I have come to the conclusion that the labor market does not price labor well. Each individual is required to sell their labor in large batches, typically, one agrees to sell a set amount of labor at set times for a single price per hour, when the amount and desirability of labor varies tremendously hour by hour.
The first example is what the potential impact of a company having no attendance policy would be, i.e., there is no penalty other than not getting paid (and not acquiring the associated experience), for not coming to work. Thus, whenever the cost of working (in foregone other activities, such as sleep), actually exceeds the rewards of working, that employee would simply not work. My prediction is that, in such a situation, the rate of absenteeism would skyrocket, and then settle towards a mean number of hours which would be relatively constant, as people discovered how much labor they needed to sell.
The second example of pricing labor would be allowing people to bid for shift premiums. Say the company divided the week up into 4-hour time slots, and people were able to bid for how much of a premium they would need to work those hours. The bottom bidders for every slot would then receive those hours.
Where I work, the rewards for taking unusual shifts is very small. Whenever a position opens on the day shift, it is normally filled from another shift within a day. If a position opens on a regular 8-hour shift, it is normally filled within a week from another shift. We can be quite confident, therefore, that the premiums for working 12-hour, night shift are considerably less than they should be, and the workers on 8-hour day shifts are overpayed. Even within shifts, certain days and hours are more valulable than others. I, for example, am overpayed working Thursdays, Fridays, and Saturdays, since I am willing to perform my work on those days for less than I'm being paid. On Sundays, however, I am definitely underpaid, and if left to myself, would never work a Sunday again without a tremendous increase in shift premium.
My prediciton for how this would work out, is that those people who work 8-hour day shifts would bid themselves negative in order to retain their current time slots. The premiums for night and weekend work would corresponding jump, until the premium for every period of time matched the personal cost to the individual for working those hours.