Kevin Drum flags a new study on the impact of unions:
Among men, if you account only for the effect of individual membership in unions, it would be about a fifth lower, which agrees pretty well with previous estimates. But if you also account for the effect of unions on surrounding nonunion employers (who often raised wages to compete with union employers and to avert the threat of unionization in their own workplace), the effect is larger: Unionization at 1973 levels would decrease income inequality by a full third. You can see this in the chart below. For intragroup differences (which account for nearly the entire effect of unionization) the top line shows the actual rise of income inequality since 1973, while the red line is a prediction of what it would look like if union density were still at 1973 levels.
The Dish 