Increased Trust in the Finnish UBI Experiment – Is the Secret Universalism or Less Bureaucracy?

1 Introduction

Welfare states are, in many ways, machinery that aims to provide citizens with social protection in a continuum of policymaking, welfare attitudes, and conflicts shaping both attitudes and policies. Furthermore, the relationships between them being bilateral calls for not only more research but also societal experiments to understand the often very complex mechanisms that produce well-being. Under this generalisation lie different versions of welfare states or regimes, as stated in the well-known studies by Esping-Andersen (1990; 2009); (see also Castles & Mitchell, 1993; Ferrera, 1996; Lewis, 1997). One way to differentiate these regimes is to measure the degree of selectiveness or the universalism of their welfare systems.

Nordic welfare states, including Finland, are classified as social-democratic welfare states with a high degree of universalism. Social security systems are built on the idea of solidarity, that is, benefits and services are collectively financed through taxes, and the equality of citizens is promoted in every policy sector (Kvist, Fritzell, Hvinden, & Kangas, 2012). The hegemony of the social-democratic ideology is reflected in the social policy systems and in the deservingness perceptions of citizens. They assess well-being as a universal right – although selectiveness is also supported, especially when benefits and services are targeted to those generally considered less deserving (for example, the unemployed). The extent of universalism in social benefits indicates trust in citizens, which is likely to induce mutual trust between citizens and the state. Not only the citizens but all residents living in these countries have been able to trust that they are treated equally regardless of gender, education, family background or other personal characteristics. Hence, we observe high levels of both generalised and institutional trust in the Nordic countries.

The idea of folkhemmet – a people’s home – is still an underlying principle in all policymaking in the Nordic welfare states, although new distributional conflicts resulting from socioeconomic or political changes may disconnect the constituents’ welfare attitudes from current state policies based on equality and solidarity. These occurrences may affect welfare policies’ institutional legitimacy, which is reflected by the fiscal crisis and the immigration-founded demand for more selectivity and activation models, or the demand for broader social policies linked to higher unemployment levels (Pfeifer, 2009; van Oorschot, 2000, 2006). Indeed, in Finland, as well as in many other European countries, the selectiveness of social security systems has gradually increased, especially with regard to unemployment benefits and last-resort social assistance. The conditions for receiving these benefits have become tighter, and more sanctioning has been implemented in the benefits systems.

Welfare states are, of course, more than policymaking and attitudes. They are about implementation: delivering benefits and services that enable citizens’ participation in society, which typically includes street-level bureaucracy (Lipsky, 1980). These street-level bureaucrats are frontline workers that exercise discretion towards clients, who often need social and economic aid and are, thus, dependent on the state’s consideration. When implemented in a just manner, the bureaucracy itself is not disruptive to the welfare state’s legitimacy. However, negative experiences of bureaucracy can break trust relations, which are fundamental to the functioning of societies (e.g. Hummel, 1977; Kumlin & Rothstein, 2005; see also Kennedy, 2016). These negative experiences can arise if bureaucracy traps individuals or families in need of institutional support, while the system is too fragmented and complicated to understand, or involves too much paperwork.

In this study, a phone survey administered at the end of 2018 in the context of the Finnish universal basic income (UBI) experiment was used to analyse the mechanisms mediating trust and the implementation of the welfare state. Its purpose is to test whether UBI leads, directly and indirectly, to greater trust in the institutions governing income support and to greater trust in fellow citizens. Specifically, we want to know (1) whether the receipt of UBI leads to greater trust directly, (2) whether it is accompanied by an actual reduction in bureaucracy (i.e. less selectivity with less discretion on the part of delivery agents) leading to greater trust or (3) whether it leads to a perceived reduction in bureaucracy that, in turn, leads to greater trust.

As a proxy for the latter we use, on the one hand, the bureaucracy experiences of the respondents, and on the other hand, the recipiency of universal versus selective benefits, that is, the UBI as universal and unemployment benefit as well as last-resort social assistance as selective benefits. According to our results, UBI is accompanied by greater trust. However, the findings also show that selectivity does not lead to less trust, nor does it necessarily lead to a perception of less bureaucracy. In times of great uncertainty and increasing discussion on the ability of present welfare states to provide welfare, our results are most relevant and should be considered in the unavoidable reforms of social security systems.

The structure of this article is as follows. In Section 2, we discuss deservingness perceptions and how they are expressed through social policy instruments. Section 3 elaborates on how UBI, as a universal social policy, could offset deservingness channelling. In Section 4, we explain the Finnish UBI experiment settings, how data were collected, and state our research question and applied methods. Section 5 presents and discusses the results, leading to Section 6, where the general conclusions are stated.

2 Universalism, Selectively

Many characteristics of social policies derived from the institutional foundations of welfare regimes are perceived as legitimate to the extent that they reflect current welfare attitudes and corresponding deservingness criteria. The more equality-founded the regime, the more universalistic its social policies are expected to be. Similarly, the tighter the constituents’ deservingness criteria are, the more exclusive their welfare attitudes and less legitimate the regimes’ universalistic social policies. On the other hand, it is argued that the position of a given welfare state in the continuum of selectivity to universality institutionally shapes constituents’ deservingness perceptions and, hence, their welfare attitudes, as well as their support for the given welfare state system, closing the feedback loop. Welfare regimes are not only legitimised by people’s welfare attitudes but are also shaped by them (Kangas, 2003; Larsen, 2008; Rothstein, 1998; Svallfors, 2012).

As argued by Staerklé et al. (2012, p. 73), subjects’ social position (assessed by material vulnerability and education levels) is a determinant of welfare attitudes, though this path is significantly mediated by their normative beliefs, that is, ‘individually held, but socially shared, values, perceptions and expectations that provide normative support for welfare attitudes’. Kangas (2003) pictured this link through the moral takeaway on laziness in the Aesopian tale of ‘The Ant and the Grasshopper’: the normative, deep belief on human agency – and, thus, on a volitional cause of the unemployment condition – makes the unemployed, for example, less deserving of state support than elderly or disabled citizens. Hence, welfare attitudes can also be read through people’s ideas on deservingness, that is, “the public’s answer to ‘who should get what, and why?’” (Larsen, 2008; van Oorschot, 2000, p. 34).

Policy instruments, which create the relationship between state and constituents, widen the institutional deservingness pathway while being value-grounded institutions, bearing meanings and representations that they can effectively imprint on their subjects (Lascoumes & Le Gales, 2007). In this sense, policy selectivity instruments, such as means-testing, activation and conditionality, reinforce institutional deservingness, as Larsen (2008) summarises:

[T]he boundaries between ‘them’ and ‘us’ generated by a selective welfare policy highlight who benefits from the welfare state (i.e., those who pay little or no tax and receive targeted benefits) and who loses on the welfare state (i.e., those who pay tax but do not receive any benefits). Thus, the reciprocity of the system will be perceived as very low, which increases the importance of grateful, docile, and compliant attitudes among those who receive the targeted benefits or services (p. 152).

Moreover, the measure of selectivity drives stricter deservingness perceptions in ‘welfare state losers’ and increases stigmatisation and alterity towards ‘welfare state beneficiaries’, leading to less solidarity and even stricter deservingness criteria (Esping-Andersen, 1990; van Oorschot, 2002). According to Korpi (1980), tighter selectivity thresholds stimulate welfare backlash at the coalition level: marginal social policies, that is, strictly targeted programmes, weaken the rationality for coalition formation between targeted and non-targeted constituents. These mechanisms can further increase the alterity between welfare ‘losers’ and ‘beneficiaries’, fraying out the social fabric, decreasing mutual trust and strengthening even more institutional deservingness pressures.

At a closer look, the institutional deservingness path through policies is reinforced by selectivity instruments’ requirements at the implementation level as they intrinsically ask for more bureaucratic screening and applying more intricate rules, thus, complicating bureaucratic processes and hampering access to benefits (Calnitsky, 2016; Kumlin & Rothstein, 2005; Rothstein, 1998). These requirements can also make selectivity systems more reliant on the discretion of street-level bureaucrats, which, unless professionally or ideologically tied to divergent principles, is also bound to average welfare attitudes (Blomberg, Kallio, Kangas, Kroll, & Niemelä, 2017; Breit, Alm Andreassen, & Salomon, 2016; Kallio & Kouvo, 2015). Given these actors’ front-rank position in the welfare ‘losers’ versus ‘beneficiaries’ struggle (behind either the counter or the computer screen, i.e. the places where the welfare state comes into existence), the deservingness perceptions of street-level bureaucrats can effectively modulate formal welfare policy rules, channelling deservingness preferences into policy outcomes (Blomberg et al., 2017; Bovens & Zouridis, 2002; Kallio & Kouvo, 2015; Lipsky, 1980; Rice, 2013; Rothstein, 1998).

In addition to the deservingness criteria imprinted in formal selectivity instruments, street-level bureaucrats express, through their discretionary space, their own deservingness perceptions, which, regardless of whether they are tighter or looser, make effective rules more intricate and decisions less predictable. Rothstein (1998, p. 80) refers to this effect as the ‘black hole of democracy […] [it] can be difficult to hold the administrators and officials who decide about the welfare of citizens in any way responsible for their actions’. This is expected to have a substantial impact on welfare recipients’ behaviour, as their encounters with bureaucracy are not merely one way: while the bureaucrats modulate regulations according to their deservingness perceptions, intricacy and uncertainty incentivise the welfare recipients to modulate information provision by their side (Kumlin & Rothstein, 2005).

With regard to universal benefits, no space exists for rules embedding any of the dimensions of deservingness – be it need, control, identity, reciprocity or attitude (see e.g. Larsen, 2008; Rothstein, 1998). In a universal policy, as is the case of UBI, automaticity replaces application procedures and the different degrees of screening. Thus, not only is deservingness eliminated from policy design but the channels no longer exist for its expression through policy processes, and no stigma is caused by bureaucratic discretion (Gentilini et al., 2019). Of course, universal policies bring their own implementation challenges; however, in the case of UBI, those do not open, a priori, any ‘deservingness channels’ as they do not impose any bureaucratic burden on welfare recipients (De Wispelaere, 2015). In other words, from their perspective, UBI does not force any bureaucratic experience on its recipients.

UBI universality might so dispel deservingness by shutting down its channelling mechanisms in policy implementation, as it implies no need for application procedures or street or screen-level bureaucracy intervention in granting benefits (Bovens & Zouridis, 2002; Calnitsky, 2016; Gentilini et al., 2019). Then, the causal link from perceived procedural injustice to distrust, theorised by Kumlin and Rothstein (2005), can effectively be broken not by the dismissal of screening or activation procedures themselves but by the utter absence of deservingness-tainted interactions with the state, protecting not only institutional but also interpersonal trust.

3 Trust, UBI and the Welfare State

The idea of UBI is to financially support people in just and non-bureaucratic ways. While UBI includes no screening and criteria rules, it cannot be taken away, that is, people can trust that they will obtain the money as long as the system itself is not changed, for example, via legislation.

In the UBI system, the payer of the benefit (e.g. state) must trust that the beneficiaries (e.g. citizens) use that money wisely and in a way that increases their well-being. While trust is reciprocal, the receivers of UBI are inclined to trust those who trust them – in this case, the societal institutions implementing the UBI system (e.g. Ostrom, 1990). More generally, social security systems that are built on the principle of trust, as are all universal and unconditional benefits in general and UBI in particular, generate more trust than programmes based on continued screening and means-testing. In the former case, people can trust to get what they are promised, whereas in the latter case, they are more dependent on the mercy and decisions of the bureaucracy. In the end, the predictability of institutions generates procedural trust with fewer rules and simpler sanctioning systems (Sztompka, 1999; Tyler, 2001; Tyler & Huo, 2002).

In other words, if we want to strive for a society of high trust, a journey in the social security system should be so simple that citizens always know what they are entitled to, and no obstacles exist between them and receiving the cash benefit. Excessively complex procedures and arbitrary requirements make it difficult for citizens to operate in the social security system, as they cannot be sure that they will get the benefit to which they expect to be entitled. In this case, the system is considered ‘bureaucratic’ in a perverted sense.

In a highly selective social security system, it can be difficult to maintain citizens’ trust in social institutions, as beneficiaries must constantly demonstrate that they deserve and are entitled to the benefit. The high transaction costs and inefficiencies inherent in the social security system can jeopardise perceptions of state’s procedural justice and undermine trust in institutions and society at large:

First, people may draw inferences about others’ trustworthiness from how they perceive public-service bureaucrats. If social workers, local policemen, public-health workers, and so on act in such a way that they cannot be trusted, why should people in general be trusted? Second, if citizens, to get what they themselves deem necessary from public services, have to engage in cheating, distorting vital information, and other forms of dishonest behavior, why should people in general be trusted? Third, if you yourself, to get what you deem fair from public services, have an interest in engaging in questionable behavior, then not even people such as yourself can be trusted, so why should ‘other people in general’ be trusted? (Kumlin & Rothstein, 2005, p. 349)

At this point, it is essential to highlight the significant role of trust in the development of functional societies and as a fundamental requisite for socioeconomic development (Bourdieu, 1986; Fukuyama, 1995; Holmberg & Rothstein, 2020; Kumlin & Rothstein, 2005; Malecki, 2012; Putnam, 2000). Trust is essential for creating and strengthening the threads of the social fabric. While states can be ruled by fear and suppression, the most successful societies are based on a high level of trust, as we have seen in established democracies, and in the Nordic countries in particular. Although high generalised and institutional trust have not been the main goal in the Nordic countries, it has become a by-product of universal social policies.

As a universal benefit without means-testing, UBI can be expected to increase trust. However, as Calnitsky (2016, 32–33) argues, universality per se might not be so determinant for deservingness imprinting processes and stigmatisation: it is not targeting as such that generates stigma and potentially weakens programs. Following Calnitsky’s (2016) arguments, targeted policies are equally capable of enhancing both institutional and generalised trust through three different ways: (1) uniform treatment, (2) automaticity (i.e. granting does not involve bureaucrats’ discretion) and (3) moral acceptance.

For trust formation and social cohesion in the societies, this has major significance. Assuming that Calnitsky’s arguments hold, no significant differences in trust-building will be observed between UBI and a last-resort social assistance as long as the procedures for both benefits are equally automated and recipients are treated uniformly. Furthermore, Calnitsky (2016) argues that the effect of ‘tagging’ recipients could be mostly reduced by automatisation, even if the arguments for uniform treatment and moral acceptance did not hold:

Income testing is often falsely conflated with demeaning experiences with bureaucrats and caseworkers. (…) However, income testing is not inherently linked to these experiences. (…) The less an ordeal is the procedure, the more automatic and less discretionary, the less likely it is to be accompanied by stigma (Calnitsky, 2016, p. 32)

When assessing universalism as contributing to trust, we cannot draw an explicit assumption that selectivity would inevitably lead to less trust. The ‘de-bureaucratising’ of client experience could be enough to effectively constrict the deservingness imprinting channels. By utilising the survey data collected in the Finnish UBI experiment and analysing the connections between trust, bureaucratic experience, UBI and selective benefits, we can test whether Calnitsky’s (2016) arguments are supported in the Nordic welfare state.

4 Research Questions, Data and Methods

A two year UBI experiment was conducted in Finland between 2017 and 2018. A monthly unconditional amount of €560 net per month (corresponding to the net minimum flat-rate unemployment benefit in Finland) was paid to a random sample of 2000 unemployed people drawn from 175,222 individuals who received unemployment benefits from the Social Insurance Institution of Finland (Kela) in November 2016. A more detailed description of the experimental setting can be found in the reports published in the evaluation research project of the experiment (Kangas et al., 2019, 2020, 2021).

In Finland, the unemployment benefit system consists of two mutually exclusive benefits. Unemployed jobseekers who have voluntarily joined an unemployment fund, usually administered by a trade union, are eligible for earnings-related unemployment benefits if certain criteria are met. The remaining unemployed jobseekers are eligible for a flat-rate unemployment benefit paid by Kela (basic allowance or labour market subsidy), but again if certain criteria are met (more detailed description in Kela’s website, Kela, 2021). If an unemployed jobseeker does not meet the conditions for unemployment benefit, he or she is eligible for last-resort social assistance, the receipt of which is not conditional on unemployment. As a sanction, social assistance can be reduced to 40%, but never completely eliminated. It can also be paid in addition to unemployment benefits if the income level of household income falls below the minimum income limit set by the Parliament each year.

Although the limitations of the experiment design do not allow for comprehensive, system-level conclusions in the absence of a ‘saturation’ site (Calnitsky, 2020), it does allow for an examination of individual behavioural changes in response to the policy intervention, which could induce system-level outcomes in a full implementation scenario – for instance, an overall increase in generalised trust. In this paper, we are interested in the role of selectivity and the perceived bureaucracy of the social security system in building trust in the social security system itself and in other people. Based on the research presented above on universalism versus selectivity in welfare states and on trust formation, our hypotheses are as follows (see also Figure 1):

Figure 1:

The hypothesised connections between UBI treatment, trust and bureaucratic experience.

In our study, we utilised a telephone-based survey administered between 17 October and 14 December 2018. All 2000 UBI recipients were included in the survey (later referred to as treated, treatment group or TG). Among the unemployed who received benefits from Kela at the end of 2016, 5000 were randomly selected to the control group (later referred to as controls, control group or CG) of the survey. More detailed descriptions of the data can be found in the research reports of the evaluation study (Kangas et al., 2019, 2020).

As Table 1 indicates, the response rates remained low in both the treatment group receiving basic income (TG) and the control group (CG) not receiving basic income. Although there are no significant differences between these groups in terms of gender, age, level of education or area of residence, there is a small difference in household size and a more significant difference in income level. The latter difference is partly explained by the receipt of UBI. A more detailed description of the differences between these two groups is provided in Appendix Table A1. To correct possible selection biases, we use weighted data in our analysis. However, it should be noted that the use of the weighting factor does not correct the skew in the data due to the low response rates.

In the survey, the respondents were asked questions about their subjective well-being, health, income, trust in other people and various societal institutions, and their attitudes towards UBI. Furthermore, the respondents were asked about their experiences with bureaucracy related to the social security system. The complete questionnaire can be found on the website of the Finnish UBI experiment (Kela, 2020).

Although UBI is intended to replace most social benefits, in the Finnish experiment, the partial UBI of €560 only replaced the flat-rate unemployment benefits paid by Kela (read more about experimental design in Kangas et al., 2021). Therefore, some UBI recipients received unemployment benefit supplements (e.g. child increase) as they would have received if they had been covered by the normal unemployment security scheme. Further, in cases where the UBI (and possible unemployment benefit supplements) did not raise the household income above the minimum threshold, the household was eligible for last-resort social assistance.

In the survey, respondents were asked about the sources of income of their current household. Since we are primarily interested in the universality of the social security system and the associated bureaucracy experience, we constructed a new four-class variable that describes the receipt of benefits by respondents and their households: regardless of whether the respondents belonged to the TG of the CG, they were grouped according to whether they received unemployment benefits, social assistance, both, or neither. The new variable, selectivity, was constructed as follows: 0 = the respondent receives neither unemployment benefit nor social assistance (716 respondents), 1 = the respondent receives unemployment benefit (515 respondents), 2 = the respondent receives social assistance (124 respondents), and 3 = the respondent receives both unemployment benefit and social assistance (277 respondents).

Generalised trust was measured in the survey through a standard question: ‘Generally speaking, would you say that most people can be trusted or that you need to be very careful in dealing with people?’ Trust in social security was measured with the question: ‘Do you trust the social security system?’ For both questions, respondents were asked to select a value on a scale from 0 to 10, where zero indicates ‘no trust at all’ and 10 indicates ‘complete trust’.

The experience of bureaucracy was measured by a retrospective question: ‘During the last two years, have you experienced too much bureaucracy when applying for social benefits?’ Two years here refer to the duration of the experiment. The answer options to the question were ‘No’, ‘Do not know’, and ‘Yes’. We excluded from the analysis 73 respondents who answered ‘Do not know’.

To measure self-rated health, we applied the following questions: ‘How is your health in general? Would you say it is … 1 (Very bad), 2 (Bad), 3 (Fair), 4 (Good), 5 (Very good) or 6 (I do not know)?’ The six answers in the last category were coded to option 3 (Fair). Thus, we formed a five-point Likert scale.

Methodically, we built our analysis on ordinary least squares (OLS) models, with the dependent variables being trust in other people, trust in the social security system and bureaucratic experience. The main independent variable in our models is a binary variable, which describes whether the respondent belongs to either TG (0) or CG (1). Due to randomisation, background variables such as gender (1 = female; 2 = male), age (in years), education (a six-digit scale) or subjective health should not have a significant effect on the dependent variables. However, to estimate the effects of these variables on the level of trust, we controlled them in our models.

5 Results

The group means for both trust in the social security system and other people are in line with our first hypothesis (H1: Recipiency of UBI as a universal benefit drives both higher trust in social security and generalised trust). At the end of the experiment, compared to the CG, the TG expressed higher trust levels toward both the social security system and other people. Whereas the mean level of trust in the social security system of the TG was 6.65, the mean for the CG was 6.36. The difference was statistically significant (p = 0.024). Generalised trust followed a similar pattern: 6.66 was the mean for TG, against 6.32 for the CG, with a significance of p = 0.007. Using OLS regression models that were progressively controlled by age, gender, level of education, and subjective health, we can refine these findings according to Table 2.

In Table 2, we can see that the treatment, that is, receiving UBI, significantly increases trust in the social security system as well as the level of generalised trust. Controlling for background variables did not offset this observed positive association between UBI and trust. Hence, it is legitimate to say that UBI recipiency is connected to a significant positive shift in trust both in the social security system and in other people. This is in line with the theory that by implying less selectivity and, thus, less stigma and alterity in its recipients, a universalistic social policy increases trust in both the institutions linked to the policy itself as well as in the society at large (see Kumlin & Rothstein, 2005; Lascoumes & Le Gales, 2007). Additionally, good health and a higher level of education have a positive and significant association with both forms of trust.

With regard to H2 (Recipiency of UBI as a universal benefit implies less bureaucratic experience), UBI recipiency implies less bureaucratic experience, according to Model 1 (see Table 3). This relationship holds in Model 2, in which demographic variables and education are controlled. These results are in line with those of previous studies (see e.g. Simanainen, 2021). In Model 3, we add subjective health as an independent variable for which a strong negative impact on bureaucracy experiences is observed. This result may reflect the fact the health-related benefits require more bureaucracy or, on the other hand, that more bureaucracy is experienced when the client has health problems.

In Model 4, we added selectivity as an independent variable to control for the recipiency of selective benefits, i.e. unemployment benefit and/or social assistance. After controlling for the treatment and the background variables, selectivity does not explain the bureaucracy experience with statistical significance. Table 4 also does not show a strong association between scores on the selectivity scale and bureaucracy experience. This detachment seems counterintuitive at first, although we must consider that the feeling of excessive bureaucracy, which comes from experiencing procedures such as screening, conditionality, activation measures and general form filling, is dependent on the clients’ deservingness self-perception, which modulates procedural justice feelings. In other words, benefit recipients considering the screening procedures to be fair will not regard them as excessive bureaucracy and vice versa.

This could explain the less frequently reported high bureaucracy feelings among those whose households received social assistance, namely 65.0%, against 67.3% of those whose households received unemployment benefits (column ‘Total’ in Table 4). However, there is an interesting difference between the TG and the CG in the simultaneous receipt of social assistance and unemployment benefits in the households in question. Whereas experiences of bureaucracy among the controls whose households received both unemployment benefits and social assistance are the lowest, the opposite is true for the treated. In the TG, experiences of bureaucracy are the strongest (73.2%) among those whose households received income transfers from the unemployment protection system and social assistance. One explanation for this could be that in the households, where one member is receiving UBI and others (e.g. a spouse) are receiving unemployment benefits, the non-bureaucratic nature of UBI is emphasised. Another explanation could be that receiving UBI in households added a new layer to an already complex social benefit package. This is in line with the fact that Kela received many more enquiries from the TG than from the CG at the beginning and the end of the experiment (Verho, Hämäläinen, & Kanninen, 2021).

There are a couple of possible explanations for the somewhat counterintuitive finding that recipients of unemployment benefits experience less bureaucracy than recipients of social assistance. First, social assistance is not only the most selective but also a last-resort benefit, i.e. more ‘charity-like’ than unemployment benefits, which is likely to imply more ‘docility’ and leniency from its recipients towards the red tape and reinforces, as Larsen (2008) argues, perceptions of fair procedures.

In contrast, unemployment benefits are strictly targeted to the unemployed and only to those who are active enough. ‘Laziness’ is highly sanctioned, and criteria for receiving the benefit are strict, not to mention complex and difficult to understand. With the launch of an activation model at the beginning of the year 2018 amid the UBI experiment (and abolished at the beginning of the year 2020 after the parliamentary election), the screening of unemployment benefit recipients was further increased, and new criteria for receiving the benefit were introduced. As reported in the evaluation study (Kyyrä et al., 2019), the activation model made the unemployment benefit system even more complex, contributing to more bureaucracy and, presumably, also to experiences of bureaucracy.

We also hypothesised that less selectivity and bureaucratic experience drive more trust on one hand in the social security system and on the other hand in other people (H3). As regards the former, receiving selective benefits can even increase trust in the social security system (Model 1 in Table 5), while the perceived bureaucracy experience is negatively and significantly connected to it, as reported in Models 2 and 3 in Table 5.

As explained above, as long as procedures regarding social benefit receipt can be deemed fair and due, there is no reason for the benefit recipient to distrust the social security system even if the benefits are selective. In Finland, most benefits can be applied quite electronically and the procedures for processing benefits are strictly regulated. Furthermore, in Finland, citizens’ trust in the societal institutions is traditionally high due to historical reasons described above. Therefore, with regard to trust in the social security system, H3 is partially confirmed as not selectivity itself, but unfair/undue procedures hamper trust in social security. As long as bureaucratic procedures are deemed due or fair, receiving selective social benefits does not decrease trust in the social security system and likely even improves it.

Regarding trust in other people (generalised trust), the coefficients for both selectivity and bureaucratic experience are negative and significant (Models 4–6 in Table 5). It is thus legitimate to state that, when controlling for gender, age, education and subjective health, receiving UBI increases trust in other people mainly by narrowing deservingness imprinting channels, i.e. policy selectivity and bureaucratic experience. Therefore, our third hypothesis (H3) is confirmed with regard to general trust.

Lastly, we hypothesised that trust in social security drives generalised trust (H4). As can be seen in Table 6, even after controlling for selectivity and bureaucracy experience, trust in the social security system explains trust in other people with statistical significance, confirming Kumlin and Rothstein’s findings (2005) about the connection between institutional and generalised trust. However, selectivity can hamper trust in other people even if trust in the social security system is maintained, as shown in Table 5, which presents a parallel path to decline generalised trust. Therefore, those receiving selective benefits can at the same time trust the social security systems and distrust other people. This signals that tagging and alterity effects are present, if not reinforced, despite the recipients’ docility toward bureaucratic selectivity. As hypothesised by Korpi (1980) and mentioned by Larsen (2008), these can spiral up to welfare backlash.

Our results imply that the higher the selectivity of benefits, the less beneficiaries trust other people, signalling a possible stigmatising effect, despite not necessarily harming the trust in the institution providing social security benefits. Of course, this effect can be further amplified by bureaucracy experiences if the system is perceived as undue or unfair.

6 Conclusions

Building trust is a complex process involving a variety of factors. Social policy, by its relevance in the life of the people targeted by it, is capable of significantly contributing to trust – the cohesive glue of the society. Further, the selectivity of social policies, which is at the core of welfare states, reflects both their core values and real standards of deservingness. Highly selective policies can stigmatise their targets, and jeopardise trust both in institutions and in other people.

In this study, we assessed how receiving a universal, unconditional cash benefit (UBI) affects trust and whether trust can be explained by the bureaucratic experiences of individuals and the use of selective instruments in the social security system. This was done by applying additive OLS regressions to the survey data collected in the Finnish UBI experiment (2017–2018).

Our results show complex relationships between UBI and trust. We found that receiving UBI is linked to significantly higher trust both in the social security system and in other people. We also found that receiving UBI implies less bureaucracy experience. This connection is however somewhat obscure: Those UBI recipients whose household income consisted also of unemployment benefits and social assistance reported the highest level of bureaucratic experience. The most likely explanation for this is that in the Finnish experiment, UBI only partially replaced social benefits and to a certain extent made the system more complicated for the beneficiaries than before the experiment. As our findings show, it is not the level of selectiveness in social benefits, but rather the procedures connected to their application and payment that affect trust in the social security system. As long as the system is comprehensible and citizens can trust it to work in a just and fair way, selectivity is accepted.

Furthermore, our findings show that higher trust in the social security system is connected to higher generalised trust. The detailed analysis comparing the treated and controls show that, in the end, procedural justice affects the level of trust in the society too. A fair and just social security system drives institutional trust, which further drives generalised trust.

However, our models also show that selectivity negatively affects generalised trust. Hence, being subjected to screening can simultaneously lower trust in fellow citizens and maintain trust in the social security system. This implies that whereas the social security system is considered a trusted actor by individuals receiving selective benefits, other people are not, for one reason or another.

In our analysis, we had subjective health as an independent variable, although health might be affected by the recipiency of basic income (as argued e.g. by Forget, 2011, 2018). This constitutes an alternative pathway through which universal policies drive trust and is an important avenue for future research.

We can conclude that overall trust is largely related to bureaucratic experiences. Hence, while not offsetting all selectivity-bound demeaning effects, procedural justice and automaticity within social security systems can increase trust and, thus, signal the value of minimum income policies as long as their implementation involves less effort from recipients to prove their rights, refraining from constantly grasshopper-stamping them. Recently implemented policies, such as those launched in response to the COVID-19 pandemic in diverse countries, could be a relevant source for future research, to better understand how they can impact recipients’ trust, according to their procedures’ justice and automaticity.

In addition, we can conclude that the psychological mechanisms extrapolating trust in institutions and their representatives, as hypothesised by Kumlin and Rothstein (2005), do not tell the whole story; it is possible to leniently live the bureaucratic experiences provided by the agents and trust them while distrusting society at large because of stigma imprinted by mechanisms built into the social policy instruments. Likely, this can also signal that Calnitsky’s hypothesis (2016) might not hold, as self-perceived low deservingness stigma can show up independently of the perceived bureaucratic experience. Again, the very idea of benefit recipients understanding themselves as grateful for (thus, undeserving of) the received cash could already affect their trust towards society through a self-tagging, alterity-building mechanism. Automaticity could increase trust in policy-providing institutions and society at large by circumventing concrete bureaucratic experience mechanisms, although it might disdain the direct connection with the idea of being a ‘welfare beneficiary’ and have trust in other people jeopardised.