Skip to main content

Cornell University

Tata-Cornell Institute for Agriculture and Nutrition

Food Loss of Perishable Produce from Farm to Retail: Evidence from Tomato Supply Chains in South India



Reducing food loss and waste (FLW) may narrow gaps between fruit and vegetable production and recommended intake. However, FLW estimates are inconsistent due to varying estimation methods.


Using multiple estimation approaches, we examined the extent and determinants of FLW along tomato supply chains in South India, from farm to retail. We also explored tomato quality assessments.


We surveyed 75 farm households and 83 tomato traders in the Chittoor district, Andhra Pradesh, and 52 vegetable traders and 50 vegetable retailers in Hyderabad, Telangana, on harvest and market days. We calculated declared FLW values using participant-reported losses to estimate the preharvest quality FLW and quantitative FLW values at the farmer, vegetable-trader, and vegetable-retailer stages. We calculated the destination FLW based on counted crates diverted to loss destinations, using participant-reported destinations (animal feed, field discard), to estimate the postharvest FLW from farm to retail. We used pile sorting with farmers to explore on-farm quality assessments.


The average preharvest quality FLW was 13.9% of harvested tomatoes. From farm to retail, the quantitative FLW was greatest at the postharvest, farm level. Among all harvests, the median postharvest, farm-level FLW was 0.0% (IQR, 0.0%–7.9%) using the destination FLW approach (tomatoes diverted to nonfood uses) and 2.3% (IQR, 0.0%–12.5%) using the declared FLW approach (P < 0.05). Among harvests with a non-zero postharvest, farm-level FLW, the median FLW was 9.1% (IQR, 2.4%–16.7%) using the destination FLW approach (tomatoes diverted to nonfood uses) and 10.0% (IQR, 2.9%–16.7%) using the declared FLW approach. Harvesting during peak season was a determinant of postharvest, farm-level and preauction, market-level FLW values. Farmers prioritize color/ripeness attributes while harvesting and tomato size while grading.


Single-point estimates may obscure FLW patterns for perishable, indeterminate crops and depend on data collection and estimation methods. Reducing FLW of perishables requires the integration of quantitative and qualitative FLW estimation methods.

Read the article