on child rights outcomes. However, through a second analysis, we learned that child rights treaties did indirectly improve outcomes for children by encouraging CRINGO activism. Consistent with a world polity outlook, then, treaty ratification encourages national connections to INGOs, which in turn benefits youth.

On the economic side of global reforms, we found some evidence that recent global economic reforms undercut the implementation of child rights, but we also found evidence that some aspects of global economic reform had a positive effect on some child rights outcomes. In general, child labor statistics improved when a country was in an active relationship with the World Bank, while other indicators of child rights were largely unaffected. The exception was the percentage of infants immunized, which was positively associated with entering into an SAA.

In terms of the implementation of SAAs, we found only weak support for the idea that SAA implementation has a negative impact on child outcomes, such as education, that are normally funded through government spending. Although the direction of the relationship appeared to be negative, it was not statistically significant. Since Mutangadura et al. (2002) found a negative effect of debt level on secondary school enrollment, the lack of a relationship in this analysis is surprising. It suggests again the complexity of neoliberal reform effects. Apparently, debt levels and the implementation of loan conditionalities have unique impacts on enrollments. More research is necessary to sort out the differing effects of debt level and structural adjustment reforms.

We also hypothesized that trade openness—a key goal of SAA conditionalities—would have a negative effect on child rights during the time period studied. We found that higher levels of trade decreased the percentage of infants immunized and both primary and secondary education enrollments. The problem may be that developing countries' reductions in trade barriers are not matched by wealthier countries. Consequently, cheap, subsidized products from the United States and Europe can undercut the profits of developing country farms and industries. The resulting negative consequences can affect many individuals, including, according to our analysis, children.

Finally, our research suggests that both the adoption and implementation of SAAs increased the number of child rights INGOs within countries. Trade openness, arguably the most troublesome element of economic neoliberalism studied here, however, was not related to CRINGOs' presence. Since CRINGOs showed the strongest and most consistent effects in improving child rights outcomes, this means that SAAs indirectly improved outcomes for children. This highlights the importance of integrating world polity effects when analyzing the overall impact of economic reforms on non-economic outcomes. It suggests that SAAs may prompt mobilization and thereby improve outcomes for children.

In her recent presidential address to the Law & Society Association, Lauren Edelman called for greater synthesis of sociological and economic theoretical approaches (Edelman 2004). A major contribution of this research is to integrate analyses of cultural and economic globalization. To do this, we identified some key components of child rights and economic neoliberalism and theorized the interplay between these component parts. Specifically, we hypothesized first that the adoption of SAAs would coincide with improvements in child rights outcomes. International financial organizations operate in the same rights-oriented normative environment as nation-states. Failure to support child rights threatens their legitimacy. By embracing the complexity of cultural and economic global reforms, this research begins the process of understanding the interaction between global rights discourses and neoliberal economic reforms.

There are many directions in which to take this research. In terms of law, it would be useful to determine the nature and extent of national-level child rights policies (in codes and constitutions). Most but not all countries in the world have constitutions that adopt broad conceptions of rights (Blau & Moncada 2007). Starting points for such analyses would be the Child Rights Information Network (http://crin.com/) and the Children's Legal Centre (2009). National policies may be more directly effective for improving child rights than treaty ratification. Future work will use detailed case studies on the history and impact of these laws within countries.

In terms of global economic policies, to determine with even more specificity the impact of SAAs, it would be useful to identify the specific provisions of each SAA and the purposes of their related loans. In addition, recent global financial trends may signal a move away from neoliberal economics, paving the way for a comparative analysis of the impact of different types of global economic reforms. As an empirical analysis of historical trends, the current contribution does not provide alternative ways of doing or understanding economic globalization. Counter-hegemonic movements with those aims will be watched closely in the future; contemporary financial woes may create unique opportunities for influence (see de Sousa Santos 2006).

In terms of child rights themselves, this study's measure of child labor was not ideal. Future work that considers more recent time periods will be able to use a more refined indicator of this variable. Furthermore, this study considered only one dimension of child rights, the right to survive and develop. The other three areas—rights against discrimination, rights to have best interests considered, and participation rights—may have very different trajectories. The present study has demonstrated the importance of deconstructing and integrating the different spheres of globalization. Regardless of the new empirical direction taken by future studies, we hope that researchers will continue to build on these principles.