New study published in PNAS – “deforestation reduces fruit and vegetable consumption in rural Tanzania”

Using a novel modelling technique which combined two-way fixed effects modelling with covariate balancing generalised propensity score (CBGPS) weighting analyses, we were able to isolate deforestation as a causal factor in the decline in household dietary quality over the study period. Specifically, for the average household who experienced a loss of 171 hectares of forest, fruit and vegetable intake decreased by 14 grams per person per day. Given the very low intake of fruit and vegetables in these communities (130 grams per person per day on average, relative to the recommended intake of 400 grams per day), this represented a substantial proportion (11%) of daily intake. Exploration into the potential causal mechanisms of this reduction pointed towards the ‘direct consumption pathway’, indicating that deforestation reduced people’s ability to directly source wild fruits and vegetables from the forest. Given the high vitamin A content of these types of foods, deforestation was also responsible for a reduction in micronutrient adequacy of people’s diets. The results of this study have important implications for policy makers. Strategies to achieve food security very rarely attend to the role of forests, and often promote agricultural intensification and expansion which often comes at the expense of forest landscapes. This study, showing that deforestation caused a decline in dietary quality, highlights the need to better understand the importance of forests for nutrition in certain settings. Indeed, conserving forests potentially offer win-wins in terms of meeting both nutrition goals as well as conservation and environmental goals.  

This new study, published in PNAS, is the first of its kind to establish a causal link between deforestation and people’s dietary quality. While many studies in recent years have found positive associations between living in/near forests and people’s diets in low- and middle-income countries, none to date have established a causal relationship. We used food consumption data from the World Bank’s Living Standards Measurement Study (LSMS) which was collected as part of a panel study from 2008 to 2013 across Tanzania. We examined 1256 rural households that were part of the panel study and used the food consumption data to assess their dietary diversity, dietary adequacy (in terms of calorie and nutrient intakes), and consumption of nutritionally important food groups (i.e. fruits and vegetables). Over the five-year study period, there was a general decline in dietary quality across the sample households. We used the Hansen et al. (2013) tree cover dataset to calculate the extent of forest loss over the study period, measuring forest loss in a ten kilometre radius around clusters of households. The average household experienced a loss of 171 hectares of forest in the surrounding area over the five years.