Use of seemingly unrelated equations to assess changes in dietary behaviours during the UK's economic crisis

Abstract

BackgroundIn 2008, the world economy entered one of the most severe crises ever. Standards of living fell for large parts of the population in many countries. For example, UK families cut their food expenditure by 8·5%. The aim of this study was to assess whether the economic crisis was associated with any dietary change in the UK and other member countries of the Organisation for Economic Co-operation and Development (OECD). MethodsA seemingly unrelated regression (SUR) model with random effect estimators was used on a panel dataset that included data between 1999 and 2013 for 21 OECD countries. The dataset included statistics from Euromonitor-Passport, OECD, World Bank, and Eurostat. The three variables of interest were purchase of fruit and vegetables, snack-bars, and snacks. Negative growth in gross domestic product (ie, economic crisis) was the explanatory variable of interest. Foodstuff-specific inflation rates and other key socioeconomic variables (eg, population age) were also included. Compared with standard statistical approaches, SUR models provide more efficient performances when dependent variables can be substituted (eg, fruit and vegetables, snack-bars, and snacks). FindingsPeople living in countries affected by the economic crisis decreased their consumption of fruit and vegetables by 3·64 kg per head per year (95% CI −6·65 to −0·64, p=0·017) which is about one portion per week. At the same time, people slightly increased their consumption of snack-bars (22·6 g per head per year, 95% CI 1·5 to 43·8, p=0·036) and snacks (40·2, −50·5 to 130·9; p=0·385). In the UK, given the levels of consumption of these products, the above mentioned variations indicate that the economic crisis was associated with a 3·5% decrease in purchase of fruit and vegetables and an increased purchase of snack-bars (3·5%) and snacks (0·5%). InterpretationThe findings of this study support the hypothesis that tighter food budgets might have led consumers to switch to lower-priced (per calorie) and less healthy food. The economic crisis might have therefore contributed to a further growth in obesity and related chronic diseases. Previous studies suggest that small effects at the population level can mask substantial inequalities across different socioeconomic groups. This study pioneers, in the field of public health, the use of an advanced econometric approach applied to a market information database. FundingThis project is partly funded through regular contributions from OECD member countries.