In 2015, the millenium development goals (MDGs) reach their target date and countries are expected to renegotiate the next set of sustainable goals to set the development agenda for the following decades. Extensive discussion is underway and a high-level panel appointed by the UN Secretary-General has already proposed a list of goals and indicators for consideration (United Nations, 2014). It is not realistic to expect mental health to feature as one of the core goals but, given the huge contribution of mental illness to the global burden of disease, as well as the clear links between mental illness and social determinants such as poverty, inequality, lack of education and unemployment (Patel & Kleinman, Reference Patel and Kleinman2003; Lund et al. Reference Lund, Breen, Flisher, Kakuma, Corrigall, Joska, Swartz and Patel2010), it is important to clarify the causal relationships between these social determinants and mental health. At present our understanding of these relationships is vague; if we want to influence the global development agenda so as to maximise its usefulness as a strategic tool for reducing the global burden of mental illness, we need to identify the key social and economic drivers. In doing so we also need to be confident regarding which indicators best illustrate the interactions between social determinants and mental disorders.
Challenges in unravelling the relationship between poverty and mental illness
Lund et al. (Reference Lund, Breen, Flisher, Kakuma, Corrigall, Joska, Swartz and Patel2010) observe that there is no longer a debate as to whether poverty negatively impacts on mental health – the debate is about which aspects of poverty and deprivation are the strongest drivers. They note that heterogeneous findings across different domains of poverty are partly due to the use of diverse measures of poverty. Cooper et al. (Reference Cooper, Lund and Kakuma2012) argue that inconsistent and imprecise measures of both poverty and mental illness hamper this field of research.
In a recent commentary, Lund (Reference Lund2014) highlights the problems and challenges faced in unravelling the complicated relationship between poverty and mental illness in low- and middle-income countries (LMICs) and proposed a research agenda aimed at bringing clarity to the issue. He raises many important issues, several of which will be discussed here:
There needs to be more precise measurement of poverty
The measurement of poverty needs to include the breaking down of ‘poverty’ into specific indicators such as income, expenditure, assets, education, employment and food security; and these should be reported at both individual and household level Lund (Reference Lund2014).
Importantly, there are problems with most commonly used poverty indicators. For example, income-based measures of poverty may misrepresent the extent of deprivation. The meaning of individual or household income varies depending on context and may not be comparable across societies or even communities. Thus a frequently cited measure of poverty at the population level – the proportion of people living on less than $1 a day – has different implications for the extent of societal poverty depending on the socioeconomic context and relative value of this figure. Furthermore, it is important to note that simple income is not the only factor determining level of deprivation; other factors such as the availability of and access to health care, social benefits and education may exacerbate or offset the effects of low income.
Perhaps the most common indicator used to measure economic growth and the extent of population poverty is ‘gross domestic product per capita’ (GDP/c). But GDP/c does not show the distribution of growth in the population, nor does it reflect accurately the extent of poverty; since a disproportionate share of the total ‘product’ may be concentrated in the hands of a few. This is the case in most countries with a significant income inequality gap e.g. South Africa, where in 2011 the top 20% of earners controlled 70% of the wealth, while the share of the bottom 20% was 2.5% (World Bank, 2014).
Studies need to be stratified according to diverse socioeconomic strata
For these reasons, Lund (Reference Lund2014) emphasises that the poverty–mental health relationship needs to be examined at different levels of poverty/wealth, as the effects may differ at different socioeconomic levels. Specifically we need to stratify by income levels in studying the effects of poverty on mental health.
There should be a broader approach to mental health outcomes including different levels of severity of mental illness
Disorders currently under-researched in relation to poverty, such as schizophrenia, bipolar disorder, substance abuse and child and adolescent disorders need to be included in the social determinants research agenda (Lund, Reference Lund2014). Poverty may interact with different disorders in different ways. In addition, the poverty–mental health relationship needs to be interrogated across the lifespan, adopting a developmental perspective, including at different levels of severity of mental illness.
Broadening the approach to mental health outcomes in poverty research coincides neatly with contemporary conceptualisation of mental health and illness as dimensional phenomena, overlapping symptomatically and genetically, and manifesting in populations as a continuum from ‘normality’ to disorder (First, Reference First2010). Further rationale lies in the growing evidence for complex gene–environment interactions and epigenetic mechanisms underlying the genesis of mental disorders (Toyokawa et al. Reference Toyokawa, Uddin, Koenen and Galea2012). Individual and ecological level aspects of poverty are likely to have their deleterious effects on mental health, at least in part, through interactions with susceptibility genes and modification of gene expression. Such genomic research frequently reveals a dose–response relationship between environmental exposures and mental health outcomes; and it is reasonable to anticipate that the inclusion of a broader range of mental health phenotypes in relation to poverty will yield a more diverse range of poverty–mental health relationships.
Future research should be theory-driven
Lund (Reference Lund2014) maintains that research approaches need to engage with theoretical concepts such as Amartya Sen's capability framework (Sen, Reference Sen1999). The capabilities framework is a powerful construct in disability discourse related to the social and economic isolation experienced by people with mental illness (Ware et al. Reference Ware, Hopper, Tugenberg, Dickey and Fisher2007; Baumgartner & Burns, Reference Baumgartner and Burns2014). Shifting the focus from a position where disability is located in personal functioning to a position where disability is located in the opportunities provided by society for social reintegration and participation, Sen's work provides a basis for conceptualising social integration as ‘a process, unfolding over time, through which individuals who have been psychiatrically disabled increasingly develop and exercise their capacities for connectedness and citizenship’ (Ware et al. Reference Ware, Hopper, Tugenberg, Dickey and Fisher2007). Similarly, in relation to poverty and (mental) health, the capabilities framework can help shift the focus away from considering economic exclusion a consequence of individual dysfunction; towards an understanding of the structural social, economic and political forces that so often render people vulnerable to both poverty and mental illness. In thinking about mental health interventions, this then means we should adopt approaches that directly address the structural barriers limiting opportunities for economic integration and participation.
More research is required on the associations between economic inequality and mental health
Lund (Reference Lund2014) highlights the need for ‘more robust research on the association between economic inequality and mental health, at national and regional levels.’ He notes that to avoid the ‘ecological fallacy’, studies need to use multilevel methods including both population and individual level data. Interestingly Lund uses the term ‘economic inequality’ (also used by Thomas Picketty (2014) whose work is discussed below) as opposed to ‘income inequality’ – the latter concept is more commonly associated with the research on the effects of inequality on health. Although there is considerable evidence that inequality is a powerful driver of poor health outcomes, it is often overlooked in discourse relating to the social determinants of health and mental health. In the lead up to negotiations to finalise the next set of development goals, there is a lobby to include the reduction of inequality as a key target for states (Stiglitz & Doyle, Reference Stiglitz and Doyle2014). For this reason, income or economic inequality is the focus of the second part of this paper.
Income inequality as a powerful driver of (mental) health
Income inequality is a measure of the ‘rich-poor gap’ in any given society and is a concept of great relevance to LMICs, many of which are among the most inequitable in the world. There are multiple associations between income inequality and health status. In the 1980s and 1990s, Wilkinson demonstrated that the relative distribution of income in a society matters in its own right for population health (Wilkinson, Reference Wilkinson1992, Reference Wilkinson1996) and this has been replicated in multiple studies (e.g. Kawachi et al. Reference Kawachi, Subramanian and Almeida-Filho2002; Subramanian & Kawachi, Reference Subramanian and Kawachi2004; Wilkinson & Pickett, Reference Wilkinson and Pickett2006). Income inequality is associated with reduced life expectancy (Kondo et al. Reference Kondo, Sembajwe, Kawachi, van Dam, Subramanian and Yamagata2009; Chiavegatto Filho et al. Reference Chiavegatto Filho, Gotlieb and Kawachi2012), increased infant mortality (Pampel & Pillai, Reference Pampel and Pillai1986; Macinko et al. Reference Macinko, Shi and Starfield2004), poor self-rated health (Subramanian et al. Reference Subramanian, Delgado, Jadue, Vega and Kawachi2003; Mansyur et al. Reference Mansyur, Amick, Harrist and Franzini2008) and violence (Kennedy et al. Reference Kennedy, Kawachi, Prothrow-Stith, Lochner and Gupta1998; Nadanovsky & Cunha-Cruz, Reference Nadanovsky and Cunha-Cruz2009; Pabayo et al. Reference Pabayo, Molnar and Kawachi2014a ). Some studies contradict the income inequality hypothesis (IIH) – especially studies that measure income inequality at a smaller geographical scale (e.g. at neighbourhood or US county level) (Lynch et al. Reference Lynch, Davey Smith, Harper, Hillemeier, Ross, Kaplan and Wolfson2004) – and several authors have suggested that whether the IIH holds as a determinant of poor health could depend on the geographical scale at which it is measured (Chen & Gotway Crawford, Reference Chen and Gotway Crawford2012).
There is a growing evidence that income inequality is associated with increased risk for mental disorders, including common mental disorders (Weich et al. Reference Weich, Lewis and Jenkins2001), depression (Ahern and Galea, Reference Ahern and Galea2006; Pickett & Wilkinson, Reference Pickett and Wilkinson2010; Messias et al. Reference Messias, Eaton and Grooms2011; Chiavegatto Filho et al. Reference Chiavegatto Filho, Kawachi, Wang, Viana and Andrade2013; Pabayo et al. Reference Pabayo, Kawachi and Gilman2014b ), suicide (Gunnell et al. Reference Gunnell, Middleton, Whitley, Dorling and Franker2003), alcohol and cannabis use (Galea et al. Reference Galea, Ahern, Tracy and Vlahov2007), first-episode psychosis (Boydell et al. Reference Boydell, Van Os, McKenzie and Murray2004; Burns & Esterhuizen, Reference Burns and Esterhuizen2008) and schizophrenia (Burns et al. Reference Burns, Tomita and Kapadia2014). These studies have measured income inequality at both national/country-level and local ward/municipality-level. There have been some negative studies also, for example, in relation to depression from the World Mental Health Surveys (Rai et al. Reference Rai, Zitko, Jones, Lynch and Araya2013). One possible explanation for this somewhat anomalous finding is the omission from this study of the majority of countries with either very high Ginis (50 or above) or low Ginis (30 or below). As will be discussed further, the negative (mental) health impact of inequality may only become evident when the ‘sample’ has an adequate distribution of inequality measures.
Despite the weight of evidence for income inequality as a risk factor for mental illness, it is still widely under-acknowledged as a driver (Pickett & Wilkinson, Reference Pickett and Wilkinson2010). Cohen (Reference Cohen2002) argues that psychiatry has failed to focus on issues pertaining to social inequality despite the growing evidence for a strong association.
Key issues to address in relation to income inequality and mental health
If we are to advance understanding of how economic disparities, and in particular inequitable distribution of income and wealth, act as powerful drivers of mental disorder, and thus contribute to the debate on global development goals, then we must engage with several key issues concerning this relationship.
First we must consider which indicators of income/economic inequality are best suited as plausible exposures in relation to disparities in population mental health. ‘Plausibility’ – is a plausible biological mechanism that can be offered linking exposure and outcome – is a key criterion for establishing causal relationships in health epidemiology (Hill, Reference Hill1965). The Gini coefficient is most commonly used in World Bank monitoring of countries' economic status. This is a composite index derived from a ratio of two areas in the Lorenz curve diagram. The Gini however does not show where in the distribution the inequality occurs; and it also tends to be oversensitive to changes in the middle of the distribution and insensitive to changes at the top and bottom. This is problematic according to the Chilean economist Gabriel Palma who demonstrated that middle class incomes almost always represent about half of gross national income while the other half is split between the richest 10% and poorest 40% (Palma, Reference Palma2011). Importantly the share of those two groups varies considerably across countries; thus two very different income distributions can have the same Gini index. The 20/20 ratio inequality metric addresses the problem of the middle 60% statistically obscuring inequality that is otherwise present in the distribution – this is the ratio of the income of the top 20% of earners to that of the bottom 20% of earners. There is also a case for comparing deciles (i.e., 10/10 ratio) or even percentiles (1/1 ratio) rather than quintiles (20/20 ratio), since the latter may hide inequalities within distribution subgroups. Another metric that provides information about the shape of income distribution (rather than the level of inequality) is the ratio of given percentiles to the median. So for example if, over time, the ratio of the 80th or 90th percentile to the mean increases, this would indicate that a simultaneous increase in inequality (e.g. rise in the Gini coefficient) would be a consequence of disproportionate gains by the upper income earners.
The utility of different measures of income inequality in relation to poverty and mental health relates to the question of how inequality interacts with poverty, growth and other economic forces. Lund et al. (Reference Lund, Breen, Flisher, Kakuma, Corrigall, Joska, Swartz and Patel2010) asks how income inequality influences the poverty–mental health relationship; and argues that concepts of income and economic inequality should be integrated into studies of poverty and mental health. Lund notes that in more equitable societies (e.g. Ethiopia, Nigeria) there seems to be a weaker association between poverty and mental disorders; whereas in highly inequitable Chile and Brazil, this association is stronger. In other words, while poverty is independently accompanied by a myriad of noxious factors that are bad for mental health (e.g. overcrowding, food scarcity, exposure to neighbourhood stressors), poverty in the context of marked inequality has an even greater negative impact on mental health. As mentioned earlier, GDP/c is a marker of economic growth, but rising GDP/c over time (i.e., ‘growth’) does not necessarily translate into improved population mental health. In fact, as first pointed out by Simon Kuznets in the 1950s, in regions with low levels of per capita income, inequality initially increases over time with rising GDP/c as the poorest group's share of the overall income growth decreases (Kuznets, Reference Kuznets1955). Kuznets argued that in the later stages of economic development, inequality would start to fall, returning to its initial levels after 60 years (the ‘Kuznets’ hypothesis'). The work of French economists Thomas Piketty and Emmanuel Saez, using data from high-income countries (HICs) over the last 200 years, seems to disprove Kuznets' predictions, showing that increasing growth has been accompanied by steadily rising income inequality as capital accumulates in the higher income group (Piketty & Saez, Reference Piketty and Saez2003; Piketty, Reference Piketty2014). Taken together with the previous point regarding the impact of inequality on the poverty–mental health relationship, Piketty's findings have very serious and worrying implications for the likely future epidemiology of mental disorders within LMICs. Middle-income countries such as South Africa and Brazil – that have what are quaintly termed ‘emerging economies’ – have some of the highest Gini coefficients in the world. If economic ‘growth’ worsens inequality, and inequality increases risk for mental illness, then the future burden of mental disorders in such contexts is likely to be substantially greater than it is currently.
In studying the effects of income inequality on mental health, it is important to consider the geographical scale at which its impact is apparent. Although earlier research suggested that the income inequality–mental health association was more evident in studies analysing countries and states than in studies of smaller geographical areas (e.g. municipalities/US counties or neighbourhoods) (Subramanian & Kawachi, Reference Subramanian and Kawachi2004), several recent analyses also show an effect for these smaller areas (e.g. Chiavegatto Filho et al. Reference Chiavegatto Filho, Kawachi, Wang, Viana and Andrade2013).
A related question is whether income inequality affects all individuals in a society similarly in terms of risk for mental illness or whether the added burden of adverse health outcomes is partitioned to the most deprived segment of the community. The earlier discussion regarding the impact of income inequality on the poverty–mental health relationship would suggest that the risk is not evenly distributed and that the negative effects of income inequality would be most evident in the poorest section of the population. And indeed, in South London, Boydell et al. (Reference Boydell, Van Os, McKenzie and Murray2004) demonstrated that only in the most deprived wards had increased incidence of schizophrenia associated with increasing inequality. However, there is now sufficient evidence to be confident that inequality is a potent risk factor for individuals independent of their income or wealth – as Kawachi et al. (Reference Kawachi, Subramanian and Almeida-Filho2002) have put it: individual health depends not just on personal income, but also on the incomes of other members of one's community or society. If this is the case, then it follows that income inequality is likely to impact on (mental) health through setting up stressful social comparisons as well as disrupting social dynamics and support structures.
Inequality is likely to impact on mental health via a range of mechanisms; however most evidence suggest two related primary mechanisms. First, inequality causes direct stress due to social comparisons where poorer individuals develop feelings of failure, resentment, shame and ‘social defeat’ when comparing themselves with their rich neighbours (Chiavegatto Filho et al. Reference Chiavegatto Filho, Kawachi, Wang, Viana and Andrade2013). Second, inequality erodes social capital in communities and societies, leading to social fragmentation and leaving individuals vulnerable to psychosocial stressors (Wilkinson, Reference Wilkinson1996; Mansyur et al. Reference Mansyur, Amick, Harrist and Franzini2008). Indeed an inverse relationship exists between income inequality and social capital (Kawachi & Kennedy, Reference Kawachi and Kennedy1997); and lower neighbourhood-level social capital has been correlated with depression at the population level within LMIC settings (Tomita & Burns, Reference Tomita and Burns2013).
What are the causes of economic inequality?
In concluding this discussion it is important to consider the upstream causes of economic inequality, since these should also become potential targets in efforts to promote sustainable development goals and improve population (mental) health. Coburn (Reference Coburn2000) argues that we have not paid sufficient attention to the social context of the inequality–health relationship and to the causes of inequality itself. He maintains that neoliberal (market-oriented) political doctrines lead to both increased income inequality and reduced social cohesion, undermining the ‘welfare state’. The rise of neoliberalism is related to globalisation and the changing class structures of advanced capitalist societies and neoliberal policies are ‘associated with a ‘package’ of other likely health-deleterious policies (e.g. de-unionisation, fiscal austerity and privatisation)' (Coburn, Reference Coburn2004).
There are many reports of how the globalisation of neoliberalism, with its emphasis on the market, has led to breakdown of the welfare state and Keynesian economic systems and a deterioration in population health over the last 30 years in both HICs and LMICs – the latter usually in the context of structural adjustment programmes (SAPs) imposed on governments as a condition of loans from the International Monetary Fund (IMF) and World Bank (Bhutta, Reference Bhutta2001; Ikamari, Reference Ikamari2004; Oliver, Reference Oliver2006; Stuckler et al. Reference Stuckler, King and Basu2008; Shandra et al. Reference Shandra, Shandra, Shircliff and London2010; Hossen & Westhues, Reference Hossen and Westhues2012; Baker et al. Reference Baker, Kay and Walls2014).
The radical economic changes in post-Soviet Russia during the 1990s make for a tragic but informative natural experiment on the effects of neoliberal ‘shock therapy’ (Klein, Reference Klein2007) on population mental health. Between 1990 and 2003, the suicide rate almost doubled to 39.7 suicides per 100 000 people, placing Russia among the countries with the highest suicide rates in the world (Veltischev, Reference Veltischev2003; Webster, Reference Webster2003). In the words of a Russian economist: ‘The main source of suicide during the last 10 years is social and economic problems linked to people not being able to adapt to the new conditions (since the fall of Soviet Union)’ (Paton Walsch, Reference Paton Walsch2003). Linked to both suicide and the ‘new conditions’ was a steep rise in both homicide rates (to three times the global average (Leon et al. Reference Leon, Chenet, Shkolnikov, Zakharov, Shapiro, Rakhmanova, Vassin and McKee1997; UNDP Moscow Office, 2003)) and alcohol consumption (Nemtsov, Reference Nemtsov2000; Reitan, Reference Reitan2000) during this period. Not surprisingly, Russia's Gini soared from 23.8 in 1988 to 48.4 in 1993 (World Bank, 2014); and it is not unreasonable to assume that rising inequality was a key mediator between political economic change and plunging population mental health.
Concluding comments
With the negotiation of the next set of sustainable development goals only a year away, those of us who wish to see the final goals and targets focus more accurately on the structural forces driving mental illness (especially within LMIC contexts), need to improve understanding of the causal pathways linking these social determinants and mental health. The poverty–mental health relationship can only be understood meaningfully by integrating the concepts of income and economic inequality into both the discourse and research in this field. Inequality is a powerful and noxious driver of poverty, social fragmentation and human physical and mental suffering. Furthermore, we need to interrogate the upstream political, social and economic causes of inequality itself. As Vicente Navarro phrases it, there is ‘a need to establish the interactions between politics, policy and health outcomes’ (Navarro et al. Reference Navarro, Muntaner, Borrell, Benach, Quiroga, Rodríguez-Sanz, Vergés and Pasarín2006). If we wish to understand the relationships between politics, poverty, inequality and mental health outcomes, now is the time to begin to develop a robust, evidence-based ‘political economy of mental health.’
Acknowledgements
I am grateful to Crick Lund for his thoughtful discussions and feedback during preparation of this editorial.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Statement of Interest
None.
