I argue that much of what researchers have been measuring as technical inefficiency may be allocative inefficiency. When inputs are aggregated it is shown that any allocative inefficiency is manifested as technical inefficiency. Greater degrees of aggregation may produce lower measures of technical efficiency.