Debunking the Notion That Unions Hurt Productivity

A neat little analysis by Ross Eisenbrey at the Economic Policy Institute may be difficult for union foes to explain away. It shows the proportion of workers covered by collective bargaining agreements in major European countries and the US and then shows productivity growth country by country in the same group 1979-2005. Despite being the only nation in the bunch with low union representation, US productivity growth is merely middle-of-the-pack.

Now one can quibble slightly with Eisenbrey’s presentation. Rather than showing 2005 union representation, it would have been cleaner to show the average over the 1979-2005 period (for example, during this time frame, US union membership dropped from 27% to 12% today. It makes for a better comparison and in no way changes the outcome).

In fact, the dirty secret of this exercise is that, were GDP of the US computed on the same basis as in European countries included in this comparison, the US would almost certainly show lower productivity gains than any other nation in this group. The reason is that, in 1980, the US started adjusting its GDP figures to allow for the fact that computer and communications technology were becoming more powerful (i.e., even though buyers were paying less, they were getting considerably more utility). No other country makes these so-called adjustments, using a hedonic price index. And the cumulative distortion is massive. In 2005, economist/investment advisor Michael Shedlock contacted the Bureau of Economic Advisers and they supplied some dated information on hedonics (including a spreadsheet). He found that hedonic adjustment to GDP was 2.257 TRILLION dollars, or 22% of then-current GDP.

Productivity is output per unit of labor. To determine productivity of an entire economy, the numerator in the calculation is generally GDP. So if we are alone among our peers in having an inflated GDP, that says were it computed on a comparable basis, our lower GDP would also result in lower productivity growth. And since Shedlock’s work indicated that our GDP is wildly overstated, so too is our productivity growth.

Maybe more unions are just the thing we need……

From the Economic Policy Institute:

Unionization in the United States has declined since the late 1970s, when 27% of U.S. workers were covered by union contracts, to today, when only about 12% are covered. This has had substantial adverse effects on inequality, the wages of typical workers, and pension and health benefit coverage.

By contrast, most of the major continental European countries have maintained strong unions, and most of their employees are covered by collectively bargained contracts, ranging from 68% in Germany to over 90% in Belgium, France, and Sweden (see the first chart below).

There is a common myth that unions hurt productivity, supposedly because they impose work rules that make their employers less efficient. The evidence from industrial relations studies does not support this myth. A broad study of the economics literature found “a positive association [of unions on productivity] is established for the United States in general and for U.S. manufacturing” in particular (Doucouliagos and Laroche 2003, 1).1 And as the second chart below reveals, international comparisons suggest that high productivity and very high union density are entirely compatible.

The dramatic drop in unionization in the United States from 1979 to 2005 did not lead to faster productivity growth than in the seven largest European countries with union density greater than 60%. In fact, those countries’ average annual labor productivity growth of 1.7% equaled productivity growth in the United States. Output per hour worked is higher in the Netherlands, France, and Belgium,2 where more than 80% of employees have union contracts (compared to the United States’ 12% unionization).3

If Congress is concerned about protecting middle-class incomes, it should pass measures to facilitate union organizing and collective bargaining coverage, including the Employee Free Choice Act. There is no reason to fear that higher rates of unionization will impede efficiency or labor productivity.

1. Doucouliagos, Christos, and Patrice Laroche. “What do unions do to productivity? A meta-analysis.” Industrial Relations. Vol. 42, No. 4 (2003). Cited in Shaiken, Harley, “Unions, the Economy, and Employee Free Choice,” Economic Policy Institute (2007).
2. OECD estimates of labour productivity for 2005 (September 2006).
3. Mishel, Bernstein, and Allegretto, The State of Working America 2006/2007, Table 8.5, p.332, Economic Policy Institute 2007.

Print Friendly, PDF & Email


  1. Steve Diamond

    Strange – he glosses over the fact that unionionization in France is about the same as in the US. The fact that workers there are covered by nationwide collective bargaining misses the problem – employers here face much tougher workplace based unions and of course unions here face much tougher bosses on the shop floor. Unions here could easily be a factor in limiting productivity and profitability – how else to explain the higher wages that unionized workers here receive?

  2. Yves Smith

    There was a very interesting post on the Economist’s blog, Free Exchange, “Do unions increase productivity?” ( Some illustrative arguments:

    ….unions often do very good work. New York’s Local 3 (electricians) is widely known for the slow pace at which union jobs proceed, but also for the extremely high quality of their installations. This is not inconsistent with economic theory. Union electricians get paid more to do the same work, which predicts that they will get more skilled workers than non-union shops, and the workers will be more keen to keep their jobs. Plus the union, eager to prove that there is some sort of value proposition to employing their workers, enforces a higher standard on its members….

    [S]ome unions are equally well known for their lack of productivity; the American teachers’ unions are generally believed (by everyone outside of the teachers’ unions) to be the primary obstacle to improving America’s appalling public schools.

    One possibility is that, to the extent that unions do increase productivity, they do so by forcing less competent workers out of the labour market, because they are not worth union pay. In teaching, where the average wages are nothing special for the target, college educated applicant pool, this doesn’t work. Indeed, by compressing wages, it makes the problem worse. In areas where there is an oversupply of graduates, such as English and history, teaching programmes choose from the applicants who have relatively few other opportunities; while in areas like science and math, where almost any qualified applicant has higher-paying alternative opportunities, they face permanent shortages.

    I have no doubt other factors come into play, such as cultural attitudes, and the prevailing level of social safety nets.

    And the focusing on collective bargaining may miss the broader issue of labor cartelization (after all, what are doctors, lawyers, and accountants? They do appear to create more favorable employment conditions for themselves. I know of many un- and underemployed MBAs in their 50s, but no members of any true profession).

Comments are closed.