Impacts of using biopesticides on tomato yield and self-consumption: Evidence from saline-prone areas in Bangladesh

Biopesticides derived from naturally generated compounds are urgently required for sustainable agriculture, particularly in salinity-prone regions like Bangladesh. Despite their potential, the effects of biopesticides on crop yield in real-world farming scenarios remain underexplored. We aimed to examine the impacts of biopesticide application on tomato yield and self-consumption in salinity-prone areas of Bangladesh. We employed propensity score matching to assess the impacts. Robustness checks were conducted using inverse probability-weighted regression adjustment and Rosenbaum bounds sensitivity analysis to ensure that estimators were unbiased. Our results indicated that biopesticide application reduced tomato yield by 28.83 kg per decimal of land, a corresponding decrease in marketed quantity by 392.67 kg per farmer, and also reduced the selling price. However, the farmers’ self-consumption increased by 4.97 kg, which was statistically significant. We also established that farmers who used both chemicals and biopesticides experienced a loss in tomato yield by 9.58 kg per decimal of land and a decrease in marketed amount by 62.44 kg per farmer, though statistically insignificant. However, self-consumption increased by 3.13 kg per household. The combined use of chemicals and biopesticides, despite expected to enhance pest control efficiency while reducing reliance on chemical pesticides, considerably affected self-consumption and tomato yield, highlighting both benefits and challenges in their adoption. While they promote self-consumption, yield reduction poses barriers to widespread use. • Biopesticides reduced tomato yield by 28.83 kg per decimal in salinity-prone areas. • Marketed tomato quantity decreased by 392.67 kg per farmer with biopesticide use. • Farmers' self-consumption increased significantly by 4.97 kg per household. • Combined use of chemicals and biopesticides reduced yield but raised self-consumption. • Sensitivity analysis confirms results are unaffected by hidden bias.