Farm Production Diversity and Household Dietary Diversity: Panel Data Evidence From Rural Households in Tanzania

Lemlem Teklegiorgis Habtemariam,Christoph Gornott,Stefan Sieber,Harry Hoffmann

doi:10.3389/fsufs.2021.612341

Abstract

Evidence on whether diversifying farm production leads to improved household dietary diversity and nutrition remains inconclusive. Existing studies analyzing the link between production diversity and dietary diversity are mainly based on cross-sectional methods, which could be biased by omitted confounding factors. Using two waves of a panel household survey of 900 rural households in Tanzania, this paper examines the link between production diversity and dietary diversity, while minimizing potential confounding effects. We estimate four regression models with two different production diversity measures and two panel estimation methods—fixed effect (FE) and random effect (RE). In three out of the four models, production diversity is significantly and positively associated with the dietary diversity measure of the food consumption score. The production diversity indicator is represented by the total crop and livestock species count, as well as by counting only crop species. The total crop and livestock species count shows a significant positive association with dietary diversity across estimation methods while the positive association with crop species count is not significant in the FE method. Our results suggest that the selection of appropriate production diversity indicators tailored to the specific circumstances of the local agricultural system is likely one key factor in identifying a robust relationship between production diversity and dietary diversity.

Highlights

Undernutrition remains critical in many low-income countries despite the many global initiatives targeting it (Global Nutrition Report, 2020)
The descriptive statistics of variables for the 2014 and 2016 waves of survey data are presented in Table 1, while Table 2 presents the descriptive statistics of dietary and production diversity variables disaggregated by the two study regions
Production diversity Age of household head Sex of household head (Male vs. female) Education of household head Household size Agricultural land Non-agricultural income Distance to road Share of food consumption from own production Household participated in fishing, hunting, collecting food products from public Household participated in food aid/nutrition programs Region (Morogoro vs. Dodoma) Constant N_obs Wald chi2

Summary

INTRODUCTION

Undernutrition remains critical in many low-income countries despite the many global initiatives targeting it (Global Nutrition Report, 2020). We use a second production diversity indicator based on the total count of crop and livestock species that the household cultivated and reared during the 12 months reference period Several studies use these indexes or variations of them to measure the production diversity of farms (Jones et al, 2014; Sibhatu et al, 2015; Demeke et al, 2017). Based on the existing literature (Jones et al, 2014; Sibhatu et al, 2015; Bellon et al, 2016; Demeke et al, 2017; Islam et al, 2018), we include a number of control variables in our analysis to account for socio-economic and market influences This includes the age, gender, and educational level of the head of the household, as well as household size, total agricultural land area, non-agricultural income, distance from main road, region, and share of own food consumption. Participation in nutrition and food aid programs can affect diet diversity and, we control for it in our model

AND DISCUSSION

CONCLUSION AND POLICY IMPLICATIONS

Findings

ETHICS STATEMENT