Paper

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=fjhr20

The International Journal of Human Rights

ISSN: 1364-2987 (Print) 1744-053X (Online) Journal homepage: https://www.tandfonline.com/loi/fjhr20

When everyone agrees: human rights norms on women and children and their effects on health

Heather Smith-Cannoy, Wendy H. Wong, Arjumand Siddiqi, Christopher Tait & Abtin Parnia

To cite this article: Heather Smith-Cannoy, Wendy H. Wong, Arjumand Siddiqi, Christopher Tait & Abtin Parnia (2020): When everyone agrees: human rights norms on women and children and their effects on health, The International Journal of Human Rights, DOI: 10.1080/13642987.2020.1743975

To link to this article: https://doi.org/10.1080/13642987.2020.1743975

Published online: 30 Mar 2020.

Submit your article to this journal

Article views: 42

View related articles

View Crossmark data

When everyone agrees: human rights norms on women and children and their effects on health Heather Smith-Cannoya, Wendy H. Wongb, Arjumand Siddiqic, Christopher Taitc and Abtin Parnia c

aSchool of Social and Behavioral Sciences, Arizona State University, Glendale, AZ, USA; bDepartment of Political Science, University of Toronto, Toronto, ON, Canada; cDalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada

ABSTRACT What are the effects of international human rights norms? This paper links the creation of two human rights treaties, the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) and the Convention on the Rights of the Child (CRC) to health indicators that signal whether well-being for women and children is improving. These two conventions are of particular interest because nearly all countries in the world are parties to both treaties, which attests to their strength as international norms. Using two new methodologies – ITSA and Joinpoint analysis, we show that both CRC and CEDAW have demonstrable and significant effects across a variety of relevant health indicators and have had effects in countries regardless of region, level of development, and regime type. Thus, our paper is one of the first to link the existence of norms with cross-national effects on the ground within countries not as rules to be enforced, but rules that have effects outside of courts or political offices that create changes for individual well- being beyond the treaties and law themselves.

ARTICLE HISTORY Received 31 January 2019 Accepted 10 March 2020

KEYWORDS Human rights; women’s rights; children’s rights; global norms

What are the effects of international human rights norms? Although considerable energy (as we show below) has been spent articulating what norms are and showing that they exist, we still struggle in international relations (IR) research with a sense of what norms, often in the guise of international law as treaties, actually do once created. In short, while we have made giant strides in clarifying the concept of norms and demon- strating that they matter in guiding state and other actors’ behaviour in terms of the cre- ation of legal instruments, we do not have as good a sense of the effects of those instruments, once created, outside of analyses of court activities1 or anticipated effects on leaders.2 If a norm is supposed to signal consensus around certain standards, it means that agreement has been reached. But what does that agreement actually do and imply?

In this article we ask not necessarily about why states comply with treaties, or what factors might increase or decrease compliance. Rather, we are interested in the effects of treaties on domestic outcomes. Similar to Simmons,3 we focus not on how international human rights law translates into domestic law, but how individuals’ well-being might

© 2020 Informa UK Limited, trading as Taylor & Francis Group

CONTACT Heather Smith-Cannoy [email protected]

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS https://doi.org/10.1080/13642987.2020.1743975

improve at the domestic level as a result of the uptake of international human rights norms. This research question allows us to move beyond the traditional approach to treaty effectiveness in the international relations literature and allows us to assess the types of quality of life indicators, which to date, have been neglected from rigorous stat- istical analysis.

This paper links the creation of two human rights treaties, the Convention on the Elim- ination of All Forms of Discrimination Against Women (CEDAW) and the Convention on the Rights of the Child (CRC) to health indicators that signal whether well-being for women and children is improving. We are interested in these two human rights conven- tions in particular because nearly all countries in the world are parties to both treaties,4

which attests to their strength as international norms. We recognise that even with such high global rates of ratification for the CRC and CEDAW, assessing norms is not as straightforward as measuring treaty ratification because state enter a wide variety of reservations, understandings and declarations that proscribe their commitments. We use treaty ratification as a good, though imperfect measure of norms to answer larger ques- tions about their domestic effects. We are interested in health indicators, rather than typical political indicators that are used in much of IR analysis5 on human rights, as many of the articles included in both the CEDAW and CRC account for economic and social rights, which have effects on general well-being within societies that is not captured in data focused on civil and political access and/or repression. What we show here is that both CRC and CEDAW have demonstrable and significant effects across a variety of rel- evant health indicators and have had effects in countries regardless of region, level of development, and regime type. Thus, our paper is one of the first to link the existence of norms with cross-national effects on the ground within countries6 not as rules to be enforced, but rules that have effects outside of courts or political offices that creates changes for individual well-being beyond the treaties and law themselves.

What does it mean to assess the effects of norms? How is compliance with a treaty dis- tinct? Throughout this research we wrestle with a critical challenge – how can we dis- tinguish between the effects of broad global norms (like respect for women’s personal integrity rights for example) and treaty commitment. We begin with the premise that global norms are perpetually in flux. We may begin to assess the state of women’s rights in North America in the 1970s to find a certain set of norms, but those standards, rules and expectations are distinct from those we might find in 1990s Eastern Europe, or rights for women in Middle East in the 2000s. Norms are both culturally and temporally distinct. Treaties, however, when they are well ratified can serve as an indicator of norm consensus, reflecting a general consensus surrounding the rights of the population in ques- tion at the time of their creation. While treaty commitment in theory does imply that a government ought to follow through with their commitment to the norm, we know that there are many obstacles associated with compliance following commitment. We therefore use treaty ratification as a basic benchmark of support for the norms contained within the treaty, not because we expect that governments will be willing or able to faith- fully execute all of the articles within it. In this sense – though treaty commitment is expli- citly measured as de jure adoption of either the CEDAW or the CRC, we see norm compliance as a broader attempt to improve the well-being of these populations.

Below, we tie the existing literature in IR on international norms to the link between treaties and human rights outcomes. We argue that we need to expand beyond our

2 H. SMITH-CANNOY ET AL.

typical, static, and binary understanding of treaties as ratified or not, which in turn indi- cates some kind of adherence to a norm or not. Because CEDAW and CRC have near-uni- versal adherence, the appropriate question is not yes or no, but what happens next and how. We justify examining changes in health indicators as our dependent variable. We then explain our methodology using joinpoint regression, a method that allows us to assess the effects of the treaties, given that states have widely recognised the importance of norms espoused within CEDAW and CRC. We further illustrate the effect of treaties on health outcomes through a novel application of interrupted time series analysis allow- ing us to make inference about the role of treaties in the absence of a control group of non- ratifying countries. Our conclusion discusses the implications of the research approach and future steps.

Literature review

In international relations, we tend to think about norms as the spread of ideas that is static in time in the sense that there is a discrete before and after period. Before the change of norms, states (or other actors) were doing X, and after signing a treaty or being socialised, they do Y. The end of X and the beginning of Y marks the change in norms, which are commonly defined as a ‘standard of behavior for a given identity’.7 One common way, par- ticularly in the human rights-related studies of norms, is to emphasise the ratification of international treaties, the act of which is taken to indicate compliance with some common set of understandings of right and wrong.

The use of a discrete end point, such as a treaty creation or ratification, however, can result in skepticism with the effect of treaties despite the fact of the more encouraging outcome of agreement to and about the contents of a treaty. The more skeptical among us suggest that the treaties have little capacity to alter state behaviour.8 Apparent instances of compliance with treaty language are, according to these scholars merely a coincidence of state interests aligning with treaty language9 or are more aptly explained with reference to underlying power structures in the international system.10 In this regard, all that we can say about those states that ratify human rights treaties is that they are willing to offer lip service to the treaties terms, but the treaties do not themselves constrain state behaviour,11

though individual judges might utilise them in their decisions.12 Early empirical studies on the relationship between treaty ratification and human rights outcomes reinforced these pessimistic conclusions, suggesting that ratifying states tend to practice torture and geno- cide at higher rates and possess worse protections for civil, political and women’s rights than non-ratifying states.13 With respect to women’s rights specifically, economic devel- opment was linked to a greater likelihood of both economic and social rights, crowding out the role of treaty commitments.14

Some of this early skepticism about the effects of treaties on rights has been mitigated by more recent empirical research. Improvements in human rights protections following treaty ratification are often contingent on a state’s level of democracy. More democratic states are more inclined to comply with their human rights treaty commitments than their less democratic peers.15 Compliance among democracies has been explained with reference to the costs imposed by a democratic populous on their government for failing to keep their promises16 and by identifying the hallmarks of democracy (elections, independent judiciaries, and active civil societies) all of which make it more difficult for

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 3

governments to repress their citizens’ human rights.17 New research in the field suggests that democratising states in particular tend to join human rights agreements and IGOs that impose high sovereignty costs allowing them to signal their intent to consolidate democracy.18 Treaties also tend to improve human rights outcomes in countries experien- cing regime instability or regime transition.19 Treaties like the CEDAW may also be one aspect of the broader forces of globalisation that improve women’s rights.20

All of the work above focuses on the direct effect of the treaties themselves without noting that reaching a treaty agreement is often just the beginning in terms of affecting state behaviour, especially if state behaviour and law were not already in-line with the dic- tates of treaties to begin with. Norms in particular, as expressed by legally-binding docu- ments, might still leave ambiguities in enforcement and meaning that require domestic interpretation. Two approaches have prevailed to address these ‘gaps’. First, starting with Price’s work on the chemical weapons taboo21 and Acharya’s piece on why norms may be more local than global in origin and practice,22 we have been revising the ‘one- size-fits-all’, linear approach to studying norms. Indeed, some have begun pushing back on this very clean and somewhat incorrect version of norm change by arguing that the process of norm change is much more discursive and much less neat than the dominant approaches might admit.23 If change and tweaking, and not stability, are the inherent fea- tures of international norms, how do we go about understanding norms and their life- cycles beyond the stage-wise process24 or ‘teaching’25 that have become so influential in our field? Both of these approaches emphasise the ‘right’ way and the ‘wrong’ ways to do things. In response, some used critical junctures to show how these moments create opportunities for relevant actors to renegotiate the terms of appropriateness and action.26

Second, scholars have turned to disaggregating states, which has uncovered very many nuances. Where sizeable proportions of a state’s citizens are members of transnational NGOs, states are also more likely to comply with their treaty commitments.27 An impor- tant and troubling implication of this body of work is that in precisely the states where human rights improvements are desperately needed, in repressive, often autocratic regimes, the treaties themselves are unlikely to make a difference.28 Further complicating the analysis, the type of human rights treaty in question may also help to explain compli- ance – the Convention against Torture and the International Covenant on Civil and Pol- itical Rights are associated with worse respect for physical integrity, while the Convention on the Elimination of All forms of Discrimination Against Women is associated with improved respect for women’s rights.29

In short, some scholars focus on domestic legal systems, civil society, and domestic politics directly to compliance with human rights treaties. Simmons’ seminal 2009 work argues that following treaty commitment, domestic political processes are responsible for improving human rights outcomes.30 Domestic actors use treaty ratification to set agendas, litigate, and protest to get governments to comply with their treaty commitments. Her work augments earlier work that statistically linked an active civil society to better human rights outcomes by showing how civil society improves rights outcomes. Civil society groups can also engage in naming and shaming campaigns to promote treaty com- pliance through domestic courts.31

The type of domestic legal system a state possesses may also impact its willingness and ability to comply. The difference between civil and common law legal systems is well docu- mented,32 and these differences have consequences on the effect of international law at the

4 H. SMITH-CANNOY ET AL.

domestic level. Mitchell, Ring and Spellman find that states with common law legal systems, rather than civil law, Islamic law or mixed law systems do a better job at protect- ing human rights, even when controlling for other explanatory factors.33

While IR scholars have concerned themselves with whether and how treaties and norms affect human rights outcomes, a small but growing literature in the field of public health aims to assess the value of treaties in improving health outcomes. Palmer et al. demonstrate that ratification of the six, core UN human rights treaties34 does not reduce HIV prevalence or maternal, infant and child mortalities35 in a global statistical analysis.36 More modest work illustrates that targeted health indicators improve as rates of ratification of the CRC increase.37 In particular, high global rates of ratification coincide with a lower percentage of children under the age of five that are underweight, higher enrolment in primary education, and a decreased rate of mortality for children under the age of five.38 There exist approximately 90 studies examining the effects of treaties on outcomes in fields as diverse as trade, civil and political rights, environmental policy, and personal integrity rights but comparatively very few singling out the effects of treaties on public health.39

Theory

In order to understand how individuals’ well- being improves at the domestic level as a result of the uptake of international human rights norms, we focus on women and chil- dren, and the effects of the corresponding human rights treaties that protect their rights as groups: CEDAW and CRC. There are a number of reasons to focus on these two treaties as a way to address the questions raised in this paper.

First, we are interested in international norms, and if one measure of agreement with a particular set of ideas is assent towards a treaty or not, CEDAW and CRC come very close to near-perfect unanimity in terms of the number of states that have agreed to be covered under their aegis. Thus, if states all have agreed these are important norms to adhere to, and almost all of them recognise the authority of the two treaties that protect them at the international level, both CEDAW and CRC represent some of the strongest human rights norms being held globally. To be sure, ratifying states have entered a wide variety of reser- vations across both of these treaties, making the treaties a good, though imperfect measure of norm compliance.40

Second, our interest in well-being is broader than the general IR approach of measuring the effects of human rights treaties using datasets based on a minimal conception of phys- ical integrity and civil rights only. Commonly leveraged data, such as the Political Terror Scale41 or other newer datasets cover a narrow band of human well-being, largely at the very extreme ends of denial of well-being (e.g. torture, killings, etc.) and not on more mundane ways to think about human well-being, such as mortality rates, lifespan, and other frequent indicators of well-being used in epidemiologic analyses.42 Even more broad-ranging data, such as Cingranelli-Richards Human Rights Data Project (CIRI), which provides annual country-level data on a broad range of human rights indicators,43

are not typically used to their fullest extents in terms of issue area coverage. CIRI data are also limited with regard to the types of rights covered.44 In short, most of the work we have on the compliance with human rights treaties and norms has dealt with the extremes of human depravity, which certainly is a justifiable approach, but it is not the only approach.

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 5

The CEDAW in particular balances the need to protect the more basic civil and political rights of women and girls, while simultaneously emphasising how economic and social rights are needed for the full realisation of such rights.45 And since CEDAW and CRC are lengthy treaties discussing the many ways that women’s and children’s rights can be denied outside of physical integrity violations, these two treaties allow us to explore dynamics outside of those typically explored in IR research.

Third and finally, while protecting women and children certainly can be approached legally, many of the practices are also social and no amount of measuring government activity can truly get at the effects a treaty has on the population in general. As research has shown, civil society organisations take international treaties and create pressure around their compliance domestically, around government officials, but also in creating social programming, assistance, and other services that governments may not be able or willing to provide. The issues around women and children are appealing types of pro- grammes for international donors. In short, by focusing on women’s and children’s rights in particular, we can try to start getting at some of the broader effects of treaties, more specifically, the types of norms they are actually espousing and how that affects indi- vidual lives.

Because our treaties have near-universal membership, agreement to women’s and chil- dren’s rights is a constant and not a variable, and we treat it as such below in our data analysis. We argue that both CEDAW and CRC will have positive effects on well-being outcomes, and we assess the effect of these treaties post-ratification. Our expectations are consistent with IR literature that shows that ratification of human rights treaties can often improve rights outcomes.46

One straightforward way to measure well-being is the life expectancy of individuals in a state. We know from the literature linking treaty commitment to women’s rights, that women’s rights generally improve following commitment.47 Here we seek to push the analysis beyond rights to a broader consideration of their health and well-being. If women’s rights have improved such that they are not subject to physical limitations or abuse and have access to food and medical care, for example, we expect that they will live longer on average, barring accidental death.

H1: Post-ratification of CEDAW, we expect that women’s life expectancies will increase (and associated mortality rates will decline).

Another way to measure well-being is what women can do outside of the home. This in particular might come from freedom in marriage (Article 9, 16) or being able to work outside the home (Article 11). An abundance of research touts the benefits associated with female economic empowerment which includes autonomy and a greater sense of well-being.48 One consequence of female employment is increased household income, which in turn leads women to have fewer children.49

H2: Post-ratification of CEDAW, we expect that women will have fewer children.

Similarly, we can examine the effect of CRC on the well-being of children by looking at mortality rates. There is a substantial literature that links access to health care with improved quality of life for children and increased mortality rates.50 Indeed, without sufficient care in early life, children cannot survive, and the CRC’s first four articles and Article 6 allude to these concerns. Given the priority placed on mortality in CRC:

6 H. SMITH-CANNOY ET AL.

H3: Post-ratification of CRC, children should have increased chances of surviving beyond age 5.

Protecting the health of children is maximised if the child receives proper medical care. Particularly important is prevention of infectious diseases through immunizations. The public health literature demonstrates that immunisation programmes, especially in high income countries tend to improve overall health outcomes.51 We expect that if a government has internalised a set of norms associated with protecting and promoting children’s health, then they may advance this belief by taking steps to boost child immunisation rates.

H4: Post-ratification of CRC, children’s immunization rates should increase.

In terms of how characteristics would affect the changes post-ratification of either CEDAW or CRC, we propose that richer countries will have more success in achieving favourable outcomes across our indicators. Even if governments in poor countries have internalised norms regarding women and children’s health, they face a resource gap in implementing those commitments. The epidemiology literature demonstrates that poverty significantly decreases population health outcomes.52 Because we expect that richer countries have more access to resources and are better able to mobilise their resources to implement treaties, we predict:

H5: Countries classified as high-income or middle-income will see greater improvements across indicators than countries classified as low-income.

The literature, as we summarise above, also indicates that regime type should dictate to what degree well-being might increase post-ratification. Democracies have consistently been shown to respect human rights more pre-ratification, and are more likely to comply with treaties post-ratification.53 Therefore,

H6: Democracies will experience more favorable outcomes across indicators than non- democracies.

Data and methods

Data sources – independent variables

We created a data set including information on treaty ratification year, country income level, geographic region, democracy, and a variety of health outcomes across countries. The present analyses included a global sample of 192 countries with data on health outcomes ranging from 1975 to 2015. Within the context of human rights research, we are particularly interested in investigating the effect of treaty ratification on the subsequent trajectory of a priori identified health indicators measured at the country level. Specifically, we are interested in whether the CEDAW had an effect on the trajectory of health indicators pertinent to women’s health in addition to whether the CRC had an effect on the trajectory of health indictors pertinent to child health across included countries.

We obtained data on treaty ratification status and the corresponding year of ratification from the publicly available United Nations Treaty Series Database.54 Of the 193 countries for which we have data on ratification status, 187 (96.9%) countries ratified CEDAW,55

whereas 192 (99.5%) ratified CRC.56 Notably, the United States has yet to ratify either

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 7

of these treaties. Other countries that have not ratified CEDAW include Iran, Palau, Somalia, Sudan, and Tonga. The US is the only country in the world to have not ratified the CRC.57

Information on country income group classification was determined by the World Bank designations. Economies are originally divided into four income groupings: low, lower-middle, upper-middle, and high. For the purpose of our analyses we collapsed the lower-middle and upper-middle income countries into one category representing middle-income countries. Income is measured using gross national income (GNI) per capita, in U.S. dollars, converted from local currency using the World Bank Atlas method.58 Both level of democracy and income group are measured during the year of ratification.

It is important to consider whether changing economic circumstances of countries intervene and ultimately effects health outcomes. We address this issue by analyzing how many countries in our sample shift categories in World Bank data during the time-frame of our analysis. For our analysis and explanation, please see Appendix 1.

For descriptive purposes, we also classified countries according to the World Health Organization’s (WHO) 6 regions including: Africa, Americas, Eastern Mediterranean, Europe, South-East Asia, and Western Pacific.

We included a measure of the construct of political democracy, as standard in the inter- national relations literature, using the Polity IV composite index which ranges from −10 to 10, with higher values suggesting higher degrees of democracy.59 We dichotomised this variable with countries having a value less than 6 being classified as non-democratic and countries with a value greater or equal to 6 being classified as democratic states. Because we are interested in controlling for the effect of procedural democracy around the time of ratification, we used the Polity IV value that corresponded most closely with each coun- try’s CEDAW and CRC ratification year.

Dependent variable – health outcomes

All information on health outcomes was extracted from the World Bank’s World Devel- opment Indicators (WDI) database.60 In examining the effect of CEDAW on health trajec- tories across countries, we consider the following outcomes: female life expectancy at birth (years), female adult mortality rate (per 1,000 female adults), and birth rate (per 1,000 people). For CRC we consider infant mortality rate (per 1,000 live births), under-5 mor- tality rate (per 1,000), and the % of children aged 12–23 months who were completely immunised. We chose our independent variables of interest for two reasons. First, the text of both treaties suggest that the health of their respective target communities should be prioritised.61 Further, the epidemiology literature emphasises that both mor- tality and immunisation rates are key indicators of population health.62

Mortality is a measure of overall population health that is often used to understand the health of societies, and to compare societies, because (a) it is an objective measure, and less subject to inconsistencies in measurement and (b) it is a ‘summary’ measure of sorts, which gives a sense of health that might be attributable to a broad range of causes (e.g. low birth weight, undernutrition, etc.) in other words, more discrete health outcomes don’t provide a general sense of health as well as mortality. If the CRC is supposed to improve the well-being of children, we should see that a metric of the overall health of

8 H. SMITH-CANNOY ET AL.

children is also improved. Immunisation rates were used because they are a routinely col- lected basic indicator of child health, and are also a metric of a functioning health care system that is responsive to members of society. The notion is that, if the CRC is supposed to improve the well-being of children, we should see that a basic indicator of child health is also improved.

In structuring the data set for these analyses, we took the value for each country that corresponded to equal time since ratification, rather than equal calendar time. For example, for a country that ratified in 1995 we would use the value for the year 2000 as their 5-year post-ratification estimate. Similarly, a country that ratified in 2000, their 5- year post-ratification estimate would correspond to the year 2005.

Methods

We use several approaches in this analysis, most notably joinpoint regression analysis and interrupted time series analysis (ITSA).

Trends over time In studying the effects of large-scale policy change on health outcomes, we are often inter- ested in detecting changes in the trajectories of these outcomes over time – before and after the implementation of the policy (in our case, ratification of CEDAW or CRC). Accord- ingly, we applied several techniques to describe changes to health outcomes before com- pared to the 20 years after CEDAW and CRC ratification across ratifying countries.

First, we conducted paired t-tests to measure differences between pre-ratification and 5- years post-ratification values for our selected health outcomes and further differences between pre-ratification and 10-years post-ratification values. This provided preliminary insight into whether or not there were significant changes in health trajectories at selected time points following the ratification of each treaty.

We then graphically represented trends in our health outcomes stratified by country income group to provide a visual portrayal of similarities and differences in the trajectories of our health outcomes of interest. All descriptive analyses were conducted using Stata/SE version 14.

Joinpoint regression Next, we implemented joinpoint regression models, which detect significant changes in the trajectories of variables over time (in our case 0–20 years post-ratification).63 For each year of observation, we first calculated the mean level of health across all countries, by income group. A total of 18 joinpoint models were run corresponding to our 6 health outcomes of interest stratified across low, middle, and high income countries. Joinpoint models fit linear segments joined at statistically significant inflection points denoted ‘joinpoints’. As an example, for any given outcome, joinpoint models would be able to identify when there were meaningful increases or decreases in the trajectory of a health indicator over time. Interpretations of these trend periods and where the joinpoints that connect them lie, allow us to understand whether there is a constant rate change over time, or rather there are periods of more or less rapid change (or even stagnation) within a pre-defined obser- vation period. When our outcomes are fit on the log scale, the slope of each line segment can be interpreted as the annual percent change in the trajectories of our selected health

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 9

outcome.64 The purpose of joinpoint is to detect significant changes (inflections) in the tra- jectory of a variable, in our case, each health outcome. Joinpoint identifies both major changes in trends (points at which the rate of change is different across the 20 years of data), as well as the annual rate of change within each of these trend periods.

Given the 20 years of post-ratification time points we included in our observation, we did not allow for any more than a 5-joinpoint model to be tested which could result in up to 6 unique periods of change. Because we only had 20 data points of post-ratification data for any country for any given outcome, allowing for more than 5 joinpoints, or allowing a par- ticular joinpoint model to identify up to a maximum of 6 unique trend periods would result in computational inefficiencies that could preclude model convergence. The best-fitting joinpoint models for each of our 18 levels of stratification (6 outcomes x 3 income groups) were identified according to Bayesian Information Criterion (BIC). As a result, each model is capable of having a unique set of statistically significant joinpoints reflecting meaningful trend periods of change. This approach allows for investigation into the hom- ogeneity or heterogeneity in the post-ratification trajectories of various health outcomes across low, middle, and high-income countries. All joinpoint regression analyses were con- ducted using the statistical software package Joinpoint version 4.2.0.2. Joinpoint models of health trajectories can provide meaningful insight into potential short- and long-term effects of these treaties across important subgroups, and can inform more sophisticated hypotheses about whether there was a significant effect of treaty ratification and whether this significant effect is sustained over time using more advanced analytic methods described in the next section. Ultimately, joinpoint regression provides an investigation into possible benefits of treaty ratification, and whether the extent of benefit varies across income levels.

Effectively, trends over short and long periods of time within the 20-year post-ratifica- tion window were analysed by calculating the annual percent change (APC) and the average annual percent change (AAPC) in our selected indicators. Whether individual periods of change (between joinpoints) were statistically significant or not was determined through the use of Monte Carlo permutation methods with a level of significance set at p < 0.05. Statistically significant changes in trends operate in both positive and negative direc- tions corresponding to meaningful increases or decreases in health outcomes over time. The variance estimates upon which the confidence intervals were calculated were based on auto-correlated errors, taking into account serial correlation as countries’ health indi- cators in one year are naturally going to be correlated with their health indicators the fol- lowing year.

Understanding that many factors may affect the trajectory of health outcomes across countries aside from treaty ratification, the aforementioned joinpoint regression models offer more descriptive information before attempting to isolate whether significant and favourable effects are sustained over time and whether these effects could be attributed to ratification itself. However, given that prior research studying the effect of treaty ratifi- cation on health has not formally tested for changes in slopes or inflection points, we emphasise this novel methodological application in this area of research.

ITSA Because accurately characterising pre- and post-ratification trends is the basis of our study, we were motivated to conduct a more extensive modelling exercise to gain insight into the

10 H. SMITH-CANNOY ET AL.

immediate and sustained effects of CEDAW and CRC ratification on our a priori ident- ified health metrics.

We more formally tested our hypotheses about the effect of treaty ratification on health trajectories using interrupted time series analysis (ITSA). ITSA allows for evaluation of the impact of large-scale interventions (e.g. treaty ratification) when no adequate control group is available. ITSA thus overcomes a major barrier initially faced by our research question. Because nearly every country has ratified each of our treaties, we did not have a truly comparable group of non-ratifiers to aid in our inferences about the effect of ratifi- cation on health. By design however, the absence of a control group is overcome by pro- jecting the pre-intervention trend into the post-intervention period to serve as the counterfactual comparison (e.g. what would the trajectory of health look like if it contin- ued the way it did pre-intervention). ITSA therefore relies on availability of multiple observations of an outcome in both the pre- and post-intervention periods. Due to the availability of health data across a wide range of years, we were able to structure our data set with 5 years of pre-ratification data and up to 20 years of post-ratification on over 80% of countries included (recent ratifiers have less post-ratification data available). We chose 5 years as a pre-intervention window because any longer than that might capture multiple shifts in these countries that might preclude comparable intervention effects.

We ran separate models for 6 subgroups, using stratifications based on combinations of economic status and regime type: (1) low-income non-democratic countries, (2) low-income democratic countries, (3) middle-income non-democratic countries, (4) middle-income democratic countries, (5) high-income non-democratic countries, and (6) high-income democratic countries.

First, we simply graphically represented the pre- and post-ratification trends on the same plot to provide a visual representation of key differences across our subgroups. We then employed ITSA to formally model whether the post-ratification trend was significantly different (and more favourable) than the counterfactual (e.g. in the absence of an intervention). This approach allowed us to determine whether there was a significant effect of ratification and importantly if this ratification effect was sus- tained over time. ITSA also provides the parameter estimates to quantify the magni- tude of ratification effect across our 6 controlled subgroups of interest. Notably, ITSA allows us to get closer to making causal inference about the role of treaty ratifi- cation in affecting health outcomes, and whether this effect varies across country income groups and between democratic and non-democratic countries within these income groups. The interrupted time series analyses were conducted using the ‘itsa’ package in Stata/SE version 14.

Results

The ratification period for CEDAW across the 187 countries ranged from 1980 to 2015 (median ratification year = 1989). For CRC, the ratification period ranged from 1990 to 2015 (median ratification year = 1991) (Table 1).

Though entering into force later, CRC ratifiers tended to adopt this treaty much sooner than they did for CEDAW where the distribution of ratifiers is much more spread out, imply- ing greater consensus with regards to the topic and content of the treaty (Figures 1–5).

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 11

Roughly 25% of countries come from the African region, 18% from the Americas, 10% from the Eastern Mediterranean region, 29% from Europe, 6% from South-East Asia, and 12% from the Western Pacific. The vast majority of countries are middle income (52%)

Table 1. Characteristics of included countries by UN treaty. CEDAW CRC

Total Countries (n) 187 192 Median Ratification Year (range) 1989 (1980–2015) 1991 (1990–2015) WHO Region (n, %) Africa 47 (25.1) 47 (24.5) Americas 34 (18.2) 34 (17.7) Eastern Mediterranean 19 (10.2) 22 (11.5) Europe 54 (28.9) 54 (28.1) South East Asia 11 (5.9) 11 (5.7) Western Pacific 22 (11.8) 24 (12.5)

Country Income Level Low 30 (16.0) 31 (16.2) Middle 98 (52.4) 102 (53.1) High 59 (31.6) 59 (30.7)

GDP per Capita at Median Ratification Year (mean) 5369.60 6491.73 GNI per Capita at Median Ratification Year (mean) 5863.31 6236.19

Figure 1. Distribution of ratification year by UN treaty.

12 H. SMITH-CANNOY ET AL.

followed by high (32%) and low-income countries (16%). Because of nearly complete overlap in the set of countries that ratified each treaty (except for the 5 countries that differ)65 the distribution of income level and geographic region is quite similar between CEDAW and CRC-ratifiers.

The mean gross domestic product per capita was roughly 5,400 USD at the median ratifi- cation year for CEDAW and roughly 6,500 USD at the median ratification year for CRC, suggesting that the distribution of wealth had shifted in a positive direction for all countries

Figure 2. Trends in Health Indicator Post-CEDAW Ratification.

Figure 3. Trends in Life Expectancy Post-CEDAW Ratification.

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 13

in the world between these two time points. The proportion of democratic countries increased across country income groups with 35% of low-income countries, 51% of middle-income countries, and 78% of high-income countries classified as democratic.

We hypothesised that treaty ratification would result in more favourable health tra- jectories relative to the pre-ratification trend. For example, because the provisions of CEDAW target discrimination against women, countries where this treaty was ratified

Figure 4. Trends in Health Indicator Post-CRC Ratification.

Figure 5. Trends in Life Expectancy Post-CRC Ratification.

14 H. SMITH-CANNOY ET AL.

should experience decreases in birth rate (H2), decreases in female mortality, and increases in female life expectancy (H1). Similarly, countries that ratified CRC should experience favourable trajectories for child health outcomes including a decrease in infant and under 5 mortality (H3), as well as increases in the coverage of child immu- nizations at the country-level (H4). We further hypothesised that the magnitude of these favourable health trajectories might vary across country by income and regime type (H5 and H6).

CEDAW results

Overall, trends for health indicators suggest that, while high- and middle- income countries experienced a downward or flattened trajectory for all outcomes (indicating improving health status over time), health trajectories in low-income countries were more variable, with adult mortality rates experiencing both an increase and then a sub- sequent decline.

For most regions and income levels, secular trends suggest significant declines in female mortality five and ten years post ratification. However, there was a striking and significant increase in mortality rates in the African region (from 292.6/100,000 in the pre-ratification year to 321.7/100,000 five years post ratification (p = 0.011) and 334.5/100,000 ten years post ratification (p = 0.021). There was also a non-significant increase in female mortality rates amongst the middle-income nations.

Female life expectancy rose significantly after five and ten years post ratification in all WHO regions except Africa, where there were no significant gains over this period. While both middle- and high-income countries experienced significant increases, life expectancy in low-income states was not statistically different after five or ten years post ratification.

Life expectancy trends in the three income groups had similar small upward slopes, though, unsurprisingly; absolute ages of life expectancy at the mean year of ratification and at the end of the twenty-year period were higher as country income rose. Additionally, in the high-income category, there was a larger gap in life expectancy between males and females than in other income groups.

For all WHO regions and all country income levels, birth rates significantly declined five and ten years post ratification of CEDAW (Tables 2 and 3).

In addition to examining changes over larger periods of time, we also investigated more nuanced, year-to-year changes through joinpoint regression, summarised in Table 4.

Results suggest that, amongst low-income countries, there were no average annual per- centage change for any health outcomes. For female mortality rates, there emerged four distinct time trends within the twenty-year period post ratification. For all other health outcomes, analyses discerned three distinct time trends. With the exception of Trend 1, in which mortality rate outcomes demonstrated significant annual percentage changes, there were not additional within-trend annual changes in outcomes. Middle- and high- income countries demonstrated a similar pattern, but with even fewer annual percentage changes within trends, or across the entire twenty-year period (Table 5).

The ITSA allows us to examine favourable ‘interruptions’ in health trajectories away from the trend that was observed prior to ratification. Graphically, figures comparing pre- and post-ratification trends for outcomes related to CEDAW can be seen in Figures 6–8.

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 15

Similar to more descriptive findings, results of interrupted time series analyses demonstrate a rather mixed effect of CEDAW across health outcomes, income groups, and regime type (democratic versus non-democratic). For example, there was a non-significant effect of ratification for female life expectancy among non-demo- cratic, low-income countries, but a significant effect among democratic, low-income countries (see Table 6). Among middle-income countries, CEDAW ratification was significantly associated with decline in birth rate for both non-democratic and demo- cratic countries, but not in low-income countries or high-income countries (significant among democratic countries only).

CRC results

Due to strong overlap in countries, results of mean comparisons (Table 3) and joinpoint regressions (Table 5) were largely similar.

However, results of ITSA much more consistently suggested that CRC had a positive effect on children’s outcomes than was the case for CEDAW (Table 6).

Table 2. Mean value for health indicators pre- vs. post-CEDAW Ratification. Pre-Ratification 5-Years Post 10-Years Post

Birth Rate WHO Region Africa 40.7 38.9 p < 0.001 37.9 p < 0.001 Americas 31.5 28.6 p < 0.001 25.8 p < 0.001 Eastern Mediterranean 30.9 28.3 p = 0.004 27.2 p < 0.001 Europe 16.8 15.3 p < 0.001 14.3 p < 0.001 South East Asia 34.2 30.5 p = 0.003 27.9 p = 0.001 Western Pacific 28.6 26.5 p = 0.001 24.8 p = 0.001

Country Income Level Low 42.9 40.6 p < 0.001 38.7 p < 0.001 Middle 32.3 29.7 p < 0.001 28.0 p < 0.001 High 17.4 16.3 p = 0.001 15.0 p < 0.001

Female Life Expectancy WHO Region Africa 55.6 55.7 p = 0.871 55.2 p = 0.886 Americas 68.5 70.2 p < 0.001 71.5 p < 0.001 Eastern Mediterranean 69.4 71.3 p < 0.001 71.1 p < 0.001 Europe 74.9 76.1 p < 0.001 76.9 p < 0.001 South East Asia 58.1 61.7 p < 0.001 64.6 p < 0.001 Western Pacific 67.7 69.7 p < 0.001 71.1 p < 0.001

Country Income Level Low 53.3 54.5 p = 0.068 55.3 p = 0.168 Middle 64.7 66.1 p < 0.001 67.0 p < 0.001 High 76.0 77.2 p < 0.001 77.9 p < 0.001

Female Adult Mortality Rate WHO Region Africa 295.0 317.1 p = 0.065 337.5 p = 0.034 Americas 165.1 155.5 p = 0.116 151.3 p = 0.212 Eastern Mediterranean 151.9 137.8 p < 0.001 135.1 p < 0.001 Europe 107.9 102.2 p = 0.006 96.2 p < 0.001 South East Asia 260.0 228.3 p = 0.001 202.4 p = 0.001 Western Pacific 185.3 163.6 p < 0.001 153.3 p = 0.001

Country Income Level Low 308.0 310.0 p = 0.874 318.1 p = 0.625 Middle 202.3 200.0 p = 0.653 198.2 p = 0.343 High 99.7 90.6 p < 0.001 86.3 p < 0.001

16 H. SMITH-CANNOY ET AL.

Notably, significant ratification effects for under-5 mortality were observed in both middle-income and high-income democratic countries and these effects remained sig- nificant over time. The same trend was not experienced for non-democratic countries at these income levels. For child immunisation, the story was mixed in that significant ratification effects over time were only experienced in low-income non-democratic countries and all middle-income countries, whereas high-income countries did not experience any particular benefit in child immunisation rates as a result of CRC ratifi- cation. Generally, the pattern for CRC ratification held that across low-income countries, significant treaty effects were sustained over time only for non-democratic countries, both non-democratic and democratic countries experienced significant treaty effects over time, and in high-income countries, only democratic countries experienced significant treaty effects over time. This is a particularly interesting observation for CRC ratification in that, as country income level increased, there seemed to be a transition from non-democratic to democratic regime types that influenced whether CRC had a significant effect over time on child health outcomes (Figures 9–11).

Table 3. Mean value for health indicators pre- vs. post-CRC Ratification. Pre-Ratification 5-Years Post 10-Years Post

Infant Mortality Rate WHO Region Africa 87.9 82.5 p = 0.002 74.0 p < 0.001 Americas 37.0 30.8 p < 0.001 26.3 p < 0.001 Eastern Mediterranean 49.0 41.1 p < 0.001 34.6 p < 0.001 Europe 20.0 16.5 p < 0.001 13.3 p < 0.001 South East Asia 77.1 64.1 p = 0.012 51.7 p = 0.001 Western Pacific 38.0 33.2 p < 0.001 28.5 p < 0.001

Country Income Level Low 101.4 92.5 p = 0.002 80.8 p < 0.001 Middle 51.9 45.9 p < 0.001 39.8 p < 0.001 High 15.8 12.5 p < 0.001 10.5 p < 0.001

Under 5 Mortality Rate WHO Region Africa 144.4 135.0 p = 0.004 119.5 p < 0.001 Americas 48.0 39.3 p < 0.001 33.2 p < 0.001 Eastern Mediterranean 63.8 52.7 p < 0.001 43.6 p < 0.001 Europe 24.8 20.4 p < 0.001 16.3 p < 0.001 South East Asia 108.7 88.3 p = 0.010 69.3 p = 0.001 Western Pacific 51.8 44.7 p = 0.001 37.3 p = 0.001

Country Income Level Low 168.7 152.3 p = 0.001 130.8 p < 0.001 Middle 72.0 63.4 p < 0.001 54.6 p < 0.001 High 19.6 15.5 p < 0.001 13.0 p < 0.001

Childhood Immunisation WHO Region Africa 61.3 65.0 p = 0.207 67.0 p = 0.079 Americas 75.7 85.8 p < 0.001 89.4 p < 0.001 Eastern Mediterranean 80.9 80.8 p = 0.989 84.6 p = 0.225 Europe 84.8 93.3 p = 0.001 93.5 p < 0.001 South East Asia 72.0 77.8 p = 0.223 80.5 p = 0.224 Western Pacific 80.0 83.3 p = 0.255 84.5 p = 0.129

Country Income Level Low 51.7 58.2 p = 0.091 61.9 p = 0.025 Middle 75.3 81.4 p = 0.001 83.6 p < 0.001 High 88.4 92.1 p = 0.005 93.0 p = 0.007

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 17

Table 4. Joinpoint Regression Results Post-CEDAW Ratification (APC = Annual % Change, AAPC = Average Annual % Change).

Health Indicator

Mean Trend 1 Trend 2 Trend 3 Trend 4 AAPC (95% CI)

Ratification Year

20 Years Post-

Ratification Year APC (95% CI) Year APC (95% CI) Year APC (95% CI) Year APC (95% CI) Overall (0–20 Years)

Low Income Countries Birth Rate 42.36 35.02 0–11 −1.0 (−1.1, -1.0)* 11–14 −0.0 (−0.7, 0.7) 14–20 −1.3 (−1.4, −1.2)* – – −1.0 (−1.1, −0.9)* Mortality Rate (females) 307.01 281.51 0–3 0.1 (−0.3, 0.6) 3–11 0.5 (0.4, 0.6)* 11–15 −0.6 (−1.1, −0.2) 15–20 −2.2 (−2.3, −2.0)* −0.5 (−0.6, −0.3)* Life Expectancy (females) 53.78 59.11 0–4 0.4 (0.3, 0.6)* 4–13 0.1 (0.2, 0.2)* 13–20 1.0 (0.9, 1.0)* – – 0.5 (0.4, 0.5)* Middle Income Countries Birth Rate 31.77 25.24 0–6 −1.4 (−1.5, −1.2)* 6–9 −0.9 (−2.1, 0.4) 9–18 −1.4 (−1.5, −1.3)* 18–20 0.2 (−1.1, 1.5) −1.2 (−1.4, −0.9)* Mortality Rate (females) 202.18 169.57 0–10 −0.2 (−0.2, −0.1)* 10–18 −1.2 (−1.3, −1.0)* 18–20 −3.9 (−5.5, −2.4)* – – −0.9 (−1.1, −0.8)* Life Expectancy (females) 65.00 69.91 0–12 0.3 (0.3, 0.3)* 12–20 0.4 (0.4–0.5)* – – – – 0.4 (0.3, 0.4)* High Income Countries Birth Rate 17.14 13.13 0–20 −1.4 (−1.4, −1.3)* – – – – – – −1.4 (−1.4, −1.3)* Mortality Rate (females) 97.84 77.06 0–7 −1.6 (−1.9, −1.4)* 7–13 −0.7 (−1.1, −0.3)* 13–17 −2.0 (−3.0, −1.0)* 17–20 −0.1 (−1.2, 1.0) −1.2 (−1.5, −0.9)* Life Expectancy (females) 76.08 79.96 0–7 0.3 (0.3, 0.3)* 7–12 0.1 (0.1, 0.2)* 12–15 0.4 (0.1, 0.7)* 15–20 0.2 (0.2, 0.3)* 0.2 (0.2, 0.3)*

* Indicates significance at a p < 0.05 level.

1 8

H .S M IT H -C A N N O Y E T A L.

Table 5. Joinpoint Regression Results Post-CRC Ratification (APC = Annual % Change, AAPC = Average Annual % Change).

Health Indicator

Mean Trend 1 Trend 2 Trend 3 Trend 4 AAPC (95% CI)

Ratification Year

20 Years Post-

Ratification Year APC (95% CI) Year APC (95% CI) Year APC (95% CI) Year APC (95% CI) Overall (0–20 Years)

Low Income Countries Infant Mortality Rate 98.15 56.18 0–3 −0.7 (−1.2, −0.2)* 3–8 −2.1 (−2.4, −1.8)* 8–20 −3.6 (−3.6, −3.5)* – – −2.8 (−2.9, −2.7)* Under 5 Mortality Rate 162.58 83.97 0–3 −0.3 (−0.8, 0.2) 3–8 −2.6 (−2.9, −2.3)* 8–15 −4.0 (−4.1, −3.8)* 15–20 −4.7 (−4.9, −4.6)* −3.3 (−3.4, −3.2)* Child Immunisation (%) 57.1 79.9 0–5 0.6 (−0.2, 1.3) 5–9 −0.4 (−2.1, 1.5) 9–15 4.1 (3.4, 4.8)* 15–20 1.6 (0.9, 2.4)* 1.7 (1.3–2.1)* Middle Income Countries Infant Mortality Rate 50.93 28.19 0–5 −2.0 (−2.2, −1.9)* 5–10 −2.9 (−3.1, −2.7)* 10–20 −3.4 (−3.5, −3.4)* – – −2.9 (−3.0, −2.9)* Under 5 Mortality Rate 70.61 36.48 0–6 −2.2 (−2.3, −2.0)* 6–11 −3.2 (−3.5, −3.0)* 11–20 −4.0 (−4.1, −3.9)* – – −3.3 (−3.3, −3.2)* Child Immunisation (%) 73.7 88.6 0–4 1.8 (1.3, 2.4)* 4–8 0.3 (−0.6, 1.2) 8–17 1.1 (0.9, 1.3)* 17–20 −0.3 (−1.2, 0.6) 0.9 (0.7, 1.1)* High Income Countries Infant Mortality Rate 15.28 7.61 0–6 −3.9 (−3.9, −3.8)* 6–17 −3.3 (−3.3, −3.3)* 17–20 −3.1 (−3.2, −2.9)* – – −3.4 (−3.5, −3.4)* Under 5 Mortality Rate 18.94 9.34 0–6 −3.9 (−3.9, −3.9)* 6–9 −3.4 (−3.6, −3.2)* 9–18 −3.3 (−3.3, −3.3)* 18–20 −2.9 (−3.1, −2.7)* −3.5 (−3.5, −3.4)* Child Immunisation (%) 88.6 87.0 0–4 0.7 (−0.2, 1.7) 4–7 −0.9 (−4.4, 2.7) 7–15 0.6 (0.2, 1.0)* 15–20 −1.0 (−1.6, −0.3)* 0.0 (−0.5, 0.6) * Indicates significance at a p < 0.05 level.

T H E IN T E R N A T IO N A L JO

U R N A L O F H U M A N R IG H T S

1 9

Figure 6. Interrupted Time Series Analysis CEDAW (Birth Rate).

20 H. SMITH-CANNOY ET AL.

Figure 7. Interrupted Time Series Analysis CEDAW (Female Life Expectancy).

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 21

Figure 8. Interrupted Time Series Analysis CEDAW (Female Adult Mortality).

22 H. SMITH-CANNOY ET AL.

Discussion

Our analysis supports our hypotheses in part. In general, democracies and richer countries tend to perform better across the gamut of indicators we tested. These findings largely square with the IR literature on the effects of democracy, development and treaties on human rights outcomes.66 However, as discussed above, and interestingly, the effects are inconsistent across categories of states and the selected public health indicators, suggesting that both ITSA and joinpoint shed light on new dynamics associated with treaty commitment.

Contrary to our initial hypothesis, female life expectancy did not show any meaning- ful change across most subgroups, with the exception of low-income democratic countries, and high-income non-democratic countries. The ratification of CEDAW revealed a significant and sustained effect for birth rate, but only in middle- and high- income countries (H2). Both under-5 and infant mortality showed promising ratification effects across low- and middle-income non-democratic countries, as well as high-income democratic countries (H3). A similar pattern was observed for child immunisation rates increasing following treaty ratification but only in low- and middle-income countries (H4). We further hypothesised that greater improvements across indicators would be seen for higher income countries which was consistently demonstrated for CEDAW, but for CRC middle-income countries showed the most benefit in health outcomes over time relative to their low-income and high-income counterparts (H5). Lastly, we hypothesised that democracies would fare better post-ratification, but this finding was

Table 6. Summary of Interrupted Time Series Results. Low Income Countries Middle Income Countries High Income Countries

Non- Democratic Democratic

Non- Democratic Democratic

Non- Democratic Democratic

CEDAW Birth Rate Significant Ratification Effect Yes* No Yes* Yes No Yes*** Significant Ratification Effect

Over Time No No Yes** Yes No Yes***

Female Life Expectancy Significant Ratification Effect No Yes*** No No Yes** Yes** Significant Ratification Effect

Over Time No Yes*** No No Yes*** No

Female Adult Mortality Significant Ratification Effect Yes* Yes** No Yes** Yes* Yes** Significant Ratification Effect

Over Time Yes** No No No Yes*** No

CRC Child Immunisation Significant Ratification Effect Yes** No Yes*** Yes** No No Significant Ratification Effect

Over Time Yes*** No Yes*** Yes*** No No

Under 5 Mortality Significant Ratification Effect Yes*** No Yes*** Yes*** No Yes** Significant Ratification Effect

Over Time Yes** No No Yes*** No Yes***

Infant Mortality Significant Ratification Effect Yes*** No Yes*** No No Yes** Significant Ratification Effect

Over Time Yes** No Yes* No No Yes***

* Indicates significance at a p < 0.05 level, ** Indicates significance at a p < 0.01 level, *** Indicates significance

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 23

Figure 9. Interrupted Time Series Analysis CRC (Child Immunisation).

24 H. SMITH-CANNOY ET AL.

Figure 10. Interrupted Time Series Analysis CRC (Under 5 Mortality).

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 25

Figure 11. Interrupted Time Series Analysis CRC (Infant Mortality).

2 6

H .S M IT H -C A N N O Y E T A L.

only observed amongst high-income countries, whereas there was a mix of where signifi- cant treaty effects over time were observed across democracies in low- and middle- income countries (H6).

What our findings show, then, are how norms, even ostensibly widely-held ones, vary in the ways in which they manifest in different countries. Rather than more general claims that have been advanced in the past, however, we are able to identify specific well-being-related factors that appear to be affected by the ratification of human rights treaties. Moreover, these findings highlight the role of income level as a prism though which treaties affect public health outcomes. With respect to the CEDAW, better sustained health outcomes were associated with high income countries (as predicted). But the same was not true for the CRC – here, middle income countries showed sustained and significant improvements in public health outcomes associated with children’s rights.

To be sure, there are likely to be other variables that intervene between treaty com- mitment and health effects. A government that ascribes to norms surrounding chil- dren’s health, may also be more inclined to invite the World Health Organization in to conduct a vaccination campaign or take other steps to promote the health and well-being of both women and children. We cannot discount these other activities that governments may engage in which intervene between treaty ratification and health outcomes.

Conclusion

The prima facie case for norm entrenchment, made by the near universal acceptance of both the CEDAW and CRC would seem to suggest that norms of women’s and chil- dren’s health have become well-entrenched globally. Yet, to stop at this agreement would be to stop short of a full understanding of the effects of those norms. Our analy- sis illustrates the important ways that these high rates of treaty ratification can mask more interesting differences between the effects of the two norms evaluated here. But to have done this analysis, we needed to come up with ways to look at things that are ostensibly ‘the same’ – nearly all states agree to CRC and CEDAW. Perhaps one of the more important contributions of our analysis, then, is the ITSA analysis, which provides a detailed evaluation of whether ratification was significant in the short-term, and sustained in the long-term across levels of income and regime type. This analysis in particular revealed a patterning that would be difficult to tease out using other methods.

The positive effects of global norms for children’s health and the relatively weaker effects of norms on women’s health, suggests that scholars of both IR and public health should proceed cautiously when evaluating the effects of popular treaties on human rights and health outcomes. This patterning emphasises the more favourable effect of CRC ratification on child health outcomes both in the long and short term, and the dis- tribution of these favourable effects across income groups and regime types. These results provide an important contrast against the effect that CEDAW had across these same sub- groups but with respect to women’s health outcomes – an effect that is less pronounced in the short-term and often not significantly sustained over time.

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 27

Our work challenges the popularly held notion in IR, that high rates of treaty commit- ment signal high rates of norm adoption,67 which is a contention that came under fire with critics who asserted that law and norms were in fact possibly decoupled.68 Our finding seems to split this difference: despite high rates of ratification across both the CEDAW and the CRC, we find relatively distinct impacts of these norms on indicators of women’s and children’s health.

These findings also suggest that future research should carefully evaluate the influence of norms on outcomes with careful attention to the issue area of the norm in question. In particular, the body of treaties created under the authority of Article 19 of the World Health Organization’s constitution provides an ideal opportunity for expanding this analysis. New treaties on global public health can be voted into existence by a 2/3’s majority of the WHO’s governing assembly. Examples of public health treaties created under this authority include the 2002 Framework Convention on Tobacco Control and the 2005 International Health Regulations.69 We know that high rates of ratification are not necessarily correlated with improved public health outcomes across the board – future research should therefore aim to identify the factors that make some norms better than others at improving public health. Moreover, another possibility for better cap- turing norm effects, may be to distinguish between those countries that ratify these treaties with reservations and those that ratify without reservations. Ratification without reser- vation may be a more straight-forward predictor of norm compliance.70

Finally, with regards to norms, the paper raises the interesting methodological ques- tion of how we test the effects of norms. The common assumption is to point to the existence of a norm and accession to a norm, and assume that adherence will follow. But there is no good way to test the effect of norms by creating different popu- lations if indeed, that norm is strong. In fact, the stronger the norm, the greater that challenge might become if we lose all variation on our variable of interest (the norm). Our paper presents a novel approach to considering some of the methodological limit- ations in studying norm effects.

In this paper, the findings are not surprising in the sense that they reinforce some widely-held ideas about the influence of regime type and country wealth on the protec- tion of rights. In our analysis, democratic and rich states tended to improve more over time, across our health indicators, than other types of states. These findings are consist- ent with previous research. However, our foray into using public health indicators and not politically-based human rights indicators is a step towards trying to account for change among economic and social rights. By accounting for the broader effects of human rights treaties that span multiple categories, our findings and future research should generate interest in understanding less-than-immediate consequences of human rights treaties.

Notes

1. e.g., Geoff Dancy and Kathryn Sikkink, ‘Ratification and Human Rights Prosecutions: Toward a Transnational Theory of Treaty Compliance’, New York University Journal of International Law and Politics 44, no. 1 (2011): 751–90; Laurence R. Helfer and Erik Voeten, ‘International Courts as Agents of Legal Change: Evidence from LGBT Rights in Europe’, International Organization 68, no. 1 (2014): 77–110.

28 H. SMITH-CANNOY ET AL.

2. e.g., James Raymond Vreeland, ‘Political Institutions and Human Rights: Why Dictatorships Enter into the United Nations Convention against Torture’, International Organization 62, no. 1 (2008): 65–101; Courtenay Conrad and Will H. Moore, ‘What Stops the Torture?’ American Journal of Political Science 54, no. 2 (2010): 459–76; Courtenay Conrad and Emily H. Ritter, ‘Tenure, Treaties, and Torture: The Conflicting Domestic Effects of Inter- national Law’, The Journal of Politics 75, no. 2 (2013): 397–409.

3. Beth A. Simmons, Mobilizing for Human Rights: International Law in Domestic Politics (Cambridge; New York: Cambridge University Press, 2009).

4. We consider the number of states that are parties (that is, those that have signed and ratified or acceded to a treaty) to CEDAW (189) and CRC (196) (see https://treaties.un.org/Pages/ ViewDetails.aspx?src=TREATY&mtdsg_no=IV-8&chapter=4&lang=en and https://treaties. un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=IV-11&chapter=4&lang=en) (accessed February 24, 2016). This can be compared to the number of parties to the Inter- national Covenant on Civil and Political Rights, which has 168 parties.

5. For example, The Political Terror Scale or the CIRI Human Rights Data Project. 6. Following Simmons, Mobilizing for Human Rights. 7. Martha Finnemore and Kathryn Sikkink, ‘International Norm Dynamics and Political

Change’, International Organization 52, no. 4 (1998): 891. 8. John Mearsheimer, ‘The False Promise of International Institutions’, International Security

19, no. 3 (1994/95): 5–49; Jack L. Goldsmith and Eric A. Posner, The Limits of International Law (Oxford; New York: Oxford University Press, 2005).

9. George W. Downs et al., ‘Is the Good News About Compliance Good News About Cooperation?’ International Organization 50, no. 3 (1996): 379–406.

10. Hans J. Morgenthau and Kenneth W. Thompson, Politics among Nations: The Struggle for Power and Peace. 6th ed. (New York: Knopf: Distributed by Random House, 1985).

11. Jana Von Stein, ‘Do Treaties Constrain or Screen? Selection Bias and Treaty Compliance’, American Political Science Review 99, no. 4 (2005): 611–22.

12. Wayne Sandholtz, ‘How Domestic Courts Use International Law’, Fordham International Law Journal 38, no. 2 (2015): 595–637.

13. Linda C. Keith, ‘The United Nations International Covenant on Civil and Political Rights: Does It Make a Difference in Human Rights Behavior?’ Journal of Peace Research 36, no. 1 (1999): 95–118; Oona A. Hathaway, ‘Do Human Rights Treaties Make a Difference?’ The Yale Law Journal 111, no. 8 (2002): 1935–2042.

14. Clair Apodaca, ‘Measuring Women’s Economic and Social Rights Achievement’, Human Rights Quarterly 20 (1998): 139–72.

15. Hafner-Burton et al., ‘International Human Rights Law and the Politics of Legitimation’; Eric Neumayer, ‘Do International Human Rights Treaties Improve Respect for Human Rights?’ Journal of Conflict Resolution 49, no. 6 (2005): 925–53.

16. Von Stein, ‘Do Treaties Constrain or Screen? Selection Bias and Treaty Compliance’. 17. Oona A Hathaway, ‘Why Do Countries Commit to Human Rights Treaties?’ Journal of

Conflict Resolution 51, no. 4 (2007): 588–621; Emilia Justyna Powell and Jeffrey K. Staton, ‘Domestic Judicial Institutions and Human Rights Treaty Violation’, International Studies Quarterly 53, no. 1 (2009): 149–74.

18. Emilie M. Hafner-Burton et al., ‘Human Rights Institutions, Sovereignty Costs and Democra- tization’, British Journal of Political Science 45, no. 01 (2015): 1–27.

19. Simmons, Mobilizing for Human Rights. 20. Mark M. Gray et al., ‘Women and Globalization: A Study of 180 Countries, 1975–2000’,

International Organization 60, no. 2 (2006): 293–333. 21. Richard Price, ‘A Genealogy of the Chemical Weapons Taboo’, International Organization

49, no. 1 (1995): 73–103. 22. Amitav Acharya, ‘How Ideas Spread: Whose Norms Matter? Norm Localization and Insti-

tutional Change in Asian Regionalism’, International Organization 58, no. 2 (2004): 239–75. 23. Antje Wiener, ‘Contested Compliance: Interventions on the Normative Structure of World

Politics’, European Journal of International Relations 10, no. 2 (2004): 189–234; Mona

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 29

Lena Krook and Jacqui True, ‘Rethinking the Life Cycles of International Norms: The United Nations and the Global Promotion of Gender Equality’, European Journal of International Relations 18, no. 1 (2012): 103–27.

24. Finnemore and Sikkink, ‘International Norm Dynamics and Political Change’; Thomas Risse et al. eds., The Power of Human Rights: International Norms and Domestic Change (New York, NY: Cambridge University Press, 1999).

25. Martha Finnemore, ‘Norms, Culture, and World Politics: Insights from Sociology’s Institu- tionalism’, International Organization 50, no. 2 (1996): 325–47; Ryan Goodman and Derek Jinks, Socializing States: Promoting Human Rights through International Law (Oxford Uni- versity Press, 2013).

26. Wayne Sandholtz, Prohibiting Plunder: How Norms Change (Oxford, UK: Oxford University Press, 2007); Kees Van Kersbergen and Bertjan Verbeek, ‘The Politics of International Norms: Subsidiarity and the Imperfect Competence Regime of the European Union’, Euro- pean Journal of International Relations 13, no. 2 (2007): 217–38; Wayne Sandholtz, and Kendall Stiles, International Norms and Cycles of Change (New York, NY: Oxford University Press, 2008).

27. Todd Landman, Protecting Human Rights: A Comparative Study (Washington, DC: George- town University Press, 2005); Neumayer, ‘Do International Human Rights Treaties Improve Respect for Human Rights?’.

28. Emilie M. Hafner-Burton and Kiyoteru Tsutsui, ‘Justice Lost! The Failure of International Human Rights Law to Matter Where Needed Most’, Journal of Peace Research 44, no. 4 (2007): 407–25.

29. Daniel W. Hill Jr., ‘Estimating the Effects of Human Rights Treaties on State Behavior’, The Journal of Politics 72, no. 4 (2010): 1161–74; Simmons, Mobilizing for Human Rights.

30. Simmons, Mobilizing for Human Rights. 31. Margaret Keck and Kathryn Sikkink, Activists Beyond Borders: Advocacy Networks in Inter-

national Politics (Ithaca, NY: Cornell University Press, 1998); Risse et al., The Power of Human Rights, vol. 66; Thomas Risse, Steve C. Ropp, and Kathryn Sikkink, eds., The Persist- ent Power of Human Rights: From Commitment to Compliance, vol. 126. (Cambridge: Cam- bridge University Press, 2013); Simmons, Mobilizing for Human Rights; Kathryn Sikkink, The Justice Cascade: How Human Rights Prosecutions Are Changing World Politics (New York, NY: W. W. Norton & Company, 2011); Heather Smith-Cannoy, Insincere Com- mitments: Human Rights Treaties, Abusive States, and Citizen Activism (Washington, DC: Georgetown University Press, 2012).

32. Sandra Fullerton Joireman, ‘Inherited Legal Systems and Effective Rule of Law: Africa and the Colonial Legacy’, The Journal of Modern African Studies 39, no. 4 (2001): 571–96; Linda Camp Keith, and Ayo Ogundele, ‘Legal Systems and Constitutionalism in Sub-Saharan Africa: An Empirical Examination of Colonial Influences on Human Rights’, Human Rights Quarterly (2007): 1065–97; Sara McLaughlin Mitchell, Jonathan J. Ring, and Mary K. Spellman, ‘Domestic Legal Traditions and States’ Human Rights Practices’, Journal of Peace Research 50, no. 2 (2013): 189–202.

33. Sara McLaughlin Mitchell, Jonathan J. Ring, and Mary K. Spellman, ‘Domestic Legal Tra- ditions and States’ Human Rights Practices’, Journal of Peace Research 50, no. 2 (2013): 189–202.

34. These include: The Convention on the Rights of the Child, the Covenant on Economic, Social and Cultural Rights, the Convention on the Elimination of Discrimination Against Women, the Convention against Torture, the Convention on the Elimination of Racial Discrimi- nation, and the Convention on Civil and Political Rights.

35. This is child mortality for those under the age of 5. 36. Alexis Palmer, Jocelyn Tomkinson, Charlene Phung, Nathan Ford, Michel Joffres, Kimberly

A Fernandes, Leilei Zeng, et al., ‘Does Ratification of Human-Rights Treaties Have Effects on Population Health?’ The Lancet 373, no. 9679 (2009): 1987–92.

30 H. SMITH-CANNOY ET AL.

37. Arthur Wilson and Abdallah S Daar, ‘A Survey of International Legal Instruments to Examine Their Effectiveness in Improving Global Health and in Realizing Health Rights’, The Journal of Law, Medicine & Ethics 41, no. 1 (2013): 89–102.

38. Arthur Wilson, and Abdallah S Daar, ‘A Survey of International Legal Instruments to Examine Their Effectiveness in Improving Global Health and in Realizing Health Rights’, The Journal of Law, Medicine & Ethics 41, no. 1 (2013): 95–6.

39. Steven J. Hoffman and John-Arne Røttingen, ‘Assessing the Expected Impact of Global Health Treaties: Evidence from 90 Quantitative Evaluations’, American Journal of Public Health 105, no. 1 (2015): 26–40.

40. We thank an anonymous reviewer for bringing this point to our attention. 41. http://www.politicalterrorscale.org/About/ (Accessed March 7, 2016). 42. Christopher J. Fariss, ‘Respect for Human Rights Has Improved over Time: Modeling the

Changing Standard of Accountability’, American Political Science Review 108, no. 2 (2014): 297–318.

43. http://www.humanrightsdata.com/2014/05/ciris-new-address-wwwhumanrightsdatacom. html (accessed March 7, 2016).

44. CIRI’s coverage of women’s rights extends across three categories (Economic, Political, and Social), but the coding of these variables combines a whole range of rights into composite scores, and as such, gives an imprecise sense of how rights for women are improving (or not). See https://drive.google.com/file/d/0BxDpF6GQ-6fbY25CYVRIOTJ2MHM/edit?pref= 2&pli=1 (accessed March 7, 2016).

45. Gloria Schuster, Ivana Martinez, and Julie Madore, ‘Between Their Stories and Our Realities’, The People’s Decade of Human Rights Education, https://pdhre.org/cedaw/index.html. (accessed March 9, 2016)

46. Simmons, Mobilizing for Human Rights. 47. Simmons (2009), Ch. 6; Neil Englehart and Melissa Miller, ‘The CEDAW Effect: Inter-

national Law’s Impact on Women’s Rights’, Journal of Human Rights 13, no. 1 (2014): 22–47; Daniel Hill Jr., ‘Estimating the Effects of Human Rights Treaties on State Behavior’, Journal of Politics 72, no. 4 (2010): 1161–74.

48. Neal Krause, ‘Employment Outside the Home and Women’s Psychological Well-Being’, Social Psychiatry 19, no. 1 (1983): 41–8; Jane Waldfogel, ‘The Effect of Children on Women’s Wages’, American Sociological Review 62, no. 2: 209–17; A. Dharmalingam and S. Philip Morgan, ‘Women’s Work, Autonomy, and Birth Control: Evidence from Two South Indian Villages,’ Population Studies 50, no. 2 (1996): 187–201.

49. Alicia Adsera, ‘Vanishing Children: From High Unemployment to Low Fertility in Devel- oped Countries’, American Economic Review 95, no. 2 (2005): 189–93.

50. Elizabeth Frankenberg, ‘The Effects of Access to Halthcare on Infant Mortality in Indonesia’, Health Transitions Review 5 (1995): 143–163; Kathleen Falster, Louisa Jorm, John Lynch, Emily Banks, Marni Brownell, Rhonda Craven, Kristjana Einarsdottir and Deborah Randall, ‘What Factors Contribute to Early Childhood Health and Development in Austra- lian Aboriginal Children’, BMJ Open 5, no. 5 (2015); Amie Shei, Federico Costa, Mitermayer G Reis, and Albert Ko, ‘The Impact of Brazil’s Bolsa Familia Conditional Cash Transfer Program on Children’s Health Care Utilization and Health Outcomes’, BMC International Health and Human Rights 14, no. 10 (2014).

51. Alessandro Zanetti, Pierre Van Damme and Daniel Shouval, ‘The Global Impact of Vacci- nation Against Hepatitis B: A Historical Overview’, Vaccine 26, no. 49 (2008): 6266–73. Thomas Frieden, ‘A Framework for Public Health Action: The Health Impact Pyramid’, American Journal of Public Health 100, no. 4: 590–94.

52. David Peters, Anu Garg, Gerry Bloom, Damian Walker, William Brieger, and M. Hafozur Rahman, ‘Poverty and Access to Health Care in Developing Countries’, Annals of the New York Academy of Sciences 1136, no. 1 (2008): 161–71; Brian Biggs, Lawrence King, Sanjay Basu and David Stuckler, ‘Is Wealthier always Healthier? The Impact of National Income Level, Inequality, and Poverty on Public Health in Latin America’, Social Science & Medicine 71, no. 2 (2010): 266–73; Kate Strully, David Rehkoph and Ziming Xuan,

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 31

‘Effects of Prenatal Poverty on Infant Health: State Earned Income Tax Credits and Birth Weight’, American Sociological Review 75, no. 4 (2010): 534–62.

53. Apodaca, ‘Measuring Women’s Economic and Social Rights Achievement’; Hafner-Burton et al., ‘International Human Rights Law and the Politics of Legitimation’; Neumayer, ‘Do International Human Rights Treaties Improve Respect for Human Rights?’; Von Stein, ‘Do Treaties Constrain or Screen? Selection Bias and Treaty Compliance’; Hathaway, ‘Why Do Countries Commit to Human Rights Treaties?’; Powell and Staton, ‘Domestic Judi- cial Institutions and Human Rights Treaty Violation’.

54. https://treaties.un.org/Pages/UNTSOnline.aspx?id=1 55. Entered into force 1979. 56. Entered into force 1990. 57. Because South Sudan did not emerge as an independent state until 2011, it is not captured in

our dataset. 58. The Atlas conversion factor for any year is the average of a country’s exchange rate for that

year and its exchange rates for the two preceding years, adjusted for the difference between the rate of inflation in the country and international inflation. The overall objective of this method is to reduce any changes in the exchange rate that are attributable to inflation.

59. Monty G. Marshall and Keith Jaggers, ‘Polity IV Project: Political Regime Characteristics and Transitions, 1800–2012’, https://www.systemicpeace.org/polity/polity4.htm.

60. http://data.worldbank.org/data-catalog/world-development-indicators 61. The CRC makes reference to children’s: mental health (Art. 17), the rights of mentally and

physically disabled children (Article 23), the right to the highest standards of attainable health (Art 24), right to be free from economic exploitation that may impact health (Article 32), and the right to have their health prioritized while reintegrating into society after a traumatizing event (Article 39). The CEDAW makes reference to women’s: right to access information to ensure the health of the family (Article 10), rights to protection safe and healthy working conditions (Article 11), and the right to be free from discrimination in health care (Article 12).

62. On mortality rates as a key indicator of population health see: R. Gibson Parrish, ‘Measuring Population Health Outcomes’, Preventing Chronic Disease 7, no. 4 (2010): A71. See also: National Research Council (US) Panel to Advance a Research Program on the Design of National Health Accounts, Washington D.C.: National Academies Press (2010), Ch 5. Avail- able online: https://www.ncbi.nlm.nih.gov/books/NBK53336/. We reviewed literature on child health interventions which consistently pointed to immunizations as a key childhood health intervention: see: Michael J. Rigby et al., ‘Child Health Indicators for Europe: A Pri- ority for a Caring Society’, European Journal of Public Health 13, no. 3 Supplement (2003): 38–46; Cynthia Boschi-Pinto, Rajiv, and Jose Martines, ‘Limited Progress in Increasing Cov- erage of Neonatal and Child-health Interventions in Africa and Asia’, Journal of Health, Population and Nutrition 27, no. 6 (2009): 755–62; Daniel Hogan et al., ‘Monitoring Univer- sal Health Coverage within the Sustainable Development Goals: Development and Baseline Data for an index of Essential Health Services’, The Lancet 6, no. 2 (2018): 152–68.

63. Hyune-Ju Kim, et al., ‘Permutation Tests for Joinpoint Regression with Applications to Cancer Rates’, Statistics in Medicine 19, no. 3 (2000): 335–51.

64. Ibid. 65. Differing countries include Iran, Palau, Somalia, Sudan, and Tonga that ratified CRC but not

CEDAW. 66. Von Stein, ‘Do Treaties Constrain or Screen? Selection Bias and Treaty Compliance’; Hafner-

Burton et al., ‘International Human Rights Law and the Politics of Legitimation’; Neumayer, ‘Do International Human Rights Treaties Improve Respect for Human Rights?’.

67. Ryan Goodman and Derek Jinks, ‘Measuring the Effects of Human Rights Treaties,’ Euro- pean Journal of International Law 14, no. 1 (2003): 171–83.

68. See, for example, Hathaway, ‘Do Human Rights Treaties Make a Difference?’; Emilie M. Hafner-Burton and Kiyoteru Tsutsui, ‘Human Rights in a Globalizing World: The Paradox of Empty Promises’, American Journal of Sociology 110, no. 5 (2005): 1373–411;

32 H. SMITH-CANNOY ET AL.

Hafner-Burton and Tsutsui, ‘Justice Lost! The Failure of International Human Rights Law to Matter Where Needed Most’.

69. Hoffman and Røttingen, ‘Assessing the Expected Impact of Global Health Treaties’. 70. We thank an anonymous reviewer for recommending this point.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

This work is supported by Canada Research Chair in Population Health Equity.

Notes on contributors

Heather Smith-Cannoy is Associate Professor of Political Science at the New College of Interdisci- plinary Arts and Sciences at Arizona State University. Professor Smith-Cannoy’s work has appeared in The Journal of Human Rights, The Human Rights Review, and The International Political Science Review, among others. Her 2012 book Insincere Commitments: Human Rights Treaties, Abusive States and Citizen Activism was published by Georgetown University Press. Her most recent book is an edited volume titled, Emerging Threats to Human Rights: Resources, Violence and Depri- vation of Citizenship, published in 2019 by Temple University Press.

Wendy Wong is Canada Research Chair of Global Governance and Civil Society at the University of Toronto. She has published two books with Cornell University Press, The Authority Trap (2017), co-authored with Sarah Stroup, and the ARNOVA Best Book for 2014, Internal Affairs (2012). Pro- fessor Wong’s work has also appeared in numerous journals, including Perspectives on Politics and Human Rights Review.

Arjumand Siddiqi is Canada Research Chair in Population Health Equity and Associate Professor at the Dalla Lana School of Public Health, University of Toronto, and an adjunct associate professor at the Gillings School of Global Public Health, University of North Carolina, Chapel Hill. Her pub- lished work appears in journals such as Social Science & Medicine, American Journal of Epidemiol- ogy, Annals of Epidemiology, and Annual Review of Public Health. She has also published chapters in books published by Oxford University Press and Cambridge University Press.

Christopher Tait completed his PhD in Epidemiology at the Dalla Lana School of Public Health. For his doctoral thesis, Chris applied cutting edge methods to a unique population-based data source to inform how dietary patterns in the population relate to the development of chronic disease and pre- mature death. His work on the largest population-based databases of food exposures linked with health outcomes in Canada provides the opportunity to have significant impact and inform chronic disease prevention efforts in Canada.

Abtin Parnia is a Research Data Analyst at the University of Toronto’s School of Public Health. A graduate of the Masters of Public Health in Epidemiology, Abtin’s area of research has been around race, immigration, and socioeconomic disparities in health.

ORCID

Abtin Parnia http://orcid.org/0000-0003-4321-3260

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 33

Appendix 1 Do the changing economic circumstances of countries impact the effect of ratification of conven- tions on health outcomes over time? The use of interrupted time series method is best when the group under study are similar and as the event occurs the trajectory of the observations following the event is studied. The underlying assumption is that the individual observations remain mostly similar and the event is the main driver of the change. There are two issues that arise when groups of countries are categorised for this method. First is change in the categories over time. Second, is the possibility of confounding by other characteristics.

This assumption requires either large numbers of countries to balance out the distribution, or a long period of stability so that the various characteristic of the countries stay relatively stable com- pared to the event under measurement. For instance, in this study, if countries go through other extreme events during the follow up time, the observations post-ratification will be biased by these external events. In checking which countries have moved between categories, World Bank data demonstrates that there has been some movement between categories of income over time. For some countries, the change in the category below or above the threshold may indicate a severe change, and in some it could just be due to their economic growth as compared to other countries.

The second issue arises when countries are comparable on some measurements but not others that affect the outcome measurement. Countries may have similar income at the beginning of the study, but one could be a growing economy and the other shrinking. It is plausible to argue that the rate of growth is a stronger predictor of a countries gain in health indicators than income level. However, the volatility of economic growth as a metric raises the previous issue with change in cat- egory over time.

The challenge with using the economic growth indicator alone is that the list becomes even more dissimilar than income alone. One plausible solution is to conduct the analysis using both the income and the economic growth rate at the same time. Doing this will raise sample size concerns as such categorisation will at least half the groupings.

The table below describes the list of countries based on income and growth rate. In the main analysis we further broke down the countries by democratic versus un-democratic as well. We measured economic growth by the 5-year average GNI per capita growth centred on the year of ratification. We then categorised the countries by the median growth rate. Countries missing the measurement of GNI growth or GNI at the time of ratification were excluded.

High Income Middle Income Low Income Below 50th Above 50th Below 50th Above 50th Below 50th Above 50th Canada Israel Algeria Algeria Bangladesh Benin Sweden Netherlands Argentina Argentina Kyrgyz Republic Bolivia Bahrain Norway Australia Australia Mali Burkina Faso

Switzerland Austria Austria Mauritania India Finland Brazil Brazil Pakistan Papua New Guinea

Luxembourg Colombia Colombia Philippines Armenia Singapore Czech Republic Czech Republic Uzbekistan Egypt, Arab Rep.

Dominican Republic Dominican Republic Azerbaijan Eritrea France France Cameroon Mozambique Gabon Gabon Chad Rwanda Jordan Jordan Comoros Sri Lanka Morocco Morocco Kenya Thailand South Africa South Africa Madagascar Tunisia Tunisia Nigeria Ecuador Belize Senegal

(Continued)

34 H. SMITH-CANNOY ET AL.

Continued. High Income Middle Income Low Income

Below 50th Above 50th Below 50th Above 50th Below 50th Above 50th Greece Bulgaria Tajikistan Guatemala Chile Timor-Leste Honduras Congo, Rep. Togo North Macedonia Cuba Uganda Peru Cyprus Saudi Arabia Japan Uruguay Korea, Rep. Venezuela, RB Malaysia

Mauritius Panama Seychelles Syrian Arab Republic United Kingdom

THE INTERNATIONAL JOURNAL OF HUMAN RIGHTS 35