In economic theory this is fine, but much of economic theory is like string theory: consistent and rational, but utterly unmeasurable--and thus useless for anyt

In economic theory this is fine, but much of economic theory is like string theory: consistent and rational, but utterly unmeasurable--and thus useless for anything outside of philosophy. Your definition of "productive" requires subdefinitions for both "value" and "unit of labor". In a Taylorist environment natural definitions for these terms might be apparent and obvious, but for the average modern day office worker these terms are nearly impossible to define, much less compare across workers performing heterogeneous tasks in a complex environment where the final product or service is not simply the sum of its parts. When confronted with exceedingly complex or impossible measurement problems, human intution has been and remains one of the best options we have. It may not be objectively accurate, especially at an individual level, but at least falling back to human intution when confronted with the unmeasurable allows correlations and changes over time to be observed. The real question is how honest the source data is for this survey, not how precise or rigorous it is.