Obesity under affluence varies by welfare regimes: The effect of fast food, insecurity, and inequality
Introduction
Body weights have risen substantially in affluent countries during the last three decades. In the United States by 2000, nearly two-thirds of the population were ‘overweight’ (BMI > 25 kg/m2) and almost one-third were ‘obese’ (BMI > 30 kg/m2) (Wang et al., 2008). Similar increases happened quite rapidly, at different levels and rates, in most affluent countries (International Obesity Task Force, 2010). Obesity is harmful to health. It is seen as unattractive and is known to be stigmatising (Latner et al., 2008, Puhl and Heuer, 2009). The literature on obesity is large and covers many disciplines, but there is little agreement about causes. Recently there were seven different models of population obesity (Ulijaszek, 2007). A British government consultation (the Foresight report) has produced causal diagrams of staggering complexity (Butland et al., 2007, Figs. 5.2–5.5).
Country rankings of obesity indicate that a cluster of wealthy English-speaking countries have a higher prevalence of obesity than other affluent countries with similar levels of income (e.g., Delpeuch et al., 2009, Fig. 1). We seek to confirm this observation, to provide an explanation, and to support it with data. The data we use consist of 96 body-weight surveys undertaken in eleven countries between 1994 and 2004. It is an ecological regression meta-study, which pools many surveys over a short period of time. These confirm that English-speaking countries form a cluster of their own with regard to obesity. There is separate evidence that overeating may be a personal response to chronic life stress (Torres and Nowson, 2007). English-speaking countries have gone further in the direction of unregulated market liberalism than other affluent societies. Our hypothesis is that market-liberal countries have an environment of greater economic insecurity, and that this is the source of the stress that drives higher levels of obesity. The institutional structures that neoliberal societies put in place promote insecurity and inequality (Lazzarato, 2009), while work-related insecurity, including low income, poor job mobility and absence of union protection, elevates the likelihood of stress and ill health (Scott-Marshall, 2010, Nakao, 2010). Responses to stress, in turn, include overeating (Greeno and Wing, 1994) and preferences for high-energy density foods (Oliver et al., 2000), both of which are implicated in the causation of obesity (Björntorp, 2001).
In this study, insecurity is a predictor for obesity. But the opposite may also be the case. Obesity can be measured objectively and is difficult to hide. Disorders like stress and anxiety are not easy to identify, diagnose, and observe. They are not often recognised as attributable to market liberalism. Richard Sennett has written of ‘the hidden injuries of class’ (Sennett and Cobb, 1972). If the link with stress and insecurity is established, then the epidemiology of obesity might be used as a diagnostic for these less visible disorders. That is one of the promises of this line of investigation.
The concept of welfare regimes comes from Esping-Andersen (1990) who made a distinction between the Nordic social democratic model of welfare, the continental European family-oriented model, and the English-speaking liberal model (also Goodin et al., 1999). Since the 1980s, there has been a movement away from social democratic (or in the USA, ‘New Deal’) policy norms, towards more market-friendly policies. This matches the timing of the emergence of obesity as a mass social phenomenon (Ulijaszek and Lofink, 2006). Hall and Soskice (2001) distinguished two ‘varieties of capitalism’, on the one side six English-speaking ‘liberal market economies’ (the USA, Britain, Australia, Canada, New Zealand, Ireland), on the other 10 ‘coordinated market economies’, Germany, Japan, Switzerland, the Netherlands, Belgium, Sweden, Norway, Denmark, Finland, and Austria (the Mediterranean countries Greece, Italy, France, Spain, Portugal, and Turkey were all more ambiguous). Our findings do not confirm Esping-Andersen's threefold classification. We have not found that the Nordic countries form a statistically distinctive group. Our results are more consistent with the Hall and Soskice approach, which highlights the distinctiveness of the ‘liberal market economies’ from everyone else. It is this distinctiveness that we investigate.
In its basic form, our hypothesis is that economic uncertainty and unequal market and household experiences have increased stress, and that stress is conducive to weight gain; that market-liberal reforms have stimulated competition in both labour and consumption markets, and that this has undermined personal stability and security. It has affected people more strongly lower down the social scale (Drewnowski and Darmon, 2005, Wang et al., 2008). Hence the more intensive the competitive and market orientation of welfare regimes, the higher the level of body weight, at both aggregate and personal levels. Support for this view also comes from market-liberal economists, who regard shifts in relative prices within markets as a plausible explanation of weight increase. Due to their commitment to the optimality of market outcomes, they are relatively untroubled about the rise in body weight, although its choice as a subject suggests an acknowledgement of concern (Cutler et al., 2003, Lakdawalla et al., 2005, Philipson and Posner, 2008). Obesity in their view is merely an unintended symptom of otherwise benign personal preferences and policy norms. Some stylized facts that support the hypothesis are that overweight is more common among the poor (McLaren, 2007), that weight has risen overall over time, and that there is a gradient in body weight, with the highest levels to be found under liberal welfare regimes in the English-speaking countries (Pickett et al., 2005).
These observations are consistent with two interpretations. We try to discriminate between them, and to estimate their impact. The first is the ‘food shock’ hypothesis. A supply shock was driven by the shift in provision of processed food from the home and into the market, where it has become much more accessible. The relative price of food fell, and high-energy density food, which is a staple of fast-food supply, is highly palatable (Offer, 2001, Offer, 2006, Cutler et al., 2003, Drewnowski and Darmon, 2005). One individual-level study of obesity in the USA has found that the strongest explanatory factor is the geographical concentration of fast-food outlets (Chou et al., 2004).1 At the same time, the occupational transition from manufacturing to services, and increased motorization, have both reduced the opportunities for physical exercise (e.g., Philipson and Posner, 2003, Philipson and Posner, 2008).
A second hypothesis arises from a physiological association that has been observed between stress and overeating (Björntorp, 2001, Dallman et al., 2003, Dallman et al., 2005, Drapeau et al., 2003). At the socio-economic level, two stress-inducing mechanisms have been invoked. One comes out of the ‘psychosocial’ hypothesis of socially differential morbidity and mortality in affluent societies (Marmot, 2004, Pickett et al., 2005, Wilkinson and Pickett, 2009). The source of stress is the experience of subordination, and a proxy indicator is income inequality. These authors report that higher income inequality at the national (or USA state) level is associated in the aggregate with higher body weight (Pickett et al., 2005, Pickett and Wilkinson, 2007, Wilkinson and Pickett, 2009).
Smith (2009) and Smith et al. (2009) have proposed that the source of stress might be economic insecurity. Their point of departure is in the study of animal behaviour. Animals in the wild and in captivity respond to food uncertainty by putting on weight. Uncertainty gives rise to anxiety which prompts ‘self-medication’ by means of food. The notion of ‘comfort food’ is likewise familiar. In a similar way, feelings of uncertainty and anxiety encourage overeating. Smith speculates about the evolutionary basis and biochemical pathways of this mechanism.
Several writers have associated myopia (or time-inconsistency) with the rise in body weight (Offer, 2001, Cutler et al., 2003, Komlos et al., 2004). People whose long-term objective is a steady body weight nevertheless find it difficult to resist the temptations of weight-increasing foods. Myopic bias is a form of impatience, and hence implicated in anxiety. Myopia is exacerbated by the pace of market innovation in food provision (Offer, 2006, 144–148). The pursuit of materialism more broadly is associated with lower subjective well-being (Kasser, 2002, Kasser et al., 2004). Oswald and Powdthavee (2007, F443) have shown increasing levels of distress in the UK between 1991 and 2004. Twenge (2000) has found more than one standard deviation rises in anxiety levels in the USA since the 1950s, and an even larger increase in general psychopathology since the late 1930s (Twenge et al., 2009). A large international survey of mental disorder in 2001–2003 has shown prevalence in the USA at 26.3%, with the average for six ‘continental’ European countries at 11.9% – a mental-health gap even larger than the obesity gap (calculated from Demyttenaere et al., 2004).2 A study of the relation between obesity and emotional disorder (using the same survey) suggests ‘modest relationship between obesity (particularly severe obesity) and emotional disorders among women in the general population’ (Scott et al., 2008, 192).
Processed food technologies have been available equally across the developed world. Their pace of deployment and their relative penetration provide a means of comparing and testing the supply shock and welfare-regime approaches. Obesity levels differ among welfare regimes. It could be argued that some countries merely lag behind in exposure to the food shock. However, if the rates of growth in obesity differ between countries with different regimes, then that points towards a welfare-regime interpretation. If the rates of obesity growth are similar, then the ‘supply shock’ interpretation will appear more compelling.
Section snippets
Data
The surveys were identified through online databases (e.g., the WHO BMI database) and other relevant web-based searches (e.g., government statistical office websites). It was judged that for aggregated data of this type, a pooled ordinary least squares (OLS) regression was the most reliable approach. This made it desirable to restrict the selection to a relatively short time frame, namely 1994–2004. This is preparatory to a panel study, which will use individual-level data in fewer countries.
Results
The main findings are reported in Table 2. The relative strength of each variable is indicated by the standardized beta coefficients, and that is the variable of interest. The regression coefficients are, however, intuitively meaningful, although not quite comparable to each other. TOTAL_OBESE, i.e., obesity prevalence, ranges from 5% to 32% of the adult population. MEASURED raises obesity prevalence by about a third over this range. MARKET_LIB raises obesity prevalence by 4 percentage points.
Discussion
Three putative determinants of obesity are discussed in turn: fast-food shock, security, and equality.
Conclusion
This study is a snapshot of obesity in advanced countries at the end of the 20th century, and allows some preliminary conclusions. Market-liberal countries stand out as having high levels of obesity, and higher rates of obesity growth. The TIME variable is more powerful in market-liberal countries (Table 3). The United States in particular is an outlier, ranking first on both levels and rates of growth of obesity, but market-liberal distinctiveness remains when the United States is left out (
Acknowledgments
We are grateful for good comments to the journal's referees, to Michael Lipton, and to participants at the Obesity and Welfare Regimes conference at Oxford in 2009, as well as at several other seminars and conferences. Financial support was provided by the BUPA Foundation and by a British Academy Conference Grant. The Economist generously provided Big Mac index data.
References (58)
- et al.
An economic analysis of adult obesity: results from the behavioral risk factor surveillance system
Journal of Health Economics
(2004) - et al.
Chronic stress and comfort foods: self-medication and abdominal obesity
Brain, Behavior, and Immunity
(2005) - et al.
Food choices and diet costs: an economic analysis
Journal of Nutrition
(2005) - et al.
Predictors of weight increases over 7 years in fire fighters and paramedics
Preventive Medicine
(1996) - et al.
Relationship between stress, eating behavior, and obesity
Nutrition
(2007) Frameworks of population obesity and the use of cultural consensus modeling in the study of environments contributing to obesity
Economics and Human Biology
(2007)- et al.
Promise and pitfalls in the use of “secondary” data-sets: income inequality in OECD countries as a case study
Journal of Economic Literature
(2001) Do stress reactions cause abdominal obesity and comorbidities?
Obesity Reviews
(2001)- Butland, B., Jebb, S., Kopelman, P., McPherson, K., Thomas, S., Mardell, J., Parry, V, 2007. Foresight: tackling...
- et al.
Why have Americans become more obese?
Journal of Economic Perspectives
(2003)
Chronic stress and obesity: a new view of “comfort food”
Proceedings of the National Academy of Sciences of the United States of America
Globesity: A Planet Out of Control?
Prevalence, severity, and unmet need for treatment of mental disorders in the World Health Organization World Mental Health Surveys
JAMA
The Obesity–Hunger Paradox
Is visceral obesity a physiological adaptation to stress?
Panminerva Medica
Bait and Switch: The Futile Pursuit of the Corporate Dream
The Three Worlds of Welfare Capitalism
Prevalence and trends in obesity among US adults, 1999–2008
JAMA
The Real Worlds of Welfare Capitalism
Stress-induced eating
Psychological Bulletin
The Great Risk Shift
Varieties of Capitalism
Occupational factors and 5-year weight change among men in a Danish national cohort
Health Psychology
Health Survey for England 2004. Updating of Trend Tables to Include 2004 Data
Economic Security for a Better World
The High Price of Materialism
Obesity and the rate of time preference: is there a connection?
Journal of Biosocial Science
Cited by (124)
Prenatal economic shocks and birth outcomes in UK cohort data
2021, Economics and Human BiologyRecreational marijuana laws and junk food consumption
2020, Economics and Human BiologyWelfare-based income insecurity in the us and germany: evidence from harmonized panel data
2020, Journal of Economic Behavior and OrganizationThe shattered “Iron Rice Bowl”: Intergenerational effects of Chinese State-Owned Enterprise reform
2019, Journal of Health EconomicsHealth impact of the Anthropocene: The complex relationship between gut microbiota, epigenetics, and human health, using obesity as an example
2020, Global Health, Epidemiology and Genomics