One day's work of an untrained person with no tools represents one unit. For trained people, the number of days training need to be factored in (apportioned over the working life). For people working with tools, the cost of the tool (again in mandays) would need to be factored in (again apportioned over the working life of the tool).
Using such an approach to measure value would be horrendously complicated for all but the most simple of economies, but it is an interesting line of thinking.
Perhaps one weakness is to know how far back to count the mandays. For example, for trained workers should the mandays of the teacher be counted? And if so, how about the teacher's teacher? And the teacher's teacher's teacher? Etc.