Why No Left-Authoritarian Parties?

Abstract

The puzzle posed by the lack of viable parties in Western democracies in the left-authoritarian quadrant of a two-dimensional space likely has demand- as well as supply-side explanations. This paper focuses on the demand side and argues that left-authoritarian voters are internally divided by the extent to which they combine two distinct non-economic preferences: views on the socio-political order (libertarian vs. authoritarian) and views on immigration (cosmopolitan vs. nativist). Leftwing citizens holding the resulting three preference bundles – left-authoritarian-nativist, left-libertarian-nativist, and left-authoritarian-cosmopolitan – have distinct and predictable partisan leanings. This complicates party entry into the left-authoritarian quadrant. Furthermore, existing parties can try to preempt party entry by appealing to a subset of left-authoritarian citizens. Nevertheless, the lack of left-authoritarian parties is likely a fleeting historical phenomenon.

Introduction

Spatial voter theories of representative democracy assume that citizens and parties hold positions in a low-dimensional space defined by salient issue bundles. Voters prefer parties that are close on issues that matter to them. Parties maximize votes by adopting issue positions that are proximate to concentrations of citizens, but also sufficiently different from competitors so that people will not abstain from voting because of indifference or alienation. As a result, most spatial theories expect that popular preference configurations will be matched with successful parties, at least in multiparty systems that allow for low-threshold party entry. A systematic representation gap – when a large proportion of citizens hold distinctive and salient policy preference profiles unrepresented by existing parties – should not occur in the long run.

Yet, in Western established democracies, such a representation gap persists: a large share of citizens voices “redistributive” (left) preferences on economic issues and “conservative” (authoritarian) preferences on non-economic issues in the sense of traditional, authoritarian and/or nationalist dispositions (“TAN”), opposing green, alternative and libertarian (“GAL”) positions. But no electorally successful parties with decidedly left-authoritarian positions exist in these systems. This puzzling lack of left-authoritarian parties in Western democracies motivates our paper.

To illustrate this puzzle, Figure 1 shows the distribution of citizens (top row) and parties (bottom row) in the two-dimensional policy space. Existing scholarship assigns different content to the non-economic dimension, the most common being socio-political governance issues (gay rights, women rights, civil liberties, law and order, etc.), migration and national identity issues, or a mix thereof (as in the GAL-TAN conceptualization mentioned above). Therefore, both rows in the figure have three panels each, the first using a representative socio-cultural issue (abortion; social lifestyle), the second using a paradigmatic migration issue (immigration), and the third using a mix of both (∼GAL-TAN). The economic dimension is always proxied by the same issue on redistribution (voters) or economic left-right positions (parties). (We provide details on these items and sources in appendix 7.1).

Figure 1.

Citizens and parties in the two-dimensional space. Source: Joint EVS/WVS 2017−2022 Dataset (top row); Chapel Hill Expert Survey trend file, 1999–2024 (bottom row)

The top row of Figure 1 shows that in all three versions of the two-dimensional policy space, citizens are found in every quadrant, including in the quadrant of left-authoritarian issue preferences (darker areas indicate more respondents). The four quadrants defined by the white lines indicating average values contain about 21.7 to 27.7% of respondents. The percentage of citizens in the left-authoritarian quadrants is 27.5 (left panel), 23.2 (middle panel) and 24.3 (right panel). In other words, the left-authoritarian quadrant – far from empty – is roughly as populated as the other quadrants. The bottom row of Figure 1 shows that there are no parties that clearly represent left-authoritarian issue combinations. ¹ This is the puzzle we want to understand.

An obvious answer to the puzzle is that spatial theories of party competition simply fail to explain basic electoral phenomena. We disagree. Rather than throwing out the baby with the bath water, we theorize that a more sophisticated mapping of spatial dimensions of voter preferences is needed to address the puzzle. Spatial theory can help identify which parties might compete for voters who do not find a party that fully matches their preferences in the current political landscape. It also clarifies why parties that want to attract left-authoritarian voters face difficult trade-offs.

We argue that a three-dimensional conception of the policy space helps us understand the puzzle. We unpack GAL-TAN (or “non-economic” or “second”) issues into two different dimensions. One evolves around issues of socio-political governance (libertarian vs. authoritarian), and the other around issues of collective identity and immigration (cosmopolitan vs. nativist). In a then three-dimensional space, erstwhile aggregated left-authoritarian citizens are of three flavors: left-authoritarian cosmopolitans, left-libertarian nativists, and left-authoritarian nativists. This preference heterogeneity translates into heterogenous partisan leanings, with leftist voters gravitating toward parties that satisfy their preference profiles on at least one of the other two dimensions. This heterogeneity makes closing the puzzling representation gap challenging, whether it was by a novel party or an existing party altering its programmatic appeals. Wherever they may position themselves in the three-dimensional space, they could hope to obtain only a slice of what appeared as a homogeneous left-authoritarian electorate from the perspective of a two-dimensional rendering of political competition. We predict and test the partisan leanings of citizens with left-authoritarian values, based on their combination of libertarian vs. authoritarian and cosmopolitan vs. nativist preferences. We believe this approach can resolve some empirical inconsistencies evidenced by previous investigations, discussed below.

In the next section, we first review existing contributions, which have relied on a two-dimensional conceptualization of the issue space. We then develop our theoretical approach, which employs a three-dimensional conception, and derive our hypotheses. The section called “Empirics” presents our results. Based on our framework and findings, we offer speculations about how the left-authoritarian representation gap might be closed in the future in the penultimate section. The final section concludes.

Theory: Preference Heterogeneity and Partisan Choices

Existing studies of left-authoritarian citizens treat them as a homogeneous group. Scholars use different measures to operationalize the authoritarian dimension, sometimes drawing on libertarian-authoritarian governance, sometimes on cosmopolitan-nativist group membership, and sometimes on both. Contributions on the topic study the political behavior of left-authoritarian citizens in the absence of an ideologically proximate party or candidate in a two-dimensional space (Gidron, 2020; Hillen & Steiner, 2020; Wronski et al., 2018). Different studies report different, sometimes contradictory findings. As the worst-represented group, it has been shown that left-authoritarians withdraw from political participation (Federico et al., 2017; Hillen & Steiner, 2020) and become cynic and mistrusting about democracy (Hakhverdian & Schakel, 2022). In terms of voting behavior, scholars report contradictory results. Some have found that left-authoritarians prioritize their economic “left” preferences by supporting left parties (Lefkofridi et al., 2014) while others have found that they prioritize their non-economic “authoritarian” preferences by supporting conservative parties (Gidron, 2020; Loxbo, 2022). ²

Thus, existing findings on the political behavior of unrepresented left-authoritarian citizens support a wide range of possibilities, with no consensus in sight. What these studies have in common, however, is that they operate with two-dimensional policy spaces of the types we illustrate in Figure 1 above. In contrast, our account argues that the left-authoritarian representation gap can be better understood when employing a three-dimensional policy space.

This deviation from the typical two-dimensional approach requires justification, which we provide in the next subsection. Partitioning each of three dimensions into opposite sides yields eight distinct configurations, and they all represent intelligible ideological preference combinations. Each configuration is associated with specific partisan leanings in the currently available menu of Western party families. In the appendix, we also consider how electoral payoffs of political parties may vary, if they marginally adjust their programmatic appeals or if the salience of issue dimension varies (online appendices 8.3 and 8.4).

From Two to Three Dimensions

Two-dimensional conceptualizations of the policy space have increasingly replaced one-dimensional approaches (e.g., Mair, 2007) and now constitute the modal approach in the scholarship on political preferences and electoral politics in Western Europe and the US (Hall et al., 2023; Häusermann & Kriesi, 2015; Jessee, 2012; Kitschelt, 1994; Oesch & Rennwald, 2018; Poole & Rosenthal, 1985). Almost all two-dimensional approaches distinguish a first “economic” dimension – running from emphasis on equality and redistribution among members of a society to the acceptance of inequality to permit individual initiative, dominance and free-market choices – and a second dimensions of non-economic issues, often labeled as “social” or “cultural” dimension. The issue content of this second dimension varies across scholars, but it often revolves around a “libertarian” emphasis on individual autonomy, responsibility and discretion in the expression of social relations, lifestyles as well as civic rights and involvement, on the one hand, and the “authoritarian” assertion of individuals’ compliance with unquestionable collective norms restricting individual choices, on the other. ³

However, two-dimensional models often bracket, explicitly exclude (Treier & Hillygus, 2009), or subsume under another dimension a respondent’s definition of who is entitled to be a member of the social entity that institutionalizes distributive mechanisms of scarce resources (first dimension) and calibrates the relationship between individual choice and collective compliance in socio-cultural and political governance (second dimension). Yet, from Carl Schmitt (1932) to Robert Dahl (1989) political scientists have considered the delineation of “friend” and “foe” – citizen and alien, native member of the community and stranger – a vital foundational consideration of politics, and one that cannot be made in democratic terms, as democratic procedures already require a pre-existing decision on membership to establish rights of participation in the collective decision-making process: Is polity membership an exclusive privilege by birth and impermeable to the inclusion of non-members (nativism, ethno-nationalism)? Or is membership universalistically shared and open to newcomers (cosmopolitanism, multiculturalism)?

Three-dimensional conceptions add community membership as a distinctive preference consideration (Kitschelt, 1992), typically measured in terms of attitudes on immigration (Caughey et al., 2019). For example, Kitschelt and Rehm (2014) distinguish between “greed” (first dimension, economic left-right), “grid” (second dimension, libertarian-authoritarian), and “group” (third dimension, cosmopolitan-nativist). Existing accounts use either libertarian-authoritarian or cosmopolitan-nativist issues, or a mix of both, to link libertarian views of governance with cosmopolitan ideas of membership, and authoritarian views with nativist ideas of membership, based on conceptual, not empirical, connections. While these associations are often empirically borne out, we argue that keeping governance (libertarian-authoritarian) and membership (cosmopolitan-nativist) issues separate helps us empirically understand the puzzling left-authoritarian representation gap.

We would argue that a three-dimensional view of the policy space is increasingly useful for understanding current and future electoral trends. Preferences that two-dimensional models would see as strange—like libertarian nativism and authoritarian cosmopolitanism—are common among voters, even though political parties have so far rarely focused on them. As an exception, the Dutch Pim Fortuyn List offered the “unusual” combination of libertarian and nativist positions. ⁴

We assert that democracy involves collective decisions over the distribution of scarce resources (distribution), structured by rules of aggregating preferences (governance), within a defined citizenship (membership). Various theories and empirical evidence have postulated natural or “logical” connections between positions on the three dimensions, and thereby often a reduction in the dimensionality of societal conflict. In contrast, we see no compelling arguments why positions on the three dimensions are logically connected in specific ways. Tellingly, the literature provides justifications for all possible reductions of distribution/governance/membership into two dimensions: by combing governance and membership, distribution and governance, or distribution and membership into a single “natural” combination. We discuss examples of each of these reductionism arguments in turn.

First, the most common reductionism combines questions of governance and membership into one dimension. The argument goes that those who prefer top-down authoritarian collectivism – subordinating the individual – also prefer restrictive citizenship rules that keep strangers and members apart, to entrap the leadership’s mass following. Indeed, throughout history authoritarian communes with high entry and exit barriers have survived longer than libertarian communes with permeable boundaries (Kanter, 1972). However, particularist authoritarianism may produce resilient social order only under empirically contingent conditions. Also, important examples—such as multicultural world empires—contradict it.

Nevertheless, while there is widespread consensus that both governance and membership issues are electorally salient, it is common practice to lump them together, perhaps because they are often correlated. Effectively, therefore, issues related to citizenship are conventionally treated as derivative of the libertarian-authoritarian dimension of socio-political and cultural governance. One influential example is the GAL-TAN terminology, which refers to a non-economic super-dimension that pits “green alternative libertarian” (GAL) preferences against “traditional authoritarian national” (TAN) positions, although only the “right” wing pole mentions citizenship (“national”) explicitly (Hooghe et al., 2002). In our terminology, this super-dimension would range from “libertarian-cosmopolitan” to “authoritarian-nativist.” While we agree that this super-dimension describes many electoral dynamics in the past very well, we argue here that it needs to be disaggregated into its underlying two components because citizens hold libertarian-authoritarian and cosmopolitan-nativist attitudes in all possible combinations. Governance and membership issues are not logically connected, and they are often empirically uncorrelated ⁵ (and we predict, increasingly so). In sum, we do not think that there is a compelling logical connection between authoritarianism and exclusionary conceptions of polity membership (nativism).

Second, we also see no exclusive connection between market liberalism (distribution) and socio-political libertarianism (governance), as American libertarians would surmise (Friedman, 1962; Rothbard, 1978). Historically, the supposedly illogical combination of market liberalism and political authoritarianism exists frequently. Likewise, Douglas and Wildavsky’s (1982) account of preferences for redistribution as derivative of combining libertarian individualism (grid) with membership (group) particularism often flies in the face of empirical preference configurations (Malka et al., 2019).

Third, neither do we see a universal link between preferences for redistribution and immigration (Shayo, 2009). Leftists certainly can be “welfare chauvinists” and reject immigration out of concerns that immigrants have not sufficiently contributed to public social benefits through income or payroll taxes (pensions, health care, housing, etc.) to deserve such entitlements. But most people in lower income brackets are not immigrants. They often support redistribution, and they may or may not embrace immigration.

Overall, while people’s positions on the three dimensions (distribution/governance/membership) may empirically correlate in certain times and places, combinations of all of them are logically viable, substantively feasible, and historically manifest (Kitschelt & Rehm, 2014). None of the combinations is logically “consistent,” and none is logically “inconsistent.” (For assertion to the contrary, see Dassonneville et al., 2023).

Three-Dimensional Preference Configurations

Dichotomizing three dimensions yields eight combinations of political orientations. All eight configurations across the three dimensions – displayed in Table 1 – have considerable support in contemporary affluent Western knowledge societies and are consequential for political interest mobilization. This includes a large constituency of “left-authoritarians” with heterogenous preferences on governance and immigration issues.

Table 1.

Eight Combinations of Political Orientations

Groups	Label
rgt/aut/nat	Right-authoritarian nativist:
	19th and 20th century anti-democratic National Liberals, 21st century right-wing market liberalism
rgt/aut/cos	Right-authoritarian cosmopolitan:
	Free market globalist liberalism with authoritarian politics and patriarchy.
rgt/lib/nat	Right-libertarian nativist:
	Liberals favoring exclusion of “culturally distant” immigrants
rgt/lib/cos	Right-libertarian cosmopolitan:
	Classical market-liberal individualist-libertarian cosmopolitans
lft/aut/nat	Left-authoritarian nativist:
	Communitarian and authoritarian conceptions of family, state, and nation. Extreme: Left-wing fascism, more mildly “welfare chauvinism”
lft/aut/cos	Left-authoritarian cosmopolitan:
	e.g. social Catholic and Protestant universalists
lft/lib/nat	Left-libertarian nativist:
	Libertarians favoring exclusion of “culturally distant” collectivist-traditionalist (and anti-redistributive) value immigrants
lft/lib/cos	Left-libertarian cosmopolitan:
	Open borders, political grassroots participation and individual life-style choice, combined with egalitarian, redistributive economy

Each of the resulting eight configurations, mapped in Table 1, has its own ideological-philosophical coherence. The political polarization literature on Western societies typically focuses on ideological “sorting” processes that combine economic redistribution with libertarian governance and cosmopolitan citizenship (lft/lib/cos), at one extreme, and market liberalism with authoritarian governance and nationalist-nativist citizenship (rgt/aut/nat) at the other (Kitschelt, 1995b; Levendusky, 2009; McCarty, 2019). While these preference combinations may be particularly common – especially in Western societies more so than anywhere else (Henrich, 2020) – all eight configurations are theoretically plausible and empirically relevant. Let us here only elaborate on the three distinct groups that are usually lumped together into a left-authoritarian quadrant:

• Left-libertarian nativists (lft/lib/nat) support secular, social, and civic individualism (governance) and economic redistribution (distribution), but are apprehensive about labor market and/or socio-cultural consequences of globalization and immigration (membership). A potentially important configuration of preferences are redistributive, pro-welfare state libertarians who are worried about immigrants who bring with them “culturally distant” values that libertarians may perceive to threaten the foundation of a libertarian and redistributive polity. ⁶

• Left, authoritarian cosmopolitans (lft/aut/cos) may characterize mainstream Catholic and Protestant Christians who insist on traditional conformity adhering to norms concerning kinship, gender roles and law and order (governance), but also embrace a powerful universalistic desire for economic equality (distribution) and cosmopolitan inclusion (membership).

• Redistributive traditionalist nationalists (lft/aut/nat) embrace redistribution (distribution) and are also concerned about both the erosion of domestic order and norms (governance) as well as immigration (membership).

Preference Distributions and Partisan Choice

How do partisan choices emerge from preference configurations? We are employing a rule of thumb, based on simple spatial theory, to predict vote choice: the number of “considerations” in favor or opposed to a party, weighted by salience of the issue dimension, tips the balance among competing partisan alternatives. The same prediction emerges from established theories of political information processing: the more considerations on an issue dimension are spontaneously salient due to a voter’s experience—or can be made more salient by political suppliers—the more likely it is that citizens weight such issue dimensions and policy considerations more heavily relative to dimensions and issues that command lesser attention (Zaller, 1992). For those issues that rise to the top of the agenda and are associated with a great number of considerations, voters will search for parties whose programmatic positions match their own preferences.

We assume that in early 21st century knowledge societies on average all dimensions—distribution, governance, membership—have a rather high baseline of salience and generate a fair number of considerations in each partisan choice situation. In that case, it is the number of positional matches between voters’ demand functions and parties’ supply functions of preference configurations across the three dimensions that may tip the balance in voters’ calculus of partisan choice. Is a party’s preference supply matching a voter’s demand on none, one, two or three issue dimensions?

To illustrate, for a market-liberal authoritarian nativist (rgt/aut/nat) a left-libertarian cosmopolitan party (LFT/LIB/COS) provides not one single issue match, and the same applies to the mirror image situation for a left-libertarian cosmopolitan voter facing a right-authoritarian nativist party. In contrast, the demands of left-libertarian-cosmopolitan (lft/lib/cos) voters are perfectly matched by several parties that inhabit the LFT/LIB/COS octant.

Most of the eight groups of voters we identified have one or several party options that position themselves in the same octant, making them obvious party choices. The situation is different for the three “left-authoritarian” octants, which are not populated by a single significant party in Western Europe – the puzzle motivating this paper. These left-authoritarian voters, therefore, are not as well represented and are “cross-pressured” on one dimension or another. For example, how should a left-authoritarian nativist (lft/aut/nat) citizen vote? On distribution, she might prefer a Social Democratic party, which usually offers LFT/LIB/COS; on governance, she might prefer a Christian Democratic/conservative party, which usually offers neutral positions on redistribution combined with authoritarian and cosmopolitan positions (=/AUT/COS); on membership, she might prefer a Radical Right party, which usually offers RGT/AUT/NAT. Weighing all three dimensions equally, the right-authoritarian-nativist positioning of a Radical Right party might offer the best congruence, especially if it holds moderate positions on the economic dimension (=/AUT/NAT). In contrast, a typical Social Democratic party (LFT/LIB/COS) is unlikely to make the consideration set, unless it makes quite radical adjustments by adopting outright anti-immigrant stances on membership (LFT/LIB/NAT) or authoritarian positions on governance (LFT/AUT/COS), or at least neutral stances on both governance and membership (LFT/=/=). Either of these far-reaching changes, however, might alienate previous voters of left parties, particularly left-libertarian cosmopolitans, and generate an overall negative electoral “issue yield” (De Sio & Weber, 2014) for Social Democrats.

To explicate this calculus of voter-party matches, we apply a simple scoring rule: each “match” between supply and demand earns one point, each partial match earns half a point, and each mismatch receives zero points. ⁷ Table 2 shows the resulting match scores for the three left-authoritarian groups (listed in columns). In the rows of the table, we list the three party families that are most clearly competing for left-authoritarian voters: Radical Right, Social Democratic, and Center-right/Christian parties. As our baseline scenario, we assume that (1) all party positions are fixed and (2) all three dimensions have the same weight. The scores predict not just the performance of an individual party type, defined by its profile of programmatic appeals on the three dimensions. It rather predicts a whole configuration of vote shares, namely which programmatic party type is likely to attract (1) larger vote shares from voters with a specific preference profile than in the general population, (2) which type is the likely runner-up relative to the party’s overall national vote share, and (3) which type is the likely loser in relative electoral performance. The main competition for voters with a particular preference profile is between the winning party type and the runner-up parties.

Table 2.

Voter-Party Preference Matches

​		Voters*			Average
		[1]lft/aut/nat	[2]lft/lib/nat	[3]lft/aut/cos
Party positions	(1) Populist radical right party	2.0	1.0	1.0	1.33
	(RGT/AUT/NAT)
	(2) Social democratic party	1.0	2.0	2.0	1.67
	(LFT/LIB/COS)
	(3) Center-right/Christian party	1.5	0.5	2.5	1.50
	(=/AUT/COS)
​	Average	1.50	1.17	1.83	1.50

*Note. Cell entries can range from 0 to 3, with higher numbers indicating higher representativeness (or a closer match). Bolded entries are highest values in each column.

[1] lft/aut/nat: Left-authoritarian nativist voter.

[2] lft/lib/nat: Left-libertarian nativist voter.

[3] lft/aut/cos: Left-authoritarian cosmopolitan voter.

The match of left-authoritarian nativist voter demand (Table 2, column [1], “lft/aut/nat”) with the empirically most common position of right-authoritarian nativist party supply in the 1990s (RGT/AUT/NAT, row (1))—if following Chapel Hill Expert Survey scoring—is a pretty robust 2.0 (Table 2, row (1)/column [1]). The match score for the left-authoritarian nativist and social democratic party dyad would be only 1.0 (row (2)/column [1]), thereby putting Social Democrats beyond hope to ever gain many of these left-authoritarian nativist voters. Center-right/Christian parties fall somewhere in between, with a match score of 1.5 (row (3)/column [1]). Thus, the best match for left-authoritarian nativist voters would be a Radical Right party. But this configuration also yields the prediction that the Center-right parties should be most competitive with the Radical Right among voters with this preference profile.

The strategic situation of competitors is different if voters demand left-libertarian nativism (Table 2, column [2], “lft/lib/nat”). Here, Radical Right parties score only 1 (row (1)/column [2]). By contrast, Social Democrats may start out in a competitive position with a score of 2 (row (2)/column [2]). Economically centrist traditionalist cosmopolitan parties have the lowest match scores for left-libertarian nativist voters, scoring a paltry 0.5 (row (3)/column [2]). Overall, Social Democratic parties are the best match for the voters in column [2]. Radical Right parties are the runner-up, focusing inter-party competition for members of this group on Social Democrats and Radical Right parties.

For the demands of left-authoritarian cosmopolitan voters (Table 2, column [3], “lft/aut/cos”), finally, the situation of Radical Right parties is less favorable with a score of only 1 (row (1)/column [3]). For this voter category, on the supply side the center-right Christian-traditionalist parties are in front with a score of 2.5 (row (3)/column [3]). Social Democratic parties are the runner up, with a match score of 2 (row (2)/column [3]).

Overall, Table 2 reveals that the typical actual party family (rows (1) to (3)) are such that each of the three left-authoritarian voter groups (columns [1] to [3]) has the highest congruence (highest match score) with a different party family. For left-authoritarian nativists (column [1]), the best match (or closest party family) is the Radical Right (row (1)/column [1]). The score is 2 out of 3 (match on governance and membership, mismatch on distribution). For left-libertarian nativists, the best match is the Social Democratic party, which matches on distribution and governance, but mismatches on membership for a score of 2 (row (2)/column [2]). Finally, left-authoritarian cosmopolitans are closest to Center-right/Christian parties which match on governance and membership, and partially match on distribution for a score of 2.5 (row (3)/column [3]). Moreover, for each configuration there are different runner-ups and losers in relative electoral performance.

Thus, in Western European party system, the three left-authoritarian voter groups do not have a unique party option inhabiting the same precise positional octant. Instead, they face different second-best options that align them with conventional left and right political parties or radical right alternatives. In contrast, the other voter groups are well represented by party options in the same octant. As mentioned above, left-libertarian cosmopolitans (lft/lib/cos) inhabit the same octant as typical Green and Social Democratic parties (and sometimes also radical left parties). On the other extreme, right-authoritarian nativists (rgt/aut/nat) would be perfectly represented by parties in the RGT/AUT/NAT octant. This is the typical position of Radical Right parties. Therefore, in Table 3, we assign the remaining voter groups the highest match score of 3. One exception is the group of right-authoritarian cosmopolitans (rgt/aut/cos): as we show below, there are no significant parties in that octant either. ⁸ Since they have relatively close second-best options, we assign this group of voters a match score of 2.5. The three left-authoritarian voter groups receive the scores derived in Table 2, all of which are less than the maximum of 3.

Table 3.

Ideological Configurations and Best Matching Party Families

Groups	Label	Party families with high match distribution/governance/membership	Max match score *
rgt/aut/nat	Right-authoritarian nativist	Radical right (RGT/AUT/NAT) or center-right/Christian and liberals (=/AUT/NAT)	2.5−3.0
rgt/aut/cos	Right-authoritarian cosmopolitan	Center-right/Christian or liberals	2.5
		(RGT/AUT/=)
rgt/lib/nat	Right-libertarian nativist	Center-right/Christians or liberals (RGT/LIB/NAT) or radical right (RGT/AUT/NAT)	2.5−3.0
rgt/lib/cos	Right-libertarian cosmopolitan	Liberal center-right (RGT/LIB/COS)	3
lft/aut/nat	Left-authoritarian nativist	Radical right (see Table 2)	2
lft/aut/cos	Left-authoritarian cosmopolitan	Center-right (see Table 2)	2.5
lft/lib/nat	Left-libertarian nativist	Social democratic (see Table 2)	2
lft/lib/cos	Left-libertarian cosmopolitan	Left-libertarian or social democratic	3
		(LFT/LIB/COS)

Note. *Match scores are explained above.

Table 3 lists the match scores for each of the eight voter groups, along with the party families that are spatially close to them and therefore would provide the best available representation. To be sure, not all party systems feature all party families, but Table 3 serves as stylized summary of Western European party systems. In case of the effective economic-distributive positions of most center-right parties, we have again awarded half points (=).

We hypothesize that voter groups are more likely to vote for parties that are inhabiting the same octant or, if empty, for the party closest to them. With respect to the three left-authoritarian voter groups, we derive the following three hypothesis, based on the above considerations summarized in Tables 2 and 3:

• Hypothesis 1: The electoral preferences of left-authoritarian nativists (lft/aut/nat) make them favor an authoritarian nativist party, even if it takes an anti-redistributive stance (RGT/AUT/NAT). As their second-best choice, they might opt for a mainstream center-right party, while voting for a center-left party would be less attractive for them than any other preference configuration that includes demands for economic redistribution.

• Hypothesis 2: Left-libertarian nativists (lft/lib/nat) have a propensity to support left-of-center parties, even if the latter are cosmopolitan. Their second-best choice would be a radical right party, with a mainstream center-right party being comparatively least attractive for voters with this preference profile.

• Hypothesis 3: Left-authoritarian cosmopolitans (lft/aut/cos) gravitate toward center-right parties, might opt for center-left parties as their second-best choice, and are generally least available to authoritarian nativist parties.

In the appendix, we show that changing the salience of the dimensions – such as giving immigration and/or economic redistribution double weight – will not change the order of relative winners, runner-ups or trailing contenders (appendix 8.3). Moreover, incremental positional modification of the established parties on immigration will not disturb their ordinal rank in success for individual preference groups (appendix 8.4). Finally, we conjecture about the possibility of party entry and conclude that conditions are most promising for a new left-libertarian nativists party (appendix 8.4). This is a topic we speculate about in the penultimate section.

Empirics: Configurations of Policy Preferences, Partisan Choices and the Left-Authoritarian Party Void

In Figure 1 above, we employed various two-dimensional spaces but did not explore the joint distribution of governance and immigration issues. The evidence in this section will operate with a three-dimensional issue space (distribution/governance/membership) to match our theoretical approach. When dichotomizing three issues in pro and contra positions, there are eight, not four, preference combinations, with the left-authoritarian quadrant housing three of them: left-authoritarian-cosmopolitan, left-libertarian-nativist, and left-authoritarian-nativist.

Using this three-dimensional set-up, our empirical strategy proceeds in three steps. First, we show that each of the eight octants is inhabited by a large share of voters, including those with the three left-authoritarian combinations. Second, we show that, in contrast, the left-authoritarian octants are largely uninhabited by relevant parties in Western Europe. Third, we show that the eight constructed preference groups have distinct partisan dispositions that match our theoretical predictions. People with different shades of left-authoritarianism are available for a potential left-authoritarian party to varying degrees, but they could also continue to support traditional left, conservative, and radical right parties to varying degrees.

Presence of Left-Authoritarian Citizens in Western Europe

We first turn to citizens. As above, we employ the “Joint EVS/WVS 2017−2022 Dataset” (EVS/WVS, 2024), which combines data from the European Values Study (EVS 2017−2021, round 5) and the World Value Survey (WVS 2017−2022, round 7). To measure citizens’ positions in three dimensions, we rely on one survey item per dimension:

• Distribution: [e035] Income equality: 1 We need larger income differences as incentives … 10 Incomes should be made more equal.

• Governance: [f120] Abortion: 1 Never justifiable … 10 Always justifiable.

• Membership: [c002_01] When jobs are scarce, employers should give priority to [nationality] people over immigrants: (1) Strongly agree, (2) Agree, (3) Neither agree nor disagree, (4) Disagree, (5) Strongly disagree.

Appendix lists detailed question wordings. It also shows that the results are robust with respect to measuring policy preferences via scales (instead of single items), different scoring rules, and a different dataset (the European Social Survey). Moreover, online appendix 8.2 replicates the main tables and figures using the European Social Survey.

To explore the three-dimensional nature of the data, Figure 2 displays scatterplots of attitudes on abortion vs. immigration, for three (aggregated) levels of the income equality item. These scatters give a sense of the three-dimensional preference distribution. We can see that there are skews (some positions are generally more popular) and affinities (some items correlate with each other). The correlations in the different panels are about 0.15 (right panel) to 0.25 (left panel), indicating that libertarian and cosmopolitan values go together, as do authoritarian and nativist values. But the correlations and explained variances are relatively small. More relevantly, we once again find that in each panel, all quadrants are populated, containing about 17.8% to 37% of all respondents. In the left panel of pro-redistribution preferences, the size of the quadrants is 17.8% (NE), 37% (NW), 21.6% (SW), and 23.6% (SE).

Figure 2.

West European voters in the three-dimensional spaces. Source: Joint EVS/WVS 2017−2022 Dataset. Datapoints are jittered. Solid grey lines indicate midpoint of scales. Solid white lines indicate averages in sample. Sample: AUS AUT CAN CHE DEU DNK ESP FIN FRA GBR GRC ISL ITA JPN NLD NOR NZL PRT SWE USA

To explore the potential heterogeneity of respondents with left-authoritarian values in a three-dimensional space, we next explore the joint distribution of attitudes by dichotomizing each of the three survey items into pro and contra positions. This yields eight groups characterized by distribution/governance/membership bundles. Respondents are coded as having a “pro” position if their attitude is above the country-specific mean, “contra” otherwise (we show that the results are robust to different scoring rules; see appendix).

We are now able to calculate the share of respondents in each of the eight groups defined by the combination of pro- and contra-positions on distribution, governance, and membership, as listed in Table 1 above. Table 4 lists the total number of respondents in our sample belonging to each of the eight groups. In the penultimate and ultimate columns, we show the size of the groups as unweighted and survey weighted percentages, respectively. The numbers suggest that all eight groups are empirically important, covering between about 9.1 and 15.6% of the population (weighted). Furthermore, the three left-authoritarian groups are sizeable and sum to about 1/3^rd of the population (left-libertarian nativists: 12.6%, left-authoritarian cosmopolitans: 9.1%, left-authoritarian nativists: 12.2%).

Table 4.

Eight Combinations of Political Orientations

Groups	Label	N	% (unweighted)	% (weighted)
rgt/aut/nat	Right-authoritarian nativist	5,675	13.4	13.7
rgt/aut/cos	Right-authoritarian cosmopolitan	3,968	9.4	9.4
rgt/lib/nat	Right-libertarian nativist	5,812	13.8	13.7
rgt/lib/cos	Right-libertarian cosmopolitan	5,909	14.0	13.8
lft/aut/nat	Left-authoritarian nativist	5,163	12.2	12.6
lft/aut/cos	Left-authoritarian cosmopolitan	3,714	8.8	9.1
lft/lib/nat	Left-libertarian nativist	5,053	12.0	12.2
lft/lib/cos	Left-libertarian cosmopolitan	6,930	16.4	15.6
All	​	42,224	100	100

Note. Grey cells are “left-authoritarian” combinations. The last column is weighted using survey weights.

Source: Joint EVS/WVS 2017−2022 Dataset.

Lack of Left-Authoritarian Parties in Western Democracies

We now turn to parties. We expand on the two-dimensional evidence of the introductory section by providing more empirical (three-dimensional) details on the “supply side.” To measure parties’ positions, we again employ the Chapel Hill Expert Survey (CHES) Trend File, (Jolly et al., 2022; Rovny et al., 2025). ⁹ In particular, we make use of the following CHES items: ¹⁰

• [Distribution] LRECON: “Parties can be classified in terms of their stance on economic issues such as privatization, taxes, regulation, government spending, and the welfare state. Parties on the economic left want government to play an active role in the economy. Parties on the economic right want a reduced role for government.” Scale: 0 = extreme left, 5 = center, 10 = extreme right.

• [Governance] SOCIALLIFESTYLE: Position on social lifestyle (e.g., rights for homosexuals, gender equality). Scale 0 = Strongly supports liberal policies, 10 = Strongly opposes liberal policies.

• [Membership] IMMIGRATION: Position on immigration policy. Scale: 0 = Strongly favors a liberal policy on immigration, 10 = Strongly favors a restrictive policy on immigration.

Our theoretical framework operates with a three-dimensional space, and we therefore show a descriptive figure that indicates the positions of Western European parties in that space. To do so, Figure 3 displays three panels: the left panel shows economically left parties, defined as parties that fall into the 0 to 4 range on the LRECON party expert measure in CHES. The right panel displays economically right parties, which score in the range 6−10 on the measure. The middle panel shows parties that are middling on the economic left-right measure (scores 4−6). For each of the three panels, the horizontal axis refers to the second dimension of socio-cultural governance, ranging from libertarian (“lib”) on the left to authoritarian (“aut”) on the right. The vertical membership axis displays immigration positions, where values on the top indicate cosmopolitan (“cos”) positions while lower values indicate nativist (“nat”) positions.

Figure 3.

Western European parties in the three-dimensional policy space. Sources: CHES and EVS/WVS. Distribution: Economic left-right [0 left 10 right]. Governance: Social lifestyle [0 open 10 closed]. Membership: Immigration [0 open 10 closed]. Dark circles are party positions (CHES, 2006−2024 averages) for parties ≥5% only. Western Europe (AUT BEL CHE DEU DNK ESP FIN FRA GBR GRC IRL ISL ITA LUX NLD NOR PRT SWE). Gray diamonds indicate approximate average citizen positions based on the Joint EVS/WVS 2017−2022 Dataset

Parties are displayed as dark circles, proportionally sized to their vote share. Only parties with at least 5% are shown. When excluding parties with middling economic left-right positions (i.e., when excluding the middle panel), the eight octants are the left and right panel, respectively. There is clear clustering in the lft/lib/cos octant (top-left quadrant in the left panel) and the rgt/aut/nat octant (bottom-right quadrant in the right panel), where most parties reside. The other octants are more sparsely populated, or entirely empty. Turning to the three left-authoritarian octants, we find that the left-authoritarian nativist octant (left panel, bottom-right quadrant) and the left-authoritarian cosmopolitan quadrant (left panel, top-right quadrant) are completely void of parties. In contrast, the left-libertarian nativist octant (left panel, bottom-left quadrant) features several parties, some quite sizeable. However, all these parties are centrist either on governance or membership, or both (the white area in the middle of the figure). For example, the most nativist party in the octant is the Danish Social Democrats, but the party is fairly centrist on the socio-political governance dimension (see also Figure 7 in the appendix). Turning to economically conservative parties (right panel), we find that three of the four octants are clearly populated by parties. But one octant remains (essentially) empty when we ignore the white area: there are no significant clearly right-authoritarian cosmopolitan (rgt/aut/cos) parties. In appendix, we show that using several CHES items to form scales yields similar results, as does employing a different party expert survey that includes a broader set of Western democracies (Benoit & Laver, 2006).

The figure also provides a rough sense of the positions of citizens in the three-dimensional policy space (using the EVS/WVS data). Because citizen positions are measured in a different dataset using different items with different scales this is not an exact science. However, we indicate the average positions of the eight distribution/governance/membership preference groups with gray diamonds (scaled by size, but on a different scale than the parties) that are expressed in terms of standard deviations away from 5, the theoretical midpoint of the party expert survey scales. This scaling yields that the eight preference groups are each roughly two standard deviations away from the midpoints, and quite symmetrically so.

Thus, Figure 3 reaffirms the motivating puzzle, using a three-dimensional policy space: while there are citizens in all octants, they face party options only in some cases. Of most interest to us, of course, is the lack of significant left-authoritarian parties of any shade. We next explore the partisan leanings of citizens with diverging three-dimensional left-authoritarian preferences.

Preference Profiles and Partisan Dispositions

We are now ready to test our main hypotheses. To explore the partisan dispositions of the eight distribution/governance/membership groups, we code survey items on partisan preferences into party families, using the excellent Parliaments and Governments Database (ParlGov) (Döring et al., 2023) to facilitate the task. ¹¹ We distinguish six substantive party families: radical left (RL), green left (GL), moderate left (ML) – which are mostly social democratic parties – moderate right (MR), liberal (LB), and radical right (RR) parties. Painting with a broad brush, the patterns for the three “left” parties (RL/GL/ML) are similar to each other, as are the patterns for MR and LB. We therefore operate in some exhibits with a simpler party family landscape, distinguishing RL/GL/ML (left) vs. MR/LB (moderate right) vs. RR (radical right). We include all country-years in which radical right parties emerged as a serious alternative to the established mainstream parties. We define as the cutoff criterion elections in which an RR party received more than 5% in a survey.

According to our first hypothesis, among supporters of economic redistribution, leftist parties should have the least opportunities to attract left authoritarian nativists (lft/aut/nat) and Radical Right parties should do well among voters in this category. Row 5 of Table 5 reveals that this is, indeed, the case. All left parties perform below par in this preference category and collect only 31.0% of the vote, compared to their overall average of 37.1% of the vote, whereas the radical right support in this group is higher than the national average (24.1% to 17.5%). But it is not just the Radical Right that does well in this group. It is also the mainstream conservative right (MR + LB) that collects a respectable share of this electorate (44.9% compared to their 45.5% average). If there is an intense competitive exchange, it is mostly between MR and RR, something that is also confirmed by the heavy flow of voters between MR parties and RR parties in many West European countries (see Abou-Chadi & Wagner, 2024).

Table 5.

Partisan Dispositions of the Eight Attitude Groups

Groups			Partisan leanings
​		% a	Left				Moderate right			RR
			RL	GL	ML	RL + GL + ML	MR	LB	MR + LB	RR
1	rgt/aut/nat	13.7	3.5	3.6	13.4	20.5	36.5	18.1	54.6	24.9
2	rgt/aut/cos	9.4	3.4	7.7	20.0	31.2	37.5	21.7	59.2	9.7
3	rgt/lib/nat	13.7	3.8	4.3	16.5	24.6	25.5	25.4	50.8	24.6
4	rgt/lib/cos	13.8	4.6	11.3	19.8	35.7	27.5	26.8	54.2	10.1
5	lft/aut/nat	12.6	6.2	5.4	19.3	31.0	32.3	12.6	44.9	24.1
6	lft/aut/cos	9.1	7.0	13.3	26.2	46.5	28.0	15.3	43.2	10.3
7	lft/lib/nat	12.2	9.5	9.8	23.3	42.7	18.9	16.0	34.9	22.5
8	lft/lib/cos	15.6	11.9	22.7	26.1	60.8	11.3	17.0	28.4	10.9
​	Average	​	6.5	10.1	20.5	37.1	26.2	19.3	45.5	17.5
​	N = 21,925	​	​	​	​	​	​	​	​	​

Note. Row percentages: (RL + GL + ML+MR + LB + RR) sum to 100 for each row.

Sample: AUT CAN CHE DEU DNK FIN FRA GRC ITA JPN NLD NOR SWE.

^aPercent of respondents, not conditional on vote in last election. Data are weighted.

Our second hypothesis says that left parties should garner a larger share of left-libertarian nativists (lft/lib/nat) and be competitive in this group with Radical Right parties. Indeed, as row 7 shows, left parties receive 42.7% of support in this group, compared to their overall average of 37.1%, and—interestingly—it is the economically most leftist RL parties where this over-representation is most dramatic. Contrary to previous studies with a two-dimensional framework (Halikiopoulou & Vlandas, 2023), voters with anti-immigrant sentiments are not wholesale lost to parties of the left, at least not if they also embrace leftist positions on distribution and libertarian positions on governance. There is mild over-representation among ML (Social Democratic) parties, but under-representation of this group in the Red-Green Left (GL), fully committed to both libertarian governance and cosmopolitan citizenship principles. As predicted, the least competitive group for these voters are the Christian-conservative and liberal parties (34.9% compared to 45.5%), whereas the Radical Right performs well (22.5% compared to 17.5%). The left bloc (RL + GL + ML) – and especially radical left and social democratic parties – may still face a great deal of downside in this anti-immigration group, particularly whenever immigration becomes a highly salient political issue. This is the group where the contest may be focused on the left partisan camp against the radical right camp, with the Center Right on the sidelines.

Our third hypothesis predicts that among economically progressive socio-cultural traditionalists—say the labor wings of Christian Democratic parties—with universalist, cosmopolitan orientation toward inclusion of immigrants (lft/aut/cos) the Radical Right has little chance, and the electoral payoffs reported in row 6 of Table 5 confirm this pattern. Here voters flock to the Left (46.5%), and particularly the moderate Left (26.2%), but with the center-right showing strong capacity to attract voters as well (43.2%). The data confirm a tight contest between center-left and center-right for a substantial share of the entire electorate (9.1%). In contrast, the RR strongly underperforms with this category of voters (10.3% vs. 17.5% average support).

Finally, as expected (though not explicitly specified as a hypothesis), left parties—and especially the red-green left—are completely hegemonic among left-libertarian cosmopolitans (lft/lib/cos, row 8). In this group, amounting to 15.6% of all voters, the left bloc crushes all the other competitors by attracting over 60% of the vote, whereas the Center Right can collect only 28.4% and the Radical Right 10.9%.

Figure 4 summarizes the disproportionate support of the eight attitude groups for three party groups: RL/GL/ML vs. MR/LB vs. RR. The horizontal axes display disproportionate support (deviations form average support for a party group, based on a multinominal estimation with a set of controls, namely age, gender, income, education, self-employment, unemployment, union membership, church attendance, as well as country- and year-dummies; see Table 6 in online appendix 8.1 for estimation results). The figure also shows the results from various robustness checks with respect to items vs. scales, different scoring rules, and different datasets, as explained in appendix. There is a high degree of overlap in the results.

Figure 4.

Partisan dispositions of the eight preference groups. Note. Based on multinominal logit estimations that include a set of controls, including country- and year-dummies. See Table 6 in online appendix 8.1 for estimation results

Figure 4 shows that support for the left increases as one moves from right-authoritarian nativist to left-libertarian cosmopolitans, by over 40 percentage points (from −16.6% below average to 23.7% above average; top-left panel). Support for the moderate right (MR/LB) displays the mirror image pattern and changes by more than 25 percentage points (from 10.1% above average to −17.1% below average; top-middle panel). It is telling, however, that the moderate right reaches its peak among right-authoritarian cosmopolitans (+13.7%), not anti-immigration nativists, thereby demonstrating the constraints on the moderate right when considering embracing anti-immigration positions. Support for the radical right (RR) zigzags by about 15 percentage points because it is most strongly driven by nativism vs. cosmopolitanism. Nevertheless, support for radical right parties declines monotonously from the best fit (rgt/aut/nat) via intermediate fit – where voters disagree with RR parties either on their socio-political governance stance (rgt/lib/nat) or their stance on questions of economic redistribution (lft/aut/nat). Nativist voters who are least connected to radical right parties are those who disagree with them on both economic and governance issues (lft/lib/nat). ¹² Among the four groups of nativists, this latter group is also the one that displays the strongest average support for parties of the Left, potentially tempting one of the left parties to moderate its stance on immigration with the objective of reducing the share of left-libertarian nativists opting for radical right parties.

For completeness, Figure 5 shows the size of the eight groups, based on the different operationalizations discussed in appendix. In the EVS/WVS data, different operationalizations yield comparable group sizes. In contrast, because the ESS items are heavily skewed (pro redistribution, pro libertarian governance, pro cosmopolitanism), relative versus absolute scoring rules make a big difference. However, Figure 4 shows that the vote propensities of the different groups are remarkably robust with respect to different operationalizations.

Figure 5.

Size of the eight preference groups

Discussion

Given the findings of this investigation, what helps explain the absence of left-authoritarian parties in Western democracies? We have shown that citizens in the “left-authoritarian” quadrant of the typical two-dimensional policy space have divergent positions on issues related to socio-political governance and immigration. Our three-dimensional conceptualization of the policy space can uncover this preference heterogeneity. It yields three types of “left-authoritarians”: left-libertarian nativists, left-authoritarian cosmopolitans, and left-authoritarian nativists. These different preferences are associated with support for different political parties. Each type of left-authoritarian tends to disproportionally align with one party family and sometimes a second that match their views on at least two of the three dimensions: Left-authoritarian nativists tend to be overrepresented among radical right and moderate right parties but underrepresented among leftist parties. Left-libertarian nativists, in turn, gravitate toward either radical right or leftist parties, but avoid center right parties. Left-authoritarian cosmopolitans, finally, mostly ignore radical right parties and choose between left or right mainstream alternatives. All three preference dimensions matter to account for the distribution of vote choices in each preference octant.

Consequently, representation is not a matter of all-or-nothing, but of better-or-worse. Different types of left-authoritarian voters, based on their three-dimensional preference profiles, have second-best party options that matches on at least two out of the three preference dimensions. This preference heterogeneity of left-authoritarian voters makes it difficult for a single new party to attract them, especially when established parties pre-empt party entry by marginally adjusting their policy positions to improve the positional match with certain shades of left-authoritarians.

To speculate, one possibility is that multiple parties will emerge to serve each octant of voters in the three-dimensional policy space with “first best” programmatic appeals that match citizens’ preference bundles. The proliferation of parties with narrow policy profiles in the Netherlands might be a sign of what is to come in other Western European democracies (there are currently 15 different parties in the Dutch parliament). Given the difficulty of party entry in most party systems, however, a more plausible – though still speculative – scenario is that the left-authoritarian representation gap will be partially closed, either by existing parties marginally adjusting their strategies, or by new parties offering novel packages in the three-dimensional space. New parties would have to choose between combining demands for economic redistribution with authoritarian governance and nativism, or with authoritarian governance and cosmopolitanism, or with libertarian governance and nativism.

In terms of a new party hoping to attract a substantive share of the roughly third of citizens with left-authoritarian preferences, we see several challenges. First, because of the documented preference heterogeneity among left-authoritarians, there is no “first-best” package of policy positions a new party could offer, as just mentioned. Second, some shades of left-authoritarians are more reachable than others. For example, “double authoritarians” (left-authoritarian nativists) are likely least available to leftist parties as their support for left parties is much below average – and by far the lowest among all types of left-authoritarians. Such voters may continue to support a radical right party, even if it violates their economic interests. Third, our analysis suggests that left-libertarian nativists are particularly “reachable” by a (new) party close to them as they have the lowest match score with the currently available party types in Table 2 (see also online appendix 8.4). Finally, an important hurdle for hopeful new parties is that existing parties have incentives to preempt party entry by adjusting their strategies.

This brings us to the question whether existing parties will close the left-authoritarian representation gap by moving toward particular shades of left-authoritarianism. Existing parties might be tempted to do so to increase their own vote share, to increase their party camp’s vote share, or to deter party entry, or a combination thereof. Let us speculate about more likely attempts of established parties to improve the “second best” representation of particular types of left-authoritarians.

First, Radical Right parties might try to improve their standing with left “double authoritarians” (left-authoritarian nativists) by moderating their conservative economic position (=/AUT/NAT). Some evidence suggests that this is happening, but primarily by radical right parties embracing social protection for natives (“welfare chauvinism”), not more far-reaching demands for economic redistribution from the rich for which there is little evidence in radical right programmatic appeals. ¹³ And movements toward a more social welfarist position on economic policy may risk precipitating a split in the radical right camp, with new “bourgeois” free-market populist right wing parties outflanking the established Radical Right, as recently happened in Denmark and The Netherlands. ¹⁴

Second, center-right parties may modify positions on polity membership by moving away from what was at times a firm endorsement of immigration (=/AUT/=), even though there is little evidence that such moves yield substantial electoral gains (Krause et al., 2023). But there would be little space left for a new entrant.

Third, left-wing parties may adopt more centrist positions on citizenship to cater to left-nativists, especially with libertarian persuasions, as the Danish Social Democrats arguably have done (see Etzerodt & Kongshøj, 2022; Hjorth & Larsen, 2022). Our analysis suggests that this might be a promising strategy: left-libertarian nativists’ “second best” representation among the existing party families are left-libertarian cosmopolitan parties, with the Radical Right as the runner-up (see Table 2). But will left parties moderate their cosmopolitanism to attract and maintain the support of left-libertarian nativists, especially if a novel party were to appear with a “first best” programmatic profile? Such parties might not find it an unambiguously winning electoral strategy to adopt restraint on immigration (Abou-Chadi & Wagner, 2019, 2020). On the one hand, more neutral stances on immigration might lose the party more voters than it can gain. On the other hand, the strategy might strengthen the left field overall by attracting voters with distinctive leftist preference profiles, such as especially left-libertarian nativists, to stick to or return to parties of the left field (Kitschelt & Rehm, 2024). Whether a party is willing to sacrifice its individual fortunes in favor of strengthening the left field remains to be seen. ¹⁵

Overall, then, the programmatic-spatial conditions for filling the left-authoritarian void are difficult. Existing center left parties have to tread cautiously because of electoral trade-offs, while a new leftist party would have to tailor its programmatic appeal quite precisely to the left-libertarian nativist element of the population, the constituency for which the established parties offer the least representation. For new parties, these difficulties are compounded by the more general challenges of party entry, such as screening entrants into the party, building a party organization, attracting attention, gaining credibility, and so on. Hence, even a promising combination of policy positions in the three-dimensional space might not be enough to get a new party electorally off the ground. Rather, additional exogenous opportunities may be critical to successfully launching such a party.

It is too early to tell, but the case of Germany’s Bündnis Sahra Wagenknecht (BSW) may offer an empirical illustration. This new party has clearly adopted left-libertarian nativist (anti-immigrant) positions (Hoffmann, 2025; Steiner & Hillen, 2025) and it almost cleared the high 5% hurdle in the 2025 national legislative election. But in addition to BSW’s programmatic appeal, its initial electoral performance was importantly shaped by factors beyond spatial programmatic positioning: the party leader’s (and namesake’s) personal charisma; her origin in Eastern Germany that intensified the identity divide between West and East Germans; and the war in Ukraine that mobilized sympathy with Russia in Eastern Germany and with the Ukraine in the West (Wagner et al., 2023, pp. 631–632).

Conclusion

Why are there no left-authoritarian parties? The puzzle of the left-authoritarian representational void in Western democracies may be smaller when put into a three-dimensional policy context. Analyzed in two dimensions, the sizeable share of left-authoritarian citizens appears to be a homogenous, entirely unrepresented group. In contrast, a three-dimensional perspective reveals that left-authoritarians have heterogenous preferences on salient issues, namely questions of governance and citizenship. Moreover, different shades of left-authoritarians are partially represented by different existing parties. Thus, there is not only no homogeneous bloc of left-authoritarians available for a single new party to capture, but existing parties also offer partial representation that they could strengthen to discourage new party entry.

We think it is likely that party competition will become increasingly multi-dimensional and that the number of relevant parties will continue to rise. If so, there will be no easy strategies for existing or new parties. For example, radical right parties might hope to obfuscate their positions on issues unrelated to immigration (Rovny, 2013), but many voters with nativist preferences are also motivated by issues of distribution and/or governance. Moreover, parties hoping to govern will need to take positions on all relevant dimensions, at least eventually. As another example, Social Democratic parties might be tempted to improve their standing with left nativists. While a move in that direction might increase the combined vote share of all left parties, it could decimate the moving party’s own votes as its core constituency of left-libertarian cosmopolitans starts supporting other left (mainly green) parties.

Because parties can only move incrementally and because successful party entry faces steep hurdles, the representation gap for all shades of left-authoritarians is likely to close only slowly. But left-authoritarian cosmopolitan, left-authoritarian nativist, and left-libertarian nativist parties are likely to crop up in the future, perhaps in this order. Thus, the puzzle of missing left-authoritarian parties might disappear altogether, but not overnight.

Supplemental Material

Supplemental Material - Why No Left-Authoritarian Parties?

Supplemental Material for Why No Left-Authoritarian Parties? by Herbert Kitschelt and Philipp Rehm in Comparative Political Studies