When automation hits home: Exploring the nuanced relationship between job risk and support for universal basic income

Introduction

The advent of automation and its rapid integration into various industries has sparked considerable debate on the potential consequences for the job market and workers’ livelihoods (Barbosa et al., 2022). Automation, driven by advancements in artificial intelligence and robotics, has the capacity to streamline processes, boost productivity, and increase efficiency across multiple sectors (Brynjolfsson et al., 2023). However, this progress also raises concerns about widespread job displacement, as machines and algorithms replace human labor (Acemoglu and Restrepo, 2018; Arntz et al., 2016; Frey and Osborne, 2017). Consequently, there has been a growing interest in studying the relationship between automation of jobs and the attitudes toward Universal basic income (UBI), as a potential solution to address the socioeconomic challenges posed by technological advancements (Busemeyer and Sahm, 2022; Dermont and Weisstanner, 2020; Haglin et al., 2024; Martinelli, 2019; Miailhe, 2017).

UBI refers to an unconditional and guaranteed income provided to all individuals within a specific community, without the requirement of means testing or work obligations (Van Parijs, 2004). Although the idea of UBI is not new and has been discussed on numerous occasions throughout history (Bejarano Beltrán et al., 2019; Vanderborght, 2021), it has gained significant prominence in recent decades, evolving from a theoretical proposal discussed primarily in academic circles to becoming a pioneering measure in social policy, generating considerable interest in pilot experiences (Alston, 2019; Arnold, 2018; De Wispelaere, 2016; Zimmermann et al., 2020). The COVID-19 pandemic revitalized public debate and academic interest in UBI (Arnold, 2020; Ståhl and MacEachen, 2021; Thompson, 2022; Weisstanner, 2022). While some research shows that public opinion remains divided on the measure, it has been observed that, during the health crisis and its aftermath, people expressed stronger support for UBI due to the perceived bureaucratic simplicity and efficiency compared to traditional conditional subsidy schemes (Nettle et al., 2021).

The essence of basic income lies in its provision of cash payments to recipients, making it compatible with conventional conditional subsidies and enabling the coexistence of other in-kind benefits, such as basic education and public health services (Standing, 2017). Notably, basic income entails regular and recurring payments, distinguishing it from one-time universalist measures that might be applicable only at specific life events, such as reaching the age of majority (Van Parijs and Vanderborght, 2017). Numerous public opinion surveys have provided descriptions of the concept of basic income before asking people about their preferences for its implementation in their country of residence. The majority of these surveys have been carried out in recent years, reflecting the substantial surge in both public and political interest directed toward basic income (Laenen, 2023).

However, since UBI remains a potentially innovative policy in countries with consolidated welfare states, the scholarly exploration of citizen support in this arena is still in its formative stages. Nevertheless, an emerging body of research has begun to document discernible patterns in this domain. Previous studies show that individual support for UBI may be influenced by several factors. It’s a well-established understanding that attitudes toward welfare state policies, including basic income, are shaped by two key factors: individuals’ sociodemographic position within society and their ideological views (Laenen, 2023; Roosma and Van Oorschot, 2020). From a sociodemographic perspective, young people with low income, precarious employment, and income insecurity tend to be more supportive of the idea of introducing a basic income (Baranowski and Jabkowski, 2021; Kuhn, 2017). From an ideological standpoint, basic income garners greater support among individuals aligning with left-wing political ideologies, endorsing egalitarian principles, advocating for pro-migration policies, believing in the deservingness of most current welfare recipients, and holding positive assessments of the efficacy of the existing welfare system (Bay and Pedersen, 2006; Parolin and Siöland, 2020; Schwander and Vlandas, 2020; Vlandas, 2019, 2021). Conversely, several other individual characteristics appear to have minimal influence on the level of popular support for basic income, despite theoretical expectations suggesting otherwise. Specifically, these characteristics include gender, household composition, employment sector, and trade union membership (Bertomeu and Vollenweider, 2011; Cantillon and McLean, 2016; Chrisp et al., 2020; Laenen, 2023; Lombardozzi, 2020; McKay, 2001; Miller et al., 2023; Vollenweider, 2013).

In addition, there is a growing literature that explores the association between automation risks and social policy preferences (Busemeyer and Sahm, 2022; Busemeyer and Tober, 2022; Dermont and Weisstanner, 2020; Gallego et al., 2022; Häusermann and Kurer, 2022; Im, 2020; Thewissen and Rueda, 2019; Weisstanner, 2023). Technological change is causing major transformations in the nature of work and the labor market. One of the most significant changes that technology is bringing to the labor market is the automation of tasks using artificial intelligence, robotics, and other advanced technologies. This is expected to significantly affect the nature of work and the types of jobs that will be available in the future (Acemoglu and Restrepo, 2019; Brynjolfsson and McAfee, 2014; Ford, 2015; Manyika et al., 2017; Schlogl et al., 2021; Spencer, 2018; Upchurch, 2018; World Bank, 2019). Moreover, in recent years there has been growing concern about the polarization of the labor market because of the automation of certain tasks, especially routine ones. Research focusing on high-income nations, where robust evidence is available, indicates that since the late 20th century, technological advancements have automated numerous routine tasks associated with medium-skilled occupations, with some scholars positing the ‘hollowing out’ of the middle classes. These alterations in occupational landscapes have given rise to labor market polarization, characterized by simultaneous increases in wages and employment shares within both high- and low-wage occupations, while mid-level jobs witness declining wages and workforce participation (Autor and Dorn, 2013; Kurer and Gallego, 2019). This phenomenon serves as a significant contributor to economic inequality and is prevalent across many high-income nations (Adermon and Gustavsson, 2015; Fonseca et al., 2018; Goos et al., 2009, 2014; Goos and Manning, 2007; Sebastian, 2018).

Nonetheless, as Gallego and Kurer (2022) argue, individuals facing an economic downturn do not point the finger directly at technological progress for the decline in their economic well-being. Instead, they often attribute this downturn, at least in part, to related but separate economic changes, such as international trade and immigration. Likewise, those who benefit economically fail to recognize that digital economies disproportionately favor people like themselves; instead, they embrace narratives of meritocracy to justify their positions (Sandel, 2020).

Despite this, the ongoing advancement of information technology is rapidly transforming the labor landscape, with automation and artificial intelligence increasingly displacing routine and repetitive tasks. This evolution poses significant challenges to the existing welfare system, which has traditionally relied on a model of stable, long-term employment. With the potential for a substantial portion of the workforce to be displaced by technology in the near future, the need to rethink welfare state structures to ensure economic security and well-being for all citizens becomes evident (Muñoz de Bustillo, 2019a). Implementing a UBI could be an effective response to the challenges posed by automation and technological disruption in the labor market (Arat and Waring, 2022; Bregman, 2017; Ford, 2015; Miailhe, 2017; Srnicek and Williams, 2015; Straubhaar, 2017). It would provide a financial cushion for those whose jobs become obsolete due to technological advancements, enabling them to retrain or reinvent themselves professionally. In addition, it could foster innovation and entrepreneurship by providing people with a safety net to take on labor risks and explore new opportunities without the fear of lacking basic income (Feinberg and Kuehn, 2020; Standing, 2017). However, the implementation of a UBI also raises a host of challenges and important questions about how to sustainably finance it, how to ensure it does not disincentivize work, and how to ensure it genuinely reaches those who need it most (Flassbeck, 2017; Kay, 2017; Schneider, 2017). Ultimately, the exploration of a UBI and other innovative ways to restructure the welfare state has become a crucial topic in the political and economic debate on addressing the emerging challenges of the digital economy and automation.

In this changing scenario marked by uncertainty about the impact of automation, exploring attitudes toward basic income is vital for several reasons. First, public perception plays a significant role in the implementation of social policies. Understanding how people perceive basic income in the context of automation-induced job displacement can inform policymakers about potential challenges and opportunities in garnering support for such a transformative social welfare measure. Second, basic income is not a one-size-fits-all solution, and attitudes toward it may vary across different demographic groups, socioeconomic strata, and cultural contexts. In-depth research on these attitudes can help tailor policy proposals and address potential concerns. Moreover, studying the relationship between automation and basic income can shed light on broader societal implications, such as income inequality, social cohesion, and economic stability, shaping the discourse on inclusive economic and social systems.

This research delves into the complex interplay between work automation, individuals’ perceptions of the risk of job displacement, and their support for basic income as a protective mechanism. The study capitalizes on a unique juncture provided by Catalonia’s Basic Income Pilot Plan project. This project aims to allocate a basic income to 5000 people almost universally, excluding only the top 10% of incomes. In case the pilot is finally implemented, which depends on possible coalitions between parliamentary forces, its design establishes that the allocation will be randomized across two Catalan municipalities and 2500 citizens throughout the territory, with adults receiving 800 euros and minors 300 euros. On a European scale, this project exhibits a notable scope, surpassing the recent experiments conducted in Finland (Halmetoja et al., 2019; Kalliomaa-Puha et al., 2016; Kangas et al., 2017).

This circumstance makes the study of public support for the basic income in Catalonia particularly interesting, as the debate about its benefits and possible negative effects is present in public opinion. The Center d’Estudis d’Opinió (CEO), a public body of reference in opinion studies in Catalonia, took the initiative to gauge citizens’ perceptions on the proposed basic income in its Barometer of April 2023. This comprehensive survey encompasses a vast amount of information on Catalonia’s geographical, economic, demographic, political, and social realities, as well as includes attitudinal variables, habits, customs, and aspects related to participants’ way of life. In the context of such a unique natural experiment of the Basic Income Pilot, the survey represents a valuable methodological tool to study the social and political support for this measure and to explore the underlying social bases driving public sentiments toward it. The insights gained from this study have the potential to contribute significantly to the ongoing discourse surrounding basic income, offering nuanced perspectives on its potential implications for European societies.

Theoretical background and hypothesis

This paper aims to contribute to the expanding understanding of technological risks and support for social protection policies amid the backdrop of the digital age, following the trajectory of previous works such as those by Thewissen and Rueda (2019), Dermont and Weisstanner (2020), Gallego et al. (2022), Busemeyer and Sahm (2022) and Haglin et al. (2024). The welfare state encompasses an array of public interventions aimed at shielding citizens from various risks or contingencies such as unemployment, illness, disability, income loss, among others (Muñoz de Bustillo, 2019b). In the context of job automation, the debate on UBI as a social protection mechanism has gained traction in the public sphere. However, less attention has been paid to the question of which policies those most likely to be affected by automation and digitalization demand for themselves (Busemeyer and Sahm, 2022).

In this regard, the study of Thewissen and Rueda (2019), using a ‘routine task intensity’ index (RTI), find that individuals in occupations at high risk of automation express significantly higher levels of support for redistributive policies to offset anticipated income losses due to unemployment, particularly more generous unemployment insurance programs. Kurer and Häusermann (2022) also find that workers at high risk of being displaced by technological change support policies of more spending on unemployment benefits, but not on pensions or other policies. Thus, these studies suggest that changes in risk exposure, stemming from technological advancements, breed uncertainty regarding future income, fueling individuals’ demand for government redistribution. However, other papers find no relationship between automation risk and preferences for redistribution (Gallego et al., 2022; Zhang, 2022) or find that demand for redistributive policies increases only under certain circumstances (Jeffrey, 2021). Thus, the evidence on this issue so far is mixed. Similarly, the growing evidence on automation risk and labor market policy preferences is also inconclusive. Several studies find that at-risk workers do not support active employment or educational policies (Busemeyer and Sahm, 2022; Häusermann and Kurer, 2022; Weisstanner, 2023), while Im (2020) arrives at opposite results.

Nonetheless, these studies did not address whether the risk of automation impacts individuals’ support for UBI. This gap is addressed by Dermont and Weisstanner (2020), who used the same data from Thewissen and Rueda (2019) on RTI but with a different dependent variable: popular support for the introduction of a basic income scheme. They find no association between the risk of job automation and support for UBI. In the same vein, Busemeyer and Sahm (2022) find no statistically significant association between individual automation risk and support for basic income. In contrast, Sacchi et al. (2020) find that RTI is correlated with support for UBI among some subgroups of voters in Italy. Again, the evidence is mixed.

The peculiarity of the basic income is its universal nature, which makes it more appealing to the middle-class constituency of the ‘new left’ than to the traditional working-class constituency of the ‘old left’. The risks of job automation are disproportionately concentrated among routine workers in the middle of the skills and earnings distribution, who may have a vested interest in traditional income support but not necessarily in a universal system that could replace parts of the existing welfare system. Dermont and Weisstanner (2020) suggest that UBI differs from traditional redistributive policies in two crucial aspects: first, there is considerable uncertainty surrounding this measure, hence the opinion formation process is still ongoing, and second, there is resistance to UBI from groups supporting traditional redistributive policies (elderly, non-college-educated individuals, union members).

This finding underscores the importance of studying the multidimensionality of political preferences. Research suggests that it is highly unlikely that all the groups within a particular society will feel exactly the same about introducing a basic income (Laenen et al., 2023; Roosma and Van Oorschot, 2020). The literature provides mixed results on whether individuals at risk of job loss due to automation view UBI as a viable remedy. Yet, as Busemeyer and Sahm (2022) argue, it is plausible to assume that individuals in jobs at high risk of automation are more likely to support the introduction of a UBI – or a guaranteed citizenship income (GCI) – as a basic form of social protection.

UBI and GCI are both policy proposals aimed at providing financial support to individuals, but they differ in key aspects. UBI is a universal, unconditional cash transfer provided to all individuals regardless of their income or employment status. It is designed to ensure a basic level of income security for everyone. In contrast, GCI is a means-tested policy that targets individuals and families facing poverty, providing financial support based on their income levels. While UBI aims to simplify the welfare system and expand individual liberty through universal coverage, GCI focuses on directing resources to those most in need, potentially at a lower overall cost. Both policies have been proposed as solutions to the economic disruptions caused by automation, but they offer different approaches to addressing income insecurity and labor market changes.

While our study primarily focuses on UBI, we recognize that GCI also presents a viable alternative to traditional social insurance programs in addressing automation-induced unemployment. UBI has garnered significant attention in recent policy debates due to its universal and unconditional nature, which proponents argue could more effectively address the widespread displacement of workers caused by automation. However, GCI offers a more targeted approach by providing support specifically to those who are most economically vulnerable. By including both UBI and GCI in our analysis, we aim to contribute to a comprehensive understanding of how different income support policies can mitigate the challenges posed by technological advancements.

In sum, our first working hypothesis is:

Next, one of the key issues raised by previous research concerns the inadequacy of the data used to capture the potential mismatch between the objective probability of an individual’s job being automated and their subjective perception of this looming risk (Dodel and Mesch, 2020; Im and Kim, 2022; Nam, 2019). The risk associated with job automation is based on objective assessments based on previous estimates and indices that operationalize the likelihood of certain occupational roles being replaced by technological advances such as artificial intelligence and robotics (Arntz et al., 2017; Autor et al., 2003; Frey and Osborne, 2013; Nedelkoska and Quintini, 2018). In social research, survey respondents are typically assigned index scores that provide researchers with a comprehensive understanding of the objective probabilities associated with technological obsolescence of the employment positions held by those individuals. Therefore, researchers know the objective probability that the jobs of respondents will be replaced by technology, but do not necessarily capture their views or concerns about this fact.

Gallego et al. (2022) introduced and empirically tested hypotheses relating the objective risk of job displacement due to the introduction of new technologies to individuals’ subjective concerns about the impact of technological change in the workplace. The study revealed a gap between the objective risk of job automation and the subjective concerns expressed by employees. Most participants held an optimistic view, believing that the introduction of new technologies in the workplace could lead to positive outcomes, while only a minority expressed concerns. Overall, perceptions of automation may be a key mechanism linking the objective risk of technological displacement and policy preferences for compensation. Hence, our second working hypothesis is:

Finally, the literature points out that the way in which basic income is described and framed has an impact on its popularity (Laenen, 2023). Furthermore, Weisstanner (2023) notes that in some countries workers may be more aware of the risks of automation than in others, or may feel better protected by the state, which may explain variations across countries in popular support for basic income. Accordingly, it is reasonable to think that respondents who are given background information on automation prior to being asked about their support for basic income will be more likely to be in favor of such a measure, especially those who are at high risk of being displaced by technology. Therefore, our third working hypothesis is:

Methods and data

Data and sample

The survey was conducted in person, at the residence of each participant. The company GESOP was responsible for data collection, following the specifications of the CEO of Catalonia, to obtain a representative sample of the population eligible to vote in Catalonia. The sample was collected through a multistage strategy. Initially, using a stratified random sampling, 211 census units from Catalonia were selected from six sectors defined based on their sociodemographic similarity. Within each census section, random routes were designed to interview between 10 and 13 people in their homes. The sample of individuals was selected based on cross quotas of gender, age, and place of birth, according to the data from the Continuous Population Register Statistics of the year 2020. The age ranges used were 18–24 years, 25–34 years, 35–49 years, 50–64 years, and 65 years or older. Regarding place of birth, individuals born in Catalonia, Spain, and abroad were considered. The 92% of the interviews were conducted in the 211 selected sections. The remaining 8% were completed in substitute sections selected by the CEO based on geographical proximity and sociopolitical similarity. Although no data on refusal rates or dropout quotas are reported, the face-to-face interview method and the data collection strategy minimize the risk of potential biases compared to other types of survey designs. ¹

The final sample consisted of 2000 participants. Table 1 displays some sociodemographic characteristics of the sample.

OPEN IN VIEWER

Table 1.

Sociodemographic characteristics of the sample.

Variable	n (%)
Gender
Male	963 (48.2)
Female	1037 (51.9)
Age
18–24 years	163 (8.2)
25–64 years	1305 (65.4)
64 years or older	532 (26.6)
Children at home
No	990 (50.0)
Yes	989 (50.0)
Education
Undergraduate student or less	1376 (68.8)
University or more	623 (31.2)

The data were obtained through the Barometer of Political Opinion survey conducted by the CEO between February and March of 2023. The primary objective of this survey was to explore the perception of residents in Catalonia on various social issues, including politics, economy, media, electoral behavior, and the assessment of various political parties or leaders. The survey consists of 99 questions, distributed across the following sections: (1) Sociopolitical context, (2) Attitudes toward politics, (3) Political values, (4) Social issues, (5) Electoral behavior, and (6) Sociodemographic data. The response options were predominantly multiple-choice or Likert-type scale. The survey was administered face to face with an approximate application time of 30 minutes.

The survey had two distinct formats based on the framing of attitudinal questions regarding UBI and GCI. In the first format of the survey, the question about UBI stated:

Universal basic income is a monthly income of around €800 that would be unconditionally provided to all citizens as a citizenship right, regardless of their situation. To what extent would you agree or disagree with the implementation of a Universal basic income in our country?

Meanwhile, the question regarding GCI stated: ‘Guaranteed citizenship income is a guaranteed income for individuals and families facing poverty. To what extent would you agree or disagree with the implementation of a Guaranteed citizenship income in our country?’

In the second format of the survey, these two questions incorporated contextualization about task automation, concluding the aforementioned instructions as follows: ‘. . . considering the current processes of technological change and task automation?’ That is to say, half of the participants received the two questions without contextual information about automation, while the other half obtained contextualization. Response options to these questions were of the Likert-type, ranging from 0 = completely disagree to 10 = completely agree.

Results

Descriptive analysis

Figure 1 presents the distribution and descriptive statistics of all variables used in this study. In this figure we can see that the sample composition is nearly equivalent regarding sex, household composition (presence of children), and automation context. The mean age of the participants was 51.29 (SD = 17.85), and the average income was €845.08 (SD = 622.73). In addition, the sample exhibits a tendency toward a left-wing political ideology (M = 3.96; SD = 1.69).

OPEN IN VIEWER

Figure 1.

Distribution and descriptive analysis of the study variables.

It can be observed that the attitude toward GCI is more favorable (M = 7.36; SD = 2.60) than the attitude toward UBI (M = 5.79; SD = 3.44). Moreover, participants on average believe that only 30.16% (SD = 29.16) of the tasks they currently perform in their jobs could be automated by machines or algorithms in the coming years. While participants express a relatively high concern about being surpassed or replaced by a machine, program, or workers with greater digital skills (M = 3.25; SD = 0.91), there is no relationship between the objective RTI index and the perceived risk of replacement by automation (r = –.302; p > 0.05).

When comparing the attitude toward a UBI based on sociodemographic variables (see Figure 2), no significant differences are observed. In contrast, when comparing the attitude toward a GCI (see Figure 3), various differences are evident. First, individuals with university education have a more favorable attitude than those with lower education level. Similarly, those without children have a more positive attitude toward a GCI. Finally, it is observed that those who were asked about GCI without mentioning technological changes or automation have a more positive attitude toward this social policy.

OPEN IN VIEWER

Figure 2.

Comparison of attitudes toward universal basic income based on sociodemographic variables.

OPEN IN VIEWER

Figure 3.

Comparison of attitudes toward guaranteed citizenship income based on sociodemographic variables.

Regression models

Regarding the UBI model (see Table 2), there is an adequate goodness of fit of the data F(11, 901) = 9.498, p < 0.001, and an R-squared of 0.104. Regarding the coefficients, it can be observed that political orientation is significant, where those who report a left-leaning tendency tend to have a more favorable attitude toward UBI. In addition, those with more favorable attitudes toward immigrants and higher trust in institutions also tend to have a more positive attitude toward this policy. Finally, those who perceive a higher risk of being replaced by automation show greater support for a UBI.

OPEN IN VIEWER

Table 2.

Universal basic income model.

	B (SE)	β	t
(Constant)	4.985 (1.027)		4.856***
Age	−.007 (.009)	−.024	−.739
Gender	−.002 (.222)	.000	−.009
Income	.001 (.001)	.043	1.229
Household composition	−.018 (.249)	−.003	−.072
Trade union membership	−.148 (.333)	−.014	−.445
Political orientation	−.429 (.061)	−.231	−6.983***
Attitudes toward immigration	.277 (.096)	.098	2.896**
Trust in institutions	.230 (.060)	.121	3.799***
Automation context	.401 (.220)	.058	1.825
Perceived risk of automation	.008 (.004)	.064	2.007*
RTI index	−.094 (.117)	−.026	−.797

*

<.05; **<.01; ***<.001.

Regarding the GCI model (see Table 3), there is an adequate goodness of fit, F(11, 904) = 10.279, p < 0.001, with an R-squared of 0.111. In this model, a significant effect of sociodemographic variables – such as gender or income – and political orientation is observed. This indicates that women, those with higher incomes, and those who report a left-leaning political orientation tend to have a more positive attitude toward the GCI. In addition, there is a significant effect of attitudes toward immigration and trust in institutions. More positive attitudes toward immigrants and greater trust in institutions are associated with stronger support for a GCI. Finally, considering the automation index proposed by Autor and Dorn (2013), we observe that support for a GCI declines as job automation risks increases.

OPEN IN VIEWER

Table 3.

Guaranteed citizenship income model.

	B (SE)	β	t
(Constant)	5.842 (.783)		7.465***
Age	.008 (.007)	.036	1.116
Gender	.380 (.169)	.071	2.245*
Income	.001 (.001)	.084	2.407*
Household composition	−.046 (.191)	−.009	−.243
Trade union membership	.116 (.253)	.015	.459
Political orientation	−.293 (.047)	−.205	−6.259***
Attitudes toward immigration	.251 (.073)	.116	3.430**
Trust in institutions	.122 (.046)	.085	2.660**
Automation context	−.088 (.167)	−.017	−.527
Perceived risk of automation	.005 (.003)	.051	1.614
RTI index	−.215 (.089)	−.079	7.465**

*

<.05; **<.01; ***<.001.

Discussion

The results of this research show that support for a UBI is generally lower than support for a GCI. In Catalonia, the general public seems to prefer the government to provide a guaranteed income for individuals and families facing poverty rather than doing so unconditionally to all citizens as a citizenship right, regardless of their situation. This preference aligns with the broader empirical evidence from opinion polls, which mostly suggest that selectively targeted, means-tested basic income models receive greater support than their fully universal counterparts (Laenen, 2023). In Spain, Rincon (2021) shows a moderately higher level of support for a targeted basic income (with a mean of 44.8) as opposed to a universal alternative (with a mean of 41.7). Thus, the findings of our study in Catalonia resonate with the broader trend observed in empirical research on basic income preferences. All this might suggest that people want to improve the conditions of those who are worse off in their country (Roosma and Van Oorschot, 2020). If this were the case, then it is not the universal character or its unconditionality that people in Catalonia value, but the fact that it provides poor people with a guaranteed minimum income.

The debate over welfare policies hinges on whether universal or selective approaches are more effective and morally just (Korpi and Palme, 1998; Laenen and Gugushvili, 2021). Some argue that universal welfare policies, such as UBI, enjoy broad support due to their appeal across social classes. Moreover, proponents argue that universal welfare programs hold moral superiority over selective programs (Bidadanure, 2019; Rothstein, 1998). However, others suggest that selective policies may be equally or more popular because they target the poor directly, resonating with a sense of fairness (Laenen and Gugushvili, 2021). From a self-interest perspective, opting for a universal system might not appear as the most rational course, particularly for middle- and high-income earners who would probably become net contributors, paying more in taxes than receiving benefits (Laenen, 2023). In addition, some argue that targeting assistance toward those in need is inherently fairer than dispersing scarce public resources to all, including those who may not genuinely require them. Be that as it may, our results show that in Catalonia there is a more favorable attitude toward targeted policy income rather than universal income. In this sense, it is important to consider that in each area of policy formulation, basic income has to compete with other policies, including means-tested social assistance. Therefore, it is very important to analyze the relative support that basic income has compared to other alternative policies (Rincon, 2021).

However, it is possible that the lower level of support for UBI in our survey may be attributed to the perceived insufficiency of the €800 amount rather than opposition to the universality of the policy. Previous research has shown that more generous UBI proposals tend to receive greater support (Laenen, 2023; Van Parijs and Vanderborght, 2017). Furthermore, the framing of the GCI as benefiting ‘individuals and families facing poverty’ likely elicited more favorable responses due to the sensitive nature of poverty alleviation (Van Oorschot and Meuleman, 2012). In contrast, the UBI item did not include such poverty framing, potentially contributing to its lower support. Thus, the possible influence of the methodological design of the survey when asking about support for the different policy proposals should not be underestimated.

Generosity in basic income refers to the level at which it is set, with debates focusing on partial versus full basic income. Partial basic income is relatively low, often below subsistence level, and may not be supplemented by other social benefits, while full basic income is higher and intended to meet subsistence needs. Survey experiments indicate higher support for basic income above subsistence levels, though some findings suggest acceptance of lower amounts if they supplement other benefits (Laenen, 2023).

In Spain, according to the National Institute of Statistics, the at-risk-of-poverty threshold in a person’s households stood at 10,990 euros per year in 2023. For its part, the minimum wage was 15,120 euros per year. In other words, the level established in the survey of our study would be well below the subsistence level and much lower than the minimum wage, so it would not be surprising if the lower support for the UBI is related to this situation.

In addition, we recognize the importance of isolating the effects of universality versus means-testing on public support for basic income policies. To address this, future research should aim to control more rigorously for variables such as the specific amount of the basic income and the framing of the policy’s target beneficiaries. For instance, varying the amount of UBI in a range that includes subsistence and above-subsistence levels, and standardizing the language used to describe both UBI and GCI policies, can help disentangle the impact of these factors on respondents’ preferences. By adopting such methodological improvements, subsequent studies can provide clearer insights into how the core dimensions of universality and means-testing independently influence public attitudes toward basic income schemes.

A second notable finding is the importance of political ideology in determining support for UBI and GCI. Our results show that ideological leaning to the right is associated with lower support for UBI and targeted basic income policies in Catalonia. This finding is consistent with some previous studies that have already documented that support for basic income varies significantly across political views, with left-leaning individuals showing the highest levels of support and those identifying as right-wing expressing the least endorsement (Laenen, 2023; Roosma and Van Oorschot, 2020; Vlandas, 2021). Furthermore, basic income is more popular among advocates of reduced income inequality (Alston, 2019; Mays and Marston, 2016). These proponents of egalitarianism likely see basic income as a promising mechanism for narrowing the socioeconomic inequality through redistribution. Conversely, support for a basic income tends to be lower among those who harbor negative sentiments toward immigrants, with lower levels of support observed among individuals who believe immigrants have a negative impact on their country of residence, compared to those who view immigrants as enhancing the nation’s quality of life (Bay and Pedersen, 2006; Bell et al., 2023). In this regard, the linear regression model for the UBI and GCI in our study shows that left-wing political orientation is significant, as well as favorable attitudes toward immigrants and trust in institutions. We could assume that in Catalonia those who trust institutions make a positive evaluation of the performance of the welfare state, which is also in line with previous literature on the topic (Laenen, 2023).

In contrast, sociodemographic variables such as age, gender, income, and household composition have no significant correlation with support for basic income in Catalonia. Evidence from opinion surveys conducted in Europe shows that there are hardly any differences in support for basic income between men and women or in households with or without children (Baranowski and Jabkowski, 2021; Laenen, 2023). Age and income are different. In previous research, the nexus between support for a basic income and chronological age has been characterized by a negative correlation, implying that as individuals advance in age, their propensity to lend support to this social policy innovation tends to diminish (Baranowski and Jabkowski, 2021; Kuhn, 2017; Parolin and Siöland, 2020). Notably, this trend may not solely reflect a cohort effect but rather a life-cycle effect, suggesting that support for basic income among the youth might diminish as they age (Laenen, 2023). This observation aligns with the notion that as individuals mature, their circumstances tend to stabilize, leading to a shift toward more conservative values – but it is worth adding that this may strongly change in the future as it is already changing in the present, because, with the transformations in work relations and life trajectories, what is stabilizing is precariousness, which leads us to think that, increasingly, the precarious population will grow older without abandoning its precarious condition and, maybe, without putting aside its transformative sociopolitical agenda (Standing, 2011). In terms of income, in many nations, individuals with high incomes tend to express the strongest opposition to basic income, whereas those with lower incomes are more inclined to support it (Laenen, 2023; Vlandas, 2021). While in our results these sociodemographic variables are not significant in support for UBI, women and high-income people support targeted income as a social policy.

Similarly, we find no significant correlation between union membership and support for basic income, nor is it significant for GCI support. This finding is in line with previous work using data from the European Social Survey in which there appears to be no evidence to support the assumption generally made about opposition to basic income by trade unionists (Vlandas, 2019, 2021).

An interesting part of the results is that although people are highly concerned about being replaced in their workplace by technological change, we find no relationship between the RTI index (objective risk of replacement) and their perceived risk of replacement (subjective risk). Specifically, there is a dissonance between what participants think and the likelihood that their work will be replaced by machines, algorithms, or digital technologies in the coming years. Research on workers’ concerns about automation and how these influence their policy preferences is still scarce. One of the few studies that have addressed the relationship between automation risk, subjective concerns and policy preferences is that of Gallego et al. (2022). Their findings show that objective automation risks are not good predictors of workers’ subjective concerns and policy preferences.

A priori, one might expect that workers exposed to a greater risk of automation would be in favor of income redistribution policies aimed at ensuring the conditions of material sufficiency necessary for a dignified life. However, that is not always the case. While Gallego et al. (2022) report that technological concern varies with objective vulnerability, as workers at higher risk of technological displacement are more likely to negatively view technology, our results show that, regardless of the objective risk of automation, it is only when people are aware that some of the tasks they perform in their jobs could be automated that they tend to position themselves in favor of UBI.

Nevertheless, it is crucial to recognize that what we term ‘objective’ assessments, such as those derived from the RTI index, inherently include subjective elements. These assessments are based on expert-driven calculations that estimate the probability of job automation based on the codifiability and substitutability of tasks. While these measures offer standardized and systematic ways to gauge automation risks across various occupations, they are ultimately grounded in broad generalizations and assumptions that may not capture the specific realities of individual job roles. Acknowledging this subjectivity is key to better appreciate the limitations and strengths of these expert-driven assessments.

In parallel, incorporating ‘subjective’ measures – reflecting workers’ personal assessments and concerns about technological displacement – adds valuable insights into our understanding of automation risks. Workers often possess nuanced knowledge about the intricacies of their tasks and the potential for technological substitution in their specific contexts. This subjective perspective is critical in capturing the lived experiences and perceptions of those directly affected by technological change. By combining both ‘objective’ and ‘subjective’ measures, we can achieve a more comprehensive understanding of automation risks. This dual approach allows us to balance standardized assessments with individual perceptions, providing a richer and more detailed picture of how automation may impact the labor market. Ultimately, leveraging both types of measures enhances our ability to develop informed and effective policy responses to the challenges posed by technological advancements.

On the contrary, Haglin et al. (2024) examine the relationship between individuals’ beliefs about the likelihood of their jobs being automated and their attitudes toward UBI in the United States of America. Their main findings reveal a complex interplay between perceptions of automation, political partisanship, and support for UBI. While there is a general awareness of automation as a potential threat, U.S. citizens do not uniformly perceive it as an imminent danger to their jobs. Interestingly, there is a significant partisan divide in the support for UBI. Democrats tend to support UBI regardless of their perceived risk of job automation, indicating a broader ideological alignment with redistributive policies. In contrast, Republicans show conditional support for UBI, which increases only if they perceive a high risk of their jobs being automated.

Thus, the findings of Haglin et al. (2024) complement our results by highlighting the importance of political ideology in shaping public support for UBI. Furthermore, it emphasizes the need to consider how perceptions of automation risk influence policy preferences. Individuals who consider themselves economically vulnerable are more inclined to support UBI, though this vulnerability is perceived differently across political lines. In conclusion, the study suggests that awareness and perception play crucial roles in shaping support for redistributive policies. Therefore, public attitudes toward basic income policies are influenced by a combination of political ideology, perceived economic vulnerability, and awareness of technological impacts.

Surprisingly, we find no significant correlation between context information about automation and support for the introduction of a UBI scheme or a GCI (Hypothesis 3). It is well known that the framing and contextual information provided in opinion polls and surveys play a pivotal role in shaping individuals’ responses and perceptions (Berinsky, 2017). How questions are framed can significantly influence the way respondents interpret and answer them. By framing questions in a certain manner, survey designers can subtly guide respondents toward particular viewpoints or biases. Contextual information provided before asking questions can also influence responses by priming respondents with relevant information or considerations. Our results indicate that the higher the perceived risk of automation, the higher the support for UBI. Despite this, contextual information about the risks of automation does not play a relevant role in triggering workers’ concerns about this possibility, which is an unexpected result for us. Therefore, we do not find support for hypothesis 3, which posited that individuals who have been provided with context about automation are more likely to demonstrate increased support for policies such as UBI or GCI.

Finally, the observation in our regression model that the intensity of routine tasks (RTI) has a statistically significant negative impact on the GCI, that is, that support for a GCI declines as the risks of job automation increase, is indeed intriguing. This result indicates that our hypothesis 1, which posited that individuals facing a high risk of job automation are more likely to exhibit endorsement for policies such as UBI or GCI, was not confirmed but rather rejected.

One plausible explanation is that individuals at high risk of job automation may also fear that a GCI, which is conditional and means-tested, will be insufficient to address the widespread displacement caused by automation. They might believe that a GCI would not provide adequate support for the growing number of people potentially losing their jobs due to automation, and therefore, they may prefer more comprehensive and robust social safety nets. Research has shown that individuals facing economic uncertainty often favor policies that offer broad and reliable support (Brady and Bostic, 2015; Hacker et al., 2013; Rehm, 2011; Standing, 2017). Another explanation could be that those who are at higher risk of job displacement due to automation might also believe in the potential for technological adaptation and the creation of new job opportunities. They might think that policies should focus more on facilitating this transition through education, retraining, and innovation rather than providing conditional income support, which could be seen as a temporary or insufficient measure (Autor, 2015; Brynjolfsson and McAfee, 2014). Finally, individuals concerned about job automation might also worry about the stigmatization associated with means-tested benefits like GCI. Studies have shown that means-tested benefits can sometimes carry a stigma, which might lead individuals to prefer universal benefits that do not single out recipients based on need (Baumberg, 2015; Moffitt, 1983; Stuber and Schlesinger, 2006; Van Oorschot, 2002).

Before concluding, it is worth going back to one of our fundamental assumptions in this study – that the framing of the basic income proposal has a strong impact on its popularity – and putting this evidence in relation to the sociopolitical and media context that the possibility of a basic income pilot has generated in Catalonia. The history of the debate on basic income in Catalonia is not new: since the 1990s and, above all, since the appearance in 2001 of Red Renta Básica (RRB), the Spanish affiliate of Basic Income Earth Network (BIEN), the discussion on basic income left academia and reached the political sphere. Perhaps, the presentation of two bills – one in the Catalan Parliament and another in the Spanish – are the most important milestones of the first years of public discussion on basic income. Social movements that emerged in the wake of the economic crash of 2008 and, subsequently, the circumstances of the COVID pandemic have highlighted the problems that conditional cash transfer schemes drag along and the need to gradually relax conditionalities in the direction of what would mean a fully universal and unconditional basic income.

Yet the difficulties that the Catalan Pilot Project itself has experienced during its parliamentary process show that basic income is far from being a proposal that is unanimously accepted in Catalonia. It is true that the refusal to grant a budgetary allocation to the pilot by the Catalan Socialist Party and Junts, a right-wing pro-independence party, has been due to fights between opposition parties with the aim of undermining the Catalan Government, but it is also true that the arguments that the Catalan Socialist Party has mobilized have to do with the importance of rejecting universality and unconditionality and prioritizing targeted schemes assisting the poorest and most vulnerable segments of the Catalan society. Faced with this situation, it is worth asking whether the debate on the need – or not – for a pilot offers basic income advocates the opportunity to disseminate, explain and frame in a clearer way the proposal. Let us look at three dimensions of the debate that make us think this way.

First, one of the objectives of the pilot was to make the proposal known. Without undermining the importance of obtaining scientific evidence, the pilot project was also designed to give Catalan citizens precise and informed knowledge of a proposal that must be able to be collectively evaluated and, perhaps, implemented. According to a survey commissioned by the Catalan Government’s General Directorate of Analysis and Foresight, knowledge of the details of the basic income would make 80% of citizens willing to approve its implementation (Direcció General d’Anàlisi i Prospectiva, 2023). It is also interesting to note that the week during which it became known that the Catalan Socialist Party would block, with the support of Junts, the implementation of the pilot project, interest and support for basic income increased substantially in Catalonia.

Second, the design of the pilot excludes the richest 10% of the population. This is because the project, despite being based on a budgetary allocation, aims to simulate the tax reform that a real basic income would require – it is a tax reform that has the support of 73% of the Catalan population (Direcció General d’Anàlisi i Prospectiva, 2023). In this sense, although the pilot embraces the principles of universality and unconditionality inherent to basic income, it shows that, in net terms, there would always be winners – the non-rich part of the population – and losers – the most economically advantaged citizens, who would fund the program through the taxation system. This fact forces us to ask whether the defense some make of targeted conditional schemes on the grounds that such schemes favor the most disadvantaged citizens – or the non-rich segments of the population – could also be applied to (and mobilize support for) a basic income which, although received by everyone, requires a tax reform that takes the form of a massive transfer of income from those who have the most to the rest of the population. Can the understanding of these technicalities of pilot projects such as the Catalan one modify people’s attitudes toward universalist policies? Needless to say, this can only be known if the pilot is maintained and finally implemented.

Third and finally, the design and possible implementation of the pilot project offers an excellent opportunity for citizens to understand that basic income is not only aimed at combating poverty and social exclusion, but also at increasing the degrees of freedom of the population protected from such poverty and social exclusion (Casassas, 2024). This is why Erik Olin Wright (2016) called for a shift from ‘static justifications’ of basic income, which refer to the kind of world basic income constitutes – one with less poverty and precariousness, etc. – to its ‘dynamic justifications’, which are concerned with the world that basic income ‘sets in motion’. In this sense, the fact that the design of the pilot includes the study of two municipalities within which everyone would know that everyone – except the richest 10% – would be receiving a basic income could modify attitudes toward universalist policies, here understood as levers for the collective activation of lives felt to be more ours, more autonomously chosen. At this point, it is of crucial importance the understanding that unconditional access to monetary – and in-kind – resources might allow for entries and exits from the labor market not mediated by despair, but by genuine choices of professional and life trajectories, which might allow for lives with more diverse and heterogeneously organized activities – the famous ‘multi-active lives’ referred to by André Gorz (1997). Under these circumstances, the understanding that basic income is not ‘only money’, but also the bargaining power deriving from the universal guarantee of the right to one’s material existence might favor processes of reduction of working hours (Hester and Srnicek, 2023; Srnicek and Williams, 2015) that might go hand in hand with the democratization of the processes of automation of work that we are facing. If carefully designed, framed and explained, could the pilot lead us to the realization that the bargaining power that emanates from the universal guarantee of the right to existence might make us less fearful of the processes of work automation and push us to see in them true windows of opportunity to rethink, from positions of greater social invulnerability, the ways in which we work and live?

The survey carried out by the Catalan government shows that the major change associated with the introduction of a basic income would be in time management: practically one in three people would spend more time on leisure or family and half of the population would consider participating in community life and organizations (Direcció General d’Anàlisi i Prospectiva, 2023). Likewise, many of the North American basic income experiments of the 1960s and 1970s and those carried out during the last 15 years in countries of the Global North and Global South show an increase in social and institutional trust, in sociopolitical and community participation, and in social interaction (Borrell-Porta et al., 2023), which makes clear that having the material guarantee of a dignified existence – and the awareness and security that the rest of the members of the community also have it, as the design of the Catalan pilot emphasizes – makes it possible to establish and consolidate meaningful social ties in the fields of (re)production and social and community participation leading to higher levels of social integration (Casassas, 2024).

As in all the economies of the world, automation has arrived in Spain and, in particular, in Catalonia and is here to stay. In 2016, the Organization for Economic Co-operation and Development (OECD) warned that 12% of jobs could be automated in Spain (Arntz et al., 2016), affecting most severely those jobs that require fewer qualifications, which could deepen the patterns of duality, polarization, and the chronic structural unemployment of the Spanish labor market. Seven years later, in 2023, having incorporated the effect of Artificial Intelligence, the OECD itself stated that 28% of employment in Spain was at high risk of automation. In the Catalan case, we can affirm that 35% of workers perform tasks that, from a technical point of view, are susceptible to being automated in the coming decades with a high probability (Hernández Gascón and Fontrodona Francolí, 2019). In particular, the proportion of people who are very concerned about the possibility of their jobs being replaced by machines or software is twice as high among those who perform less complex tasks – administration, accounting, finance, customer and user service, sales services, logistics, quality and manufacturing (Future for Work Institute, 2019). Thus, given the inadequacy of the conditional cash transfer schemes deployed in Catalonia and Spain, which exclude the vast majority of the population due to the strict nature of their conditionalities and which, when they do not exclude them, offer them meager amounts of income; in a context marked by the social extension of precariousness, especially among young people who know that this same precariousness can be consolidated throughout their life cycle; and given the weight of automation in deepening such precariousness, it seems necessary to articulate policy-making processes based on scientific evidence available to all citizens and consisting in frames and comprehensive and understandable explanations of the causal mechanisms of the foreseeable results of these policies.

Conclusion

By testing several working hypotheses relevant to the literature, this research offers valuable insights into the dynamics of work automation, risk perception, and the potential role of basic income in mitigating job loss concerns. Our study sheds light on the nuanced preferences and determinants underlying support for different forms of income redistribution policies, particularly UBI and GCI, in Catalonia. Our findings suggest a prevailing preference for targeted, means-tested income support over universal approaches among the public in Catalonia, aligning with broader trends observed in empirical research. Our analysis also underscores the influence of political ideology on attitudes toward income redistribution policies, with left-leaning individuals generally showing higher levels of support compared to their right-leaning counterparts. This ideological divide reflects differing perspectives on the role of government intervention and social welfare in addressing societal challenges. In addition, our study highlights the importance of perceived job automation risks in shaping attitudes toward UBI, indicating that individuals who perceive their work tasks as automatable are more likely to support universal income measures. However, we find no significant correlation between contextual information about automation and support for UBI or GCI, suggesting that framing and contextual cues may not strongly influence individuals’ policy preferences in this context – but we wonder whether a stronger and clearer framing strategy such as the one that emerges from the design of the Catalan pilot and the debate it has sparked could boost the extension of political support for basic income. Overall, our findings contribute to a deeper understanding of the complex factors influencing public attitudes toward income redistribution policies and underscore the need for further research to explore these dynamics in greater detail.