The difference made by unions can be cumulative too -- as in the more unionized the workforce the EVEN better everyone does. Unfortunately the only example I can offer is a bad one, given the current decline of American labor muscle: the presumably super-strong union of the NYC police and fire departments (sounds like high testosterone, militant type employees to me) have not had a raise above inflation since 1974 (still at the same $65,000/hr) though average income has grown 70% since. This in a city which features the only island in history where wealth is a plateau not a pinnacle.

This could only happen in an -- overall -- workforce made up of the bargaining power clueless -- who have no idea how high their pay could have climbed. More on that in the next post.

Almost always and almost at all times our top economic progressives cannot identify the solution to all our income deficits as RE-EMPOWERING American labor -- most effectively by instituting sector-wide labor agreements and by returning to LBJ's minimum wage ($10/hr) with a significant premium to reflect the doubling of average income since (at least $12.50/hr?). This takes two steps:

First, American labor -- blindered by oceans and an economically retarded press needs to be woken up to its plight: how about telling everyone that what was once the minimum wage is now the 25 percentile wage (double the average income later) -- and that even if all medical care became paid by the gov that could still leave 25% of American families living below the minimum needs line w/o food stamps and other helps -- UP from 15% in the LBJ era (before the war on poverty shifted more income around but got out shifted by the wholesale loss of labor's bargaining power).

Second, we then educate American workers to the easy way to get the WalMarts of the world back under proper labor market checks and balances, the easy way that is used on the other sides of both oceans (Germany, Norway, Indonesia) and even on the other side of the equator (Argentina) and even on this side of the equator (French Canada): SECTOR-WIDE labor agreements (collective-collective bargaining). (Europeans do it, Asians do it, even North and South Americans do it; let's do, lets fall in love with fair and balanced labor markets -- again.)

That's all it takes: the old one-two.

Ezra, though I'm a huge fan of unions, I have to say that your chart overstates the case. The "right" way -- or at least, the standard economic way -- to measure the causal impact of union membership on earnings requires that you control for selection bias. Looking at the raw earnings data can be misleading, because workers who join unions tend to be different from those who don't in some important ways. Often, for example, they are more highly skilled than their non-union counterparts. Or they come from areas of the country where the prevailing wage tends to be higher.

To control for self-selection effects, the way most economists would measure the impact of union membership on earnings would be to run a regression. The regression would control for factors like race, sex, age, education, geography, occupation, etc. When that is done, most studies have shown that union membership is associated with an increase in earnings of about 20%. I should check that again, but the 20% is the figure that's stuck in my brain.

Of course, unions have benefits that go way beyond increased earnings for their members. There is some evidence that their presence is associated with increased pay for non-union workers as well. And the union movement as a whole helps get many progressives elected to higher office, which has made America more liberal, more egalitarian, and more genuinely democratic than it would have been otherwise.

If you want to read an excellent book about the economic impact of unions, do yourself a favor and check out Richard Freeman's What Do Unions Do? Written by a Harvard economist, it's the classic work on the subject, and it's completely accessible to nonspecialists. It's 20+ years old but its basic insights and findings still hold up.

A few rather obvious questions suggested by this post:

1. What effect do unions have on the wages of non-union members working in the same industry?

2. What effect do unions have on the rate of employment in the industry?

3. What effect do unions have on the broader economy?

I read all the way back to the bls.gov website and I don't see anything saying they adjusted for cost of living. If your claiming Union workers in CA, NY, WA, MI etc make more then non union workers in the South it's an obvious statement, it's also menaingless without adjusting for average wage where they live. States with labor friendly laws tend to have higher cost of living and wages.

You are for more government control of healthcare so they can drive down reimbursement to providers, but will you still be for unions if all of the providers join one to push back?

Nate appears to be correct. According to this article, both the cost-of-living and taxes tend to be higher in states with high rates of unionization than in states with low rates of unionization. After adjusting for these differences, average weekly earnings are lower in high-union states than in low-union ones.

This is like 60 minutes, Erza post something that people immediately call suspicious and question. A little bit of research later and bam there are the studies and facts to prove it wrong. Just like 60 minutes you have to ask; if normal people with jobs can tell right off the bat that something is wrong and within 7 ½ hours have the evidence to conclusively prove the post is inaccurate, why can’t the people getting paid to do this for a living get it right? Do you not care it’s wrong because it supports beliefs you hold or do you really somehow not know how ridiculous these claims are?

FYI, and in response to some of the comments made by others in this thread: according to the research Freeman summarizes in What Do Unions Do?, and according to subsequent studies of the economic impact of unions, unions do the following:

--They increase the compensation (wages and benefits) of union members to a large and highly significant degree.

--They have no impact on overall employment and unemployment.

-- They have a mixed impact on the earnings of non-union members (it sort of depends which studies you look at, and, frequently, on the ideology of the researcher).

-- They decrease wage inequality.

-- They lower the profits of unionized firms.

-- They increase productivity. Though standard neoclassical theory does not predict this, the empirical findings show that this is indeed the case. One possible reason for this: unionized employees may be more likely to suggest more productivity-enhancing innovations in workplace practices, since, unlike their nonunion counterparts, they can do so without fear of being fired or discriminated against. In general, unionized firms often have higher levels of cooperation between employees and management.

-- One more counterintuitive finding: although union members tend to report high levels of satisfaction with their union, and they are far less likely to quit their jobs than nonunion employees, they also consistently express lower levels of job satisfaction than their nonunionized counterparts.

Kathy G,

The book you mention is over 20 years old. It does not represent the views of mainstream economists. More recent research contradicts many of its claims. You can find a lengthy critique of the book here.

There was a time when I might have taken this post and comments seriously, but then I read Ann Coulter's column on the death of her father and this gem:
"He hated unions because of their corrupt leadership, ripping off the members for their own aggrandizement. But he had more respect for genuine working men than anyone I've ever known. He was, in short, the molecular opposite of John Edwards."
Now I realize those who truly care about the working man spend their lives destroying unions.

Jason, the Freeman book is one of the canonical texts in the field of labor research and is widely regarded as the best on its subject. Economics profs at the University of Chicago use in in their grad-level courses (I first learned about it in one of my courses. And btw it was also recommended in a class I took at the law school on labor and employment law). The author, Richard Freeman, is an econ professor at Harvard is very much a mainstream, neoclassical economist.

Yes, as I mentioned, it's over 20 years old. But its findings have been confirmed by more recent research in the field. For example, there was a symposium on the book in the summer 2004 issue of the academic journal, the Journal of Labor Research. The economists who participated in that forum concluded that subsequent research confirmed Freeman's main findings.

Kathy (no longer address Jason),
Freeman came out in 2007 with America Works which needless to say does nothing to undermine the conclusions of his earlier book.

The trolls talk about 'lower cost of living' as if it has no downside. Sure places with lower unionization have lower costs of living. They also, on average have higher crime, lower quality education, poorer services, more poverty...basically a lower quality of living.

There are reasons why housing costs are higher in some places...supply and demand. People want to live in these places, not only because they will get a higher wage with better job security, but also because the people who already live there have invested their tax dollars into high quality schools, better police and firefighters, and better roads and transit.

What these data don't show is that the union wage premium has fallen dramatically over time. It used to be much higher. Thus while there is still some benefit to working in a unionized business versus one that is not unionized (other things being equal), that benefit is gradually disappearing. One reason is that many employers are happy to pay the union wage rate just to be free of all the stupid work rules that unions often impose, which do nothing except hamper flexibility and efficiency.