For a number of editions of his principles text, Paul Samuelson (1964) starts the first chapter with a gestalt diagram. The lines and points are drawn such that some people look at it and see antelopes while others see flamingoes.
The previous issue of this journal featured an article by Mitchell Langbert, “The Left Orientation of Industrial Relations” (2016). Langbert presents his paper as a case study portrait of near-universal left dominance among academics in the social and behavioral sciences and an alert to “students, parents, taxpayers, policymakers, and citizens,” who “need to know” about this (2016, 46). If I can paraphrase, Langbert looks at social science professors in universities and sees a large flock of bright pink flamingoes which stand on their left legs, bury their heads in the academic water, and slowly ruin the pond with counterproductive, left-biased research and teaching. Against this overall background, Langbert focuses attention on the industrial relations (IR) field and locates it on the distinct left side of the flock, considerably past the pale-pink economist flamingoes and near the flaming-pink sociologist flamingoes. The IR flamingoes stand together in a relatively small and shrinking collective group, distinguished by distinct traces of reddish social-democratic plumage.
But beauty is in the eye of the beholder, all human stories have two sides, and sometimes one person’s flamingoes are another person’s antelopes. When I read Langbert’s account of industrial relations, I see some of the same flamingoes, although since I am closer to the pond (Kaufman 2004) many of the flamingoes look well on in years (few under 40 human years) and are less pink and active than in days gone by. For the same reason, I do not feel as alarmist about the flamingoes’ left-winged habits nor am I as put off by their clannishness. Actually, as the flock gets smaller each year, some pond watchers wonder whether the flamingoes are headed for an extinction event, brought on by a harsher environment, more aggressive right-winged predators, and failure of the species to evolve and reproduce. Langbert, I get the sense, is happy (or not unhappy) to see the flock get smaller, perhaps as a healthy and overdue culling. But my sense is the union of flamingoes is necessary to maintain ecological balance and, without them, it could be a backward step to the law of the jungle where the powerful creatures dominate the weak, might makes right, and too much habitat competition can precipitate a destructive race to the bottom.
For context, I helped Langbert organize a session on ideology in industrial relations at the 2014 annual meeting of the main professional group, the Labor and Employment Relations Association (LERA, formerly Industrial Relations Research Association, IRRA). Then, early this year, he sent me a copy of his published paper. After reading it, I sent Langbert an email with congratulations and a set of plus and minus comments, presented as ‘if I had been the referee.’ He forwarded my email to EJW editor Daniel Klein, who invited me to turn it into a comment paper. Shortly after the editor’s invitation, Langbert emailed and also encouraged me to take on this assignment. As the old adage says, however: Be careful what you wish for!
I divide my reviewer’s report into three sections. The first section deals with the paper’s conceptual and theory side; the second looks at its empirical and inferential side; and the third examines the role of external funding and ideology. Part of my review directly engages with Langbert’s paper as he wrote it; another part broadens and reinterprets his analysis with an alternative conceptual framework meant to not only sharpen and reframe the critique but also engage and take forward the larger literature on academic left ideology. However, what I believe provides extra insight may look to him as reframing his paper in a way neither intended nor accurate.
Plot redux:
Wrong turn at the ideological fork in the road
We know that at an automobile crash scene the police officers often hear quite divergent eyewitness accounts of which driver did what, and when. Langbert looks at the industrial relations field, sees an academic crash scene, and reports to the readers of EJW that the guilty parties are the IR professors who persistently take the left-hand fork of the road even though it is clearly marked with flashing red lights and signs marked “Road to Serfdom” and “Poor House Dead Ahead.” Langbert mostly leaves it to his readers to infer what industrial relations professors should have done if they had instead practiced invisible-hand, value-free driving, although he cites professors in human resource management, marketing, and business strategy as a better role model (2016, 52).
But one surmises (with some tongue in cheek), certainly from some of the citations, that Langbert thinks IR professors should have instead taken the right-hand fork and traveled down the free-market, free-enterprise interstate (Friedman-Hayek Expressway, Inc.?) where all the signs and flashing green lights point to the shortest route, and also fastest without speed and safety laws and other ill-designed and intrusive government interferences, to the twin-cities destination of Libertyville and Wealthy Nation.
And, by taking this route, then, as a serendipitous unintended consequence, the IR professors would eventually see rising in the distance, as the union scales fall from their eyes, the Rocky Mountain-like majestic beauty and natural law truth of Adam Smith’s obvious and simple system of natural liberty. With the benefit of 20/20 hindsight, wisdom of the free-market, free-enterprise saints (Gospels of Adam, Ludwig, Friedrich, Milton, etc.), and born-again faith in the beneficence of the competitive Holy Ghost, these professors belatedly realize that the entire subject of industrial relations, and all of its counterproductive and coercive unions, minimum-wage laws, and other restrictions and exactions, is in the end an ill-considered and anti-social detour into leftist defective thinking that can be dropped from universities with savings for all.
Langbert has here an interesting plot. To shift metaphors, it is like going back to the Cold War period and imagining the socialists/communists take over Radio Free Europe, rotate the transmitters from east to west, and start beaming left-wing propaganda into American universities. Either because they are collectivist true believers or social-democratic dupes, industrial relations professors are among the most active and enthusiastic front-men (not too many front-women in IR) for the leftist, totalitarian cause. Before you know it, formerly contented Walmart workers are out on strike and the Tea Party is demonstrating in front of the White House for single-payer healthcare. Maybe as a really clever plot twist, the European socialists send a coded radio message that activates a sleeper cell of radical sociologists who late at night switch the highway signs so the arrows to Libertyville and Wealthy Nation point down the left fork.
OK, I overdramatize, but the reader gets the idea. With plot summarized, the question is whether Langbert can fill it in with a compelling script that convinces readers and satisfies critics. Here is my reviewer’s report.
IR left ideology: Concept and theory
Langbert says the ideology of industrial relations “leans overwhelmingly to the political left” (2016, 46). Other descriptors are left-oriented, tendentious, social democratic, and impervious to opposing viewpoints. The main objective and contribution of his paper is to empirically document IR’s political-ideological leftism, with some conceptual undergirding and explanation. A pre-existing literature on normative values and ideology in IR exists but does not get included (e.g., Godard 1995; Kochan 1998; Budd and Bhave 2008; Darlington 2009). It too documents a left orientation, so no challenge to the Langbert thesis here—it’s conventional wisdom—but with considerably different framing and explanation which could have usefully informed and rounded out his story.
A larger lapse, after crunching all the data and demonstrating left ideological orientation, is that Langbert does not provide a better-developed and more compelling answer to the ‘so what?’ question. He opens his article with the question “Why bother?” and answers in the next sentence, “One reason is that students, taxpayers, policymakers, and citizens, all influenced by academics, need to know that practically all academic fields in the United States are preponderantly left oriented” (2016, 46). But why do they need to know?
Langbert tells us that left ideology leads to academic groupthink, intellectual blinders and biases, and preaching to the converted and unsuspecting. But this conclusion only kicks the proverbial can down the road. The most we get are several scattered and unsupported claims that leftist ideology in academe leads to “defective” and “wrongheaded” thinking (2016, 46, 50)—manifest in IR as support for labor market regulations and labor unions. A schematic representation is (1) more left ideology → (2) more defective, wrongheaded thinking → (3) more support for unions and regulation.
We have here a model and it leads to hypotheses and policy implications. But I see some problem areas that weaken the argument or need strengthening in the next version.
It is important to carefully define and operationalize key concepts. A case in point is ideology. Langbert does not define ideology. General descriptors he uses are “sentiments” and “value system.” For the empirical analysis, he operationalizes ideology as “sentiment, argument, or evidence” (2016, 63). The alert reader or referee spots here a serious positive apples and normative oranges problem. Ideology is conventionally defined as a belief system that structures choice and guides action.The Cambridge Academic Content Dictionary gives this definition: “a theory or set of beliefs, esp. one on which a political system, party, or organization is based.” It is a normative and subjective concept grounded in psychology, sociology, and ethics. The first two elements in his ideology concept, sentiments and argument, fit this conceptualization. The third element, evidence, does not. Evidence is fact-based information and knowledge; it is a positive, value-free input that goes into creating an ideology. By mixing positive and normative within ideology, Langbert introduces serious conceptual and empirical problems. But more later.
Another lacuna with Langbert’s conceptualization of ideology is that it seems one-sided. A person reads the paper and sees left ideology on practically every page, but the term ‘right ideology’ (or ‘rightist,’ etc.) is not mentioned. If ideology is the main explanatory variable, per the three-step model above, the paper in conceptualization and design seems to suffer from severe truncation bias and penchant to frame the subject as a one-tail test.
A way to visualize the problem is to take some group of people, such as across social science departments or across regions of a nation, create a single-number empirical measure of their ideologies, and graph the data as a frequency distribution. Let’s for simplicity assume it resembles a normal distribution. The left-hand tail contains extreme left ideologies, the right-hand tail contains extreme right ideologies, and in the middle is the mean/average ideology in the population. The mean defines a natural centerline separating left ideologies on one side and right ideologies on the other. Moderates cluster around the centerline; as people become more left- or right-leaning they move out toward the tails. Evidently, each population has its own frequency distribution so that what may be right ideology in one (France, sociology departments) is left in another (the U.S., business schools).
The peculiar feature of Langbert’s conceptualization is that universities generally, and IR specifically, are not given their own ideological centerlines. Instead, Langbert judges them to be predominantly left, presumably relative to the centerline of some other frequency distribution that he has in mind, such as that of all American voters. But Langbert never makes clear the nature of the distribution or reference point he uses to establish left orientation in academe. His resulting depiction of IR scholars is that they generally adhere to a consensus ideology labeled ‘leftism.’ Langbert then moves to characterize non-left IR scholars as dissidents from consensus, rather than simply as comprising the right wing of IR’s ideological distribution. But that is not an accurate characterization. The Journal of Labor Research, and its former editor James Bennett—whom Langbert describes as a “classical liberal” rather than using a directional term—are regarded within IR as being right-of-center rather than dissident. Making explicit reference to a right-hand side of the IR distribution would have been a more accurate and balanced way for Langbert to underscore that non-left IR scholars are not brave escapees from ideological influences but are instead merely possessed of their own, rival ideologies.
Even if Langbert brings in both left and right sides of the ideology distribution, it is still too blunt to satisfactorily frame the location of industrial relations. A popular theory framework in industrial relations is called IR Frames of Reference (see Budd and Bhave 2008). It distinguishes four, not two or one, belief systems (ideologies) about how labor markets and employment relationships work—and should work. The four would be placed along the horizontal axis of the frequency distribution, going in order from right to left. The individualist frame is an atomistic, competitive, free market view; the unitarist frame is an organic, unity of interest, mutual-gain view of organizations; the pluralist frame is a mix of structurally imperfect and power imbalanced markets and organizations with a mix of cooperation and conflict; and the radical frame is monopoly capitalism with class conflict and systemic exploitation and subordination.
I earlier said that Langbert’s study lacked nuance because it did not bring in IR theories and concepts. Here is a prime example. Industrial relations is often defined broadly to include all employment relationships and all four ideological frames; other times, however, it is defined narrowly to include only the pluralist component. Thus, by treating IR as having a consensus left-side ideology Langbert commits two types of specification error. If IR is broadly defined and includes all four frames, Langbert mistakenly makes IR too narrow and left-biased by excluding the right-leaning individualist and unitarist frames. On the other hand, if IR is defined narrowly to include only the pluralist frame, Langbert mistakenly makes IR too leftist by including the radical frame in the tail of the distribution. Regardless of which way IR is defined, most people in the field would regard it as much-too blunt and inaccurate to characterize IR—of the American variety—as subsuming the left side of the ideology continuum. British IR, on the other hand, mostly excludes the individualist and unitarist frames but includes the radical frame. Hence Britain is a better fit with Langbert’s left-side IR version (Darlington 2009).
To better indicate the full extent of IR’s leftward tilt, Langbert compares the IR ideology score he calculates, based on political party affiliation, with similar scores for economics (Econ) and human resource management (HRM). The details are covered in the next section; the point of interest here is that all three are to the left of Langbert’s centerline with HRM least left, Econ moderately more left, and IR considerably more left. As a reviewer, I put a plus mark on the paper’s scoresheet for presenting an interesting and new empirical finding. Then, I look in the paper for some kind of theory or model that gives insight or formal explanation for this ordering. For example, instead of HRM, Econ, and IR, why isn’t the order Econ, IR, HRM? Likewise, later in his paper Langbert is able to empirically rank IR journals and publications by degree of left orientation: Journal of Labor Research (JLR), the LERA annual proceedings, Industrial Relations (IR), Industrial and Labor Relations Review (ILRR). What accounts for this order? On this aspect I have to put a minus on the scoresheet.
Langbert devotes several pages to theory ideas, centered principally on a model by Daniel Klein and Charlotta Stern (2009a). As referee, I enter another plus on the scoresheet for giving some theoretical structure to the empirical analysis and, likewise, I give a plus mark to the model itself for illuminating why, once an ideology gets established in an academic field or organization, it tends to get locked in as part of the DNA of the group and resists internal change or external challenge. Key factors are majoritarianism in departmental hiring and promotion, which allows the dominant faculty group to maintain and expand control. Then, ideological conformity is solidified by factors such as left groupthink, left elite control of journals and top-ranked departments, and status emulation by people in the lower ranks. The explanation, as one can tell, is largely sociological. This aspect is fine for any IR person since sociology is part of the field’s multidisciplinary coalition. A drawback, however, is that the model is largely silent on the antecedent question of how and why a particular ideology first gets established as the dominant belief system. Once the group has been cemented together, the social glue of solidarity also seems remarkably resistant to familiar forces of competitive undercutting, free-rider problems, and defection or non-cooperation.
A good referee not only points out shortcomings but also helps the author strengthen and clarify the theory part of the argument. Here are some suggestions with not only Langbert’s paper in mind but also the broader literature on academic left ideology.
The schematic representation, sketched above, is (1) more left ideology → (2) more defective, wrongheaded thinking → (3) more support for unions and regulation. As Langbert says on the bottom of the first page, the sociological model does not explain why a left ideology gets established but it does explain why it persists. In effect, then, Langbert takes dominance of left ideology as an exogenous stylized fact and devotes his analysis to the rest of the sequence.
I find it helpful in working through the ideas in the paper, and to also make progress on the antecedent issue of what factors shape a right-leaning vs. left-leaning research stance, to express them in terms of a regression model. The equation I come up with is of the form: Truth Gapi,j = β0 + β1Dj + β2Ideoi,j + β3Xi,j + ε.
The dependent variable, Truth Gap, is measured across i professors working in j social science/humanities fields and disciplines. It represents the deviation, right toward excessive individualism and left toward excessive collectivism, between what they publish in research journals and teach to students and what represents value-free, objective truth on the subject. The notion of objective truth corresponds to Milton Friedman’s (1953) conception of positive economics; that is, a value-free representation of the actual process in the economy generating the data. It also seems to correspond to Langbert’s goal of getting ideology (or just left-wing ideology?) out of academic research so it is no longer defective and wrongheaded.
The control variable D is a dummy variable across the j fields to control for field-specific fixed effects, say with D1 = IR set to zero as the omitted field. If, for expositional purposes, the only other field is HRM (D2 = HRM), then the expected sign is β1 < 0. That is, ceteris paribus, the (alleged) truth gap in HRM is smaller (β0 − β1D2) relative to IR.
The central explanatory variable is ideology (Ideo). Following Langbert, it is measured in two ways. In the first specification, ideology is measured using a discrete interval scale centered on 0 (value-free, or pure positive science) and bounded on one side by −5 (extreme right ideology) and +5 (extreme left ideology) for each of the i,j professors. In the second specification, Ideo is a categorical 0, 1 measure, say with Republican = 0 and Democratic = 1. (Measurement details in the next section.) The predicted sign is β2 > 0; that is, ceteris paribus, the larger the ideological factor the greater the distortion in research and teaching and the larger the size of Truth Gap (in absolute value). Likewise, if Democrat = 1 then Truth Gap increases. The variable X is a vector of control factors, such as gender, public/private university, etc. The error term ε is not further considered.
This regression model reinforces some of the points made above, and leads to additional insights and areas for further work. For example, the referee’s expectation is that when Ideo = 0 (value-neutral) then Truth Gap = 0 (no distortion from defective thinking) and, as Ideo takes on more extreme values, say −5 or +5, then thinking becomes progressively wrongheaded and the size of Truth Gap deviations increase. The problem, however, is a disconnect with the prose, conceptual part of Langbert’s argument where, as noted above (the one-sided problem), only left ideologies (apparently) cause defective thinking and Truth Gap deviations. In other words, positive values of Ideo (left IR) cause positive Truth Gap deviations but negative values of Ideo (Friedman? Hayek? Bennett?) cause no similar defective thinking and thus Truth Gap remains at value-free 0.
Looking back at the regression model, a person also has to wonder if Langbert’s findings and conclusions are distorted by a serious omitted variable problem. The only explanatory variable in the model is Ideology and, in particular, left ideology. But, really, is social science research, even in industrial relations or sociology, so cravenly political and reflexively groupthink-ish that support for unions and regulation has no other credible explanation than ideologically distorted, defective thinking? An alternative idea, at least worth thinking about, is that IR support for unions and regulation comes not from normative ideology but positive science. For example, the IR literature (and portions of labor economics) for the last century conclude that labor markets are among the most competitively imperfect and non-commodity like in the economy (Budd 2004; Kaufman 2010). A further conclusion is that the forces of demand and supply, or call it bargaining power, are in the general case tipped in favor of employers’ interests and against labor’s interests.
A reasonable person may well apply a discount/correction factor to this research to remove ideological bias. It seems unreasonable, however, to set the correction factor at 100%, or near it. Here I appeal to authority. It may be unusual to consider Adam Smith an IR economist, but the depiction of labor markets in The Wealth of Nations accords well with the description just given (Kaufman 2016). Smith seems to take a very IR leftist stance, for example, when he declares, “When the regulation, therefore, is in favor of the workmen, it is always just and equitable” (1976/1776, 157–158). He comes to this seemingly radical conclusion because employers in the normal case have the bargaining advantage in wage determination. He states, “It is not, however, difficult to see which of the two parties must, upon all ordinary occasions, have the advantage… the workman may be as necessary to his master as his master is to him; but the necessity is not so immediate” (ibid., 83–84). The tipped nature of labor markets is also noted by Richard Ely, a root stem founder of industrial relations. Ely observes, “Through contract the actually existing economic forces manifest themselves with all their inequalities and injustices. When economic forces make possible oppression and deprivation of liberty, oppression and deprivation of liberty express themselves in contract” (1908, 139). Langbert mentions Ely, but as an illustration of the symbiotic connection in IR between leftist ideology and support for unions and regulation. My reading of the Ely quotation, on the hand, suggests that Adam Smith could have equally well written it.
With these ideas in mind, I would encourage Langbert in a future paper to add an IR science variable to his model. Call it Scie. Conceptually, Scie measures, perhaps on a −5 to +5 scale, what IR theory, empirical evidence, and historical experience indicate is the optimal level of regulation and union density (−5 = very low, +5 = very high). The expanded regression model is: Truth Gapi,j = β0 + β1Di + β2Ideoi,j + β3Sciei,j + β4Xi,j + ε.
Truth Gap is the difference, for example, between what IR professors claim about the optimal level regulation and union density and what the actual optimal level is. In the Langbert model, positive and typically large values of Ideo lead IR professors to substantially inflate the socially optimal level of regulation and density and, hence, Truth Gap is large. However, it surely is going too far to suppose that the IR professors give 100% weight to ideology and none to science in making recommendations on optimal regulation and union density. Thus, if Scie has any truth content, and if IR academics give positive weight to Scie in their mental models, the inclusion of an IR science variable must improve the fit of the model—that is, reduce Truth Gap. Likewise, if Scie and Ideo are positively correlated, omitting Scie from the regression leads to upward bias in the estimated effect of ideology. Langbert, of course, may save his case by arguing that Scie is 100% driven by Ideo (but, note, it could also go the other way), making the two variables completely collinear so Scie has no explanatory power. I am going to guess, however, that not even for the most ardent sociologist is the relationship 100%.
IR ideology:
Empirical measurement and evidence
The larger part of Langbert’s paper, and its most important area of contribution, is empirical measurement and analysis of IR ideology, and analysis of its variation across journals, universities, and leadership/status positions. Let me acknowledge, before going further, that it took Langbert a good deal of time and effort to assemble these data, and problem areas notwithstanding, he has made a contribution.
The empirical part of the paper is to work up measures of IR ideology. Here Langbert uses voter registration (after the manner of Cardiff and Klein 2005), political donations (after the manner of McEachern 2006), and his own scoring of content in the four IR periodicals.
Langbert measures the ideological position of IR academics on two different dimensions. The first is their Democratic vs. Republican political party affiliation and support; the second is the extent to which published IR research demonstrates a favorable (pro) vs. negative (anti) stance on, respectively, labor market regulation and labor unions. Consider first the party affiliation measure.
Langbert uses voter registration data from 30 states to determine if IR academics belong to the Republican or Democratic parties (counting Libertarians as Republican and members of two minor left parties as Democrats). He also uses information from the Federal Election Commission to identify the IR academics who have made financial contributions to either party. The sample of IR academics is people with at least one publication between 2009 and 2013 in ILRR, IR, JLR, or the LERA annual proceedings, and also LERA officers and journal editors and editorial board members (720 person observations, 916 population observations).
Langbert finds a Democratic-to-Republican ratio for authors publishing in the three “mainstream” IR outlets, ILRR, IR, and the LERA proceedings, plus LERA officers and members, that is 13:1; for the JLR authors, it is 3.4:1 (JLR is a newer and more conservative-leaning journal). When Langbert drops from the data set people included two or more times because of multiple publications, the ratio falls to 8.2:1. The data also show that the Democratic-to-Republican ratio for authors is an increasing function of school ranking (higher ranking → more Democratic) and is also higher if a person is a journal editor, editorial board member, or LERA officer.
The implications are, first, that the IR mainstream is quite leftist (Democratic) and becomes more leftist as one ascends the field’s institutional/status pyramid. The leftist orientation of IR is reinforced by comparisons with other research fields. A study of faculty in California universities found, using similar voter registration data (for 2004), a ratio of 1.3 Democrats to every 1 Republican for business school professors (Cardiff and Klein 2005). For economics the ratio is roughly three Democrats to one Republican and for all humanities and social sciences is in the range of 8 to 10 Democrats per Republican (Klein and Stern 2009b; Langbert 2016, 46 n.2). With regard to financial donations by participants in ILRR, IR, and LERA, the Democratic-to-Republican party ratio is 22:1; for the JLR it is 1.1:1.
The take-away conclusion is that for every one Republican among traditional IR academics there are roughly eight Democrats and, if weighted by number of publications, the ratio rises to thirteen Democrats for every one Republican. The implication, seemingly unavoidable, is that here is a sharply leaning left-oriented field. But really? Here are two pieces of contrary evidence.
If IR academics, and academics in general, are left-leaning then the policy/political positions they support presumably are noticeably left of the electorate. But, with respect to labor regulations and labor unions, the IR academics and the general electorate are not far apart.
Illustratively, Americans have for three decades (1987–2015) given an average 78% favorable response rate to a question that asks about an unspecified increase in the minimum wage. The favorable response rate only slightly declines (74%) when the question specifies an increase of 26%–50%. One reason for such a favorable response is that 60% believe that a minimum-wage increase would either have no impact on jobs or increase jobs; another reason is that 52% believe a minimum-wage increase would improve the economy (Roper Center 2015). A recent Gallup poll (Saad 2013) found that even among self-rated political conservatives 54% favor an increase to $9 per hour; among liberals the proportion is 94%.
Next consider public opinion on labor unions (Saad 2015). Over the same three-decade period, the favorable rating given to labor unions remained basically the same: 60% in 1987 and 58% in 2015. More respondents also said they wanted unions to have more influence than said they wanted them to have less. As with the minimum wage, Democrats are far more favorable toward unions (79%) but a substantial minority of Republicans also gave a favorable rating (42%).
I am not trying to marshal data in support of either unions or minimum-wage laws. What I am raising for consideration is whether characterizing IR academics as far to the left in the political spectrum is a meaningful generalization when large majorities of Americans express similar relatively strong favorable union and minimum-wage preferences. Actually, most at odds with Langbert’s left-dominated ‘defective thinking’ thesis is that 62% of employers, when asked their opinion on raising the minimum wage, give a favorable response (CareerBuilder 2014). The proportion rises to 70% among employers who hire minimum-wage workers.
I cannot locate a poll of management professors on unions or minimum wages so unfortunately it is not possible to compare them relative to the American public. The Booth Business School at the University of Chicago, however, has a panel of nationally recognized economists and they were polled about the minimum wage (IGM Economic Experts Panel 2013). A slight majority (52%) gave support to a minimum-wage increase, albeit of small size (to $9 per hour). Asked about the employment effect of an increase in the minimum wage to $15 by 2020, 38% were uncertain, 26% said it would reduce employment. Only 4% agreed that a minimum-wage increase would be a significant stimulus to output growth. Compared to the general public, therefore, these poll numbers suggest that the average economist is noticeably more ideologically-politically conservative with respect to labor/employment policy issues. Recall, however, that Langbert cites a study of California universities that concludes economists are in the moderate political left—because the Democratic-Republican ratio was 3:1. But, compared to the public’s position on minimum wages, economists are clearly to the conservative right of the public.
Also recall that the left-right ratio for economists in the California study is 3:1 but not far from 1:1 for business school professors. So, by logical inference, if economists are actually significantly right-oriented relative to the national average, and if economists are three times more liberal than business school professors, where does this put management professors, such as Langbert? I think the conclusion is, with some irony, that IR professors are actually located in the broad middle of the American ideological spectrum (plausibly moderately center-left) while economists are located toward the right side and management professors, as a group, are probably a clear-cut case of ‘right dominance.’
An anomaly lurking in the last several paragraphs also needs to be resolved. The anomaly is that party affiliation data indicate economists are moderately left, given the 3:1 ratio of Democrats to Republicans, yet as demonstrated above the polling data indicate economists are to the right of the public on labor/employment issues. My proffered resolution is that the Democratic vs. Republican dichotomy is not a close approximation of the 50-50 political-ideological center line in the American electorate. In particular, the ideological center of gravity of the Republican Party has steadily drifted rightward to the point that its right-side boundary has moved into the right-extremist tail of the frequency distribution and the left-side boundary, once near or overlapping the center (thinking of Dwight Eisenhower, Nelson Rockefeller, etc.), has also shifted rightward and no longer is near the center, at least in a numerically significant sense (Ellis and Stimson 2012). Thus, I conjecture, some center-right academics, two or three decades ago comfortable members of the moderate-liberal wing of the Republican party, have since the 1980s and 1990s seen the party shift increasingly right to the point that their ideological position no longer fits and they become independents or Democrats. Party affiliation, therefore, is an increasingly misleading measure of left vs. right ideology.
Adding to the inferential problem, educational attainment and voting Democratic are positively associated so one would naturally expect Ph.D. holders to be predominantly Democrats (Pew Research Center 2015). Presumably these Democrats, with their high level of education, are better informed and make better decisions. However, Langbert suggests just the opposite—that the Ph.D. holders get hired into leftist groupthink university departments and the warped ideology they acquire leads to deteriorated, incorrect thinking.
A second measure of academic leftism presented by Langbert is a scoring of 539 IR journal articles on a −5 to +5 scale with respect to ideological orientation on labor unions and employment regulation (−5 = most anti, +5 = most pro). The journals, identified earlier, are ILRR, IR, JLR, and the LERA proceedings. Here too was invested much time and effort, as Langbert read each article, wrote a summary description of the main findings (in Excel files on a web link), and determined respective union and regulation scores. Inevitably this process has a significant subjective element and an error term.
He finds a ratio of pro-regulation articles to anti-regulation articles of about 5:1 and a pro-union vs. anti-union ratio of about 7:1. Authors with a Democratic party affiliation wrote more of the pro-regulation and pro-union articles. The three mainline IR publications grouped together with high pro-regulation and pro-union scores, while the newer and more right-oriented JLR (established 1981, compared to ILRR in 1947, LERA proceedings in 1948, and IR in 1961) was a neutral outlier. As with the other parts of the paper, these data tell a plausible story and reveal left vs. right regularities that are probably expected, or certainly not unexpected. And, again, probably the most eye-catching finding is the quite large pro-regulation and pro-union left tilt in the articles.
On the other side, no one in the IR field will be surprised by the right-side positon of the JLR. It was created to provide a research focal point and publishing outlet for right-of-center IR/labor researchers, but with a proactive outreach to people from across the field. A look at JLR tables of contents, particularly in earlier years, reveals an ideological distribution of authors with large, ecumenical variance but right-of-centerline mean and Langbert’s data nail down the conclusion. If Truth Gap is the difference between mission statement and tables of contents, the JLR scores a close-to-zero value.
If another reader considers using this type of scoring system, I counsel some revision. I looked through the spreadsheets where Langbert very helpfully lists every article, provides a brief summary review, and gives the two scores on regulation and unions. What stands out is a major interpretative problem. The intended purpose of the scores is to measure the normative, political-ideological dimension embedded in the articles. But this is not what is actually measured. The crucial sentence says, “I scored each article on a scale from −5 to 5, where negative scores indicate anti-regulation or anti-union sentiment, argument, or evidence” (Langbert 2016, 63, my emphasis). As noted earlier, ideology and normative values are present in sentiment and argument, but evidence, certainly in the spirit of scientific research, is by nature objective, factual, and positive. As an example, a free-market researcher may be resolutely opposed to minimum-wage laws and unions, and yet in good faith summarize in the literature review section of a journal article all the arguments in favor of minimum wages and unions and then, in the evidence section, fully report regression results which, unexpectedly, seem to support the pro-minimum wages, pro-unions side. The question is: How does this article get scored, and will the score give a meaningful measure of left vs. right ideology?
As examples of the problem, consider the articles by Hristos Doucouliagos and Paul LaRoche (2013) and Boris Hirsch, Thorsten Schank, and Claus Schnabel (2010). The Doucouliagos and LaRoche paper is titled “Unions and Innovation: New Insights from Cross-Country Evidence.” The authors of the paper did a meta-regression analysis of the effect of unions on firm-level innovation. My evaluation is that the article is entirely factual and has no ideological or normative content. But the main finding from the empirical analysis has a definite negative ring. In the conclusion, Doucouliagos and LaRoche (2013, 487) summarize “All the available evidence indicates that unions are associated with a depressed level of innovation.” Based on this statement and others, Langbert gives the paper a maximum anti-union score of −5.
The Hirsch, Schank, and Schnabel paper is titled, “Works Councils and Separations: Monopoly, Voice and Insurance Effects.” Langbert scores it as maximum pro-regulation and pro-union, +5, +5 (a works council is not a union, but perhaps this is a detail). It is a standard-type IR empirical paper with literature review followed by regression analysis of German works councils. This paper is also entirely factual and has no normative slant that I can detect. In addition, the +5 and +5 scores seem inflated and ambiguous. For example, the paper finds that works councils reduce turnover, have a monopoly-like positive effect on wages, and a weak voice effect on productivity. These effects are mixed, not uniformly positive for unions, and also not clear-cut for any particular ideology. For example, lower turnover may signal the works council or union has made the firm a better place to work or, alternatively, signal they are protecting the deadwood and restricting management’s flexibility to terminate and discipline.
To my surprise, I ranked in the worksheets as the most consistently ‘maximum left’ researcher in industrial relations. Langbert gave five of my papers scores of +5, +5. I am happy to rank #1 in something, but I never would have guessed it is as the most leftist person in the field. I am in fact one of the more ideologically conservative members, per my several previous joint projects with JLR editor Bennett. The AFL-CIO will also find this ranking perplexing since among organized labor’s most despised employer practices is the company union, and I have raised many IR eyebrows to the −4 or −5 level by taking up the cause of company unions in books and articles. My articles are also being scored +5, +5 by Langbert in part for content/evidence reasons that have no correlation with my personal ideology. Illustratively, one of my articles on the worksheets is a history of thought piece on the labor market theory of Fabian socialists Sidney and Beatrice Webb, and the +5, +5 score seems a case of guilt by association.
So, my own case study points to substantial measurement error. As another case study, Morley Gunderson of the University of Toronto, a well-published labor economist and IR participant, also has multiple entries in Langbert’s worksheet. Gunderson’s mean regulation score, modestly rounded up, is +4. The score strikes me as a considerable over-estimate. To double-check, I emailed Gunderson and requested that he self-rate his regulation and union ideology on the −5 to +5 continuum. (I did not reveal the purpose nor the article.) He replied with a 1, 0 score. I would probably self-rate as 2.5, 1. Perhaps, one concludes, too much Truth Gap error exists in these data to be comfortable.
IR ideology: The financial side
Langbert devotes a section of his paper to “The financial dimension” and, in particular, the influence of external research money in shaping left ideology in academe. He observes, for example, “academic research is influenced by the markets for intellectual products” and “If these interests [the people giving the money] tend to favor government intervention, then a market-based explanation of left-orientation will reinforce the groupthink-based and historical explanations” (2016, 53). The reader notes, however, that this section contains only two paragraphs and the only data provided are a list of nine organizations (three companies, three unions, three joint company-union training programs) which, as sustaining members, contributed between $5,000 and $10,000 per year to LERA. One interpretation is that financial considerations have little explanatory role; my counter-argument is they are important but underdeveloped.
Translated into the microeconomic theory of rational choice, Langbert’s model explains IR’s ideological position as a choice outcome largely driven by left-leaning and sociologically structured indifference curves with little attention to the explanatory role of relative prices, incentives, and endowments, represented by the slope and position of budget constraints. This treatment is problematic because it creates a potential omitted variables problem in his explanatory model, particularly given the widespread presumption in economics that more of the variation in choice outcomes is explained by differences in budget constraints than indifference curves (Stigler and Becker 1977).
A useful exercise is to incorporate into the regression model the budget constraint factors of relative prices and incomes. For expositional convenience, lump them together into a composite variable Fund (dollars of external funding) that varies across the i professors and j academic units. The expanded model is Truth Gapi,j = β0 + β1Di + β2Ideoi,j + β3Sciei,j + β4Fundi,j + β5Xi,j + ε. Incentive effects generated from external funding include both pecuniary and nonpecuniary factors. Examples include: many or no academic job offers, yes or no on promotion and tenure, low or high salary, small or large research resources, few or many external grant opportunities, endowed chair or named school, block grant from a foundation, funding for a research or teaching program, and a large annual travel budget.
Changes in prices and incomes, by rotating or shifting budget lines, may induce (or bribe?) a professor to move the content and tone of research and teaching to the right or left. Likewise, in an era of tight budgets and shrinking government support, universities are in sharp competition to bring in external money which gives the donors, be it company, foundation, or wealthy individual, leverage to shape the research and teaching agenda toward left or right interests. For example, if the big check to the university comes from George Soros, one can confidently predict the research output has, on average, a left orientation to fit the agenda of, say, the Institute for New Economic Thinking. On the other hand, if the big check comes from Charles and David Koch, the research program with high probability tilts to the right to align with the policy agenda of, say, the Cato Institute. Note that the researchers at the two institutions may have identical ideological value systems (indifference curves) but yet, because they face different left vs. right incentive structures, produce research and teaching that exhibits noticeable left vs. right variation. Langbert’s paper runs afoul of this identification problem, leading to probable over-emphasis on the role of ideological predispositions and sociological preference conformance.
Given theory, what evidence can be gleaned from available empirical data? First consider the role of external money in the IR field—where ‘external’ means from a source outside the university system. If the research focus is kept solely on the IR field, the amount of external funding in support of research and teaching is tiny by business school standards (the top twenty B-school gifts since 2000 total more than $1.5 billion) and pennies on the street relative to total university charitable donations ($40 billion in 2015 alone). The external funding trend for IR, unlike business schools and universities, is also trending downward.
As useful historical context, it is interesting to note that the first autonomous IR programs in American universities, starting with the IR Section at Princeton in 1922 and continuing with similar programs in the 1930s at MIT, Michigan, Cal Tech, and Stanford, were largely financed with charitable gifts from John D. Rockefeller Jr. and Rockefeller-connected foundations and philanthropists (Magat 1999; Kaufman 2004). In keeping with the ‘money matters’ hypothesis, all five of these IR programs emphasized the management side of industrial relations (a term then broadly defined to include HRM in nonunion companies), including management-oriented research and executive training.
After World War II, several dozen additional autonomous IR schools and institutes were established, exemplified by the School of Industrial and Labor Relations at Cornell. These new IR programs, although typically outside business schools, also had a management component. But the management component was distinctly secondary in size, prestige, and ideological preference relative to collective bargaining, union-management relations, labor law, social insurance, and dispute resolution, reflecting the post-New Deal surge in union density and labor market regulation. This pattern is also evident in the membership and organizational structure of the Industrial Relations Research Association. It was established in 1947 as an academic-practitioner association with practitioners representing unions, management, and neutrals but with ‘management’ implicitly defined to favor managers from unionized companies.
Both the IR programs and IRRA gradually drifted over the 1950s–1970s to a more focused research/teaching emphasis, and normative support for, the union and collective bargaining sides of the employment relation. In reaction, the management side of IR among academics broke away and reestablished itself in business schools as the largely separate and competitive subject of personnel/HRM with the Academy of Management as the annual go-to conference. Management practitioners, in turn, shifted from IRRA to the Association for Personnel Administrators, later renamed Society for Human Resource Management (SHRM). The relatively few management-side academics that remained in IR, particularly if on the right of the ideological spectrum, often came to feel unwelcome and out of place, much as Langbert describes from his own experience. For the record, I also deplore this narrowing in IR’s positive and normative frame of reference but, unlike Langbert, come to different conclusions and explanations.
This snippet of IR history is critical for appreciating why Langbert can only find two paragraphs of material for his section on the external money dimension. As documented below, very few companies, wealthy business people, and foundations have found it serves their purposes to give more than token amounts to what in practice is a pro-union, pro-regulation field. Similarly, while the union side in its heyday occasionally gave gifts of modest size to IR/labor studies programs (e.g., the Fraser Center and Reuther Library at Wayne State University), the main part of organized labor’s financial support came indirectly by using its political clout in state legislatures to ensure these programs received budget funding. However, as union density in the private sector plummeted over the last four decades from 35 percent to 7 percent, the labor movement lost much of its ability to make gifts to universities and much of the political clout to protect the IR programs’ budget allocations.
Here are a few empirical indicators of the anemic flow of external money to IR. Companies and wealthy individuals have given universities gifts from $10 million to $300 million to endow and name business schools. The most visible example is the $300 million gift to the business school at the University of Chicago from David Booth, CEO and founder of Dimensional Investment, Inc., and which is now renamed the Booth School of Business. I scanned the membership list of the seventy IR/HRM programs which are members of the University Council of Industrial Relations and Human Resource Programs (list provided by the LERA office) for programs that are named after an external funder. Only one appeared—the W. T. Beebe Institute of Personnel and Employment Relations at Georgia State (my university; I was institute director for four years). The university received a gift of approximately $500,000 from Delta Air Lines and, in return, renamed its Institute of Industrial Relations in honor of former Delta CEO Tom Beebe and, in addition, changed the title and research and teaching focus to emphasize personnel/HRM. The take-aways are twofold: only one named program out of seventy points to very low external funding and, second, the Georgia State example illustrates that, inevitably, ‘money matters’ in shaping program content and ideology. (I can testify that Delta practiced complete hands-off regarding institute activities; nonetheless, the gift would not have been made if the program stayed IR-oriented.)
I also asked deans/directors of three well-known IR/HRM programs for information regarding the amount and type of external funding received. Two people replied. Both programs, as broad ballpark figures (neither respondent wants to be identified nor have specific figures cited), receive in the neighborhood of $500,000 per year of external funds, mostly from research grants and contracts, followed by typically small-sized donations from alumni, with no gifts of significant size from companies, unions, or wealthy individuals. Most of the research grants and contracts, they said, are not related to unions and, increasingly, reflect the management/psychological training of the organizational behavior (OB) faculty who are gradually taking the place of IR-trained economists. The two take-aways from this data snippet are, first, that external funding for IR programs is very modest and substantially for non-union research projects and, second, the research/teaching content and ideological orientation of the programs are shifting rightward, due to student enrollment and new faculty hiring favoring HRM over IR, and thus are not as rigid and locked-in as Langbert portrays. Seeing these trends and career implications, few doctoral students major in IR—another ‘money matters’ effect—which reinforces the process of cumulative decline and ideological reorientation.
A third and final piece of evidence comes from external funding received by LERA, the main IR professional association. The organization, like the IR field in general, has suffered serious long-term loss of members with consequent necessity of budget cuts and retrenchments. Also like the field in general, LERA receives very modest-to-minimal external funding. As a ballpark figure (data provided by LERA but with request to not give specific numbers), LERA receives roughly $50,000 per year from around fifteen companies, unions, and non-profits. The amount provided by organized labor is a modest fraction of the total. The take-aways from this example mirror those above.
Now let’s put the external funding situation for industrial relations into broader perspective. Here is where Langbert again presents only one part of the story. The main university competitors to IR programs are business schools, economics departments, and HRM programs. Arguments can be made back and forth on whether these three rivals are also left-oriented as Langbert suggests (citing previous studies); however, what is not in dispute is that these other programs are less left and more right oriented. Data snippets also indicate that these three IR rivals, on balance, receive far greater external funding.
Level of external funding of business schools has already been indicated. I next did a Google web search using the words donations economics departments. Of the ten entries on the first page, eight deal with external funds received by economics departments from a network of foundations and initiatives financed by Charles and David Koch, wealthy oil company founders and libertarian activists. The leading recipient of Koch money is reported to be George Mason University. One of the entries (link) reports that the university received slightly under $90 million over 2005–2014 from Koch foundations, approximately $46 million went to the George Mason University Foundation, and states in the report, “The bulk of this funding [to the foundation] has gone to GMU’s economics department and GMU’s law and economics center.” From 2008–2013, Florida State University received $1.5 million from Koch foundations, largely for the economics department (link). I next did a Google search using the words donations human resource management. The first entry is SHRM Foundation, grant-giving arm of the Society for Human Resource Management. The foundation’s web site provides a list of research grants funded in earlier years and the average annual amount for 2010–2014 is $487,000.
The implication, admittedly based on only a few data points, is that academic and ideological rivals to industrial relations receive much greater levels of external funding. This funding surely has a greater-than-zero influence on the programmatic content and ideological complexion of the universities and individual departments and professors receiving it. If IR is pro-union and pro-regulation, it seems equally certain that the large amounts of external funding for business schools, selected economics departments, and HRM programs shifts the mix of research and teaching toward the neutral-to-anti position. Whether this funding-induced shift (the β4Fundi,j term in the regression) leads to a smaller or larger value of Truth Gap is open to debate, with Langbert no doubt taking the affirmative side but many IR researchers taking the negative side. What I think can fairly be surmised, however, are two additional implications.
First, giving fuller consideration to external funding paints a much different picture than Langbert provides. In particular, it suggests that the funding disparity between IR and more right-leaning rival fields leads both at a point in time and over time to a shift in research and teaching toward an ideological position that is more aligned with free-market economics and union-free management. Second, Langbert should be greatly concerned about the small, declining, and perhaps soon-to-disappear place of industrial relations in universities if, as he says, in academe “uniformity is dangerous” (2016, 47) and causes students to “lack models as to how respectful deliberation can occur” (49). Without IR, where will students learn about the positive side of unions, market regulation, and the social interest in protecting and advancing the conditions of labor? Thus, in my referee’s report, I would ask Langbert to better reconcile his plea for pluralism, inclusiveness, and avoidance of groupthink in academe with the opposing theme that one-half of the ideological frequency distribution—the left-hand side—has (apparently) little-to-no constructive role in universities (per his blanket indictment “political leftism is wrongheaded,” p. 50).
Lastly, and to wrap up, Langbert devotes a section of his paper to “The Journal of Labor Research is different.” He says the JLR “stands apart from the rest of the IR field” (2016, 54). My own view is that the JLR has enriched the labor/IR field and I particularly applaud the journal’s founder and former editor, James Bennett, for his demonstrated commitment to soliciting and publishing papers from across the ideological spectrum. Here is pluralism in word and deed—positively cited by Langbert but with some ambiguity as to whether the reason is the virtue of academic pluralism per se or, alternatively, its instrumental value as a conduit for more of the conservative, right-leaning, and correct-thinking perspective he favors.
Whatever the case, the JLR provides yet another example of money helping to shape the ideological configuration of a field of study. Langbert, for example, features a quotation from Bennett who says, “I went to the Olin Foundation in 1979 seeking funds to start a labor journal….” Langbert, however, does not pursue the Olin connection. The story is that from its 1953 founding the foundation spent hundreds of millions of dollars to promote conservative causes. An article in the New York Observer (Ratner 2005) notes the foundation distributed “grants to conservative think tanks and intellectuals—the architects of today’s sprawling right-wing movement for a half century.” Thus, without the Olin Foundation there would be no JLR and, hence, less progress in minimizing Truth Gap.One also notes that JLR founder Bennett is a professor in the aforementioned economics department at George Mason University.
In closing, this observation leads back to the question raised at the beginning of the paper; that is, whether the small and diminishing flock of pink IR flamingoes can get the protection and resources needed to survive into the future. If I were to emulate Bennett and seek funds for a new IR/labor journal, but pitched in the opposite direction—inclusive but progressive or left-leaning—could I find an Olin substitute to fund the start-up? I suspect it is not likely, and that even in a growing IR market it would be quite difficult to find an external funder. Thus, while Langbert looks at the IR field and sees an overwhelmingly left-dominated gestalt, the gestalt I see is a field increasingly tipped to the free-market/union-free right with the flamingoes in danger of sliding off the edge to extinction, at least in individualist America if not social democratic Europe.