Tracking Homo Oeconomicus: Development of the Neoliberal Beliefs Inventory

Neoliberal Policy and Ideology in the United States

Neoliberalism commonly refers to global and domestic economic and social policies that expand and enrich capital markets while curbing governments’ regulatory and social welfare systems. As indicated by the very names of two cornerstones of U.S. neoliberal legislation, the North American Free Trade Agreement and the Personal Responsibility and Work Reconciliation Act (commonly referred to as “welfare reform”), such policies operate under the pretense of promoting free choice and fair consequences. They are founded on an ideology that supposedly neither favors one group over others, nor presumes equal talents or worth among them. Instead, neoliberalism rests on the appealing premise of rational, self-interested systems – ranging from individual actors to entire nations – vying for resources, opportunities, and success on the same, level playing field. Under such conditions, all consequences, whether rewards or losses, are deemed equally deserved (for thorough delineations of neoliberal ideology, see: Brown, 2003; Duggan, 2003; Harvey, 2005).

Neoliberal policies are often transnational in their scope, yet their reception and influence vary widely across nations and constituent communities and therefore require context-specific analysis. In the U.S., the focal point of our research, neoliberal tenets hew close to hallmarks of national identity, such as rugged individualism and the mythic figure of the self-made man. Neoliberalism’s resonance with these long-standing and prized values may partially explain why it took hold so rapidly and easily in the U.S. It was first introduced as a concerted movement to reverse, through legislative and rhetorical means, social welfare and economic policies instituted to buffet the Great Depression of the 1930s and prevent its recurrence (Duggan, 2003; Harvey, 2005). Beginning in the 1970s, neoliberal advocates mounted the argument that government oversight was no longer required in order to ensure fair competition, and that its regulatory practices were in fact to blame for the nation’s economic stagnation. In the U.S., neoliberalism’s economic claims intermingled with objections to domestic social policies and programs, specifically provisions aimed at reducing social inequalities (e.g., affirmative action and means-tested anti-poverty programs such as Aid to Families with Dependent Children; Duggan, 2003). Social welfare expenditures and civil rights protections were cast as antagonistic to a robust work ethic and a functional meritocracy, and therefore fundamentally antithetical to the nation’s core values. Through a neoliberal lens, state programs and policies were viewed as threatening citizens’ self-determination, whereas previous government interventions (e.g., policies enacted during the Progressive Era and under the New Deal) had been perceived as enabling citizens’ liberty by increasing their access to resources (Zucker & Bay-Cheng, 2010). This position was solidified through the 1980s, and by the 1990s saturated dominant discourse in the U.S. (Duggan, 2003; Harvey, 2005). This manifested not only in the common parlance of key neoliberal phrases such as “personal responsibility” but also in the neoliberal cooption of concepts such as “empowerment,” which were gradually individualized and dissociated from radical, politicized critique (Bay-Cheng, 2012; Riger, 1993; Simon, 1994).

Brown (2003) saw the ramifications of neoliberalism as so pervasive that it resulted in “human being[s] configured exhaustively as homo oeconomicus [such that] all dimensions of human life are cast in terms of a market rationality” (Section 9; see also Harvey, 2005; Stringer, 2014). A growing body of social science research substantiates the cultural suffusion of neoliberal ideology, showing its impact on language (Nafstad, Blakar, Carlquist, Phelps, & Rand-Hendriksen, 2009), social cohesion (Coburn, 2000), and workplace organization (Williams, 2013). Neoliberal ideology also exerts influence on individuals’ behavior (e.g., voting [Allen & Ng, 2000], academic cheating [Pulfrey & Butera, 2013]) as well as their identity and reasoning. For example, neoliberal emphases on self-interest, achievement, and personal responsibility are echoed in contemporary constructions of youth (Harris, 2004; Kelly, 2001), sexuality (Adam, 2005; Bay-Cheng & Eliseo-Arras, 2008; Evans & Riley, 2014), response to trauma and victimization (Stringer, 2014), and feminism (Fitz, Zucker, & Bay-Cheng, 2012). Thus while explicit references to neoliberalism may remain tied to debates about finance and markets, it appears that at least in the U.S., neoliberalism has trickled down and out beyond the bounds of economic discourse and policy.

Related but Distinct: Neoliberal Beliefs and Political Attitudes

Despite the empirical evidence and theoretical grounds for neoliberalism’s influence on psychology, researchers lack a means for measuring its embrace by individuals. We argue that a distinct, explicit measure of neoliberal beliefs is warranted not only because of its salience in contemporary U.S. discourse and culture, but also because it is easily mistaken for substantively different positions. For instance, neoliberal ideology may appear to be aligned with various forms of conservatism (e.g., right-wing authoritarianism [RWA; Altemeyer, 1981], social dominance orientation [SDO; Pratto, Sidanius, Stallworth, & Malle, 1994]), but it also differs from these constructs in its prioritization of self-interest over preservation of the status quo, its relative amorality, and its consequent objection to regulatory practices (Brown, 2006). The neoliberal position is a meritocratic, post-feminist, post-racial one that presumes institutionalized discrimination to be largely eradicated (Harris, 2004). Thus neoliberal and RWA ideologies may take equally skeptical stances vis-à-vis policies to promote the interests of marginalized groups; however, RWA-based opposition – whether derived from authoritarianism, conservatism, or traditionalism – is likely to be framed in terms of defending existing political and social systems against diverse threats (Duckitt & Bizumic, 2013), whereas from a neoliberal perspective such efforts are unnecessarily meddlesome at best and obstructions of meritocracy at worst. Neoliberalism also concurs with SDO’s endorsement of competition and hierarchy; neither aims to achieve equal outcomes or status among groups or individuals. However, SDO’s groupist foundation, that one’s own group is inherently superior to others and deserves dominant status, contradicts the strict individualism of neoliberal ideology, which supports hierarchies that result from free and fair competition among individual constituents. SDO is also associated with an adversarial, disparaging disposition toward others (Duckitt & Bizumic, 2013; Duckitt & Fisher, 2003), while neoliberalism assumes a seemingly sympathetic stance: individuals should be free from all external constraints, whether social bias or governmental interference, in order to strive at will and succeed if deserving.

From such a perspective, the U.S. is often cast as a functioning, post-prejudice meritocracy in which individuals may be masters of their own destinies. The outstanding success of gender and/or racial minority individuals (most notably President Barack Obama and First Lady Michelle Obama, who are not only racial minorities but also were born without class privilege) is often used to confirm that the U.S. largely operates as an equitable, level playing field. This premise also reveals a critical distinction between neoliberal and libertarian ideologies: whereas the latter is predicated on a singular devotion to “negative liberty” (Iyer, Koleva, Graham, Ditto, & Haidt, 2012), U.S.-based neoliberalism makes some allowance, at least rhetorically, for system intervention. The very title of U.S. welfare reform legislation, the Personal Responsibility and Work Opportunity Reconciliation Act, exemplifies this state-citizen accord: the government ensures equal access to opportunity and it is then up to individuals either to seize the opportunity and flourish, or miss it and fail. Unlike libertarianism, which opposes policies such as affirmative action and welfare entitlements as a matter of principle, neoliberalism takes the practical position that they are no longer necessary in the contemporary, post-prejudice U.S. (Duggan, 2003).

Related but Distinct: Neoliberal Beliefs and Perceptions of Control

Even when enduring disparities are acknowledged, these are often described as conquerable through focused, strategic, individual effort (e.g., Sandberg’s [2013] advice that women ought to “lean in” if they wish to excel in business). Duggan (2003) identified the assertion of “‘multicultural’ neoliberal ‘equality’ politics” (p. xii) as the current front for advancing neoliberal ideology in the U.S. The conceit that individuals in the U.S. are generally free agents operating within a fundamentally equitable social system suggests high degrees of both internal locus of control and environmental mastery. However, measures of these constructs do not capture the dimension of neoliberal ideology that links one’s self-concept to views of the social environment. Neoliberal tenets also closely resemble the core constructs of system justification theory (Jost, Banaji, & Nosek, 2004) and belief in a just world (Lerner, 1980; Rubin & Peplau, 1975). These theoretical frameworks tap the deep-seated motivational drives and aspects of personality that compel individuals to express faith in a fair and just world (Carney, Jost, Gosling, & Potter, 2008). Despite many parallels, established measures of system justification and just world beliefs are imprecise proxies for neoliberal ideology, which, as a context-specific worldview, should not be equated with underlying traits or dispositions.

Related but Distinct: Neoliberal Beliefs and Feminist Attitudes

Neoliberalism’s intersection with various forms of conservative thinking and policy is fairly unsurprising and has been critiqued vigorously (e.g., Brown, 2003; Duggan, 2003; Harvey, 2005). Far less expected and investigated are points of ideological coincidence between neoliberalism and feminism (for an analysis of feminism’s assumption of neoliberal tropes vis-à-vis victims and victimization, see Stringer, 2014). Many studies of feminist identity reveal that most women endorse feminist positions without explicitly identifying as feminist (for a review, see Zucker & Bay-Cheng, 2010). These women are often treated as a single group that is inclined toward feminism but wary of being labeled as such. Recent research casts doubt on the presumed ideological comparability among “non-labelers”: Fitz and colleagues (2012) discovered that while some appeared feminist in ideology if not in identity, this characterization did not fit a significant proportion. Instead, the authors speculated that this subgroup of non-labelers was best described as neoliberal insofar as they viewed gender equality as a matter of individual freedom and fair competition (i.e., neoliberal principles), not as a feminist or social justice position regarding women’s rights. Without explicit attention to neoliberalism, it would have been easy for this ideological diversity among non-labelers to go unnoticed, or for support for gender equality to be misattributed to feminist intentions.

The Need for a Measure of Neoliberal Beliefs

There are compelling theoretical and empirical grounds for believing that neoliberalism has taken hold as a pervasive and influential ideology in the U.S., one that comprises beliefs both about system-level functioning (e.g., as a meritocracy in which government intervention runs counter to the productive and natural momentum of competition) as well as individual-level attributes (e.g., the premium placed on personal responsibility, individual freedom, and self-interest). Despite the apparent salience of neoliberalism to attitudes, behaviors, and relationships, we lack a means of formally testing its psychological impact. Given its manifestation in the U.S. as a widespread and diffuse cultural discourse, one that intermingles with an array of other attitudinal and ideological sets, neoliberalism is not fully or accurately captured by discrete indicators (e.g., political party identification, support for particular policies) or measures of related but non-equivalent psychological constructs. Without a direct measure of individuals’ neoliberal beliefs, studies thereof are confined to tentative and impressionistic speculation. We developed the Neoliberal Beliefs Inventory (NBI) as a tool to use in directly studying neoliberal ideology and its psychological impact. Our goal was not to test individuals’ conscious endorsement of neoliberal policies or political platform or to label a particular group of individuals as neoliberal. Instead, we focused our efforts on developing a means for detecting the infiltration and manifestation of a U.S.-specific variant of neoliberal ideology in individuals’ attitudes and views.

Participants and Procedure

We collected data during three semesters (Spring 2012, Spring 2013, Fall 2013) to develop and refine the NBI. All three samples comprised distinct groups of undergraduates at a private university in Washington, DC. We provide basic demographic characteristics of each sample in Table 1. All samples were composed predominantly of white women in their 1st or 2nd year of undergraduate studies who averaged between 19 and 20 years old and who came from affluent families. Participants’ ages and year in school is typical of those drawn from psychology course subject pools, as in the case of the current research, and their socioeconomic background is reflective of students at the private university where they studied. Of those participants in each sample who did not identify as white, 5-7% identified as Black; 3-4% identified as Latino/a; 12-18% identified as Asian/Pacific Islander American; 3-5% identified as Middle Eastern; 3-6% identified as biracial or multiracial; and 3-6% did not select any racial or ethnic category. Sample 2 participants only were asked about their political party affiliation, with 52% identifying as Democrat. Of the remainder, 16% identified as Republican, 5% identified as Libertarian, 25% as Independent, and less than 1% of participants identified as members of the Constitution Party and Green Party, respectively.

NBI development proceeded in three stages: 1) item pool generation and exploratory factor analysis (EFA; Sample 1); 2) confirmatory factor analysis (CFA; Sample 2); and 3) assessment of the final scale’s reliability (Sample 3), internal consistency (all 3 samples), and validity (Samples 2 and 3). At all stages, we collected data using online surveys and participants received course credit. To examine test-retest reliability, Sample 3 participants completed the NBI twice (referred to as Time 1 and Time 2), separated by a two-week interval.

Measures

Preliminary Item Analysis and Exploratory Factor Analysis (Sample 1)

Initial statistical analyses indicated 21 items with low item-total correlations (i.e., rs < .30); we deleted these items. We deleted two more that were redundant with others, leaving 60 items.

Using IBM SPSS 20.0 (IBM Corporation, 2011), we conducted principal axis factoring with direct oblimin rotation to examine the factor structure of the 60 items. Kaiser-Meyer-Olkin’s measure of sampling adequacy (MSA) revealed a high level of shared variance (MSA = .92) and Bartlett’s test of sphericity was significant (χ² (1,770) = 9807.93, p < .001), indicating that the data met multivariate normality assumptions and were suitable for factor analysis (Tabachnick & Fidell, 2001). We based factor retention on eigenvalues, scree plots, and factor interpretability. We retained factors with eigenvalues greater than one and that contained three items or more (Tabachnick & Fidell, 2001).

Six factors had eigenvalues greater than 1. The first factor was comprised of items asserting that discrimination is a thing of the past; however, these items also loaded onto other factors, including one with items related to the prioritization of hard work, not discrimination per se. A second factor included conceptually divergent items, including some linked to competition and others linked to personal responsibility. Many of the competition-related items were also cross loaded on Factor 5, which seemed focused on beliefs about competition. The third and sixth factors both included items pertaining to government interference [e.g., “A problem with government social programs is that they get in the way of personal freedom” (Factor 3); “Social programs sponsored by the government reduce individuals’ ambition” (Factor 6)]. In all, the six-factor solution produced considerable conceptual and item redundancy (e.g., between Factors 2 and 5 and between Factors 3 and 6), making it difficult to interpret the individual factors and distinguish among them.

We therefore pursued additional approaches for determining factor retention. Following O’Connor’s (2000) model syntax, we conducted parallel analysis using randomly generated data and a principal components approach. We also specified the parameters of 1,000 iterations and a 95% eigenvalue percentile. Results suggested that six factors be retained. However, we also examined the scree plot of the data, which indicated a four-factor solution. These four factors were conceptually discrete and there were fewer cross loadings than among the possible six factors. On these grounds, we adopted a four-factor solution, retaining items with factor loadings of .50 or greater. We discarded items that loaded onto multiple factors if the difference between loadings was less than .10 (Kahn, 2006). To complement these statistical criteria, we examined the conceptual meaningfulness and integrity of the factors and constituent items. Based on this, we eliminated one item (“People need to learn from their mistakes”) that did not cohere with other factor items and opted to retain four items with factor loadings of .49 and that aligned conceptually with other items on their respective factors.

The resulting scale included 27 items loading on four factors, which we dubbed: Social Inequality (i.e., beliefs regarding the existence and implications thereof); Natural Competition (i.e., the belief that competition is fair, natural, and beneficial); Personal Wherewithal (i.e., the belief that personal attributes of strength and skill yield success); and Government Interference (i.e., the belief that state intervention violates personal freedom and meritocratic principles). These factors closely resemble the themes identified in our literature review. System Inequality accounted for 28.81% of the data’s variance. The Competition, Personal Wherewithal, and Government Interference factors accounted for 6.16%, 5.05%, and 4.60% of the variance, respectively. Altogether, these four factors explained 44.62% of the data’s variance.

Confirmatory Factor Analysis (Sample 2)

We used Lisrel 9.1 (Jöreskog & Sörbom, 2012) to conduct CFA of the 27 retained items using Sample 2 participant responses. Consistent with EFA results, we hypothesized a model with four latent factors representing the four subscales. Sample 2’s size (n = 446) exceeded the recommended minimum for CFA (e.g., MacCallum, Browne, & Sugawara, 1996), thus increasing the accuracy of models and fit indices. Additionally, data met univariate and multivariate normality assumptions (Weston & Gore, 2006): Tolerance scores ≥ .10; VIF scores ≤ 10; skew values ≤ 3; and kurtosis values ≤ 8.

We determined absolute model fit using the root-mean-square error of approximation (RMSEA) and the standardized root-mean residual (SRMR). We also tested incremental model fit with the comparative fit index (CFI). We did not employ chi-square since it is less discerning in models with more than 400 cases (Kenny, 2014). Following Weston and Gore (2006), we used RMSEA and SRMR ≤ .10 and CFI ≥ .90 as standards of acceptable fit.

To evaluate the four-factor structure, we specified that items would load onto their appropriate factors (based on EFA results) and we allowed the four factors to correlate. The CFA demonstrated that the four-factor model provided an acceptable fit to the data, RMSEA = .07; SRMR = .06; CFI = .96. However, based on modification indices and factor loadings, we deleted two additional items whose loadings were .50 and below and allowed two error variances (between Item 1, “Affirmative action is an outdated policy now that people are generally treated as equals” and Item 2, “Discrimination does not exist today to such a degree that affirmative action policies are necessary,” and between Item 3, “Affirmative action does not help eradicate discrimination. Instead it exacerbates it by promoting people on the basis of minority status instead of merit” and Item 4, “Affirmative action is a problem because it treats people unequally”) to co-vary. This yielded the best fit: RMSEA = .07; SRMR = .06; CFI = .97. We also examined a unidimensional model wherein all items loaded on a single latent factor. This was a poor fit, RMSEA = .13; SRMR = .11; CFI = .87, suggesting the modified four-factor model was superior. See Table 2 includes the final set of 25 items and their respective CFA loadings.

Internal Consistency, Reliability, and Validity (Samples 1-3)

We assessed internal consistency for the NBI by calculating Cronbach’s alphas for the 25-item scale and its subscales for each of the samples. All alphas were acceptable (see Table 3), falling within the fair to excellent range (Ponterotto & Ruckdeschel, 2007).

Based on Sample 3, two week test-retest reliability for the NBI and subscales were: overall NBI, r(146) = .84, p < .001; System Inequality subscale, r(146) = .83, p < .001; Competition, r(146) = .62, p < .001; Personal Wherewithal, r(146) = .79, p < .001; and Government Interference, r(146) = .83, p < .001.

Descriptive statistics and correlations among the NBI subscales are presented in Table 4.

NBI scores and socially desirable responding were unrelated, r(443) = -.03, p = .48. Indicating convergent validity, Sample 2 participants’ NBI scores correlated significantly with an internal locus of control, r(441) = .26, p < .001, greater environmental mastery, r(438) = .11, p = .02, and stronger belief in a just world, r(438) = .53, p < .001. Additionally, an independent samples t-test showed Democratic participants (M = 2.79, SD = .60) reported significantly lower NBI scores than the other participants (M = 3.29, SD = .67), t (436) = -8.16, p < .001. NBI scores were also significantly (p < .001) inversely correlated with all three scales of feminist beliefs: Liberal, r(317) = -.39; Radical, r(318) = -.33; and WOC, r(317) = -.53. Furthermore, Sample 3 NBI scores, measured at Time 1, correlated positively with SDO, r(176) = .51, p < .001, and ACT scores, r(146) = .37, p < .001.ⁱⁱ Lending support for criterion validity, NBI scores were negatively related to collective action on behalf of women, r(314) = -.35, p < .001, and stigma consciousness, r(107) = -.34, p < .001, and positively related to rape myth acceptance, r(178) = .52, p < .001. Table 5 includes descriptive and correlational data for the NBI and measures of validity.

To explore the distinctiveness of the NBI, particularly in relation to social dominance and right-wing authoritarianism, we compared the NBI’s correlations with stigma consciousness, and rape myth acceptance to SDO and ACT correlations with the same two measures. Collective action was not included since it was used with a different sample of participants than the SDO and ACT. Correlational analyses indicated that although the respective associations of the NBI, SDO, and ACT with the criterion validity constructs resemble one another, there are also some differences. Specifically, SDO was uncorrelated with stigma consciousness, r(107) = -.17, p = .09, whereas the NBI and ACT were negatively correlated with it: r(92) = -.24, p =.02 and r(92) = -.32, p = .002, respectively. The NBI and ACT were similarly correlated with rape myth acceptance: r(148) = .43, p < .001 and r(148) = .48, p < .001, respectively. SDO was also positively associated with rape myth acceptance, but at approximately half the magnitude of the NBI and ACT: r(178) = .20, p = .008.