I've been reading Barry T. Hirsch's paper, "What Do Unions Do for Economic Performance," and while it's probably not the last word on the matter, some of his findings seem worth sharing. According to Hirsh, most economic studies show that, on average, the effect of unions on productivity is zero, although this varies from sector to sector. (Unions do good things for productivity, such as improving a firm's personnel policies, and bad things, such as restrictive hiring rules.) Unions also lower corporate profits, as expected, though they do not seem to have any effect on business failure. Some finer points:

1. The ability of unions to win benefits for its members depends, predictably enough, on the degree of competition facing both unions and workers. A company in a fairly competitive, largely nonunion industry can't just pass wage gains on to consumers in the form of higher prices. This explains why unions have taken firmer hold in less competitive settings, like the public sector.
   
   
2. Positive union effects on productivity are highest in sectors where competitive pressure exists—because management needs to respond to an increase in labor costs by organizing more efficiently, etc. But these are also the sectors in which there is the least scope for union organizing and wage gains. (Because of #1).
   
   
3. Says Hirsh: "This implies that steady-state union density in the U.S. private sector must remain small, absent a general union productivity advantage. By the same token, introduction of unions or the strengthening of other instruments for collective voice into highly competitive sectors of the U.S. economy is unlikely to have large downside risks for economy-wide performance."
   
   
4. Some theories have held that the union wage premium will provide an incentive for employers to upgrade the skill level of the work force, thus increasing productivity. But, Hirsh says, "empirical evidence for skill upgrading is weak." (One theory for this: employers may reason that if they were to train their workers, the union would just bargain for even higher wages next time around, thus restoring the premium, so the employers decide not to bother.)
   
   
5. "Union wage increases can be viewed as a tax on capital that lowers the net rate of return on investment," says Hirsch. Unionized firms tend reduce investment in physical and innovative capital, such as R&D, leading to slower growth in sales and employment.
   
   
6. There is some evidence that union companies in the U.S. have performed poorly relative to nonunion companies, which has led, to some extent, to a shift of production and employment away from the former sector to the latter. Since unions tend to lower profitability, this could explain the decline of unionized industries in the 1970s and 1980s. But Hirsch claims this is far from settled. (See #7).
   
   
7. Importantly, Hirsch emphasizes that many of these empirical findings don't always identify the causal relationships at play. Do unions really cause X effect on productivity? Sometimes it's hard to say. For instance, older plants tend to have lower productivity, but older plants also have, on average, higher union density. Plus, union status is often an endogenous, rather than random, variable. Nor is it clear, moreover that union effects on, say, sawmills can be generalized to union effects among industries of the future.

So what does all this mean? If Hirsch is right, union representation is likely to continue to decline in the private sector. Firms just won't happily follow, say, the CostCo model and expect that collective bargaining will enhance productivity or profits. So labor will have to rely, increasingly, on the government for support. Now Hirsch thinks we should be thinking of more flexible ways to give workers representation and participation; regulations that capture the gains and considerable benefits from giving employees a greater voice in the workplace, while limiting the economic losses that come with "excess" union rent-seeking. Maybe. I get more than a little bit skeptical, though, when he starts touting an alternative form of organizing called "conditional deregulation." Beware the Greeks bearing gifts and all...