Access to fast food has often been blamed for the rise in childhood obesity; our econometric evidence suggests that fast food exposure had no effect

Childhood obesity, is fast food exposure a factor? Peter J. Dolton, WiktoriaTafesse. Economics & Human Biology, June 26 2022, 101153. https://doi.org/10.1016/j.ehb.2022.101153

Highlights

• Access to fast food has often been blamed for the rise in childhood obesity.

• Possible links to obesity has motivated policies to curb the spread of fast food.

• Spatial and timing data on early fast food outlets from the UK from 1968-86 is used.

• Medically measured data on Body Mass Index for a British Cohort is exploited.

• Our econometric evidence suggests that fast food exposure had no effect.

Abstract: Access to fast food has often been blamed for the rise in obesity which in turn has motivated policies to curb the spread of fast food. However, robust evidence in this area is scarce, particularly using data outside of the US. It is difficult to estimate a causal effect of fast food given spatial sorting and ever-present exposure. We investigate whether the residential access to fast food increased BMI of adolescents at a time when fast food restaurants started to open in the UK. The time period presents the study with large spatial and temporal differences in exposure as well as plausibly exogenous variation. We merge data on the location and timing of the first openings of all fast food outlets in the UK from 1968 -1986, with data on objectively measured BMI from the 1970 British Cohort Survey. The relationship between adolescent BMI and the distance from the respondents’ homes and time since opening, is studied using OLS and Instrumental Variables regression. We find that fast food exposure had no effect on BMI. Extensive robustness checks do not change our conclusion.

JEL: I120I190

Keywords ObesityDiet

8. Conclusion

This paper studied the relationship between exposure to fast food and adolescent BMI using the BCS and historical data relating to the inception of fast food in Great Britain. The data on the timing of establishment and location of all fast food outlets prior to 1986 allowed us to investigate whether fast food proximity, duration since opening, as well as a generated intensity measure taking into account the the proximity and durations of multiple outlets, affects BMI. We do not find any evidence of a positive association between numerous measures of exposure in the home environment and adolescent BMI. This study has filled a gap in the existing literature which has mostly focused on the distance to ever-present fast food restaurants using American data.

Our results are robust to instrumenting for the distance to one’s closest fast food outlet with the distance to a fast food distribution centre. Additionally, one company, Wimpy, suddenly increased its number of fast food outlets which did not allow for a strategic timing and citing of their outlets. Restricting the analysis to Wimpy outlets confirms the zero results. The lack of a relationship is supported by previous research, see; Anderson and Matsa, 2011, Fraser et al., 2012, Lee, 2012, Dunn et al., 2012 and Asirvatham et al. (2019).

There are several potential explanations for our null findings. Firstly, the effect may be highly context specific, see Dunn, 2010, Anderson and Matsa, 2011, Dunn et al., 2012, Grier and Davis, 2013. Adolescents may have less control over food choices in their home environment compared to their school environment. Our findings might differ from studies using American data where fast food is eaten more frequently than in the UK (Fraser et al., 2012). In fact, our analysis does not find support of fast food proximity having a meaningful impact on the frequency of takeaway consumption. Alternatively, the time period of our study may not translate to large effects on obesity as only a small proportion of our sample was exposed to fast food very near their home, particularly at younger ages. Specifically, the lack of weight gain during our study period may be explained by fast food being a novelty or not being comparably cheap relative to other foods as they are today which may cause different consumption patterns (Wiggins et al., 2015). Moreover, it is uncertain how well diets and caloric expenditure during the 1980’s compare to current levels and how this may interact with access to different size and scope of fast food establishments. Additionally, the population studied might not have a large propensity to gain weight if exposed to fast food given that positive effects have been found for specific sub-groups such as ethnic minority urban youth (Currie et al., 2010, Grier and Davis, 2013 and youth living in the poorest and one of the least healthy American states (Alviola et al., 2014). Our reduced sample size does not permit us to do a heterogeneity analysis.

We contribute to the existing literature, often based on data for smaller geographical areas, by using nationwide data. Our results are based on the sample of BCS respondents who did not relocate in the last 6 years and for whom anthropometric and postcode information is available. However, it should be noted that this population could differ from the overall nationally representative sample. Various types of food outlets have been shown to cluster together (Hobbs et al., 2019a). Therefore, a limitation is that we are unable model the direct effects of community determinants of body weight, such as the commercial food environment or access to exercise inducing spaces, due to limited area-level information in the BCS and the lack of supplementary historical data. Furthermore, our small sample size and rich set of controls does not allow us to control for local area fixed effects.

Keeping these caveats in mind, there is no evidence of that the introduction of fast food induced any behavioural change which resulted in weight gain amongst adolescents in the UK in the 1980s. Our overall findings are supported by there being a decrease in total calories purchased since the 1980’s (Griffith et al., 2016). Thus, we suggest that it is unlikely that the access to fast food caused the British obesity epidemic. Half of local government areas in England have enacted policies to curb takeaway food outlets which for example restrict new outlets from opening in designated exclusion zones around places used by children (Keeble et al., 2019). However, despite such policies being common, there is a scarcity of literature evaluating these effects besides Sturm and Hattori (2015) which showed that zoning interventions do not deliver the expected results (Keeble et al., 2019). Thus, our paper supplements the evidence base regarding the lack of a relationship between a changing access to fast food and childhood and adolescent obesity which suggests that complementary interventions need to be considered.