The Agenda

A Follow-Up on Ramesh’s Unions Column

At the end of Ramesh’s column, he observed the following:

The shift toward a more competitive economy has not hurt workers in general: Total employee compensation as a share of the economy held fairly steady during the second half of the last century even as unions were shrinking. (It’s true that wages as a share of the economy fell, but that was a result of the increased cost of benefits.) The shift has, however, increased inequality among workers, with more rewards going to those with higher skills. [Emphasis added]

This didn’t raise an eyebrow when I first read it. I’m of the view that rising wage dispersion reflects a number of interrelated phenomena: (1) rising gains from cooperation among skilled workers; (2) the rise of pay-for-performance compensation schemes, particularly in knowledge-intensive services; (3) demographic factors, including the aging of the population, etc.

Henry Farrell objects by pointing to increased inequality among workers as a source of “immiseration” for the middle class, using a term drawn from Ramesh’s column. He points to a paper co-authored by Bruce Western and Jake Rosenfeld:

This is a reasonable summation of their findings, although Bruce Western, in a recent paper with Jake Rosenfeld appears to have changed his mind in the intervening years, noting that “an influx of corporate donations influenced policymakers to oppose pro-union reform of labor law in the 1970s” and that “[p]olitical defeats in the 1970s and 1980s yielded an ‘enervated’ labor law that enabled employers to block organizing campaigns and weaken existing unions.” Perhaps more to the point, Western and Rosenfeld provide substantial supporting evidence for the ‘immiseration of the middle class’ thesis that Ponnuru is seeking to disprove.

Farrell’s evidence for this proposition is the following, drawn from the Western and Rosenfeld paper:

Counting the union and nonunion wage effects, deunionization explains about a third of the rise in men’s earnings inequality. Increasing returns to education and increasing wage inequality among highly educated workers explains a similar share of the rise in wage inequality. Among women, union decline explains about a fifth of the rise in wage inequality. Rising educational inequality in pay explains nearly twice as much. In short, the effects of deunionization on inequality are only half as large as the effects of education for women, but union and education effects are equally large for men.

Western and Rosenfeld go on to make a broader point:

The analysis suggests that unions helped shape the allocation of wages not just for their members, but across the labor market. The decline of American labor and the associated increase in wage inequality signaled the deterioration of the labor market as a political institution. Workers became less connected to each other in their organizational lives, and less connected in their economic fortunes. The de-politicization of the American labor market appears self-reinforcing: as the political power of organized labor dissipates, economic interests in the labor market are dispersed and policymakers have fewer incentives to strengthen unions or otherwise equalize economic rewards.

I should stress that, as Farrell acknowledges, Ramesh explicitly said that inequality has increased, and I suspect that he thinks that the “immiseration” of the middle class and a deterioration of the position of middle earners relative to workers at the top aren’t the same thing. An alternative definition of “immiseration” might include declining living standards for comparable workers from period 0 to period 1.

Farrell’s interpretation of Western and Rosenfeld leads him to the following conclusion:

In short, there is good prima facie support for the claim that deunionization has hurt the middle class by contributing to a particularly top-heavy form of increased income inequality, from an author whom Ponnuru cites and presumably takes seriously. Very likely he wasn’t aware of this recent work.

It is certainly possible that the middle class is hurt by rising inequality, e.g., via intensified positional competition. But I’d want to have a better sense of the counterfactual. The U.S. prime-age population of the present is different from the prime-age population of thirty years ago. Household composition has changed, and we’ve seen a broad trend in which affluent families embedded in social networks with high levels of human, social, and cultural capital tend to have fewer children than the less advantaged. The youngest cohorts entering the workforce have educational attainment levels comparable to much older cohorts, which speaks to the stagnation of educational attainment levels. This has occurred against a backdrop in which we’ve seen dramatic dispersion in productivity levels across firms within the same industry, as economist Chad Syverson has observed. If we assume that all of these factors have spillover as well as individual effects, the fact that we haven’t seen pronounced deterioration in absolute terms represents a more interesting puzzle than is commonly understood. (Yes, that’s right. I’m arguing that in a competitive global economy in which capital and goods move relatively freely, and given the extraordinary inefficiency of the U.S. public sector, the extent of family disruption and mass incarceration and the impact of both on human capital accumulation, etc., it is easy to imagine economic outcomes for Americans at the median and below having been much worse over the past thirty years.)

Moreover, one wonders about whether Western and Rosenfeld are giving us an adequate picture of the counterfactual labor market in which high levels of unionization persisted. Would the economy have experienced the same or greater productivity gains, only with a greater compression of wages? Or would domestic firms have accelerated the offshoring of core production in tradable industries, leaving employment even more heavily skewed towards nontradables like health, education, accommodations and food, and miscellaneous public employment?

Farber and Western circa 2000 are not the only social scientists to have argued that competition has played a role in unionization levels. Adam Ozimek has summarized some of the relevant issues here.

As I read the Western and Rosenfeld paper, I noticed the following passage:

US union decline is often explained by changes in the economy and intensified political conflict in the workplace (Farber and Western 2001; Freeman and Medoff 1984; Goldfield 1987). In this account, union firms could not respond to 1970s stagflation, industry deregulation, and eco- nomic globalization. The biggest driver of decline in the percentage unionized was employment growth outside of the traditional union strongholds of manufacturing, construction, and transportation, utilities, and commu- nications. Faced with a newly competitive economic environment, em- ployers in unionized industries also intensified their opposition and union employment and new organizing—at least through union elections—plunged through the 1980s (Tope and Jacobs 2009; Hirsch 2008).

Employer opposition unfolded in an increasingly adverse political context for labor. Hacker and Pierson (2010, chap. 5) report that an influx of corporate donations influenced policymakers to oppose pro-union reform of labor law in the 1970s. Union reform efforts were defeated, and the rightward shift of the National Labor Relations Board under Republican administrations made organizing more difficult in the 1980s. Political defeats in the 1970s and 1980s yielded an “enervated” labor law that enabled employers to block organizing campaigns and weaken existing unions (Cowie 2010, 288).

Of the citations above, I’ve only read Hacker and Pierson. Though Hacker and Pierson are well-regarded scholars, there is some dispute as to whether they gave a complete picture of how important the influx of corporate donations was relative to broader cultural changes that may also have contributed to deunionization. As with all qualitative work, the facts in question are subject to a number of interpretations. One doesn’t need to believe that the community of sociologists focused on the role of organized labor is fairly homogeneous in its interpretive instincts to believe that it helps to draw on alternative perspectives.

One quick point, with regards to the following from Henry:

However, ‘workers in general’ seems a rather unusual proxy for ‘immiseration of the middle class’ given the high degree of variation in employee compensation (those in the highest income segments, where inequality has increased most dramatically since the late 1980s, cannot plausibly be described as ‘middle class’). Nor is inequality, as Ponnuru seems to suggest, the product of more money going to those with higher skills. Western and Rosenfeld’s figures suggest that the decline of unions has been just as important a factor as education in explaining the rise of inequality among men.

I agree that “workers in general” is an unusual proxy. But as for the latter claim, that inequality is “the product of more money going to those with higher skills,” it strikes me as an amusing oversimplification of the real sources of wage dispersion, which, again, reflect organizational spillovers, among other things. A skilled worker in isolation, or even in a firm that hasn’t done a good job of investing in organizational capital, isn’t likely to fare well in a modern economy.

One good place to look for (non-ideological) insights into managerial practice and how it varies across economies is a 2010 paper by Nicholas Bloom and John Van Reenen in the Journal of Economic Perspectives on “Why Do Management Practices Differ across Firms and Countries?” (Actually, I recommend rooting around Bloom’s Google Scholar page in general.) Among other things, Bloom and Van Reenen find that the uses of incentive-based strategies and monitoring-based strategies vary across countries, depending on the nature of labor market regulation.


The Latest