The effect of ‘Traffic-Light’ nutritional labelling in carbonated soft drink purchases in Ecuador

Overweight and obesity have become global concerns in developed and developing countries due to their rise in recent years and their association with the prevalence of non-communicable diseases including diabetes, hypertension and cardiovascular diseases. In fact, it is estimated that roughly 39% of adults worldwide are overweight and 13% are obese. Ecuador is an example of a developing country concerned with the overweight and obesity problem, where it is estimated that 30% of children, 26% of teenagers and 63% of adults are either overweight or obese and where 1 in 4 deaths are attributed to chronic diseases. To address the overweight and obesity problem via the promotion of healthy eating habits, in 2013 the country approved technical regulation for the labelling of packed processed food products. The regulation included a mandatory traffic-light (TL) supplemental nutritional information labelling system to be displayed on the package of all processed foods for sale in the country. This new labelling system displays a traffic light panel for the product content of sugar, fat and salt in addition to the traditional nutrient declaration label. The objective of this paper was to evaluate the effect of the TL supplemental nutritional information on consumers’ buying behavior in Ecuador. More specifically, we concentrated on the purchasing behavior of carbonated soft drinks. For our analysis, we used monthly aggregated purchase data (total expenditures, quantities and average prices) of carbonated soft drinks from January 2013 to December 2015 obtained from Kantar World Panel—Ecuador. We estimated a non-linear Almost Ideal Demand System where we model the demand for high sugar and low sugar carbonated soft drinks. We found that the introduction of the traffic light supplemental nutrition labelling did not have the expected effect of reducing purchases of carbonated soft drinks during its first year of implementation, especially those high in sugar. Additionally, we found that lower income-status households tend to spend more on and consume more calories from CSD than households with higher socio-economic status. Finally, we identified that over time purchases of high sugar soft drinks decreased while purchases of low and no sugar soft drinks increased. Beyond our contribution of evaluating the effect of the traffic light on the purchases of carbonated soft drinks, we also estimated price and income elasticities of carbonated soft drinks which can be useful in the evaluation of fiscal policies.


Introduction
•Overweight and obesity are now problems in high and low-and middle-income countries (WHO 2016a). In Ecuador, a middle income country, it is estimated that 63% of adults, 26% of teenagers and 30% of children are either overweight or obese (Freire et al. 2014).
•To help reverse its overweight and obesity problem, Ecuador implemented in 2014 a 'Traffic-Light' (TL) supplemental nutritional labelling system. The objective of the policy is to help consumers make educated choices with regard to their consumption of sugar, fat and salt in processed food products.
•Ecuador is the first country to implement a 'Traffic-light' nutritional labelling policy at the national level. Therefore, little information is available with respect to the effectiveness of this type of policy at influencing the consumption habits of the population towards healthier food products.

Objective
•The objective of this research is to evaluate the impact of the 'Traffic-Light' supplemental nutritional labelling system in the consumption habits of carbonated soft drinks (CSD) of Ecuadorean consumers.
•We focus our analysis in carbonated soft drinks because of the high level of consumption of these products in the country. Moreover, CSD can be classified in two distinct high and low sugar groups which facilitates the analysis of potential substitution effects between the two groups as a result of the policy (Freire et al. 2014).

The Traffic Light Label
•Approved in November of 2013.
•Applies to all processed food products.
•Medium and large companies had until August 29th of 2014 to comply with the regulation and small companies until November 29th of 2014.
•Additional to the nutrient declaration/facts label.
•Assigns a traffic light color to the content of sugar, fat and salt.

Data
•Aggregated monthly food and drinks purchase data (volume in litters (L) and value in US$) from a Panel of 1,646 Ecuadorean households from January 2013 to December 2015 obtained from Kantar World Panel.
•The data set contains purchase information of 13 food groups and 17 drinks groups, including information on 23 brands of carbonated soft drinks.

Demand Model
•We estimated a unconditional non-linear Almost Ideal Demand System (Deaton and Muelbauer 1980). •The demand system consists of the following 5 equations: 1) Coca-Cola, 2) Dark colored high sugar Coca-Cola substitutes, 3) Low and non-sugar CSD, 4) Fruit flavored and all other CSD, and 5) a numeraire good that includes all other foods and drinks.
•A dummy variable was included to evaluate the effect of the 'Traffic-Light' label, in addition to dummy variables for socio-economic status, quarter and a time trend.

Discussion
•The Ecuadorian CSD market is clearly dominated by one brand. Out of the 1.67 L. per-capita per month that are consumed at home, Coca-Cola accounts for 58% of consumption.
•Own price elasticities suggest that Ecuadorean are sensitive to price changes in dark colored and low-and non-sugar CSD categories but not for Coca-Cola and all other CSD categories. Cross price elasticities suggest that Coca-Cola is a substitute of low-and non-sugar CSD and that low-and non-sugar and all other CSD are substitutes of dark colored CSD.
•The expenditure elasticities suggest that CSD are necessary goods.
•Relative to low income households, households in the highest socioeconomic status group consume less high sugar CSD and more lowand non-sugar CSD.
•During the period of observation, we observed a downward trend in the consumption of Coca-Cola (-0.27%/month) and all other CSD and an upward trend in the consumption of low-and non-sugar CSD (+1.30%/month).
•A joint test of effect of the 'Traffic-Light' label dummies in the demand system (F 4,108 =4.645) suggest that the labeling policy did have an effect in the consumption of CSD. The effects are small relative to the total consumption of drinks. Low and non-sugar CSD are estimated to have increased by about 0.0081 L/per-capita after the policy was implemented. Contrary to expectations, high sugar CSD are also estimated to have increased by 0.0238 L./per capita after the introduction of the policy; however, this estimated aggregate effect is not statistically different from zero.

Conclusions
•We find some evidence that the introduction of 'Traffic-Light' labelling policy had an effect in the consumption of CSD in Ecuador; however, relative to the overall level of CSD consumption the estimated effects are very small. Moreover, we do not find evidence of a reduction in the consumption of high sugar CSD which was the main policy objective.