Coyoteblog has an interesting table regarding the role of hours worked vs. wages earned in total incomes.
This is relevant, I think, to both the minimum wage issue and the overtime pay issue.
The regressions I did using 2 year periods before, during, and after minimum wage hikes suggested a decline in the Employment to Population ratio of about 0.17% for each 1% increase in the ratio between the minimum wage and the average wage of production and non-supervisory workers. California and New York seem to be phasing it in, generally, at about $1 per year. At current wage levels, that will be about a 4% rise in the MW/AW ratio per year, so that, over time, each year should cut trend EPR by about 3/4%, eventually leading to disemployment of over 3% of the population. For states with nearly 60 million residents, this should add up to more than 2 million job losses over the next several years.
The trend in EPR is strong now, many of these displacements will be marginal workers moving out of the labor force, and most of it will be at the end of a complex series of capital allocation decisions over a period of years so that much of the dislocation will be unseen. There is a decent likelihood that some contraction in general economic growth will confound analysis. Maybe migration will mitigate some of the employment dislocation in the high cost areas. With all that said, I wonder if the natural experiment aspect of this will really amount to that much, if evidence of past episodes of dislocation isn't enough to be convincing. Although, I suspect that the fact that even the New York Times was against the minimum wage in the 1980s is because minimum wage levels were high enough at the time to tip the scales of public opinion, so maybe it is possible.