Hacker News Reader

Comment by js8 | Hacker News Reader

To everybody who thinks this is just correlation and not causation - read the post properly. They mention:

"Labor unions both sustained prosperity, and ensured that it was shared; union bargaining power has been shown to moderate the compensation of executives at unionized firms."

And give a link to a study. This result has been seen across countries, and is the CAUSAL link you're looking for.

View on HN

Mr Always-grumbles-about-correlation-studies reporting in here. Maybe it's a nice causal study, but the jump from "moderates compensation of top executives" to "reduces equality across the whole income spectrum" is a big stretch.

Update: also, I went and scanned the linked article (because I'm that type). They're credible authors, but they themselves explicitly say: "In reporting correlations between unionization and managerial pay, we are not necessarily able to establish whether unions cause differences.... We try a variety of strategies to understand these issues, however each is imperfect."

"In reporting correlations between unionization and managerial pay, we are not necessarily able to establish whether unions cause differences"

Yes, but that's a different correlation vs causation issue. The point is, the phenomenon of unions equalizing pay, on the company (micro) level (which has been reproduced in other countries and time periods), can be considered to be a causal explanation for the phenomenon of rising inequality when unions decline, in the society at large.

Right, the paper cites a lot of other papers which looked more relevant on the broader topic of what unions do to pay.

Let's just read the title: "As unions decline, inequality rises"

That is very much a statement about correlation. Otherwise it would say, "Unions' Decline CAUSED Inequality". But, to anyone with a passing knowledge of statistics, that is clearly an impossible claim to make. This is one time series realization for one country.

Let's go further. Is there any mention of causality in the article. And I do mean quantified causality. The answer is "no". Is there even any mention of correlation? Again, the answer is "no".

Let's keep going. Let's evaluate the aims and intent of the article. Let's start with EPI, the publisher of this article. The Chair of EPI is "Elizabeth H. Shuler ... president of the AFL-CIO, a federation of 58 unions and 12.5 million working people". Other board members include current and former union leadership from the United Steelworkers, the UAW, the Nonprofit Professional Employees Union, Service Employees International Union, the International Association of Machinists and Aerospace Workers, the American Federation of State, County and Municipal Employees, the Communication Workers of America, and the American Federation of Teachers. All huge unions.

In other words, this publication is not a work of science. It is not a credible academic publication. It is the motivated, self interested rhetoric of union leadership.

I don’t have any real opinion about the direction of causality. If anything the idea of unions contributing to more wage equality makes sense to me. But, linking to a single study as “proof” of anything is a big no no. It is very easy to cherry pick studies so that you get the “true” outcome. The only way to make a halfway convincing argument is to look at every piece of evidence on the question and aggregate them. This is why we believe global warming is man made and severe despite a few studies that suggest otherwise. One other caveat I would add is that if you read the linked study, they’re only reporting correlations. There doesn’t appear to be a causal identification strategy there. Which isn’t to say it isn’t interesting and worthwhile. Only that one should take the results with an appropriate level of caution.

> To everybody who thinks this is just correlation and not causation - read the post properly.

A quick glance at the chart and one could also show a link between federal income tax rates and executive compensation.

If you are taxed 73% you simply don’t have any incentive to “increase inequality” — you pay someone to show your income as a lot less than it really is. The real irony is the Reagan tax cuts resulted in more revenue because it just wasn’t worth the trouble to dodge paying taxes anymore.

Correlation, causation, lies, statistics, who knows?

Oh wow, they have a study! Definitely definitive then, case closed.

More seriously, my wife and I got into an argument about the effects of unions once and this was the best summary I could find: https://www.heritage.org/jobs-and-labor/report/what-unions-d...

(yes it is by the Heritage foundation so their top-line summary probably has some ideological bias but the list of study summaries at the bottom is great) Generally, studies found that unions benefitted the unionized workers at the expense of the most and least skilled workers. And was slightly bad for company investment/earnings long term. But by far my favorite result (and the best study design from the list, imo) was a regression-discontinuity analysis that found no effect of unionization whatsoever:

"Compares companies whose workers voted narrowly for a union with companies whose workers voted narrowly against a union. Since the difference between winning and losing is close to random, this provides an estimate of the causal effect of randomly organizing a given company. Finds that workers who vote to join a union do win certification but that unions have essentially no effect on the firm or the workers. Wages do not rise, and employment and productivity do not fall. Unionized companies are no more likely to go out of business than are non-union firms."

DiNardo, John, and David S. Lee, "Economic Impacts of New Unionization on Private Sector Employers: 1984-2001," The Quarterly Journal of Economics, Vol. 119, No. 4 (November 2004), pp. 1383-1441.

My theory is unions thrive when there's a labor shortage, so workers will get paid more just because there's a labor shortage, decreasing inequality. If there's a labor glut, the union model falls over because it can't maintain its monopoly, and executives pat themselves on the back for lowering production costs, even though it was mostly driven by the macro environment.

Reply on news.ycombinator.com