Thursday, February 27, 2014

If Not Unions, What?

The piece I wrote about unions still seems to be in discussion. It's received responses from Brad DeLong and Kevin Drum in agreement, and Michael Hiltzik and Michael Wasser in disagreement. The discussion seems to have crystallized into two questions: Are unions finished? Are unions necessary?

That's how I framed my response to Michael Wasser, and I don't think any of their new critical responses actually address those questions. This is problematic. As I wrote before, if all you argue is that unions are irreplaceable, and you don't address the primary issue that there's a hole where your union should be, you aren't really arguing anything. You're just saying that the U.S. is screwed.

Let's go to Hiltzik's piece, which I think suffers from this analytical flaw. Hiltzik writes:
Wasser is certainly correct in arguing that Soltas' argument that the U.S. is better off without unions and "unions can't be saved" reflects the limitations of textbook-learning. A few specific issues:

To think that federal labor law has had "little to do" with union decline, as Soltas puts it, is hopelessly naive. He's misled by the fact that union membership has fallen even though we have laws guaranteeing the right to collective bargaining, and by the failure to recognize how inadequately those laws are enforced.

"Soltas doesn’t even consider the ramifications of broken labor law," Wasser observes, and he's right. "Without any real penalties to fear, employers have an economic incentive to violate federal labor law. Research shows that indeed they regularly do, using a variety of often unlawful tactics to coerce and intimidate workers during union organizing campaigns."

When the employers don't do so, political representatives of the capital-holding class will, as was seen in Chattanooga, where politicians used the threat of the withdrawal of government subsidies, and the impact that would have on the workforce, as a weapon against the union.

Over the years, employers have developed an exquisite arsenal against union organizing. For a succinct description of how the war is waged, Soltas needs to examine "Confessions of a Union Buster," the heartfelt memoir Martin Jay Levitt published in 1993.
First, you've got to appreciate the potshot at the fact that I know my economic theory, as if calling something "textbook-learning" is an actual response to a thorough argument on my part. It's the same thing with his charge that it's "hopelessly naive" to think that economics explains the decline of unions. And your evidence for that is...?

Let's review the evidence I gave for the decline of unions. First, you've got an uninterrupted decline in the unionization rate since the 1950s:

Looking at the chart, it's pretty clear that unionization began to fall as U.S. labor encountered increasing global competition. Why? The U.S. in the 1960s and 1970s wasn't exactly a hotbed of anti-union sentiment.

But let's say that you aren't convinced by the historic slide of unions into nothingness. Then you should reckon with the consensus of research that the decline of unions can be explained by global economic competition. This is something Hiltzik simply refuses to do. An actual rebuttal would have responded to the evidence I put forward, not just repeat the view that employer opposition matters.

That's the same way that Wasser's response proceeds:
Given that workers are still trying to organizing unions, one must ask why they don’t succeed. The answer, as Kate Bronfenbrenner describes convincingly in detail, is intense employer resistance. Much of this resistance violates federal labor law. Employers also commonly use procedural delays to lengthen the union election process...This is not a natural outcome. It is the result of a deliberate decision to not address the deficiencies in our labor law.
So it's great that they can cite research that shows the use of the illegal anti-union tactics, and so on. The more damning problem with their line of argument is that they simply has no evidence that, in the aggregate, the factors they cite actually matter. The stuff that we normally call evidence is one-sided. Again, from Farber and Western:
We find that most of the decline in the union membership rate is due to differential employment growth rates and that changes in union organizing activity had relatively little effect. Given that the differential employment growth rates are due largely to broader market and regulatory forces, we conclude that the prospects are dim for a reversal of the downward spiral of labor unions based on increased organizing activity.
And two from Hirsch:
The paper’s principal thesis is that union decline has been tied fundamentally to competitive forces and economic dynamism...Of course there are other explanations for declining private sector unionization – strong management opposition, a less favorable labor law environment, high levels of unfair labor practices, changes in worker sentiment, and weakened demand by nonunion workers for union services due to the increased protection offered by employment laws (antidiscrimination laws, pension insurance, etc.). But these explanations are less fundamental and of second-order importance compared to competition and dynamism.
In competitive and dynamic economic environments, a union tax on company earnings and slow response to economic shocks combine to produce poor performance by union companies.
And from Slaughter:
The main finding is a statistically and economically significant correlation between falling union coverage and greater numbers of inward FDI [foreign direct investment] transactions...Because U.S. affiliates of foreign multinationals have higher unionization rates than U.S.-based firms do, this correlation does not reflect just a compositional shift toward these affiliates. Instead, it may reflect pressure of international capital mobility on U.S.-based companies, consistent with research on how rising capital mobility raises labor-demand elasticities and alters bargaining power.
The truth is I couldn't find one bit of research arguing what Hiltzik or Wasser say: that union decline is primarily the result of public policies that undermine labor or increased employer hostility.

What Hiltzik and Wasser are more interested in doing, as Hiltzik's title question "Are unions necessary?" suggests, is saying that you need unions to get liberal policies. I have no doubt that unions would make it easier for those on the political left to get the policies they want. And I think Kevin Drum's assessment, which describes the political vacuum left by the decline of unions, is totally on point.

But I don't see what unique weight this carries in the argument. If we can't save unions, then it doesn't count for anything that unions help win whatever policies you desire. The thesis of my original piece was that since the decline of unions, the U.S. political system has done a poor job of sharing prosperity with workers -- and that since unions aren't a real option going forward, we have to look for other ways to build employee bargaining power.

It's entirely non-responsive to my position to say, as Timothy Noah does, "To think you can tackle inequality without empowering workers is fantasy." Or, as Jeff Spross does: "[Evan] wants to replace unions w/ policies only passable w/ the political muscle unions historically provided." Or, as Hiltzik does:
Where does he think the impetus for these advances will come from, if not the labor movement? He may not have noticed, but Congress today is in the grip of the employer class.
 Or, as Wasser does:
I still don’t see how we get to policies like full employment, which Soltas believes will address inequality better than unions. Who, if not unions, has the resources, both in terms of money and supporters, to support this policy agenda and take on any opposition to it? And assuming these policies become law, who ensures that future administrations and congresses do not repeal them or render them ineffective? Unions aren’t just a check-and-balance on employers in the workplace. They also serve as a check-and-balance on corporate power in politics.
Why is it non-responsive? Because if it's a "fantasy" to think we can do good for workers without unions, then you better have some compelling evidence that unions can be brought back to life. The left can't get by just pretending that employers and politicians are to blame for killing off unions. It needs to actually reckon with the reality that unions have a substantial disadvantage in a globally competitive economy -- where a large union wage premium, for example, is simply unaffordable.

Yet I think my role in this debate is to say that the left's emphasis on the union isn't merely a flawed argument in this way, it's also wrong. We have no other choice but to look for another way to restore labor's share, but I also think another way exists.

I don't think, for instance, it takes unions to push the political system towards full employment. Clinton did it in the 1990s. A team of economists -- you may have heard of them, they're called the Federal Reserve -- is doing it now. I don't think it takes unions to push the political system to deal with inequality. The second largest U.S. antipoverty program (the earned income tax credit, after Social Security) was born in a de-unionized America, and mainly with the backing of Republican presidents. These are not difficult or new policy ideas.

We can rebuild the power of labor without labor unions. We just have to try.


One further comment on Hiltzik's piece. He really doesn't get the economic analysis in Freeman and Medoff about unions, which he waves off as "two academic economists" (but is in fact the single most important thing anybody has written about unions in the last 30 years) and then says, all of a sentence later, that I "cite no authority for these statements."

"The only vantage point from which union power can be seen as inefficient and bad for the economy is that of rent-seeking management," he says. So let me explain the basic economics of why the way unions help workers is less than ideal.

Unions build power by restricting the supply of labor -- quite literally, they say that, for any given wage rate, they'll supply less labor than would have otherwise been provided. (See the chart on the right.) That's not the way I'd like to do it, because it means we employ fewer and produce less, even if it drives up wages. I'd like to maintain strong demand for labor. This goes two-for-two on outcomes: Labor gets higher wages and higher employment. (See the chart on the left.)

If someone cares, I could actually write out a macroeconomic model for this, but it's going to get to the same result as above.


  1. 1) Your first graph, which you post after you wrote: “First, you've got an uninterrupted decline in the unionization rate since the 1950s: “You attribute this to globalization: “Looking at the chart, it's pretty clear that unionization began to fall as U.S. labor encountered increasing global competition.” The first thing I’d say to this is that correlation does not prove causation. The second thing I’d say is that it’s pretty clear that the rapid growth in unionization stopped after the passage of the Taft Hartley act, roughly stabilized for a bit, and then began an accelerating decline. This is exactly what you’d expect if Taft Hartley were what killed unions- it wouldn’t immediately cause a huge drop, and the decline in unions would accelerate over time, rather than declining at an even pace, because it would take time for employers, the anti-union industry, and the political landscape to change in ways less favorable to unions.

    2) Farber and Western. I looked at this quickly, but again, this isn’t my job. From the abstract: “… we examine data on NLRB representation elections to determine if changes in the administration of the NLRA during the 1980s reduced the level of organizing activity and success…” Well of course they didn’t find any change after 1981. That’s because the key year isn’t 1981, it’s 1947.

    3) Hirsch I didn’t even read. Sorry. But your quote doesn’t seem to prove what you say it does. “The paper’s principal thesis is that union decline has been tied fundamentally to competitive forces and economic dynamism…” Well of course. If you can’t unionize new industries, or unionize expansion (including, especially foreign expansion into the U.S.) of existing industries, more dynamism will hurt unions. The fact is that in a capitalistic economy, full of creative destruction and all, unions need to continually organize in new firms just to stand still. If they can’t organize new firms, and only hold on to where they’re already organized, they’re bound to shrink. There’s no contradiction between this thesis about dynamism and the thesis that if the U.S. had more pro-union labor laws we’d have more unions.

    4) “The main finding is a statistically and economically significant correlation between falling union coverage and greater numbers of inward FDI [foreign direct investment] transactions…” Yes, firms will race for the bottom, if they can. And yes, certain examples of FDI might not have happened if the U.S. was more unionzed. (E.g.- who knows if Japanese and German car companies would have set up in the U.S., or to the extent they did, if they knew that their factories would be unionized by a strong union.) This is an economic argument against unionization. But it doesn’t say much about whether unions would represent a greater proportion of the U.S. labor force if the laws were more favorable to unions.

    (continued in next post, character limit)

  2. 5) “The truth is I couldn't find one bit of research arguing what Hiltzik or Wasser say: that union decline is primarily the result of public policies that undermine labor or increased employer hostility.” Truth is, I can’t directly find that either, partly because I haven’t looked and don’t have time. But I’d point to two things: the unionization rate in foreign countries (in Canada, much higher than the U.S. In Scandinavia, even higher. Aren’t those places exposed to globalization too?), and the research showing the effectiveness and extent of employer anti-union campaigns in the U.S.

    6) The 1990s. Or really, the late 1990s. Maybe just 1998-2000? It’s unrealistic to expect the United States to have full employment forever, because we haven’t done that. As far as I know, the only countries that have done that in modern times are Israel and Australia, and they’re both small countries that may have had unique circumstances. But even if we could maintain full employment, inequality still grew in the late 1990s. Yes, wages grew for everyone, but they grew faster at the top. Is that really the best we can hope for?

    7) Finally, your silly graph at the bottom. Econ 101 level stuff. The same type of argument can be used against the minimum wage, btw. The problem with this (and the anti-minimum wage arguments) is that the labor market is a lot more complicated than the market for bushels of indistinguishable wheat. For starters, there’s a pretty good argument that just about every employer has monopsony power vis-à-vis his workers, and this is especially true for large employers and low-skill workers.

    Time for me to go to sleep so I can go to my real job in the morning.

  3. I think what unions mostly do is provide institutional and political support for Democrats in between "wave" election years. You can still get a pretty good set of reform policies even with weak unions (see New Deal in the 1930s, reforms in the 1990s and in Obama's first term), but it's harder to maintain political strength in between those elections.

    You make a good point that it's disingenuous to just talk about how necessary unions are without addressing the elephant in the room: unions have declined ever since Taft-Hartley passed in 1947 with a veto-proof majority (including tons of Democrats), and they've been completely unable to get it seriously weakened or repealed despite decades of filibuster-proof majorities of Democrats in Congress. The closest they came was in 2009 with the Employee Free Choice Act, and that couldn't get through the Senate.

    To me, that seems like there's a deeper weakness that unions have to address, namely the lack of actual support for union-friendly policies at the ballot box among most voters.

  4. This comment has been removed by the author.

  5. I have one more thing to add about the wage growth in the late 1990s.

    That growth took place in a certain context. Unions had the level of strength that they had- it was not zero. Labor law protected unions to the extent that it did- it was not as anti-union as it could have been. The possibility of unionization likely increased wages for non-union employees. Low unemployment may have greatly increased the risk of unionization, because employees had less to fear if they lost their jobs. So the effects of unions and low unemployment reinforced each other and are difficult to separate.

    If you take away that context, wage growth might have been lower. Imagine that in the late 1990s the private sector union membership rate was only 1%, we had a national RTW law, and federal law was otherwise less union-friendly. Would wage growth in 1999 have been the same? Maybe, but you shouldn't just assume it, and there are intuitively good reasons to believe otherwise.

    The path that you're advocating- allowing unions to wither away- necessarily implies getting closer to that no-union hypothetical I just sketched out. You may be happy with unions' level of penetration and power right now, or their levels in the late 1990s, but keeping things static like that is not really an option. The choices are continued union decline, or making some changes to the law to turn that around and encourage growth. You can't just bring 1999 back and keep it that way forever.

  6. Evan is missing some other counterfactuals to his argument. First, there's the international comparison between Canada and Germany. Both countries have had to contend with international competition from low wage countries as much as the US and their unionization rates have been far more solid. In addition, Evan's macro argument fails to distinguish the dramatically different sectors labor organizes.

    While manufacturing has shrunk as a share of the national economy (along with corresponding unionization rates), a variety of other sectors that face no international competition have grown and organizing in those sectors has been difficult to say the least. I would argue that much of the reason for that difficulty is the deadly combination of terrible labor laws (especially their enforcement), a 40-year propaganda campaign against labor, and enough of a power imbalance that essentially leads many workers to cower in fear when it comes to standing up to their boss.

    Lets think about retail. There you have several dozen of centrally controlled national chains that are incredibly hard on their workers and yet there is virtually no penetration into these firms. Large stores like Wal-Mart, Target, Kohl's, etc employ hundreds of thousands (millions?) of people, yet labor is thoroughly unable to get a real foothold. Hard to fault international competition and capital mobility when you're talking about location dependent firms that are pretty well anchored in place in order to serve their market segment. Yes, some of that is due to the ineptness, insularity, and whipped dog mentality of some senior labor leaders. But the fact of the matter is every worker in those stores knows that any attempt to organize a union will be met with an incredibly hostile response that will inevitably result in fewer or more irregular hours, severe harassment, and probably dismissal. They want more rights, greater control of their schedules, wage increases, and benefits. But they also want jobs. The fact that fast food and other retail workers recently did what they did without the trappings of a union organizing drive is a testament to both how bad things have gotten and the reality that the NLRA route leads off a cliff. Things aren't too dissimilar with home care workers and skilled nurses, although they've had much more success

    Evan also doesn't discuss construction, which is by its nature fixed in place, and has high union densities organized in the top metros. While there's certainly a tremendous amount of technological deskilling and labor saving happening in the industry, capital mobility is minimal. So its not surprising that the most aggressive and creative unions can secure work for their members. Its also not surprising when thinking about the construction industry's relentless short term focus and susceptibility to booms and busts that unions serve a stabilizing role that actually enables investment in workforce skills and safety, the long term wellfare of the workforce, and jobsite empowerment while simultaneously permitting a dynamic flexibility that retail or manufacturing could only dream of.

    My point in all of this is to say that no, the decline of unions does not have to be inevitable . It is however inevitable if we shrug our shoulders at the massive obstacles facing unions and move on to envisioning a world after their demise. To paraphrase Paul Krugman and a few others: 1. Destroy Unions (or let them be destroyed) -> 2. ?????? -> 3. Workers gain voice into workplaces, wage increases, and job quality improvements.

  7. This statement is incredibly naive and just not in keeping with reality "Unions aren’t just a check-and-balance on employers in the workplace. They also serve as a check-and-balance on corporate power in politics."

    Unions are run by men (and a few women) who make decisions based on 2 things and 2 things alone: What can I do to increase membership which will increase dues and bring more money to me and What can I do to increase my personal political power.

    That is it folks. Anyone who believes otherwise has no grasp on what the modern union movement has become in cushy America 2014, nor any understanding of human motivation. Union bosses now earn well past $300,000 a year. The idea that they are welders, steel workers, postal workers, carpenters, kindergarten teachers, janitors, etc, like traditional low to middle class workers is ridiculous.

    They will do anything and everything to keep their power and grow it, especially if that means selling out their own membership or causing a plant to close 10 years out.

  8. Unions are necessary, or rather, what is necessary is an institution, or institutions, powerful enough to balance the concentrated powers of capital, in order to create a robust economy, and a robust democracy.

    In the absence of this power, capital unchecked accumulates and concentrates, to the detriment of the economy and society. As capital concentrates, the growth of the buying power of labor, (and we see with the austerian policies promoted by the Right, government, also) first slows, then stops, then starts to decline. And we are starting to enter the period of decline, after years of stagnant income by labor. Capital is destroying its market, and with it, opportunity for continued profit, and the incentive to invest. This leads to a gradual decapitalization of the economy, although this will first be concealed by investment bubbles in enterprises which do not produce real goods and services. That is, the economy will continue to appear to grow, even while in real terms, in terms of industry and infrastructure, it actually declines.

    Government itself, and certainly after Citizens United, is not strong enough to provide this countervailing power. Indeed, its policies show it has clearly become the captive of capital, and so aggravates the already mal-distribution of power.

    So this countervailing power once, inadequately perhaps, partly provided by unions, but now largely absent, is necessary even for the survival of capital.

    Cross-posted to my blog: