On March 5, Bloomberg View contributor Evan Soltas concluded a wide-spanning, 3-week econ-blog debate over the long-term decline of unions. While the discussion as a whole has been illuminating, following the constituent parts has been difficult. Berkeley economist Brad DeLong did an excellent job of capturing the debate as it was happening, but it’s worthwhile to extend that job up to and including the recent (apparent) conclusion. So, here is the format – I will give a very concise summary of the arguments, followed up with brief comments on where my own thinking has ended up after reading everyone a couple of times, and finally I give a long-form account of of the debate (an endeavor largely indebted to Brad, through the 28th).
So, we begin with the concise version:
On unions’ long-term decline in the U.S.
Causes of decline
Economic forces (primarily globalization)
Farber & Western, Hirsch (a), Hirsch (b), Slaughter
Broken labor law
Bronfenbrenner, Western & Rosenfeld, Levitt
Country’s choice of political economy
Schmitt & Mitukiewicz, Warner
Reversibility of decline
Bronfenbrenner, Western & Rosenfeld, Levitt, Schmitt & Mitukiewicz, Warner
Uniqueness & Necessity of unions
Unique tool in the political arena; their death leaves a vacuum
No evidence cited
Necessary to promote pro-labor policies (e.g., the EITC and full-employment)
No evidence cited
Not necessary to pursue pro-labor policies
Abandon unions; pursue full-employment, other labor-empowering policies
Save unions by fixing broken labor law; this will be necessary to pursue policies like full-employment
Personally, I am equally convinced of the globalization story and the political economy story – I think it’s in all likelihood a mix of the two, and that means broken labor laws have their day too. Everybody gets a trophy. Ultimately, though, I find myself where Bruenig wants Soltas to be – that there was political-economic, as well as purely economic determinism at play. In that case, regardless of the mix of economic vs. political-economic determinism, I find myself opposed to Wasser and Hiltzik in that I don’t believe the trend to be reversible.
Where I’d like to see some concrete evidence is with regards to unions’ unique political role in promoting pro-labor policy. Wasser, Hiltzik, and Bruenig all repeat some version of the claim that organized labor’s consolidated political voice is necessary to push through national policies like full-employment, or the EITC. Yet, I cannot find empirical evidence cited in any of their arguments, and so I find myself not leaning one way or another on the issue. This is important because it determines whether I end up with Soltas, or with Drum.
If we can achieve policy goals like an expansion of the EITC, full-employment, etc. without the political voice of unions, then we don’t really have a problem. Let’s just pursue those policies, and others, to increase wages for low-income workers. That is essentially Soltas’ view.
On the other hand, if unions really are unique and necessary on the national level in order to push through pro-labor policy, then we’ll need to find a replacement for that consolidated political voice (given my belief that their decline is irreversible). That is essentially Drum’s view.
It may be that this debate has run its course (in its current iteration, on this set of blogs), but I personally find myself in need of more evidence before reaching for a final policy implication, prescription, or conclusion.
The inciting event: workers at a Chattanooga, Tennessee Volkswagen plant closely vote “no” to joining the United Auto Workers (UAW) union (712 no, 626 yes).
Using the Chattanooga vote as the impetus for discussion, Evan Soltas kicks things off. He starts from the premise that unions have been dying a long, slow death in the U.S. – this is clear enough in the data, and no one will argue this point.
More controversially, however, Soltas claims that the death of unions was assured by economic forces (most notably global competition), and that we should therefore accept the decline of unions as irreversible fact.
Soltas yet wants to fill the void left by unions. He raises a few alternate solutions, but ultimately dismisses them as too-incremental, and so he finishes by voicing his actual preference – the pursuit of full-employment, in the hope that tight labor markets will raise wages as much or more than unions would.
Michael Wasser responds to and disagrees with Soltas on two major points.
First, Wasser argues that broken labor law has allowed employers an unfair upper hand in undermining union formation. He does not offer an idea of how much this mattered in the decline of unions, especially relative to Soltas’ story of globalization. Still, this marks the first challenge to Soltas’ original premise that the death of unions was preordained, and that the long-term trend is presently irreversible.
Secondly, Wasser notes that unions not only perform the purely economic function of bargaining in the workplace, but they are also crucial to the political process: “unions consistently advocate for expanded access to health care, the protection of Social Security and Medicare and a more fair tax code.” Wasser argues that none of the alternate solutions to unions that Soltas offered can fulfill this political role.
Soltas replies directly to (some of) Wasser’s criticisms.
First, Soltas tries to keep intact his story that the trend is irreversible, and so he attempts to counter Wasser’s claim about the importance of broken labor law. He cites two papers – one from Farber and Western, and another from Hirsch, which combined tell the story of union decline as a consequence of globalization (this essentially reiterates Soltas’ original position).
Secondly, Soltas brings in a graph showing broad-based wage increases in the 90s, which he attributes to full-employment. In making this argument, Soltas is trying to emphasize that unions are not unique in their ability to reduce inequality via a rise in labor’s wages. Still, this does not amount to an explicit rebuttal about unions’ unique political role (maybe the implication is that wage growth can be achieved without a consolidated voice in politics, but Soltas does not say so).
Wasser posts a short update to his original article, in reply to Soltas’ February 20 response.
First, Wasser cites additional, specific research to back up his claim that broken labor law has been a factor in the decline of unions – and that societal neglect of that breakage was not inevitable, nor is it irreversible now (unlike globalization). Nonetheless, he still does not muster evidence on the relative importance of broken labor law. It is a factor, but he does not assert how much (i.e., Wasser could be right that broken labor law is a factor, but without saying how much, the globalization story could still dominate).
Wasser also lends some specificity to his stance on unions’ political role, and tries to counter Soltas’ invocation of full-employment as wage-raiser. As he says, “I still don’t see how we get to policies like full employment” without the consolidated political voice of unions.
Finally, he cites research by Western and Rosenfeld that union-decline is a direct cause of inequality. Soltas thinks that globalization caused both, but Wasser’s citation directly refutes that claim.
The following week, new voices enter the debate, including Kevin Drum, who comes down partially with Soltas, and partially with Wasser.
First, Drum buys Soltas’ story that the decline of unions cannot be stopped. Very simply, he asserts that “The decline of union power is irreversible,” and “that as a broad-based force that provides a countervailing force against the power of the business community, labor's day is over.”
However, Drum reemphasizes one of the points that Wasser made, and which Soltas largely bypassed in his initial response. That is, unions were a consolidated power in the political arena which could oppose business (that is, employer) interests. And, like Wasser, he does not see an immediately obvious replacement for that unique role.
It is important to note that Drum does not really introduce any new evidence to the debate, and that his outlook actually combines the worst of both worlds, in sum – unions are unique and necessary, but their decline is irreversible.
Hiltzik enters firmly on Wasser’s side, although he brings a somewhat unsettling anti-academic tone to the debate. Hiltzik’s primary contribution is in providing minor buttressing evidence for Wasser’s claims about broken labor laws. Hiltzik first cites a memoir by a former union buster, and then points to one of the same papers that Wasser brought up originally.
Hiltzik also makes claims similar to Wasser that unions are needed for their unique role in the political process to push ideas like full-employment in the first place, but he offers no additional evidence.
He makes some other points, but they are logically flawed (e.g., implying correlation to be causal proof) and don’t really add to the debate.
Later that same day, Soltas mounts a reply.
New to this post is a second paper from Hirsch, as well as a separate paper by Slaughter, which Soltas uses to reinforce his point about the irreversibility of unions’ decline.
Interestingly, Soltas concedes the point that the death of unions leaves a political vacuum – he just doesn’t think this matters. He points to Clinton in the 1990s and the Fed in present-day as non-union forces pushing for full employment, and to the EITC being born in a de-unionized America with the backing of Republican presidents.
Matt Bruenig brings a data-driven and political-economy point of view to the debate, and is critical of Soltas’ argument that economic forces primarily killed unions.
Bruenig points to a CEPR paper which highlights cross-country differences in unionization rates, and they do indeed vary wildly. He makes the salient point that since these (economically advanced) countries were all exposed to the same globalization market forces, there must be another factor which accounts for the varying unionization rates.
His second graph is fairly convincing – it breaks up those same countries by type of political economy (e.g., Liberal Market, Social Democratic, etc.), which demonstrates a strong correlation between type and unionization rates. There is not a causal argument here, but there does not need to be – this set of facts is at odds with Soltas’ globalization argument, and something with which he will need to reconcile.
Finally, Bruenig brings in a final graph demonstrating the split between Canada’s and the United States’ unionization rates over the last 90 years – since Canada is also a liberal market economy, Bruenig argues that a claim of political economic-determinism has to contend with the split between these two countries in particular. In other words, a country can be a liberal market economy and still encourage/protect unionization.
Wasser returns to the debate, and tries to address some of the evidence that Soltas has cited thus far.
First, he criticizes the Farber & Western paper for being reliant on NLRB election data, which Wasser claims are subject to bias because the NLRB election process itself is subject to bias. Next, Wasser points out that Hirsch himself caveats his and others’ empirical research by saying that the literature’s conclusion in favor of a globalization story “rests heavily on evidence that is dated and (arguably) unable to identify true causal effects.”
Wasser goes on to re-cite the research that Bruenig brought in, agreeing with all of Bruenig’s points.
Finally, Wasser argues against Soltas’ claim that pro-labor policies like the EITC would be pushed forward without an organized-labor interest in politics. For example, he points to Democrats’ recent push for a higher minimum wage, and claims that “the reason Democrats are rightly doing that is because of the efforts of unions and their members to make addressing wage stagnation and inequality a national priority.” He does not, however, cite evidence or research to back this claim.
Hiltzik reemerges with a tone toward Soltas that waffles between adulatory and acrid. Primarily, he clarifies his own position that broken labor law is not necessarily primarily the culprit in the decline of unions, just a big player, and traces the root back to the Taft-Hartley Act of 1947. Along the way, he restates and finds himself in agreement with Bruenig’s excellent points on cross-country differences.
Hiltzik goes after Soltas’ observation that the EITC was “‘born in a de-unionized Amierca,’” citing that union membership at the time was near its peak. He also points to the birth of Social Security, which came at a time when union membership was on the rise.
Finally, there are a couple of flaws in Hiltzik’s arguments worth noting. He incorrectly attacks Soltas for never identifying “‘other ways to build employee bargaining power,’” when in fact Soltas did just that in his original post (works councils, stop businesses from classifying employees as independent contractors, and recommendations from Freeman to encourage employee ownership and profit sharing agreements).
Hiltzik ends his post by taking issue with Soltas’ use of “a couple of macro charts that I don’t recognize,” which are in fact econ 101 supply and demand curves (albeit labeled for the labor market, rather than goods), and essentially says that anyone believing that such charts represent the real world “needs to get out more.” Unfortunately, such basic misunderstanding of how economists intend and apply even intro-level models like these somewhat undermines Hiltzik’s own credibility.
Soltas finishes off the debate, and to his credit revises prior beliefs, although only to an extent. He admits that Bruenig had an excellent point that political economy helped determine the U.S.’s failing unionization rates, and he even supplements this admission with additional evidence from Jake Rosenfeld (who has a new book out, and who sent Soltas an email on the topic).
However, Soltas sticks to his guns in regards to his belief that global economic forces were the primary driving factor in the death of unions, and he believes that the types of policy choices that Europe made to keep unions strong would be unwise for the U.S. In sum, Soltas ends where he started, but he does manage to make room in his worldview for political economy considerations (and, implicitly, broken labor law considerations), even if those factors do not ultimately sway his conclusions.
Latin American Startup Champions Gear Up for Moscow Gathering
A Look Back at Innovation Daily