It's pretty easy to find research coming to a different conclusion than the Card/Kreuger paper that most pro-minimum-wage-increase people cite: http://gregmankiw.blogspot.com/2006/06/sperling-on-minimum-w... http://cafehayek.com/2009/09/new-study-casts-doubt-on-card-k... http://rjwaldmann.blogspot.com/2009/12/in-which-i-disagree-w... But the best reasoning I've seen on the subject is Bryan Caplan's: "Suppose you have…
http://www.deakin.edu.au/buslaw/aef/workingpapers/papers/200...
Three things to observe from the graph on page 31 (adobe reader page 33):
* There is publication bias in favor of the hypothesis that minimum wage increases decrease employment.
* If one takes the liberty of selecting only a handful of studies to argue with, one can find "strong" support for either side of this argument.
* Neither the "true mean" (mean at large y) nor the "biased mean" (mean at small y) of the studies lies anywhere near the break-even point (x value of -1) where the proposal of raising the minimum wage begins to hurt poor people in aggregate (i.e. the sum total of money flowing to poor people is less than it was before the raise).
Also, observe that the graph on page 33 (adobe reader page 35) suggests a trend in elasticity over time which has a plausible argument to back it up. Signed elasticity should increase (minimum wage increases become more beneficial) when labor is overabundant. In this case, labor tends to be sold nearer to the minimum price laborers will sell at and further from the price at which employers will stop purchasing so that a minimum wage increase results in more "win" scenarios (wages are increased without being pushed over the "cliff" that results in negative productivity) than "loss" scenarios (where wages are pushed over the cliff). This is consistent with the narrative of American employees losing leverage over time.
If both the narrative and its theoretical+observed impact on elasticity are true, minimum wage increases are becoming more beneficial over time.