In today's WaPost, political scientists Jacob S. Hacker and Paul Pierson argue that not only are unions good for their members but also that unions promote a strong middle class. Their main argument is that unions are a strong political force that lobbies for pro-middle-class policies. Dealing with that argument, which I think has some correct points and some incorrect, would take me too far afield from their more-narrowly-economic argument about the direct impact of unions on wages, which is what concerns me here. Here is that argument:
The reason [that unions reduce income inequality] isn't just that unions defend their members. They also create changes in social norms, such as pressures for nonunion employers to match union gains.
What Hacker and Pierson have done here is point to one effect of unions--labor economists call it the "threat effect"--but left out another that is stronger. But to see why, we need to back up and think about what unions have been in the United States since the 1930s.