Most unions are found in manufacturing, but the new pro-union arguments emphasize the creation of unions in service industries. I can think of a few differences between manufacturing and services:
1. Labor costs are more important in service industries, so unions have less scope to raise wages. This is Megan McArdle’s argument.
2. There is more long-term fixed capital in manufacturing, and that gives unions greater scope to confiscate those quasi-rents. This is related to #1. (In a service industry, would the transfer be taken away from the return to brand name capital?)
3. On average there is more market power in manufacturing, which again gives unions greater room to raise wages in those sectors. In a perfectly competitive industries, extra wage demands will bankrupt the firm.
4. Many service sector firms face less foreign competition, but I believe they nonetheless face more competition overall. Lower fixed costs mean a more competitive industry, which brings us back to #3.
5. Jobs have shorter duration in service industries, which tightens the link between wages and the current state of the labor market. That also means a smaller role for unions.
6. We have a mental model of service sector companies such as Wal-Mart, which try to get by on the cheap in labor markets. It is harder to make the same claim about General Motors.
The bottom line: Except for #6, most of the effects imply that unions will be less effective in the service sector. You’ll all think of some mechanisms I didn’t, but my tentative conclusion is that unions bring both lower costs and lower benefits in service industries.
I might add I once belonged to a service sector union, in a supermarket as a teen. It was not a pretty picture. I paid high dues and received no apparent benefit, relative to the workers in non-unionized supermarkets. I even heard rumors of corruption.