Disentangling determinants of insecticide use to manage production, food security, and health risks in Cambodia and Vietnam: evidence from household surveys and risk-assessment experiments

Abstract

Background: Insect pest problems are among the main causes of crop yield losses in global agriculture. Insecticides protect households from food-security and income shocks, but can induce human health and environmental risks. Semi-subsistence farm households (SSFHs), which farm for both consumption and market, make decisions about crop management and output allocation to maximise food consumption, income, and health outcomes, depending on their risk preferences and other household and community characteristics. In this study, we aimed to disentangle the determinants of insecticide use by SSFHs and identify whether health consideration has had any effect on insecticide use. Methods: In this econometric analysis, we used field data collected from household surveys and risk-assessment experiments in 2014 in Cambodia and Vietnam to analyse insecticide use among more than 1000 SSFHs. Findings: We found that crops (except for rice) whose outputs were used to a greater degree for consumption were less likely to be sprayed with insecticides or were sprayed fewer times. Health-conscious households (as indicated by the use of modern-fuel cooking stoves and reported concern over food safety as a main reason for maintaining home gardens) consistently refrained from spraying, but this tendency diminished as output allocation shifted toward commercial use, suggesting a possible moral-hazard phenomenon. Farmers were more likely to apply insecticides to crops of high food security or dietary importance, such as rice, although the difference between fresh produce and grain produce in risk of insecticide residue exposure might also have contributed to the difference in insecticide use between rice and non-rice crops. The two samples from Cambodia and Vietnam had similarities regarding the deterring effect of health consideration and differences in other factors affecting insecticide use, such as risk preference, landholding size, household head's education level, and land tenure security. Interpretation: We identified both push and deterring factors for use of insecticides at household, plot, and crop levels. These factors varied between the samples from Cambodia and Vietnam, indicating the importance of addressing specific contexts. Although the extent of crop output used for consumption was negatively correlated with insecticide use, caution needs to be taken when drawing causal interpretation from this correlation because of the complexity in SSFH decision making. Improved understanding of these issues is essential to inform policies that address multiple sustainable development objectives cutting across poverty reduction, food security, health, and the environment. Funding: CGIAR Fund Donors.