« The demand for Trudie | Main | Africa »
What do unions do for economic performance?
I've spent the last few days perusing the economics of unions. I've unearthed Barry T. Hirsch's useful and serious piece, which looks at whether the Freeman-Medoff pro-union work from the 1980s has held up. Here are a few select quotations from the article:
...while it is true that much of the negative relationship between unions and growth is not causal, slower growth is partly attributable to the lower profits and investment resulting from union rent seeking.
Empirical evidence on unions and productivity was rather sketchy in 1984; it remains less than clear-cut today.
...union firms reduce investment in physical and innovative capital, leading to slower growth in sales and employment and shrinkage of the union sector.
...empirical evidence for skill upgrading [through union participation] is weak.
The thesis that unions substantially increase productivity has not held up well. Subsequent studies are as likely to find negative as positive union effects on productivity.
...employment declines have been concentrated in the unionized sectors of the economy.
...the empirical evidence finds that U.S. unions are associated with slower employment growth...
I am genuinely puzzled why the highly intelligent segment of the left-wing blogosphere is so attached to the legal encouragement of labor unions.
Posted by Tyler Cowen on March 7, 2007 at 05:50 AM in Economics | Permalink
Comments
mostly because unions prevent things like dangerous working conditions, below poverty level wages, that sort of thing. these may not help economic performance - but they are necessary for quality of life.
similarly, from the perspective of a worker, higher productivity is not necessarily a good thing, if it means they are being forced to work unpaid overtime, have no home life, use drugs to keep up, etc.
Posted by: barry at Mar 7, 2007 7:44:00 AM
barry,
you're spot on, but for the economist, there is no "worker's perspective"; entertaining that would be just too leftish....
Posted by: Rob at Mar 7, 2007 8:19:31 AM
The answer couldn't, of course, have anything to do with the fact that lefties hang around with union bosses who would rather see the whole country go down the tubes than lose their positions of power. Nah. See Peter Sellers in "I'm All Right, Jack" from fifty years ago. What number will the percentage of unionization have to drop to before people realize that unions are the enemy of the working man?
Posted by: Robert Speirs at Mar 7, 2007 8:24:05 AM
but also, in answer to Tyler's question, the left may be attached to unions because it's the easiest, most visible real manifestation of left-wing principles, despite the relative failure of the union movement (in the US); the "highly intelligent segment of the left-wing blogosphere" probably doesn't have a majority of members whose origins are working-class (or in many of the other identity groups supported by the movement). But they still champion the underprivileged/"oppressed", exhibiting this to themselves by their writings, to their target audience through support of unions - why not political parties? Well, the political mainstream (in the US) is rather right-leaning...
Posted by: Rob at Mar 7, 2007 8:30:33 AM
Matthew Yglesias is way too young to remember what union power did to Britain in the 1970s.
Posted by: Steve Sailer at Mar 7, 2007 8:39:07 AM
I'm curious what the limiting principle is -- in both directions. I'll go ask "the highly intelligent segment of the left-wing blogosphere" what powers unions shouldn't have. But over here, what organizing rights, if any, should there be? Should attempted union organization be recriminalized? Should we allow specific performance of employment contracts? Should there be any antitrust restrictions on employer conduct?
Posted by: James Grimmelmann at Mar 7, 2007 8:53:04 AM
The ideology behind government granted union power is the same ideology behind the entire leftwing (socialist / etc) history. Even if most on the left in America today have given up the extreme version or are unaware of the relationship.
James: I doubt anyone thinks people should not be able to unionize, but that is quite a separate question from whether government should grant them special powers. My take is that unions are a cartel just as firms form cartels; both should be legal, there should be no antitrust laws at all, on employer or employee.
Posted by: liberty at Mar 7, 2007 9:07:26 AM
Intelligence has nothing to do with their support; emotion everything.
Posted by: Russell Nelson at Mar 7, 2007 9:11:23 AM
James Grimmelmann sensibly asks:
I'm curious what the limiting principle is ... [W]hat organizing rights, if any, should there be? Should attempted union organization be recriminalized?
liberty responds:
James: I doubt anyone thinks people should not be able to unionize ...
Well, why not?
Posted by: alkali at Mar 7, 2007 9:37:54 AM
Capital is allowed to organize into state-chartered companies; why should labor not also be allowed to organize? It would seem unfair to put higher restrictions on labor's ability to organize than on a company's.
Has there been analysis that treats a union as a firm that sells labor?
Posted by: sema at Mar 7, 2007 9:58:41 AM
I think it has to do with conceptions of power and who should have it. Libertarians and right wingers fear the guy with the gun, left wingers fear the guy with the money.
Progressives aren't worried about efficiency they are worried about democracy. As if the purpose of finding a job, or creating one for that matter, was to enter into another arena in which one could cast a vote. It is an issue that strikes at the very core of positive vs. negative rights.
Posted by: steveintheknow at Mar 7, 2007 10:06:35 AM
the smarter parts of the left-wing blogosphere realise how catastrophic continued union decline would be for the Democrats' Get Out The Vote efforts.
Posted by: Chris at Mar 7, 2007 10:07:17 AM
Tyler
In response to your last question, one might then also wonder why the intelligent right is so enamored of the legislative supports (read defense) of large corporate interests. I suspect it is for similar reasons.
Posted by: Martin at Mar 7, 2007 10:09:35 AM
Hmm...I wonder exactly what effect unions have on productivity, quality, and customer service?
Posted by: monkeypower at Mar 7, 2007 10:58:43 AM
The other, perhaps bigger, problem with unions is that they tend to be so illiberal and jingoistic. People who cross pickets lines are scabs and rats. People who want to trade with foreigners are traitors. The foreigners themselves are calculating predators with funny names and dark skin, who want to steal food off of the American Working Man's plate.
Posted by: David J. Balan at Mar 7, 2007 11:31:38 AM
mostly because unions prevent things like dangerous working conditions, below poverty level wages, that sort of thing.
Actually, they haven't done these things since the 20s and 30s, that's why the decline of unions has not actually increased measures of inequality that check the disparity between the poorest and richest deciles. Unions since the 1950s have only raised some portions of the lower middle class into the higher middle class. Not a bad deal for those in the unions, but hardly a magnificent public service. It would be nice if unions could actually start to help those at the bottom of the income distribution who could use it, but then you have major problems with it being too easy to hire outside the union so the bargaining power isn't very strong.
Posted by: JAC at Mar 7, 2007 11:34:22 AM
DiNardo and Lee QJE 2004 uses close union elections(regression discontinuity) to look at the effects of unionization in the manufacturing section from 1984 onwards. The effect of unions on everything (including wages and productivity) is a big fat, precisely estimated 0.
I think if you look at the service sector, e.g. SEIU and HERE unions over the same period, you'll find substantial gains to unionization. Hasn't been done, AFAICT.
Posted by: idlehands at Mar 7, 2007 11:34:53 AM
Work generates an economic surplus. Some of that goes to labour and some to capital. Unions push for more of the surplus to go to them, whereas employers try to retain it. This need not result in increased prices.
Posted by: Peter Clay at Mar 7, 2007 11:42:40 AM
There should be no government granted monopolies, either to corporations or to labor. AT&T broke up and everyone benefitted. Break up the UAW and the same will happen with the automotive industry.
Posted by: Matt at Mar 7, 2007 11:51:37 AM
yep thats right monkey - the only reason I am not a teacher is that they can't fire those lameos that are teachers now. It has nothing to do with that the pay is 1/3 of mine and the working conditions stink. If it weren't for those bad teachers that they should fire, I would be a tetacer right now as those bad teachers are taking my place. Somehow, the fact that there isn't really a huge amount of talented people who want to be teachers doesn't enter into the thinking.
And comparing the UK in the 60s and 70s is nothing but a strawman. Unions aren't an either or proposition, they are a matter of degree. Its totally acceptable to determine a preferred rate of unionization even better would be a preferred band of union participation.
Now, on to Tylers posit - Unions are bad for the economy and productivity. Yes. Yes they are. However, the economy is a subset of our society. Every day, each one of us does things that are not important economically or actively hurt our economic standing. Additionally, we as a society make decisions that effect others every day. At this point, its pretty clear, its not a matter of either/or. its a matter of degree. I want a slightly higher level of union participation than our current levels, however, I want a level that is much lower than UK in the 70s.
In direct economic terms, I am willing to trade a small amount of growth for slightly greater worker happyness, given current U.S. levels of growth and worker happyness. I believe that most US workers would agree with me.
And then, I decided to read the actual paper that Tyler recommends. Its actually quite good. The conclusions they draw are not as clear as what Tyler presents them. In fact, one of the major conclusions of the paper can be be interpreted as "Unions don't seem to affect productivity. What they impact is the managment/labor relationship, and they do this in a large and positive manner. This in turn raises productivity for shops that use modern labor/managment relations" This, I think is a more accurate description of what happens than simply stating that unions don't increase productivity. It's true, they don't, but they make it more likely that the company will be well managed.
One thing that jumps out at me about this paper is that it doesn't even begin to quantify the economy wide impact of unions. This is the most important data point. There needs to be a model where we can say, well increasing union membership by this much is going to cost us x% of growth. Even a bab model is far better than what we have now. Now, nothing from AEI is going to cut it for me - lets be clear, we want somebody with a small amount of credibility at least. Until we have an idea of this, we're all just jawing over drinks at the corner bar. Or even more enjoyably, over fine food and good wine. :)
Posted by: mickslam at Mar 7, 2007 12:02:23 PM
Rob,
I think you have it backwards. Paul Fussell notes in his classic Class the university professors, particularly at state schools, tend to have lower-middle class backgrounds. Their attachment Unions developed around the dinner table, not in the classroom. It's certainly true in my family: my pro-union College Professor father was the son of a unionized Auto worker.
Posted by: tylerh at Mar 7, 2007 1:07:43 PM
Try this for starters:
http://www.epinet.org/content.cfm/book_unions1991
Posted by: Miracle Max at Mar 7, 2007 1:08:41 PM
"James: I doubt anyone thinks people should not be able to unionize ...
Well, why not?"
-alkali
Because in a free market both firms and workers can form coalitions. They can do so because it is their right to do so, and those outside the coalition can see the profit and underbid in order to gain some of it. This is the way that competition and innovation ultimately are induced.
Posted by: liberty at Mar 7, 2007 1:44:31 PM
If the measured effect of unions on productivity is zero, I'd say that's a point in their favor, because the unmeasured effects are almost certainly welfare-enhancing. If you measured input in units of worker disutility (or "effort", if we can understand "effort" to mean something unpleasant) rather than hours worked, the input would surely be less when there is a union.
Posted by: knzn at Mar 7, 2007 2:53:28 PM
liberty said:
My take is that unions are a cartel just as firms form cartels; both should be legal, there should be no antitrust laws at all, on employer or employee.
I think this is reversed. I feel that the problems with unions come largely from their monopoly status, and competing unions should be allowed.
Companies competing for your money are careful to keep your interests in mind and pursue them when offering products and services. That's not true of unions, cafeterias in big campuses, or any other effective monpoly.
That's why we only see unions in the worst of workplaces. Because they often don't act in the interests of those they represent.
Posted by: Jon Kay at Mar 7, 2007 3:51:05 PM
