« How to read Thomas Pynchon | Main | Prospect theory »
Unions really really don't matter that much these days...
Idlehands points to this paper (QJE 2004) by John DiNardo and David Lee. Neither author is a crazy right-winger, let's hear their message:
Economic impacts of unionization on employers are difficult to estimate in the absence of large, representative data on establishments with union status information. Estimates are also confounded by selection bias, because unions could organize at highly profitable enterprises that are more likely to grow and pay higher wages. Using multiple establishment-level data sets that represent establishments that faced organizing drives in the United States during 1984-1999, this paper uses a regression discontinuity design to estimate the impact of unionization on business survival, employment, output, productivity, and wages. Essentially, outcomes for employers where unions barely won the election (e. g., by one vote) are compared with those where the unions barely lost. The analysis finds small impacts on all outcomes that we examine; estimates for wages are close to zero. The evidence suggests that-at least in recent decades, the legal mandate that requires the employer to bargain with a certified union has had little economic impact on employers, because unions have been somewhat unsuccessful at securing significant wage gains.
Keep in mind what this means. Once we control for endogeneity in where unions are formed, there may not be a union wage premium at all. (A few posts ago I was telling you it was 10 to 20 percent, learn something new every day, etc.) I learned also that when we look for the wage premium in establishment-level data, rather than household data, it usually isn't there. And that's without considering the contribution and method of the authors.
I would like the highly intelligent left-wing part of the blogosphere to respond to this paper and to the Hirsch piece. Here is an NBER version, here are other copies. By the standards of labor economics, it does not suffice to note that the 1950s had both a more equal income distribution and more unions, or to call Western Europe a kinder, gentler place. Those citations don't sort out cause and effect, and in fact we do have more advanced ways of scrutinizing the data.
It is fair to say that these papers do not support the "right-wing scaremongering" scenarios about unions. So a Kevin Drum might claim: "well, it can't hurt to try more unions." That still represents a significant downgrading of the original vision. Unions are an emotional issue for the left, much as free trade and the fall of communism are so for the right. Would it not be meaningful and rallying for the left to have the battle over collective bargaining once again? But I am telling you all, there simply isn't that much there.
By the way, here is Bloggingheads.tv with Megan and Matt on unions, can you imagine that?
Posted by Tyler Cowen on March 8, 2007 at 07:21 AM in Economics | Permalink
Comments
If union have no significant impact it is just as illogical for the right to oppose them as it is for the left to favor them.
Posted by: spencer at Mar 8, 2007 9:01:25 AM
If union have no significant impact it is just as illogical for the right to oppose them as it is for the left to favor them.
You are forgetting the crucial distinction between left and right: the collective vs. the individual. Folks on the right bristle at the very notion of tying our compensation to everyone else we work with.
Posted by: Christina at Mar 8, 2007 9:21:06 AM
This would seem to imply that workers are all idiots for voting for union representation, employers are idiots for working against union representation and economists are the only bright people to see the reality of the situation. I suspect there might be a flaw in the reasoning.
Wasn't there an article on here about how non-skilled wage earners are earning substantially similar wages to technically skilled workers? There must be a reason for that massive historical change.
Posted by: Steve Roberts at Mar 8, 2007 9:26:26 AM
"well, it can't hurt to try more unions."
This is the part that I find really interesting - aside form new and better data in general. The language seems to be changing dramatically. Down from a Marxian like reduction where the labor market essentially creates slaves in the absence of unionization, to something along the lines of "well, I'll concede a little efficiency for workers being a little happier".
Perhaps it would be wise, in response, for fans of choice to temper the ant-union language a bit to avoid the appearance of hyperbole – or actual presence of. The arguments for unions seem to be dying; opposition hysteria may have the opposite affect.
Posted by: steveintheknow at Mar 8, 2007 9:40:43 AM
I couldn't get to the full article. However, it seems from the abstract to focus on wages. Unions bargain for much more than just wages and often will take a wage hit to get some other benefit such as preserving retirement or health care. It would be interesting to determine if the total work compensation, including benefits, was helped or hindered by the presence or absense of unions.
IMHO, unions have a societal impact that other institutions don't seem to have. They tend to act as curbs on excesses of corporations. Many of the benefits now accepted as standard were originally supported and proposed as union ideals. Viewing unions narrowly as mere negotiating mechanisms is as false as is viewing environmental organizations merely advocates of animals. By forcing the debate, both institutions change the nature of the debate. Policy decisions and the societal environment as a whole result from that changed debate.
Posted by: Steve Popkes at Mar 8, 2007 9:47:09 AM
This would seem to imply that workers are all idiots for voting for union representation, employers are idiots for working against union representation and economists are the only bright people to see the reality of the situation. I suspect there might be a flaw in the reasoning.
Me too. Especially since economists love to assume that people behave in economically rational ways.
Posted by: bernard Yomtov at Mar 8, 2007 11:36:21 AM
Can't it hurt? Unions do charge dues, you know. They can also cost efficiency in the long run, expansion, future raises, etc (not sure how long term the study was).
On the other hand, it is true that unions do bargain for more than just wages, including shorter hours, easier work, nicer working conditions, job security, etc etc.
Posted by: liberty at Mar 8, 2007 12:11:43 PM
well, it can't hurt to try more unions
Well, by the same token it couldn't hurt to try fewer unions either.
Then Steve Popkes said:
IMHO, unions ... tend to act as curbs on excesses of corporations
But who acts as a curb on the excesses of the unions? In fact, corporations have competition from other firms that act as a (perhaps insufficient) curb on their excesses. But unions, by their definition in most western countries, have none. The problem with unions is just this - their universality. If they were more like businesses, where multiple unions formed in the same industry and competed with each other on price and services, and where those who didn't wish to participate could just as easily work from outside a union, then I think most people on the Right would have no problems with them. The fact that they require an employer to hire 1 and only 1 federation of workers and that they force all workers to join - regardless of whether they wanted to in the first place - is what gets most Conservatives upset. People joining together voluntarily to increase their bargaining power should be a choice that people are free to make. But forcing people to join a union against their will, and forcing an employer to spend his money on that union is something quite different. So I say again, it couldn't hurt to try fewer unions either.
Posted by: Josh at Mar 8, 2007 12:31:04 PM
Based just on the excerpt here, I don't think this says much about whether unions raise wages overall. I'd suspect that employers make significant concessions in establishments threatened with unionization. The employer still gains from blocking the union, even if he offers the same wage the union would push for, because he maintains future flexibility and control.
Bottom line--workers who nearly vote to unionize have quite a bit of barganing power (and conversely, a barely-successful election suggests a weak union that can't threaten much of a strike). I'd still guess that making unionization easier in general would make workers better off.
Posted by: Ben at Mar 8, 2007 12:40:08 PM
Economists can wallow in so much data and still have no concept of reality.
Although there is not likely a 100% cause-and-effect, trace the decline of
unions versus the decline of solvent defined benefit plans.
Academic economists have the most effective trade guild (union) of all,
with tenure as the ultimate leverage against employers.
Not too many economists advocating more competition in academia. Tenure
is such a nice perk.
Posted by: save_the_rustbelt at Mar 8, 2007 12:45:34 PM
This would seem to imply that workers are all idiots for voting for union representation, employers are idiots for working against union representation and economists are the only bright people to see the reality of the situation. I suspect there might be a flaw in the reasoning.
How so? Just because something has no effect on wages (if the study is correct) doesn't mean it wouldn't have an effect on other variables employers might care about (e.g., flexibility) or that employees might care about (e.g. protection of longer-tenured employees).
Not too many economists advocating more competition in academia. Tenure is such a nice perk.
Does Steve Levitt count? http://www.freakonomics.com/blog/2007/03/03/lets-just-get-rid-of-tenure/
Anyway, it's a little strange to use tenure as an argument on the pro-union side, since I don't think Prof. Cowen is a member of the "American Economists Union" or anything...
Posted by: Steven Vickers at Mar 8, 2007 1:00:52 PM
FYI - Freakonomics author Steve Levitt advocates getting rid of tenure:
http://www.freakonomics.com/blog/2007/03/03/lets-just-get-rid-of-tenure/
Posted by: m g at Mar 8, 2007 1:05:36 PM
This comment is VERY off topic. My apologies. I have some things to add about union organisation but am too busy for it now. I am a BIG fan of marginal revolution but am finding the IQ wank too much to take without comment.
I am really sick and tired of the whole "highly intellegent left", thing. I don't think IQ is everything though "I value it in my friends" bullshit.
I, by standard tests, have an above average IQ (not by a great deal) but find it to be more irrelevant each day. I know this is an economics blog and i am no expert in the field. I have an undergrad major in econ and am working towards an undegrad major in psych + sociology (to futher my inerest in econ by the way). I am still amazed at the interest in IQ. Given the authors of this blog both seem to have a serious and lifelong interest in 'the Arts', I am surprised how much weight is put on the subject. Clearly some of it is hornblowing (and they both deserve a bit of room there) but lets call it what it is.
IQ tests are NOT a reasonable test of potential contribution to almost anything worthwhile in life.
Again, sorry to bust in on the topic. ;^>
Posted by: Clinton Mac at Mar 8, 2007 1:17:38 PM
trace the decline of
unions versus the decline of solvent defined benefit plans.
Insolvent defined benefit plans are usually a consequence of decisions made decades ago, when unions were strong.
Posted by: Douglas Knight at Mar 8, 2007 1:17:44 PM
[quote]This would seem to imply that workers are all idiots for voting for union representation, [/quote] If the median wage is less than the average wage and unions bring everyone to the average then it would make sense for the majority of workers to be pro-union.
Also the authors are looking at situations where the union vote was close, and some of them failed. It's hard to say unions must be good for workers because some workers vote for unions, when half of the workers looked at vote against unionizing.
Posted by: badger at Mar 8, 2007 1:18:48 PM
In general, when a union vote is close, management has taken steps to discourage workers from voting for a union. Those steps usually include increasing wages as a short term cost because it is an incentive. I would regard a dataset bases on this criteria to be inherently biased.
Posted by: Mike at Mar 8, 2007 1:28:22 PM
Although there is not likely a 100% cause-and-effect, trace the decline of
unions versus the decline of solvent defined benefit plans.
I'd think it's probably closer to 0%, in that they're both the effect, and the cause is a more volatile and dynamic work environment. The more people change jobs, the less likely they'll join a union and the less likely they'll accrue a vast benefits package.
Posted by: Josh at Mar 8, 2007 1:44:49 PM
I can see three possible explanations: (1) The study doesn't measure the change in productivity. If two identical firms are compared after one goes union and the other does not, it is not enough to compare the salaries after the change. Suppose the union firm now needs 20% more workers to produce the same output (perhaps because of work rules, intended to increase employment), then clearly there is a reason for the firm to have opposed unionization and the employees to have supported it. (2) It may be that case that following a union defeat, firms becomes more conscious about keeping the union out in the future, and therefore becomes more generous in its wage package. (3) These elections may have been close because the firms provided union scale benefits before the election.
Posted by: Hari at Mar 8, 2007 1:48:17 PM
Having just read this paper in my labor economics class, the message should actually be that labor unions formed between 1984 and 1999 are very weak. Older data shows a pretty wide range of things depending on the study (does employment go down is the big question, wages do seem to go up back in the day). I don't think that showing the Reagan and Clinton weren't good for unions is a big accomplishment of a study though.
Posted by: Noah at Mar 8, 2007 3:33:51 PM
For reasons why the Left might promote unions:
1. There are claims that they help left-wing parties get elected. See the "power resources" school exemplified by Walter Korpi http://muse.jhu.edu/journals/world_politics/v058/58.2korpi.pdf. From a different political persuasion, Troy's "The Twilight of the Old Unionism" also highlights how much money and in-kind resource flows from unions to Democrats in the USA.
2. Unions appear to compress the (bottom half of the) wage distribution, potentially yielding more income equality (when combined with other agency): http://users.ox.ac.uk/~polf0050/Rueda%20BJPS.pdf
Posted by: Tim Hicks at Mar 8, 2007 3:39:22 PM
Note, though, that the election outcome, including the margin of victory/loss, is engogenous. Workers at firms at which unions are more likely to have an impact on wages and benefits are also more likely to vote for a union. In relatively evenly divided votes, my prior is that workers are less certain (and, in a rational expectations world, correctly less certain) that unions can provide benefits.
I would think that comparing unions that win close certification elections with unions that win easily might come closer to the issue. Controlling, of course, for the other factors we are likely to think matter (capital per worker, industry/market structure--yes, I'd expect unions to be able to caputre some monopoly profits, worker characteristics, and so on).
Posted by: Donald A. Coffin at Mar 8, 2007 4:02:33 PM
My experience has been that the agency costs for unions are very high - that actions taken by union management often are at odds with the best interests of the membership. This creates a big friction in the economy with a cost that exceeds any benefits the union may provide.
The impression that union organizers often give is that they are rent seeking. They want a union so that they can shirk or channel funds towards their political party, or otherwise benefit from the union presence. Altruistic statements to the contrary are unconvincing.
The legal channels that exist today to protect workers' rights, and the fluidity of the labor market, make unions unattractive in this era.
Posted by: bastiat at Mar 8, 2007 4:12:04 PM
Abstract:
"estimate the impact of unionization on business survival, employment, output, productivity, and wages. Essentially, outcomes for employers where unions barely won the election (e. g., by one vote) are compared with those where the unions barely lost. The analysis finds small impacts on all outcomes that we examine; estimates for wages are close to zero."
So they looked at cases where unions were VERY close to organizing - i.e., those cases where if conditions or wages at the firm got even marginally worse, the union would win the next election. In those situations, is it any surprise that the firm would increase wages after an election to match union promises, trying to demonstrate that management is sufficiently understanding without the need for any 'outside agitators'?
Moreover, few union drives are successful solely because of wage inadequacies. Its usually more a matter of working conditions, benefits, and fairness. But the authors didn't look at things like seniority, job security, safety records, scheduling and overtime policies, health coverage eligibility, grievance or harassment reports, or anything else that reflects the real conditions of most workers. Just dollar-per-hour compensation, as if every job was equivalent to running an economic database through an underdeveloped model on a bucolic tree-lined university campus.
Anyways, what they really found was not that unions aren't successful, but rather that unions are so effective that even the THREAT of a union pushes employers to make substantial changes in wages and working conditions.
Posted by: EthanJ at Mar 8, 2007 4:16:25 PM
Has anyone seen any data on the turnover rate for employees in union and non union firms. It seems reasonable that firms with unions would have lower turnover and since new employees are costly for a frim, the extra costs that a union imposes might be canceled. I have read that the low turnover at COSTCO is part of the reason they can pay much higher salaries than Walmart and still make a profit.
Posted by: joan at Mar 8, 2007 4:21:47 PM
George Mason economist blog at several places. But the one constant is that you can be sure on any issue they will tell you why anything issue or policy that increase the returns to labor or the share of gdp tht goes to labor is a bad thing and will harm the working poor. If you learned your economics from these blogs you would conclude that the best way to increase the economic wellbeing of the working poor is to increase the share of the pie that goes to capital to 100%.
Now, I recognize that this is an extreme statement. But I have seen nothing to contradict at the various blog where the George Mason economist blog.
So I suggest that you write up as one of the 50 response from readers a single thing you would find that increasing the share of the pie, or returns to labor is a good thing and actually makes the working poor better off.
Posted by: spencer at Mar 8, 2007 6:10:27 PM
