Consumer food waste as household technical inefficiency across countries and time

Abstract Despite its implications for global food security and sustainability, quantitative analysis of consumer food waste across countries and over time is scarce. We examine food waste dynamics across 145 countries from 1995 to 2013 by modeling uneaten calories as a form of household‐level technical inefficiency, using an input distance function within a panel stochastic frontier framework. Unlike prior studies that rely on production functions, our input‐oriented specification reflects consumer behavior by treating body mass as a predetermined level and analyzing the calories required to sustain it—aligning the specification with consumer behavior. The panel structure allows decomposition of inefficiency into time‐invariant (persistent) and time‐varying (transient) components and clarifies development‐related factors associated with food waste. Our results show that persistent inefficiency is the dominant source of household food waste, with fruit and vegetable consumption associated with higher overall levels of transient inefficiency. We find evidence of an inverted‐U (Kuznets‐style) relationship, where food waste tends to rise at lower income levels but declines as countries become wealthier. Between 1995 and 2013, global food waste increased by an estimated 336 million metric tons, with middle‐income countries—particularly Brazil, Russia, India, and China—accounting for a large share of the rise. This novel application establishes a foundational, generalizable framework for longitudinal, indirect measurement of food waste across spatial scales and levels of aggregation. Additionally, by decomposing inefficiency into persistent and transient components, this framework enables dynamic computable general equilibrium and partial equilibrium simulations, integrated with life‐cycle assessment, that distinguish structural changes from short‐run shocks.