The Political Ecology of Alcohol as “Disaster” in South Africa’s Western Cape
While attention to the socio-ecological and political economic influences on health grows, there remains a paucity of political ecological analyses of health (King, 2010). At the same time, the growing burden of non-communicable diseases (NCDs) in the Global South demands new conceptual and pragmatic engagements with their modifiable risk factors. Drawing on the example of South Africa, this paper argues that alcohol consumption might usefully be theorised in political ecological lexicon as a “disaster”. To do so, it draws attention to the upstream causes of vulnerability, rather than just the downstream effects of risky drinking. This reorientation is needed for sustainable, publicly acceptable alcohol policies. To realise this, it draws on Blaikie et al’s (1994; 2003) political ecological approach to risk, vulnerability and coping and, more specifically, applies their Pressure and Release model to explore liquor as a situated “disaster” in South Africa’s Western Cape province. In so doing, it aims to mark out an under-explored research agenda that considers alcohol as a pervasive governance dilemma. In addition, it also reflects on the model's utility as a means of communicating findings that might reorient policy discussions on alcohol control in both South Africa and countries of the Global South.
This paper emerges from two points at different ends of the research process. The first concerns how best to theorise and conceptualise the intersections between alcohol, poverty, development and urban space. The second involves the question of how best to convey the complexity of these interrelationships to a series of disparate “end users”. The importance of generating impact from research is undoubtedly clear, especially where research concerns risk behaviours that result in disproportionately high mortality and morbidity rates and that undermine progress towards developmental objectives. Yet, despite this burgeoning of demand for impact (Pain, Kesby et al. 2011; Williams 2012) from the UK’s Research Excellence Framework (REF), research grant criteria, and government Higher Education and research policy, trying to effect impact may necessitate reframing how the research problem itself is viewed. For example, the UK’s Department for International Development (DfID) asks that the primary area of impact be poverty alleviation. The challenge is therefore clear: how to translate social science research in ways that hold the potential to alleviate poverty and generate ‘useful, legible and relevant’ research findings (Wilton and Moreno 2012, 107). In the case of the research project explored in this paper which concerns behaviours and practices that both stem from and aggravate poverty, this task is all the more important, but no less demanding. This paper therefore ruminates on the theoretical and empirical problem of how best to translate findings of complexity into messages that can be communicated simply and which, ultimately, could have the intended outcome of alleviating the vulnerabilities that drive and exacerbate poverty. In so doing, it draws on the example of the risks associated with alcohol (production, retailing and consumption) in the Western Cape Province of South Africa (SA) to consider two impact objectives: (1) Shifting governmental agendas from treating effects to identifying, acknowledging and addressing the causes of hazardous drinking and its (intended and unintended) consequences, and (2) injecting development aspirations into alcohol control policy and, no less importantly, alcohol control into development policy.
To do so, it argues for the value of adopting a political ecological approach to alcohol as an emergent and serious challenge to development, an area of research that is largely absent from the study of alcohol within Geography with the exception of Lawhon’s recent work (2012). The disproportionately high rates of alcohol-related harm in low and middle-income countries (Parry 2000; Room, Jernigan et al. 2002) represent the kind of “disaster” for which the social (pre)conditions ‘arise out of “normal” life’ (Blaikie, Cannon et al. 1994, 49).However, in spite of the political attention now being devoted to alcohol harm reduction; these conditions of normal life court less interest. This doubtlessly undermines the efficiency of efforts to reduce alcohol-related harms as the prevailing risks associated with ‘normal life’ remain unaltered and vulnerabilities persist. If risky drinking practices and their effects are to be mitigated to achieve ‘conditions that improve the health and wellbeing of all people, and equity in the distribution of these conditions between people’ (Labonte, Polanyi et al. 2005, 10), then it is fundamental to reframe discussions on alcohol to consider the causes of risky drinking. Doing so also responds to Blaikie’s recent interjection over the ‘use’ of political ecology (Blaikie 2012) by reflecting on the approach’s application to alcohol framed, somewhat controversially, as “disaster” (Blaikie, Cannon et al. 1994). In contrast to hyperbolic accounts of “epidemics” or even “pandemics” of disease or risk behaviours now common in media and academic accounts alike, theorising and communicating alcohol as “disaster” serves a valuable analytical purpose. In this case, the notion of disaster presents an opportunity to theoretically engage with the intersections of alcohol production, retailing and consumption practices with macro and micro-scale social, economic, and political processes and materialities that shape and reinforce vulnerabilities. Thus, adopting a political ecological approach to alcohol-as-disaster augments nascent work on the political ecologies of health and disease within geography (Turshen 1977; Mayer 1996; King 2010), adds a new perspective to the emergent field of critical geographies of alcohol (Wilton and Moreno 2012) and also enables an engagement with contemporary behavioural risks and their political economic drivers in the study of development.
In Cape Town, capital of the Western Cape, the high prevalence of alcohol-related harms is layered over abject poverty and profligate wealth. The province and city have long struggled with the need to develop an effective alcohol control strategy and, only in April 2012 was the provincial liquor bill made law, after a decade of contestation, debate and delay (see Lawhon and Herrick, 2012). Against an urban backdrop of endemic risky drinking practices, this paper asks how theorising drinking in Cape Town as “disaster” might, somewhat counter-intuitively, deflect tendencies to hyperbole to instead engage with the multi-scalar processes driving risk and vulnerability that underpin socio-ecological models of health. To do so, the paper is structured in three sections. First, it explores how political ecologies might be fruitfully applied to alcohol as an instance where ‘the activities of daily life comprise a set of points in space and time where physical hazards, social relations and individual choice converge’ (Blaikie, Cannon et al. 1994, 13). Second, it argues that Blaikie et al’s under-utilised Pressure and Release (PAR) Modeloffers a powerful tool to interrogate the ‘social production of vulnerability’ (1994, 21) which, in turn, creates the conditions for disaster when they intersect with hazardous substances and behaviours. The search for appropriate frameworks to organise ‘work from analysis to action’ (WHO, 2010: 8) prioritises structuralist interpretations of vulnerability and, consequently, interventions. Socio-ecological models of health also draw attention to the contexts in which health is produced or reproduced and must also, this paper argues, inculcate post-structural interpretations of the progressions of vulnerability if they are to be effective. The PAR model is explored in four sections that examine: (1) the root causes; (2) dynamic pressures; and (3) unsafe conditions that produce (4) hazards and disaster. Third, it discusses the implications of these theoretical engagements for alcohol policy in SA, before offering a brief conclusion.
Disasters, alcohol and political ecologies of health
Non-communicable diseases (NCDs) and their manifold risk factors (e.g. drinking, smoking, inactivity, poor diets) are frequently cast unproblematically as “epidemics” or even “pandemics” within the public health literature, by international organisations (see for example World Health Organisation 2000; World Health Organisation 2011) and the media. Notwithstanding the fact that in strictly epidemiological terms, an epidemic or pandemic is of infectious origin; counter-narratives from within critical social science approaches to health and illness have recently started to challenge the justificatory bases for such semantic choices. For example, many within the emergent field of ‘critical obesity studies’ have deconstructed and challenged the hegemony of obesity’s purported “epidemic” status (Astrup, Larsen et al. 2004; Gard and Wright 2005; Flegal 2006; Oliver 2006; Mitchell and McTigue 2007; Saguy and Almeling 2008; Chiolero and Paccaud 2009), and, in the process, highlighted the consequences of such a framing for both policy and the obese. However, those studying alcohol use and abuse have not yet thought through the consequences of similar metaphorical usage. This paper asserts that alcohol, especially in countries of the Global South (GS) where alcohol control policies are often nascent and poorly enforced, is a disaster on a number of levels and scales. However, rather than applying this label unreflectively and uncritically, it is appended purposefully in order to actively explore how such “long-wave” (Barnett and Blaikie 1992; Adger 2006) disasters are constituted and unfold. Unlike an “epidemic”, disaster is not merely a statement of magnitude or of mode of transmission, but rather represents a systematic starting point for an analysis of the upstream causes and constructions of risk and vulnerability. For, as Parry et al have asserted, ‘attention must now be directed towards addressing the drivers of alcohol use, especially of heavy use, and particularly those drivers operating at the social and environmental level’ (2011, 1722).
The increasing urgency with which alcohol is being viewed by some countries of the GS mirrors bold shifts in the burden of disease from infectious to NCDs such as coronary heart disease, diabetes and cancers, for which lifestyle is a primary risk factor (Beaglehole and Yach 2003; Boutayeb 2006). As such, public health paradigms and policies are being fundamentally challenged by the changing burden of disease and, even more so in countries like SA, where infectious (especially HIV and TB) and chronic diseases coexist (World Health Organisation 2010; UN Secretary General 2011). The report of a recently convened high-level meeting of the UN General Assembly NCDs highlighted that 80% of deaths from NCDs occur in the GS and the burden is by far the highest in middle-income countries (Beaglehole, Bonita et al. 2011). Both the UN and WHO are now calling for greater global attention to the burden of NCDs and, as a result, the causal role played by alcohol as one of four modifiable risk factors which also include: tobacco; physical inactivity and unhealthy diets. Given that alcohol consumption among drinkers in middle-income countries is fast approaching that in high-income countries, the interrelated focus on alcohol and NCDs is a laudable, especially since both have been shown to have serious and complex relationships to poverty. In 2012, the WHO set voluntary targets for countries to reduce per capita adult alcohol consumption by 10% by 2025 (World Health Organisation 2011). It is notable that SA has gone one step further than this, setting a target of a 20% reduction in per capita adult consumption by 2020 (Payne 2012).
In SA, ‘poverty, gender inequalities, crime and violence play a major role in exacerbating the health problems of the South African population’ and, as a result, ‘efforts to improve health will have to extend to the very core of [...] society and cultures, with refurbishment of [the] social fabric and comprehensive strategies to reduce poverty’ (Bradshaw, Groenewald et al. 2003, 687). With these aspirations in mind, political ecological approaches to the study of alcohol are of extensive value given their concern with:
…the need to set a problem or phenomenon into its broader social and economic context, and the need to relate both the phenomenon and its socioeconomic context to a variety of scales ranging from the local to the global (Mayer, 1996: 447)
Such thinking mirrors recent thinking at the WHO on addresses the social determinants of health inequities, building on the work of Michael Marmot (2005; 2010). Given this, it is still curious that while political ecologies have grown in scale and scope (see Bryant 1998; Robbins 2004; Muldavin 2008 for reviews), their application to health remains limited despite concerted calls for greater engagement (Turshen 1977). Political ecologies of health remain scarce even within the burgeoning field of health geography (King 2010) and its concern with the complex relations between health (outcomes, practices and understandings), place, meaning and experience (Kearns 1993; Kearns 1997; Rosenberg 1998; Kearns and Moon 2002; Parr 2004). While mounting concern with so-called ‘diseases of comfort’ that stem from physical inactivity and/or obesity (Choi, Hunter et al. 2005) have again reframed how we should best conceptualise vulnerability in relation to health (Yach, Leeder et al. 2005; Daar, Singer et al. 2007; Robert, David et al. 2010), the notion that behaviours are driven by certain attributes of place mean that they require critical attention to the ecological. By extension, if we agree that ‘the major determinants of health are social, [then] so must be the remedies’ (Marmot 2005, 1103), then critical reflection on the political is of equal importance.
Over the past thirty years, ecological approaches to public health have increased in importance and application as the environmental contexts of behaviour have come under renewed scrutiny when designing intervention strategies. At root, ecological approaches aim to identify and produce environments or ‘social contexts’ (Blankenship, Bray et al. 2000) that enable healthy lifestyles (Sallis, Owen et al. 2008) with ‘ecology’ here representing a human/non-human ecosystem within which human behaviour is shaped (Kickbush 1989). Ecological approaches to health have been prolific in relation to physical activity (Bauman 2005; Blanchard, McGannon et al. 2005; Sallis, Cervero et al. 2006) and obesity (Egger and Swinburn 1997; Reidpath, Burns et al. 2002; Pepin, McMahan et al. 2004; Lang and Rayner 2007), however they have not yet been explicitly applied to drinking despite a burgeoning of geographical engagements with alcohol (see for example Jayne, Valentine et al. 2008a; Jayne, Valentine et al. 2008b; DeVerteuil and Wilton 2009; Wilton and Moreno 2012). This is notable given that the factors producing risky drinking practices exist at multiple levels of influence (e.g. individual, community, regional and national policy, global political economic systems) and require multi-level interventions to change individual behaviour and its environmental determinants (e.g. location of drinking places, licensing, availability etc). Moreover, as models prioritise the role of external drivers, they consequently problematise the politicised notions of individual responsibility and choice that most frequently characterise debates on alcohol policies in the Global North (Sallis, Owen et al. 2008). As Blakenship et al (2000, S16) have argued ‘structural interventions operate to promote public health without necessarily altering individual behaviour’. Thus, such an approach may inspire productive engagements with vulnerabilities and their mitigation, rather than deferring to the tendency to castigate and blame “irresponsible” drinkers in the GS. Thus, the paper responds to Wilton and Moreno’s assertion that geographers should ‘problematize taken for-granted assumptions about the nature of, and motivations for, [...] alcohol use, and direct attention to both the intended and unintended consequences of regulation’ (2012, 106).
While ecological models of health remain primarily structuralist, political ecology itself has embraced post-structuralism (Forsyth 2008). The limited application of political ecology to contemporary health challenges (King 2010) thus seems curious given the clear uptake of post-structuralism to explore the production of health (Lupton 1995; Petersen and Lupton 1996). Furthermore, calls for an explicitly urban political ecology (Bryant 1998; Keil 2003; Swyngedouw and Heynen 2003; Keil 2005), mindful of the ways in which meanings of health and illness are constituted in and by social processes situated in city spaces, make the need for political ecologies of health all the more valid. This is particularly the case in relation to those health concerns located in cities where rapid urbanisation and shifting livelihood patterns can quickly enhance vulnerabilities and undermine coping strategies. Obesity may have been earmarked as a ‘growth area’ in urban political ecology (see for example Guthman and DuPuis 2006; Heynen 2006; Marvin and Medd 2006), but this growth has not yet spilled over into the study of liquor (Lawhon 2012). Questions consequently still need to be asked of how alcohol, as practice and product, is best conceptualised within a political ecological framework and, in turn, how such frameworks might be usefully deployed. Spelt out another way, what kind of punitive ecologies might alcohol reveal and in what ways can political ecology’s concern with ‘political economy and power’ (King 2010, 42) add to our understandings of alcohol-related harm in SA? To begin to approach these questions, Blaikie et al’s PAR model is used as a starting point and framework through which to critically interrogate alcohol-as-disaster in SA and, in so doing, both strengthen the case for political ecologies of alcohol and provide a clear point of communication that gives necessary weight to the importance of causal influences on drinking-as-disaster.
South Africa and the Alcohol Pressure and Release Model.
With a population of 49.3 million and GDP per capita of $5,786 in 2009, SA is Africa’s most powerful economy. Since 2000, GDP per capita has risen by $2,736 or 90% (World Bank 2010). However, with “official” unemployment in excess of 25%, average life expectancy of 51.5 (which represents a decline of 4.5 years since 2000) and average HIV prevalence of 18% (Ibid), the country faces clear barriers to achieving equity in health. Added to this, SA’s high rates of alcohol consumption are layered over a post-apartheid ‘protracted and polarised health transition’ characterised by the ‘persistence of infectious diseases, high maternal and child mortality and the rise of non-communicable diseases’ (Chopra, Lawn et al. 2009, 1). SA may be an exception in the region for its reasonably and relatively well-developed alcohol regulations and a vocal public health lobby (Parry 2010); but its fine line between ‘sociable and unsociable drinking’ (Mager 2010, 3) presents clear policy challenges. Moreover, the vulnerabilities to the multiple effects of alcohol consumption are not limited to the country’s drinkers, but rather endured by the majority of the population. Furthermore, the co-existence of regulated (i.e. licensed and legal) and unregulated (i.e. unlicensed and illegal) drinking spaces present clear governance challenges. This is due to the extent to which the delineation between public and private drinking (mirrored in legislative terms by the on/off trade distinction) that guides the political logic of alcohol control interventions (i.e. public order v. public health) in the Global North becomes even more of a “grey zone” of “ambiguity” (Blum 2011) when transposed to contexts in the Global South. Indeed, where prevailing policy thinking has been to tightly regulate alcohol supply to reduce average drinking rates among the whole population (Kneale and French 2008), the dominance of the informal unlicensed sector in alcohol retailing and consumption (and, increasingly production) raises important questions about the broader effects of supply-side interventions on either risk or vulnerabilities. For this reason, the paper suggests a need to reorient thinking from addressing effects to mitigating the causes of consumption.
The development rationale for this is evident: In 2010, SA ranked towards the lower end of the “medium” development category (110/169) in the Human Development Index. The persistence of poor health indicators (e.g. under-nourishment, the mortality rate of under fives and life expectancy) means that this ranking has not demonstrably improved since the measure’s inception in 1990 (UNDP 2010a). The importance of health outcomes is further revealed by the country’s progress towards meeting the MDGs. The ‘MDG Monitor’ suggests that SA is “on track” to eradicate extreme poverty and hunger, but looks less likely to satisfy the criteria needed to combat HIV/AIDS, malaria and other diseases, reduce child mortality and maternal health (UNDP 2010b), largely because of the rising burden of NCDs. It should be noted that the MDGs have been subject to intense criticism for their ‘performance measurement’ logic which is ‘poorly and arbitrarily designed to measure progress against poverty and deprivation’ (Easterly 2009, 26). This logic, Sen (2009, 229) argues, tends ‘to focus specifically on the enhancement of inanimate objects of convenience’, at the expense of enhancing opportunities for people to improve their wellbeing. Moreover, the downstream thinking encouraged by the target-driven logic of the MDGs risks losing sight of the upstream focus of rights and capabilities-based approaches to development (Sen 1999; Sen 2009). For this reason, political ecological approaches and the PAR model serve as valuable tools for identifying those upstream causes of vulnerability that so often actually make drinking a rational response to a seemingly irrational or hazardous environment.
The PAR model therefore serves as an ‘explanatory or organisational device’ (Blaikie, Cannon et al. 1994, 59) within which different theoretical constructs can be emplaced. At root, it systematises how “disaster” emerges when ‘a significant number of vulnerable people experience a hazard and suffer damage/ disruption of their livelihood system’ (Blaikie, Cannon et al. 1994, 21). Recovery is compromised because hazards impinge upon household resources, resilience and capacities. Vulnerability is thus the culmination of three stages: root causes (e.g. the unequal economic, demographic and political processes that influence resource distribution); dynamic pressures (e.g. processes that translate root causes into unsafe conditions); and unsafe conditions (e.g. the specific expression of vulnerability in space and time). The PAR model highlights those upstream causes of vulnerability that are often deeply unpalatable to policy makers because ‘any fundamental solutions involve potential change, radical reform of the international economic system, and the development of public policy to protect rather than exploit people and nature’ (Blaikie, Cannon et al. 1994, 233). However, such reform may be necessary if the ‘catastrophic expenditure’ associated with alcohol-related harm and poor health are to avoid becoming ‘a substantial drain on society’s economic potential by adversely affecting the four main factors of economic growth – i.e. labour supply, productivity, investment, and education’ (Beaglehole, Bonita et al. 2011, 450-451). Yet, despite the model’s clear potential for exploring the risks and vulnerabilities associated with alcohol, its application has thus far been limited to HIV/AIDS (Tsasis and Nirupama 2008).
In contrast to most European counties where drinking is the norm, ‘many Africans abstain from alcohol...[but] those who do drink, drink a lot’ (Jernigan and Obot 2006, 58). South African drinking is characterised by high rates of heavy episodic drinking at weekends, underage drinking and alcohol dependence among men (World Health Organisation 2004). While 55% of men and 83% of women do not drink, this does not make them any less vulnerable to the effects of unsafe sex, interpersonal violence, accidents and injury precipitated by others drinking. In SA, therefore, alcohol might be viewed as a disaster that ‘affect[s] people in varying ways and differing intensities’ (Blaikie, Cannon et al. 1994, 5), with distinct spatial manifestations. SA is notable in this respect as apartheid-era urban planning has carved cities such as Cape Town into distinct pockets of privilege and privation (Parnell and Mabin 1995; Parnell 1997; Lemanski 2004; Lemanski 2007; Tucker 2008), which often map directly both onto the unsafe conditions that generate vulnerability and the normative social relations to drinking that both legitimate and castigate risky behaviours. Thus, place – understood as having distinct materialities and produced through ‘particular sets of social relations, networks and experiences’ (King 2010, 42) – is inextricable from the constitution of unsafe conditions.
Blaikie et al (1994) originallyapplied the PAR model to famines, biological hazards (i.e. the Irish Potato Famine), floods, coastal storms and earthquakes. However, they offered little methodological guidance as to how to apply the model to alternative contexts. Thus, for the research project explored in this paper, the model represents the research team’s collective analysis of both secondary (i.e. literature review and health survey data) and primary data sources (i.e. focus groups, stakeholder interviews, ethnographic fieldwork) according to the categories in Blaikie et al’s original model (1994, 23). This is shown in figure 1. Building from this, figure 2 details the PAR model as applied to the political ecology of liquor in the Western Cape. The model thus represents the systematisation of the research participants’ assertions of the most significant causes of risky drinking practices and their consequences by the research team, bringing together the broad findings of a multi-method research project in a way that enables further, in-depth exploration of each category. Each stage of the model will be explored in turn below before turning to a broader discussion of the policy significance of such an approach.