> Why is that so hard to believe? Economics isn't a game of make believe for me. I have professional standards to uphold, and this is probably where my actuar

> Why is that so hard to believe? Economics isn't a game of make believe for me. I have professional standards to uphold, and this is probably where my actuarial background makes me uncomfortable with the argument as posed. Supply, demand, and marginal productivity are great and useful components of an economic model, but I cannot take their relationships as they appear in a freshman micro/macro sequence to be axiomatic of an empirical system. I cannot infer that the resulting equilibrium must occur in the real world. In actuarial work, a model's assumptions are not observations. If I posit, what, a hazard function or assumptions about the independence of variables or a particular claim distribution, I can't treat the consequences of those assumptions as empirical facts about the financial portfolio; instead, I have to establish that the model adequately represents (and simplifies) the phenomenon I'm modeling. > Wages (in a free market) are determined by the Law of Supply & Demand. I'd apply the same standard here. This is a model specification, not an observed law of nature. We can express labor demand as a relationship to marginal productivity under certain assumptions. Neither of these alone establishes sufficiency for an increase in productivity to result in an increase in workers' wages, let alone that it must result in shorter working hours. My remark about necessity holds within the proposed mechanism though. If marginal productivity is the mechanism by which wages rise and worked hours fall, increased productivity is necessary before wages can increase and worked hours can decrease. > (You see this in non-union shops all the time.) "All the time" is a frequency claim. Where's the data? > The unions only accomplished what was inevitable. This is a counterfactual claim. In undertaking economic analysis of such a claim, I must specify the market structure, identify the constraints and frictions, and establish that the model's assumptions are sufficiently good approximations of the historical labor market. I must also try to estimate the causal effect of unions, which amounts, more or less, to constructing the counterfactual labor market in which they did not exist.