Elsevier

Social Science & Medicine

Volume 70, Issue 8, April 2010, Pages 1185-1193
Social Science & Medicine

Decomposing cross-country differences in levels of obesity and overweight: Does the social environment matter?

https://doi.org/10.1016/j.socscimed.2009.12.011Get rights and content

Abstract

This paper examines the influence of environmental factors on weight gain and obesity. We take advantage of a markedly different pattern of obesity between Italy in Spain to undertake a non-linear decomposition analysis of differences in the prevalence of overweight and obesity between the two countries. The analysis is based on cross-sectional national surveys for 2003. We have attempted to isolate the influence of lifestyle factors, socio-economic and socio-environmental effects in explaining cross-country differences in BMI status. Our findings suggest that when the social environment (proxied by different measures of peer effects and regional BMI) is not controlled for, we explain about 27–42% of the overall Spain-to-Italy overweight and obesity gap. Differences in eating habits and education between the two countries are the main predictors of the gaps in obesity and overweight. However, when social environment is controlled for, our estimates explain between 76 and 92% of the obesity and overweight gap and the effect of eating habits are wiped out. These results suggest healthy body weight depends on cultural or environmental triggers that operate through individual level health production determinants.

Introduction

In the last two decades obesity rates have tripled in Europe (Branca, Nikogosian, & Lobstein, 2007) where obesity is estimated to be responsible for up to 6% of total health care expenditure (WHO, 2006). This feature parallels the increasing globalisation and modernisation of southern European societies, which impacts on individual's lifestyles, especially through time constrains. Importantly, vast differences are found among different countries although very little empirical analysis has been gathered on what underpins cross-country differences in obesity across Europe.1 Cross-country differences in obesity point towards the importance of institutional (e.g., education systems) along with socio-environmental factors (e.g., cultural treats and social values that influence the importance of body image or the amount of food ingested), both of which can influence the diet and physical activity (or the caloric balance) that are ultimately responsible for weight gain.

Common economic approaches to weight gain tend to explain obesity as dependent on individual's decisions. Although individuals receive utility from being healthy, at the same time they experience sacrifices in terms of opportunity cost from foregoing other consumption activities such as pleasure gained from food intake and, in general, energy saving incentives that prevent unnecessary physical activity. Current research into obesity emphasises several factors that impact on obesity. Examples of these include the effect of technological change (Cutler, Glaeser, & Shapiro, 2003), sedentary lifestyles, impatience and time preference, alongside food price drops resulting from agricultural innovations and availability of fast foods (Chou et al., 2004, Finkelstein et al., 2005, Komlos et al., 2004, Lakdawalla and Philipson, 2002, Philipson, 2001, Rashad et al., 2006).2 However, food-related health behaviours rely typically on information that is socially learned, both from mimicking each others behaviours, and from prevailing social norms (see Glaeser et al., 2002, Lewitt, 1999). Hence, against the backdrop of standard economic theory, we argue that cost and benefit comparisions are undertaken in the specific community of reference rather than at the individual level, and accordingly depend on local culture. To shed some light into this question, this paper has two broad aims.

First, we take advantage of an empirical puzzle that arises from comparing patterns of obesity and overweight between two Mediterranean countries, namely Italy and Spain. Indeed, both countries share some socio-cultural and geographical treats; have a similar GDP per capita and the same level of dietary compliance with the World Health Organization nutrition targets (Mazzocchi, Brasili, & Sandri, 2008). Hence, this leaves us with other underlying factors including education (Costa-Font & Gil, 2008) and differences in cultural adaptation to a more global world as underpinning reasons for the obesity gap. Hence, further analysis need to be carried out to examine the extent to which traditional predictors like education and lifestyles and other social environmental factors explain cross-country differences in the prevalence of obesity. Paradoxically, when we examine obesity rates in the countries under study, Fig. 1 reveals that whilst in 1990 the prevalence of obesity was between 6 and 7% in both countries and obesity rate gaps were not statically significant, in 2003, the prevalence of obesity in Italy remained at 8% but grew to 14% in Spain. In such a short period of time, the respective genetic variation cannot provide a reasonable account for such differences. Yet, other explanations point towards cultural and environmental factors that might have influenced individual's energy balance (Bleich, Cutler, Murray, & Adams, 2007, Drewnowski & Darmon, 2005). Furthermore, Eurobarometer data on Italy and Spain reveals significant differences in social norms and values regarding body shape and food behaviour (Eurobarometer, 2005). Hence, we examine socio-environmental factors as underpinning differences in obesity and overweight between both countries.

Second, we address the paucity of literature undertaking cross-country comparisons of obesity which is quite limited. Contoyannis and Wildman (2007) use relative distribution analysis to examine changes in the distribution of BMI using non-parametric methods. Interestingly, in a comparison of England and Canada, their results indicate that the expansion of BMI increased at a much faster rate over time in England. However, the factors that contribute to creating this gap remain unclear. Costa-Font et al (2009) adopt a counterfactual quantile decomposition approach to the analysis of Spain-to-Italy BMI gaps. They find that such gaps are largely explained by cross-country variation in the returns to each covariate, especially for younger women, while adverse underlying determinants do not contribute. Cross-country behavioural differences appear to be key.3

This paper employs an econometric decomposition of cross-country differences in the prevalence of overweight and obesity using different measures of socio-environmental effects that have not been employed before which avoid individual sorting problems. Given that the variable of interest is binary, we draw upon methods of non-linear decomposition (Fairlie, 1999, Fairlie, 2005) to examine the underlying factors (including age, education and lifestyles) that might explain cross-country differences in overweight and obesity in two otherwise similar Mediterranean countries. We do so by defining sub-samples according to different criteria of interest (age and gender), to identify among observable factors – aggregated into broad groups – those that account for the cross-country differences. Note that the non-linear decomposition is undertaken with and without the inclusion of socio-environmental effects. Social environment is examined using two different measures of peer effects and social norms on the one hand, and regional BMI on the other.

Our findings suggest that eating habits and education are the main predictors of total cross-country differences in BMI status between Italy and Spain (27–42%). Among men, eating habits explain up to 35% of differences in levels of (class 1) obesity in early adulthood, whilst education explains 41% of these differences in later adulthood. Among women, education appears to explain about 25% and 39% of the obesity difference in early and middle adulthood respectively, while eating habits only explain these differences in older adulthood (23%). Paradoxically, when social environmental controls are included, we find that social norms and environmental effects override previous results (basically regarding eating habits), regardless of gender, and explain between 70 and 96% of total differences, a pattern that rises with age. The rest of the paper is structured as follows. Next section describes our data, then the paper reports our empirical strategy, followed by the results and a final section discusses these findings.

Section snippets

Data

We use cross-sectional data drawn from representative surveys for Italy and Spain. The Italian data are taken from the 2003 edition of the National Survey on Daily Life (“Indagine sugli Aspetti della Vita Quotidiana”), a survey conducted by the National Institute of Statistics (ISTAT) that collects multipurpose individual data including health conditions, health care access, dietary habits and body weight and height. The original sample contains information on 20,547 complete households

Methods

Our aim here is to compute differences in the prevalence rates of overweight and obesity between the two countries by subpopulation groups and, then, to decompose these differentials into their separate underlying factors. To do this, rather than apply the traditional Blinder-Oaxaca (Blinder, 1973, Oaxaca, 1973) decomposition method to determine differences in measurable characteristics on a continuous variable, we employed Fairlie, 1999, Fairlie, 2005 decomposition technique, as it is

Descriptive evidence

In Fig. 2 we present the age distribution of the prevalence of self-reported overweight and obesity (classes 1 and 2) by gender and adults aged 18–65 in the two countries. According to this evidence, rates of overweight and obesity prevalence tend to increase as the individuals age for both genders in the two countries, although this pattern is much more pronounced in the case of Spain. Obesity rates in Italy always falls below the levels reached in Spain with a cross-country gap that remained

Discussion

Drawing upon evidence of two countries that are subject to significantly different social norms and values regarding body shape and behaviour towards food (Eurobarometer, 2005), this paper takes to the data the hypothesis that social environment matters by exploring the role played by our proxies of such effects, namely, social norms (peer influences) and regional BMI, along with education and lifestyles. We have undertaken an empirical analysis of the cross-country gaps in overweight and

References (37)

  • F. Branca et al.

    The challenge of obesity in the WHO European region and the strategies for response

    (2007)
  • J. Cawley et al.

    The impact of state physical education requirements on youth physical activity and overweight

    Health Economics

    (2007)
  • P. Contoyannis et al.

    Using relative distributions to investigate the body mass index in England and Canada

    Health Economics

    (2007)
  • D.M. Cutler et al.

    Why have Americans become more obese?

    Journal of Economic Perspectives

    (2003)
  • Encuesta Nacional de Salud
    (2003)
  • Eurobarometer # 64.3: Foreign languages, biotechnology, organized crime and health items, November–December 2005, ICPSR...
  • R.W. Fairlie

    The absence of the African–American owned business: an analysis of the dynamics of self-employment

    Journal of Labour Economics

    (1999)
  • R.W. Fairlie

    An extension of the blinder-oaxaca decomposition technique to logit and probit models

    Journal of Economic and Social Measurement

    (2005)
  • Cited by (46)

    • The weight of patriarchy? Gender obesity gaps in the Middle East and North Africa (MENA)

      2020, Social Science and Medicine
      Citation Excerpt :

      In some world regions, manual tasks that were traditionally carried out by women may have been automatized more quickly, leading to a decrease of physical activity among women. Other explanations relate to culture and body type preferences in different societies (Costa-Font et al., 2010). In some cultures, female weight is associated with high social status, maternity and nurturing, leading to a preference for high body weight.

    • Geographic determinants of individual obesity risk in Spain: A multilevel approach

      2017, Economics and Human Biology
      Citation Excerpt :

      First, adult obesity rates in Spain are relatively high compared to those in the rest of Europe and the OECD (OECD, 2012), while among EU countries the impact of obesity on avoidable mortality is particularly high in Spain (Costa-Font and Gil, 2008). Additionally, obesity rates have increased rapidly compared to those in countries of similar characteristics (Costa-Font et al., 2010). Second, because of the country's decentralized health care system with health competencies having been devolved to the ACs, disparities are conspicuous even within Spain, making the country even more interesting to study.

    • ‘Globesity’? The effects of globalization on obesity and caloric intake

      2016, Food Policy
      Citation Excerpt :

      That is, diets change towards greater consumption of fat, added sugar, and animal food products, but reduced intake of fiber and cereals (Bray and Popkin, 1998; Duffey et al., 2010). Another source of influence refers to changes in the exposure to global socio-cultural environments (Egger et al., 2012; McLaren, 2007; Monteiro et al., 2000; Costa-Font and Gil, 2004; Costa-Font et al., 2010; Ulijaszek and Schwekendiek, 2012; Ulijaszek, 2007). Such social environmental changes increasingly are recognized as responsible for an “obesogenic environment” (Lake and Townshend, 2006; Swinburn et al., 1999), which predispose people to becoming obese if they follow some prevailing environmental norms.

    View all citing articles on Scopus
    View full text