Abstract
Building on the work of Inglehart and colleagues, Welzel sets out a step-by-step theory explaining how democracies arise from processes of modernization. The intermediary stages in the causal chain he sets out explain the connection between action resources and emancipatory values. In short, Welzel provides strong evidence that people must first have the material, intellectual, and connective resources to exercise certain freedoms before they develop values that will motivate them to seek out those freedoms. Although we are convinced by much of Welzel’s argument, we also note a substantial overlap between these intermediary stages of Welzel’s theory and the individual-level authoritarianism literature. Integrating current theory on authoritarianism into Welzel’s thesis at the point of overlap provides for a distinct set of hypotheses and a more nuanced understanding of how individual differences work in Welzel’s theory. Analyses of data from Wave 5 of the World Values Survey and Vanhanen provide some initial evidence in support of our amended view of the intermediary stages of Welzel’s modernization thesis.
This article integrates two currently distinct but strongly related literatures: the current incarnation of modernization theory as theorized and evidenced by Inglehart and Welzel (2005; Welzel, 2013) among others and individual-level authoritarianism, as recently conceptualized by Feldman and Stenner (Feldman, 2003; Feldman & Stenner, 1997; Stenner, 2005) and Hetherington and colleagues (Hetherington & Suhay, 2011; Hetherington & Weiler, 2009). Specifically, we argue that the concept of individual-level authoritarianism provides a more nuanced understanding of the evolution of political culture and how societies respond to threat than does Welzel’s (2013) emancipative values; effectively, we argue that the concept of individual-level authoritarianism (or libertarianism, if one prefers to focus on the opposite pole as we do here) is better suited to explaining the rise of attitudes that favor human autonomy, choice, equality, and voice, and how perceived threat may temporarily interfere with such preferences.
Welzel’s (2013) theory of human empowerment elucidates a step-by-step process in which individuals gain the means, motivation, and guarantees to exercise universal freedoms and civic agency. The human empowerment process is initiated with the attainment of action resources, which provide people with the means to overcome constraints imposed by resource scarcity. With an increase in action resources comes an increase in existential security and the motivation to look beyond the fulfillment of physiological needs. Existential security encourages a shift in values from those that focus on and facilitate existential security, to those that focus on and facilitate psychological fulfillment. This shift in values creates increased pressure on governing institutions to guarantee society the freedoms necessary to accommodate the pursuit of psychological fulfillment.
We generally agree with the sequence Welzel sets out as described above. However, we suggest a fundamental adjustment to specific details in the human empowerment sequence. This adjustment proceeds in two steps and focuses on what Welzel (2013) refers to as emancipative values. In brief, Welzel argues that emancipative values will decline during existential crises as people become more focused on existential security. Following previous research, we argue that people’s value orientations are more stable than Welzel suggests and that although attitudes may rapidly change in response to existential crises, value orientations will not. First, then, we argue that Welzel’s concept of emancipative values does not satisfactorily explain people’s responses to existential crises. The concept of authoritarianism (e.g., Feldman, 2003; Hetherington & Weiler, 2009; Stenner, 2005)—a concept that focuses on the trade-off between social conformity values and individual-autonomy values—more accurately accounts for shifts in attitudes in response to existential crises. Second, we argue that the battery of items that Welzel uses to create his emancipative values scale are more accurately described (at least partially) as emancipative attitudes and, in many ways, conceptually overlap with libertarian (nonauthoritarian) attitudes (cf. Feldman, 2003; Feldman & Stenner, 1997; Hetherington & Suhay, 2011; Hetherington & Weiler, 2009; Stenner, 2005). These two alterations allow the human empowerment process to more capably handle existential threat while maintaining the relative stability of value orientations.
The purpose of this article is, therefore, twofold. First, we seek to unite two literatures that have rarely, if ever, interacted with one another: the modernization strand of the political culture literature and the individual-level authoritarianism literature. As we discuss below, uniting these literatures provides the modernization literature with a more nuanced understanding of individual-level processes, and the authoritarianism literature with a more expansive consideration of the causes and consequences of authoritarianism. We discuss these benefits in greater detail in the conclusion.
In what follows, we outline the beginning of Welzel’s human empowerment sequence, as this serves as the basis of our modified theory. We then explain how current theory on authoritarianism ties into Welzel’s theory and how authoritarianism serves to better explain attitude change under existential crisis conditions (i.e., in response to existential threat). Although our argument is primarily conceptual and rooted in evidence provided by previous research, we nevertheless provide analyses of individual-level data from the fifth wave of the World Values Survey (WVS5) and aggregate-level data from Vanhanen (2003) to test a small number of the hypotheses that follow from our discussion. These models provide some base-level evidence in support of our suggested adjustment to Welzel’s theory.
Specifically, we provide evidence that libertarianism (the opposite pole from authoritarianism) is grounded in the interaction between personal and societal resources during personal development, whereas emancipative attitudes are responsive to not only resources during development (though to a lesser degree than is libertarianism) but also present-day resources. Furthermore, libertarianism is a strong predictor of emancipative attitudes. This evidence supports our conjecture that libertarianism fits into the human development sequence between resources and emancipative attitudes.
Action Resources and Value Development
As we are primarily concerned with the stage at which action resources affect value formation, we take Welzel’s (2013) argument regarding the development of action resources in a society for granted, and focus specifically on how action resources influence individual value orientations. Welzel argues that action resources take three forms: material resources (equipment, tools, and income), intellectual resources (knowledge, skills, and information), and connective resources (networks of exchange and contact interfaces; Welzel, 2013). Until societies control a modicum of such resources, freedoms have little utility as most, if not all, people are simply unable to make use of them. As societies increase their access to action resources, freedoms gradually gain utility. The acquisition of action resources allows people to focus on something other than mere survival; the attainment of material, intellectual, and connective resources gives people the ability to both imagine and pursue improved living conditions.
The key thrust of Welzel’s theory is the human drive toward emancipation. People perpetually strive against any force that curtails their ability to control their own lives. As noted above, the first condition that limits control over people’s own lives is a lack of material resources. To control one’s own life, one must first be capable of sustaining it. With the attainment of more material resources than is necessary to sustain a population (or some, usually arbitrarily—cf. Pratto, Sidanius, & Levin, 2006; Sidanius & Pratto, 1999—designated subsection thereof), societies begin to invest those materials in ways intended to improve their circumstances further; to improve their control over their own lives to a greater and greater degree as access to resources allows. Such investment frequently results in increased education and technological advancement; the latter of which, especially over the last few decades, directly contributes to improved connective resources. As social and individual resources grow to allow increasing control over our lives, our value priorities shift in response to that control.
Of considerable importance to Welzel’s theory is that it is not only individual-level resources or only societal-level resources that matter for this value shift; rather, it is the interaction of the two. As the values among those with greater levels of action resources shift toward those that support control over one’s life, aggregate values also begin to shift in this direction. This, in turn, feeds back into societal norms, which increasingly support those values that facilitate control over one’s own life. Continued increases in action resources further facilitate changes in individual values, building on constantly evolving social norms that are supported by ever-increasing aggregate levels of action resources, especially increased education and social connectivity. As support for such values spreads, the drag that opposing values exert against the expression of those values conducive to control over one’s own life decreases. The increasing expression of such values thereby encourages the further expression of such values, creating a positive feedback loop. Welzel refers to this as the confirmation mechanism. Crucially, even those whose current levels of resources do not encourage such values are nevertheless exposed to such, and with increasing social pressures are more likely to adopt such, values. Welzel refers to this as the contagion mechanism. Both mechanisms feed into one another, facilitating the spread of those values that are conducive to gaining greater and greater levels of control over one’s own life. Simply put, higher levels of action resources both increase the base level of values that facilitate control over one’s own life and increase the effect that personal levels of action resources have on the development of those values that increase one’s control over their own life. In quantitative terminology, there is a positive interaction between societal and individual action resources, whereby greater societal levels of action resources increase the intercept and slope of the effect that individual levels of action resources exert on those values that lend themselves to increased control over one’s own life.
Change and Stability in Value Orientations and Attitudes
Although we are convinced by the ideas underlying Welzel’s theory, we note that the research literature on values indicates that there is a distinct difference between the development of value orientations during an individual’s initial development and later in life (e.g., Bardi, Lee, Hofmann-Towfigh, & Soutar, 2009; Konty & Dunham, 1997). The resources available during initial development, in general, up through early adulthood, will partially determine an individual’s value orientation (Gecas & Seff, 1990; Kasser, Koestner, & Lekes, 2002). Once an individual reaches adulthood, her value orientation will be relatively stable (Konty & Dunham, 1997). This means that although a prolonged increase or decrease in action resources later in life may somewhat shift an individual’s value orientation (Bardi et al., 2009), temporary fluctuations are unlikely to have a substantial impact. This, of course, is a considerable deviation from Welzel’s (2013) theory, which treats value orientations as fluid, and suggests that even brief changes in resources can have a large impact on current value orientations.
Welzel’s (2013) theory implies that values are equally malleable across an individual’s life span. Therefore, his suggestion that value orientations will directly and immediately respond to changes in societal levels of action resources makes perfect sense; as action resources change, there will be a corresponding change in individuals’ value orientations. Welzel (2013) makes this argument with the support of evidence that his measure of emancipative values appears to respond to changes in existential conditions: Where societal levels of action resources expand, there is a corresponding aggregate-level increase in those values that lend themselves to increased control over one’s own life; where societal levels of action resources contract, there is a corresponding aggregate-level reduction in those values that lend themselves to increased control over one’s own life.
However, values are partially defined by their stability (Hechter, 1992; Rokeach, 1973) and trans-situational properties (Cieciuch, Schwartz, & Davidov, 2015; Schwartz & Bilsky, 1987). Welzel’s observation, therefore, contradicts not only certain definitional presumptions about the nature of values but also previous research, which finds values are largely stable over time (Bardi et al., 2009; Konty & Dunham, 1997). We suggest that this contradiction stems from how Welzel (2013) operationalizes his values measure.
Key to our argument is the fundamental distinction between values and attitudes; a distinction well established in the literature (cf. Hitlin & Piliavin, 2004; Maio, Olson, Bernard, & Luke, 2006; Mayton, Ball-Rokeach, & Loges, 1994; Rokeach, 1968, 1973). As already noted, values are a relatively stable personal attribute that predisposes one to respond to certain situations in certain ways. Cieciuch et al. (2015) define values as “transsituational goals, varying in importance, that serve as guiding principles in the life of a person or group” (p. 43). Attitudes, however, are more transient and situation specific (Banaji & Heiphetz, 2010; Canetti-Nisim, Halperin, Sharvit, & Hobfoll, 2009; Konty & Dunham, 1997; Wood, 2000). Drawing on a general consensus among attitude researchers in psychology, Ajzen (2001) defines an attitude as “a summary evaluation of a psychological object captured in such attribute dimensions as good-bad, harmful-beneficial, pleasant-unpleasant, and likeable-dislikeable” (p. 28). Values are more abstract and central to the self than are attitudes, and ultimately serve to guide attitudinal valance. The primary problem with Welzel’s (2013) empirical investigations is that a substantial portion of the component items used to construct his values measure are more accurately classified as attitudes rather than values.
Welzel (2013) calls those values that lend themselves to increased control over one’s own life emancipative values. His measure of emancipative values is an average of four subindexes: autonomy, equality, choice, and voice. Autonomy, perhaps the single subindex that could be considered a measure of values, is based on the following question: “Here is a list of qualities that children can be encouraged to learn at home. Which, if any, do you consider to be especially important?” The respondent is then given a list of “qualities” of which the coding scheme takes interest in three: independence, imagination, and obedience. We agree with Welzel that this question does tap respondents’ values preferences and, therefore, take no issue with this subcategory as a values-based measure. The remaining three subindexes, however, are more problematic. For the equality subindex, the survey instrument inquires to what degree the respondent agrees or disagrees with the following statements: “When jobs are scarce, men should have more right to a job than women,” “A university education is more important for a boy than for a girl,” and “On the whole, men make better political leaders than women do.” The choice subindex derives from a question that asks the respondent how “justifiable” abortion, divorce, and homosexuality are. These six items from these two subindexes are clearly not values as they are neither abstract concepts nor likely to be essential to the self. They are, however, attitudes, as they are evaluative of specific psychological objects. The remaining subindex, voice, falls more into the gray area between values and attitudes. This subindex is based on Inglehart’s (1971, 1997) measure of postmaterialist values, and asks the respondents to rank their first and second most important national goals on each of two separate lists of four goals. The three responses coded for are “Giving people more say in important government decisions,” “Protecting freedom of speech,” and “Seeing that people have more say about how things are done at their jobs and in their communities.” These items may represent values as they are at least somewhat abstract and could be argued to be essential to the self. However, the phrasing of the question may interfere with such an application as the question applies them to a specific context—national goals over the next 10 years—rather than keeping them abstract, trans-situational goals capable of acting as guiding principles to the person. When these four subindexes are combined into the full emancipative values index, at best we can say that the measure is an even combination of values and attitudes.
The predecessor to Welzel’s emancipative values measure, Inglehart and Baker’s (2000) self-expression values, a measure containing similar items to Welzel’s emancipative values measure, has previously been referred to as an attitudinal measure, which measures “values indirectly by inferring them from responses to attitude items” (Fischer & Schwartz, 2011, p. 1130). Although the self-expression and emancipative values scales may be appropriate proxy measures for values in certain circumstances, their responsiveness to potentially transient situations makes them suspect as measures of values. Attitudinal proxies for values also suffer from three further problems: First, unless attitudes are explicitly connected to values, there is often little correlation between the two (Maio et al., 2006); second, attitudes toward a psychological object can change based on which value(s) they are explicitly connected to (Chong & Druckman, 2007). Third, the relationship between attitude and values can be conditional on the current social and/or political environment (Hetherington & Weiler, 2009; Oyamot, Fisher, Deason, & Borgida, 2012; Stenner, 2005). Given these rather critical problems, we hold that using an (at least partially) attitudinal measure as a proxy for values should be avoided where possible.
To sum up, it is our contention that the emancipative values scale utilized by Welzel is, in actuality, more accurately described as a mix of values and attitudes. As a combination, it is, therefore, more akin to an attitudes index than a values index, as its attitudinal component certainly decreases the measure’s level of abstractness and likely also its trans-situational and relatively stable intraindividual properties that the measure should demonstrate as a values-based measure. Therefore, from this point forward, we will reference Welzel’s emancipative values scale as Welzel’s Emancipative Attitudes scale.
Authoritarianism in the Human Empowerment Sequence
Similar to Welzel (2013), we are concerned with a concept that sets one set of values against another. Whereas Welzel (2013) focuses on his relatively newly generated concept of emancipative values, we instead rely on the concept of authoritarianism/libertarianism. Authoritarianism/libertarianism centers on the trade-off between competing value sets. Authoritarianism is defined as an enduring predisposition based on the relative balance between social conformity and personal autonomy (Feldman, 2003). An authoritarian is one who prioritizes social conformity over personal autonomy. A libertarian is one who favors personal autonomy over social conformity. Conceptually, this dimension seems to fit just as well, if not better, into Welzel’s theoretical setting than does emancipative “values” for three reasons. First, the trade-off between personal autonomy values and social conformity values speaks directly to how much control one should have over one’s life, the key conceptual focus of emancipative values. Second, the authoritarianism/libertarianism scale used herein relies on well-established theory and a cross-nationally validated values measure from the cross-cultural values literature: the Schwartz Values Inventory (Cieciuch et al., 2015; Fischer & Schwartz, 2011; Schwartz, 1992, 2014; Schwartz & Bilsky, 1987). This measure avoids certain controversies that prove problematic for a theory that speaks to both individual and societal-level value orientations (cf. Alemán & Woods, 2016; Welzel & Inglehart, 2016). Third, as we explain further below, authoritarianism/libertarianism accounts for the stability of values while explaining situational variation in attitudes.
Hetherington and Colleagues (Hetherington & Suhay, 2011; Hetherington & Weiler, 2009) follow Feldman and Stenner (Feldman, 2003; Feldman and Stenner, 1997; Stenner, 2005) in conceptualizing authoritarianism not as a personality trait that unconditionally determines attitudes (e.g., Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950; Altemeyer, 1996) but as a predisposition, which is expressed to varying degrees as a function of the perception of threat. Threat causes stress and fatigue, leading to a reliance on emotion and instinct over cognitive reasoning. Those with an authoritarian predisposition are more likely to express authoritarian attitudes as they developed fewer cognitive tools to cope with ambiguity and political/social diversity, and are thus hypersensitive to perceived threat. Perception of threat among authoritarians is likely to be high relative to libertarians, as there is “always some measure of threat to social cohesion” (Hetherington & Weiler, 2009, p. 41). By contrast, those with a libertarian predisposition are less sensitive to threat, as they have developed cognitive abilities that allow them to more effectively deal with ambiguity and diversity. In nonthreatening situations, libertarians tend to express attitudes that reflect their prioritization of personal autonomy over social conformity, and thus, in our adjustment of Welzel’s terminology, express more emancipative attitudes. However, when libertarians feel threatened, they express more authoritarian (nonemancipative) attitudes, as the stress of threat leads to an increased reliance on emotion and instinct and a need to protect the social order so as to reduce increased ambiguity and confusion. In threatening situations then, those across the authoritarianism/libertarianism scale express remarkably similar, and relatively nonemancipative, attitudes. When this occurs, aggregate levels of emancipative attitudes in society decline, not because everyone is becoming more authoritarian but because libertarians are expressing more authoritarian attitudes. An example of such convergence in attitudes among authoritarians and libertarians is evident in the shift in U.S. politics toward more authoritarian attitudes immediately following 9/11 (cf. Hetherington & Suhay, 2011). When all feel threatened, all seek safety and security, and attitudes become much less emancipative.
Theoretical Synthesis
People possess an inherent desire for self-determination. However, circumstances rarely allow us to determine the course of our lives. The primary inhibitor of our self-determination is a lack of material, intellectual, and connective resources. Where we lack such resources, we develop values that lead us to pursue those very resources. Where we possess such resources, we develop values that emphasize our desire for ever greater control over our lives. Those values that we develop and internalize during our developmental years stabilize and form an enduring part of who we are, changing only slightly along with our circumstances unless subject to a strong exogenous shock. Therefore, those whose values develop under conditions of existential need prioritize safety and security (embodied in social conformity values), whereas those whose values develop under conditions of existential security prioritize self-determination (embodied in autonomy values).
As the number of people in society who benefit from greater levels of action resources increases, the number of people who also prioritize autonomy values over social conformity values also increases; that is, the number of libertarians will increase whereas the number of authoritarians will decrease. As the proportion of libertarians increases, the resistance to autonomy values from those who prioritize social conformity values will decrease, paving the way for autonomy values to be normalized and openly expressed in society (Welzel’s confirmation mechanism). The more socially acceptable the expression of autonomy values becomes, the more these values will be inculcated, through various social and political institutions, in those who do not necessarily directly benefit from the action resources that set this process in motion to begin with (Welzel’s contagion mechanism). The combination of these pathways (direct and indirect) results in a society where autonomy values are accepted by most of society, even if they are not broadly dominant; however, they are held most strongly by those who were not just exposed to such norms but who also directly benefited, in their developmental years, from the existential security provided by action resources.
Under status quo conditions in any given country, to varying degrees, some individuals will prioritize autonomy values whereas others will prioritize social conformity values; that is, societies will be constituted by a mix of libertarians and authoritarians. Importantly, although libertarians do not prioritize social conformity values, they nevertheless still hold social conformity values. Similarly, in any society where autonomy values have normalized, authoritarians, though prioritizing social conformity values, also hold autonomy values.
Prioritizing one set of values over the other loosely translates into possessing attitudes reflective of the dominant value set. The strength of an individual’s value priority, that is, how much the person prioritizes one set of values over the other, corresponds with the extremity of the associated attitude. This, in turn, results in a societal distribution of attitudes similar to the societal distribution of value priorities: Libertarians express their preference for autonomy values via relatively emancipative attitudes; authoritarians express their preference for social conformity values via relatively nonemancipative attitudes.
Where conditions diverge from the status quo and threaten individuals’ existential security, individual attitudes, and therefore aggregate attitudes, shift toward expression of social conformity values. This does not imply that those who under status quo conditions prioritize autonomy values suddenly shift to prioritizing social conformity values when threatened, rather, the stress of the situation causes the cognitive association between value priorities and attitudes to break down. Under stress conditions, individuals across the value–priority distribution respond in similar fashion: by emotively and reflexively moving to protect the stability of the social order so as to reduce situational ambiguity and insecurity. In other words, regardless of whether one is an authoritarian or a libertarian, under existentially stressful conditions, social conformity values will guide social and political attitudes. For authoritarians, there may be little, if any, change in attitudes. For libertarians, who are usually inclined to frame social and political issues in terms of autonomy values, existential threat will reframe value-relevant attitudes in terms of social conformity values, thereby resulting in a shift away from emancipative attitudes toward nonemancipative attitudes. In terms of aggregate-level attitudes, during existential crises such as economic recessions or terrorist attacks, societal expressions of social and political attitudes will become less emancipative.
Hypotheses
A great deal of the many hypotheses derivable from the above synthesis have, as noted above, already been tested and confirmed in previous research, particularly in the work of Inglehart and Welzel (2005; Welzel, 2013) and Hetherington and colleagues (Hetherington & Suhay, 2011; Hetherington & Weiler, 2009). There are, however, a few hypotheses that are needed to connect these two literatures and that can be tested with available data.
First (Hypothesis 1a: H1a), both personal and societal action resources during development will be positively related to an individual’s level of libertarianism. Similarly, as values are loosely predictive of attitudes (H1b), action resources during development will also be positively related to an individual’s level of emancipative attitudes.
Second (H2a), societal postdevelopment action resources will not affect an individual’s level of libertarianism. This will also apply to personal levels of action resources, but only insofar as they substantively diverge from those personal action resources available to the individual during development. Unfortunately, as the data to measure an individual’s personal resources during development are not available, we suspect that (H2a1) personal action resources at present, given their likely high correlation with action resources during development (cf. Bowles & Gintis, 2002; Breen & Jonsson, 2005), will demonstrate a significant and positive correlation with libertarianism. As attitudes, unlike values, are more transient and subject to current circumstances (H2b), current personal and societal action resources will be positively related to emancipative attitudes.
Third (H3), the interaction between personal and societal action resources during development will be positively related to libertarianism.
Fourth (H4), libertarianism will be positively related to emancipative attitudes.
Method
Sample
With the exception of our authoritarianism/libertarianism variable, which Welzel does not make use of in his analyses, the following analyses use variables measured in similar, if not identical, fashion to Welzel. As with Welzel, then, we rely on data from the WVS (http://www.worldvaluessurvey.org) and Vanhanen (2003). The data necessary for the following analyses are available for 37 of the 58 countries surveyed in the WVS5 (see Figure 1 for a list of the countries used in the following analyses).

The mean level of libertarianism in each country in the study.
The WVS draws a representative sample of at least 1,000 respondents between the ages of 18 and 85 for each country involved in the study. Full probability sampling is done where resources allow; otherwise, strictly governed quota sampling, or some combination thereof, is used. Interviews with each respondent are done face-to-face. Every reasonable effort is made to reduce nonresponse.
Measures
Individual libertarianism
We largely follow Feldman (2003) in constructing our measure of authoritarianism/libertarianism using Schwartz’s (2012) “motivational continuum” of values. Feldman, in his “SCA Values” (social conformity–autonomy values) scale, measures the relative importance of conformity and tradition values against self-direction and stimulation values. For his measure of authoritarianism/libertarianism, he combines this SCA Values scale with his “SCA Beliefs” scale as they appear to measure the same underlying dimension. We adopt the general idea behind the SCA Values scale, that authoritarianism reflects the relative importance of social conformity values against individual-autonomy values, and vice versa for libertarianism, but measure it slightly differently. First, we do not combine our equivalent of the SCA Values scale with an equivalent of his SCA Beliefs scale as the equivalent does not exist in the WVS5, nor is it necessary as both scales arguably measure the same underlying concept. Second, we extend our SCA Values scale to reflect the full motivational clusters suggested by Schwartz’s Openness to Change and Conservation values clusters. The Openness to Change values cluster contains the hedonism, stimulation, and self-direction values. The Conservation values cluster contains the security, conformity, and tradition values. In effect, then, we are adding an additional value to each of Feldman’s value clusters. We do this for two reasons: First, from an empirical perspective, Schwartz has found the Openness to Change/Conservation dimension (along with the self-transcendence/self-enhancement dimension) to be nearly universal in his research (Schwartz, 2012). Second, the additional values fit nicely into the authoritarian/libertarian conceptual construct, with prioritization of individual autonomy for hedonism 1 and social conformity for security.
Specifically, we average each individual’s scores on the hedonism, stimulation, and self-direction values items to produce the Openness to Change scale. We do the same for the security, conformity, and tradition values to produce the Conservation scale. Each scale ranges from 1 to 6 with a higher score indicating a higher reported identification with that value cluster. We then subtract the Conservation scale from the Openness to Change scale. This produces an authoritarianism/libertarianism scale ranging from −5 to +5, rescaled to range from 0 to 1. Those scoring above .5, the midpoint of the scale, prioritize individual autonomy (Openness to Change values) over social conformity (Conservation values), whereas those who score below .5 prioritize social conformity over individual autonomy. Those who score precisely .5 prioritize neither. Higher scores on this scale, then, indicate a greater degree of libertarianism, whereas lower scores indicate a greater degree of authoritarianism. Figure 1 presents the mean level of libertarianism in each country used in the following analyses.
Personal action resources
As noted above, there are three forms of action resources: material, intellectual, and connective. We measure these three concepts in identical fashion to Welzel (2013). Material resources are gauged via reference to an individual’s income level, measured in deciles for each country. This variable is transformed to range from 0 to 1 with a higher value indicating a higher income level.
Intellectual resources are gauged via reference to an individual’s level of education. The education question from the WVS is recoded to range from 1, indicating the lowest level of education, to 7, indicating the highest level of education. The various categories, from low to high, are no formal education, incomplete primary school, complete primary school, incomplete secondary school (technical/vocational or university preparatory), complete secondary school (technical/vocational or university preparatory), some university-level education without degree, and university-level education with degree. This variable is transformed to range from 0 to 1, with a higher value indicating a higher level of education.
Connective resources are derived from eight questions in the WVS5 ascertaining an individual’s recent use of various sources of information “to learn what is going on in their country and the world”: newspapers, news broadcasts (television or radio), print magazines, in-depth reports (television or radio), books, Internet/email, talk with friends/colleagues, and computers. Each of these items is scored 1 if it was used in the last week (or frequently in the case of the computers question) and 0 otherwise. These responses were then summed for each individual and rescaled to range from 0 to 1 with a higher value indicating the use of a wider range of connective resources.
These three action resource variables are then averaged to produce a single measure of personal action resources ranging from 0 to 1 with a higher value indicating a higher level of personal action resources.
As alluded to in the “Hypotheses” section of the article above, we use this variable as a proxy measure for personal action resources at birth (i.e., personal action resources during development) as well as an actual measure of personal action resources at present (i.e., personal action resources postdevelopment). Although using this variable as a proxy measure for personal action resources at birth produces a less accurate measure than we would prefer, it is unfortunately necessary given data limitations. Importantly, we believe the relatively high correlation between resources available during development and later in life (cf. Bowles & Gintis, 2002; Breen & Jonsson, 2005) provides sufficient justification for using this variable in this way. Nonetheless, the below findings that result from using this as a measure of personal resources at birth should be interpreted cautiously.
Societal action resources at present
This is a country-aggregated measure of personal action resources intended to reflect the average resources available to the society at present. Societal action resources at present are the mean level of personal action resources in each country.
Societal action resources at birth
This variable is a measure of the societal-level action resources in each country during the decade in which each individual was born (for those born in 1940 and onward). As the above measures of action resources are difficult, if not impossible, to replicate at the societal level for each individual across all birth cohorts, Welzel uses a proxy variable to measure this concept. We use the values provided by Welzel (2013, online appendix). This proxy measure of societal-level action resources is calculated by weighting a society’s urbanization rate by its literacy rate, a measure found to be highly correlated with a society’s technological advancement in previous research (Welzel, 2013, online appendix). The constituent measures for this variable, urbanization rate and literacy rate, are taken from Vanhanen (2003). Due to the source data, this variable is constant within decades; for example, all those born between 1940 and 1949, inclusive, possess the same value for this variable.
Emancipative attitudes
Although Welzel (2013) refers to this measure as emancipative values, for those reasons discussed above, we interpret this as a measure of emancipative attitudes. We construct this variable in line with the instructions provided in Welzel (2013, online appendix). As discussed above, this variable is constructed from four subindexes; each of which is constructed from responses to questions from the WVS5. The first subindex, autonomy, is derived from a question asking about the perceived importance of specific qualities (independence, imagination, obedience) that children can be encouraged to learn at home. The second subindex, equality, relates to whether individuals believe males should receive priority over females in specific contexts (jobs, university places, political leadership). The third subindex, choice, measures how justifiable individuals believe certain actions (homosexuality, divorce, abortion) are. The fourth subindex, voice, is a measure of how important individuals consider having a say in their communities and societies. These subindex measures are equally weighted in the final measure of emancipative attitudes. The emancipative attitudes variable ranges from 0 to 1 with a higher value indicating a higher degree of emancipative attitudes.
Individual-level control variables
In our individual-level analyses, we also account for the age and sex of the respondent. Age is a measure of an individual’s age transformed to range from 0 to 1 with a higher value indicating an older individual. Sex is a measure of the interviewer’s perception of a respondent’s sex (female or male) where 0 indicates a female and 1 indicates a male.
Results
The following analyses use a linear mixed-model approach to account for the nonindependence of within-country observations and to allow simultaneous analysis of both individual and aggregate-level data, without having to worry about artificially inflating the significance of aggregate-level variables, where necessary (Rabe-Hesketh & Skrondal, 2008). The following analyses are divided into three sections to distinctly address what we see to be three sets of hypotheses: those dealing with demonstrating the differential reliance of values and attitudes on different temporal resources (H1a, H1b, H2a, H2a1, H2b), that dealing with the interactive relationship between societal and personal action resources in predicting libertarianism (H3), and that dealing with the ability of libertarianism to predict emancipative attitudes (H4).
Distinct Effects of Actions Resources on Libertarianism and Emancipative Attitudes
Our theory states that value orientations should only be influenced by action resources during development unless there is a substantial and sustained change in resource levels later in life. Attitudes, however, will be influenced by resources during development and those available in the present. Table 1 presents evidence that supports the idea that action resources at birth and action resources in the present have varying effects on value orientations and attitudes. The outputs of these analyses show the impact each of the three types of action resources exert on both libertarianism and emancipative attitudes. Moving through our hypotheses one at a time, we find support for H1a: Both personal resources and societal action resources at birth correlate with libertarianism; an increase from the lowest to the highest level of personal resources corresponds with an increase of 8% of the range of libertarianism, whereas an increase from the lowest to the highest level of societal action resources at birth results in an increase of 10% of the range of libertarianism. H1b is also confirmed: Both personal and societal action resources at birth correlate with emancipative attitudes. An increase from the lowest to the highest level of personal resources corresponds with an increase of 20% of the range of emancipative attitudes, whereas an increase from the lowest to the highest level of societal action resources at birth results in an increase of 7%. In agreement with our stated hypotheses, then, action resources at birth predict both libertarianism and emancipative values. However, there is a substantial difference in the coefficients on personal action resources in the two models. We would expect, given our theory, that personal action resources at birth would have a greater impact on libertarianism than emancipative attitudes. We assume, an assumption we unfortunately cannot test given our data, that the greater impact of personal action resources on emancipative attitudes is due to the variable actually being a measure of present-day personal action resources—a relationship that is consistent with our theory, as noted below.
Linear Mixed Models Predicting Libertarian Values and Emancipative Attitudes.
Action resources in the present demonstrate partially different effects on libertarianism and emancipative attitudes. Consistent with H2a, societal action resources at present show no significant relationship with libertarianism. Personal action resources at present, in line with H2a1, do predict libertarianism. As noted, this is inconsistent with our theory, but consistent with our expectations due to measurement issues. H2b is also confirmed: Both personal and societal action resources at present predict emancipative attitudes. As expected, societal action resources at present demonstrate diverse effects across libertarianism and emancipative attitudes, showing no relationship with libertarianism and a significant and quite substantial relationship with emancipative attitudes: An increase from the lowest to the highest level of societal action resources at present corresponds with an increase of 52% of the range of the Emancipative Attitudes scale. Personal action resources, as already noted above, demonstrate a significant and small but substantive effect on libertarianism and a significant and moderately substantive effect on emancipative attitudes.
In sum, Table 1 provides evidence that suggests libertarianism is influenced by resources at birth whereas emancipative attitudes are influenced by resources at birth and in the present. 2
Action Resources and Libertarianism
We expect that a greater level of action resources will promote a more libertarian disposition. Both societal levels of action resources at birth and one’s personal level of action resources affect an individual’s level of libertarianism. Table 1 provides confirmatory evidence on this point. The coefficients on both personal action resources and societal action resources at birth are positive and significant. Those who were raised in societies with the highest levels of action resources possess a libertarianism score of ~9% of the range of the scale higher than those who grew up in societies with the lowest. Those who personally had access to the highest levels of action resources possess a libertarianism score of ~8% of the range of the scale higher than those who had access to the lowest.
However, we also expect that the action resources available in a society will constrain the degree to which personal action resources will affect the development of libertarianism during development. The effect of individual action resources on one’s libertarianism during development will, therefore, be conditional on the action resources present in society at the time. In other words, we expect a significant interaction between the level of societal action resources at birth and the level of individual action resources. Table 2 speaks to this prediction. The coefficient on the interaction term is a positive .123. As interpreting coefficients on interaction terms and the constituent variables of such is not always intuitive, we plot the interaction in Figure 2.
Linear Mixed Model Predicting Individual Libertarianism.

The impact of personal action resources on libertarianism conditional on the societal level of action resources at one’s birth.
Figure 2 illustrates this interaction. The figure illustrates the marginal effect of individual action resources on libertarianism at three distinct values of societal action resources at birth: 1 standard deviation below the mean (.16), the mean (.46), and 1 standard deviation above the mean (.76). The slopes for each of these values are statistically significant at p < .001: The slope for 1 standard deviation below the mean is 0.047 with a standard error of 0.005, the slope for the mean is 0.084 with a standard error of 0.004, the slope for 1 standard deviation above the mean is 0.121 with a standard error of 0.006.
First of all, the figure demonstrates that, all else equal, a higher level of societal action resources at birth predicts a higher level of libertarianism. This is a statistically significant relationship at all values of personal action resources. For example, those who score highest on the action resources variable and live in a society that falls at 1 standard deviation above the mean on the societal resources variable score 0.09 points higher (9% of the range of the scale) on libertarianism than those with the same personal action resources but from a society that falls at 1 standard deviation below the mean on the societal resources variable. Second, all else equal, a higher level of personal action resources predicts a higher level of libertarianism. This is a statistically significant relationship at all values of societal action resources at birth. Third, the greater the societal action resources at birth, the greater the impact of personal action resources on libertarianism. The difference in libertarianism between those who score the lowest and those who score the highest on the personal action resources variable is 0.05 points (5% of the range of the scale) for those who live in a society that falls at 1 standard deviation below the mean on the societal resources variable, 0.08 points (8% of the range of the scale) for those who live in a society that falls at the mean on the societal resources variable, and 0.12 points (12% of the range of the scale) for those who live in a society that falls at 1 standard deviation above the mean on the societal resources variable. All this indicates that whereas a higher personal level of action resources does facilitate libertarianism, a higher level of societal action resources at birth increases the impact of personal action resources on libertarianism, confirming the interactive effect posited in H3.
Libertarianism and Emancipative Attitudes
We expect libertarianism to be predictive of emancipative values. In the full scale model of Table 3, this is precisely what we find. In line with H4, libertarianism is positively, substantially, and significantly related to emancipative attitudes. An increase from the lowest to the highest value on the libertarianism scale corresponds with a ~25% increase across the range of the Emancipative Attitudes scale. Similar patterns are found for each subscale: An increase from the lowest to the highest value of the libertarianism scale predicts an ~40% increase across the range of the autonomy subscale, a rather small ~3% increase across the range of the equality subscale, an ~33% increase across the range of the choice subscale, and an ~23% increase across the range of the voice subscale.
Linear Mixed Models Predicting Emancipative Attitudes and Subscales.
Discussion
Our primary goal in this article is to provide a theoretical synthesis between the human empowerment sequence and current theory in the individual-level authoritarianism literature. Toward this end, we provide detailed argument as to how the concept of authoritarianism can add a valuable additional level of detail to Welzel’s (2013) human empowerment sequence. Our primary argument interjects Feldman’s (2003) conceptualization of authoritarianism/libertarianism in between Welzel’s (2013) conceptualization of action resources and emancipative values and reconceptualizes Welzel’s emancipative values as emancipative attitudes. This provides a number of beneficial adjustments to Welzel’s current theory. First, in line with theory on values development, value orientations are presumed to be largely static after their initial development during an individual’s formative years. As with Welzel’s current theory, we expect that both individual and societal levels of action resources contribute to the development of value orientations for the reasons that Welzel currently proposes, although we expect these to primarily work through socialization processes during an individual’s formative years. Second, individuals’ attitudes are a conditional outcome of the interaction between value orientations and threat. For authoritarians, who consistently perceive some form of threat to societal well-being, attitudes will consistently tend toward nonemancipative attitudes. For libertarians, attitudes will vary from emancipative to nonemancipative, conditional on how threatened they feel. Therefore, although values will usually remain steady, attitudes will shift dependent on the perception of threat.
In addition to providing a theoretical synthesis of these two theories, we also provide some initial, base-level evidence to support this synthesis. First, we examine how action resources at birth and in the present differentially predict libertarianism as opposed to emancipative attitudes. Second, we examine whether the interaction between personal and societal action resources during an individual’s initial formative years predicts their level of libertarianism. Third, we examine whether libertarianism predicts emancipative attitudes.
Our analysis of 32,072 individuals within 37 countries provides support for each of the seven hypotheses derived from our theory (accounting for measurement issues). Libertarianism and emancipative attitudes derive from different temporal sources of resources. Libertarianism is affected by action resources at birth though not from those in the present (though see below), whereas emancipative attitudes are affected by resources at birth and in the present. Importantly, personal and societal action resources at birth interact to predict libertarianism. A higher level of societal action resources at birth increases the impact that personal levels of action resources exert on an individual’s level of libertarianism. This corresponds to Welzel’s elevator and amplifier effects: Greater societal levels of action resources at birth both increases (elevates) an individual’s predicted level of libertarianism and amplifies the impact that personal levels of action resources exert on the same. Libertarianism, in turn, strongly predicts an individual’s emancipative attitudes. The more libertarian one’s value priorities, that is, the more one prefers individual autonomy over social conformity, the more one expresses support for emancipative attitudes: attitudes regarding autonomy, equality, choice, and voice.
The importance of integrating these two literatures cannot be overstated. As is, Welzel’s theory somewhat contradicts theory on the difference between values and attitudes and completely ignores individual differences in how we respond to threat. The difference between values and attitudes and the differences in individual threat response have a great deal of importance to political opportunities and outcomes as has been well demonstrated in the authoritarianism literature. One particularly relevant example that emphasizes the importance of these differences is as follows: Welzel’s theory suggests that authoritarian politicians in advanced economies need to wait for resources to contract before they will find an increase in support for any suggested nonemancipative policies, such as military action against foreign targets or restrictions on civil liberties among one’s own citizenry. Our adjustment to this theory suggests that existential threat from any source might do. This is certainly more in line with the threat caused by terrorism and the resulting widespread support that momentarily arises, and then recedes, for authoritarian/nonemancipative policies among mass publics in advanced economies. Although Hetherington and Suhay (2011) provide fairly strong empirical evidence on this point, one need only look to the responses to terror attacks we have seen from mass publics in numerous advanced economies not undergoing resource contraction to understand why we need to account for more than just resource threat. Integrating the authoritarianism literature into the human empowerment sequence gives us the conceptual and theoretical tools to understand, and possibly predict, how and why mass publics respond to not only economic recessions but also terror attacks in the way they do and how this can influence party support, public policy, and even the health of liberal democracies.
As already implied, there is still a good deal of work to be done to fit these two theories together and demonstrate how action resources, values, threat, and attitudes are tied together. Welzel’s (2013) current framework already deals with a temporal and change element and yet has been forced to rely on cross-sectional data to provide evidence of its viability in explaining the rise and success of democracy. Integrating authoritarianism into this framework makes the temporal and change element even more critical. As we attempt to move beyond descriptions and predictions of societal change to those of individual-level change in response to changing action resources, societal values, and existential threat, longitudinal cross-national panel data are ever more necessary to examine how well our theory fits empirical evidence. Cross-national, long-term data periodically surveying individuals from adolescence well into adulthood would speak volumes in support or refutation of our proposed theory. Cross-national survey experiments determining how libertarianism, emancipative attitudes, and threat interact would also be particularly helpful in determining the level of nuance the authoritarianism literature can add to Welzel’s theory.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
