Smallholder agriculture in sub-Saharan Africa is commonly characterized by high levels of technical inefficiency. However, much of this characterization relies on self-reported input and production data, which are prone to systematic measurement error. We theoretically show that non-classical measurement error introduces multiple identification challenges and sources of bias in estimating smallholders’ technical inefficiency. We then empirically examine the implications of measurement error for the estimation of technical inefficiency using smallholder farm survey data from Ethiopia, Malawi, Nigeria, and Tanzania. We find that measurement error in agricultural input and production data leads to a substantial upward bias in technical inefficiency estimates (by up to 85 percent for some farmers). Our results suggest that existing estimates of technical efficiency in sub-Saharan Africa may be severe underestimates of smallholders’ actual efficiency and what is commonly attributed to farmer inefficiency may be an artifact of mismeasurement in agricultural data. Our results raise questions about the received wisdom on African smallholders’ production efficiency and prior estimates of the productivity of agricultural inputs. Improving the measurement of agricultural data can improve our understanding of smallholders’ production efficiencies and improve the targeting of productivity-enhancing technologies.