Why Do We Run Basic Income Experiments? From Empirical Evidence to Collective Debate

1 Introduction

Over the last decade, the interest in basic income has spectacularly spread amongst the mass media, politicians and the public in general. The ‘utopian’ idea of granting each individual of society a regular, individual, and unconditional cash transfer was never so popular. Indeed, it has turned as popular amongst the scholarly community as it almost is in our daily life. Sometimes it is argued for as a tool to fight poverty and inequality, while it is also seen as a way to redress the actual dysfunctionalities of traditional social policies and welfare systems. In light of the scenario, basic income (BI) has also been invoked as a mechanism to deter the most dramatic consequences of the COVID pandemic. Basic income is now into the political agenda and, consequently, its hypothetical consequences are at stake. To which extent would it reduce inequality? Will people quit their jobs? Although there might exist robust-theoretical arguments proving the virtues of a basic income, the main problem is that we lack empirical evidence to advance its possible real impacts, both at individual and at collective or structural level. As a result of this empirical requirement, various governments are now willing to test different modalities of (more or less) unconditional cash transfer schemes in the form of basic income experiments (BIE). Which is the ultimate goal behind rolling out these experiments? Do we conduct them because more robust empirical evidence that BI properly works is required or, on the contrary, do we conduct them to prove that basic income is a just and ethically desirable measure?

If the reason is the former, the so-called “empirical claim” of BIE, it might be argued that sufficient evidence has already been collected. Indeed, several BIE implemented across quite different contexts – from North America to Europe, from Asia to Africa; from the 1970s until nowadays – have gathered a large amount of empirical data. The majority of these experiments have proven that basic income does generally produce positive effects. Not all BIE are equally designed, nor are they intended to alter the very same indicators. But regardless of their differences on their design and aims, all of them got some ‘positive’ considered impact on several indicators, such as reducing malnutrition, stress and household’s indebtedness or increasing well-being, enhancing subjective well-being and happiness, or improving social trust among citizens and on public administration, amongst others. Overall, these results would prove that BI is an effective measure in tackling some major individual and societal problems. Therefore, in this case, the “empirical claim” would seem solved as we would already have collected enough evidence demonstrating that BI actually works.

Nevertheless, if the reason for conducting BIE is the second one, the so-called “ethical claim”, things may be slightly more complex. In this case, we must first agree which would be the moral ends we must prioritize, for example reducing poverty, as we all do likely agree in understanding it as one of the most morally-reprehensible issues of our society. Then, assuming that, we should concur that BI is a proper measure to advance this moral end (or others). Consequently, BIE would be seen as a mechanism to prove that basic income is ethically desirable (or even superior to other measures) for advancing our moral claims. Assuming prima facie that we all find poverty as morally reprehensible, then the question is i) whether we really need a BIE to validate this moral claim, and ii) whether BI is an ethically-adequate manner to cope with that. As one of the researchers involved in one BIE told to us, poverty is morally condemnable in any case and hence, he claimed, “we do not need an experiment to prove that”. As in the case of slavery, he concluded, “a pilot was not required to demonstrate that [to] release people is … good for them”. No social experiment is required to demonstrate that slavery, like poverty, is morally condemnable. Likewise, no pilot nor experiment is really required to demonstrate that putting an end to slavery or eradicating poverty is ethically justified therefore.

This paper shall address both the “empirical” and the “ethical” claim of basic income experiments by illustrating the discussion with the recently concluded experiment in the city of Barcelona, the so-called B-MINCOME pilot, alongside other basic income experiments recently conducted in Europe and elsewhere. When considering Barcelona’s and others’ experiences, the answer to the question of what (if anything) can justify basic income experiments may appear ambivalent. Nonetheless, the empirical discussion is always necessary and informative to address some fundamental critical ethical questions that basic income does always entail. Therefore, and as conclusions shall sustain, although we might think we have already accumulated enough empirical evidence to justify the “empirical claim”, they are always necessary and a powerful tool to deal with ethical debates which are in turn pivotal for advancing the cause of basic income.

2 The B-MINCOME Pilot

Which is the reason for conducting basic income experiments? Which is their ultimate goal? This question was pretty present early in 2016 when firstly designing the B-MINCOME pilot in the city of Barcelona (Laín & Torrens, 2019). Understood as an attempt to test a new municipal cash scheme, this social experiment declared as its priority reducing poverty and social exclusion. However, its experimental goal was rather to test the very same effectiveness and cost-efficiency of this cash transfer schema. So, while the political aim behind that was to use the resulting empirical evidence to justify implementing a new cash transfer schema in the city of Barcelona, the scientific goal was to collect robust evidence to test the same experiment. Thus, the B-MINCOME project seems to have been driven by an evidence-based or “empirical” claim which, in light of the increasing poverty rates in Barcelona over the last decade, was also seen as “ethically” understandable and justifiable.

The B-MINCOME pilot began in October 2017 and was running for 24 months until November 2019. Its experimental design was based on a randomized control trial methodology because a lottery randomly selected 1500 households from a universe of almost 5000 eligible ones living in the so-called Eix Besós area, which comprises the three poorest districts in Barcelona. The lottery picked up the 1000 households who were about to take part in 10 different treatment groups, while the remaining 500 were assigned to the control group. The former were receiving a cash transfer (the Municipal Inclusion Support or SMI) of a maximum of 1675 euros a month depending on their composition, their overall income and living costs. Among them, 450 households did solely receive this SMI, while the remaining 550 also took part in one in four active social-inclusion policies (training and employment; entrepreneurship in the social and cooperative economy; refurbishing flats to rent out rooms; and communitarian participation). This way, the pilot combined four modalities of participation: i) conditional and ii) unconditional – whether participation in one of these policies was mandatory to keep receiving the SMI; and iii) limited and iv) unlimited – whether any additional income proportionally reduced the SMI amount or was net added. As explained, the overall experimental goal was to test which of these four modalities of SMI and which of the ten-treatment groups of combining the SMI and active policies was the most effective (regarding their results) and most efficient (regarding their cost) in order to inform future municipal social policies (Laín, Riutort, & Julià, 2019).

Like other similar projects, the B-MINCOME pilot was not really on basic income. Among other reasons, it differentiates itself from a real basic income because; i) it was time-limited (lasting 24 months); ii) was not universal (only former social services users were eligible); iii) most of recipients were not unconditionally granted (just one modality was really unconditional); iv) was not individual (but household based); and v) the SMI amount was not constant (it was actually varying according to households’ costs and income). In any case, these or similar non-basic-income-features are somehow found in all of the so-called “basic income” experiments. In all cases, they are also time-limited. In some cases, participation is restricted to unemployed people, while in others cash transfer is not really unconditionally granted. Although assuming that a pure BI cannot be tested for these and other reasons (Laín, 2021; Torry, 2019: 11–13; Widerquist, 2018: 37–42), the complexity and the results of the B-MINCOME pilot and of other similar pilots raised a myriad of empirical and ethical issues that are worth stressing if we want to shed light on the debate of what (if anything) can justify basic income experiments.

3 Empirical Claim When Conducting Basic Income Experiments

In spite of their methodological differences and the particularities of their distinct national contexts, almost all BIE conducted over the last years have reported positive or pretty positive results which, at least theoretically, governments would like to extrapolate to an eventual long-term, national basic income scheme (Standing, 2012: 134). For instance, in Finland, a decrease in personal stress and an improvement in subjective well-being were observed. Finnish recipients found their financial situation more stable, while their trust in their fellow citizens and public institutions ameliorated (Kangas, Jauhiainen, Simanainen, & Ylikännö, 2020). In Utrecht, the Netherlands, recipients’ autonomy and time for care and housework increased (Verlaat, de Kruijk, Rosenkranz, Groot, & Sanders, 2020). In Ontario, although the experiment was abruptly cancelled, the results were also positive in terms of rising social and communitarian participation, improving health status and the quality of sleep, and reducing stress levels, which was translated into an increase in satisfaction in the workplace and entrepreneurship (BICN, 2019; Ferdosi & McDowell, 2020).

In other contexts, results were quite positive too. For example, in Iran, “the first country in the world to provide a de facto basic income to all of its citizens’’ (Tabatabai, 2011: 2), regional, urban-rural, household and individual income inequality declined notably within the first year of the schema (Karshenas & Tabatabai, 2019; Salehi-Isfahani & Mostafavi-Dehzooei, 2018). In Kenya (Haoushofer & Shapiro, 2016) and Namibia (Haarmann et al., 2019; Osterkamp, 2013), the results did also reveal an important decrease in extreme poverty, a considerable increase in school attendance rates, and in economic and self-employment activities. In the case of Delhi and Madhya Pradesh in India, economic and entrepreneurship activity performed by women boosted notably. Simultaneously, there was also an increase in school attendance and access to health services, and in the quantity and quality of individual and household nutrition patterns (Davala, Jhabvala, Mehta, & Standing, 2015).

In the case of the Barcelona pilot, the impact evaluation reported quite positive impacts too (Riutort, Julià, Laín, & Torrens, 2021). The assessment of results was performed through an interdisciplinary technique which made it highly robust in methodological terms. On the one hand, a quantitative evaluation was performed through three survey waves, including a baseline survey, a follow-up one by the end of the first year, and a last survey two months before finishing. In addition, information collected through surveys was validated, completed and criss-crossed with administrative and register data from the Spanish Tax Agency and the Social Security System, the Catalan Health Institute, the Barcelona Consortium of Education, and the Municipal Social Services. Overall, these quantitative results pointed out a substantial reduction in severe material deprivation, in going to sleep hungry, in individual and familiar indebtedness, and in having mortgage or housing rental debts. This was followed by a significant reduction in households’ financial stress as well as being more confident with private financial situations and having more time for care and housework. On the other hand, the evaluation design did also include a rigorous qualitative evaluation. Two in-depth surveys conducted in 2017 and 2018 with about 680 active participants reported that their subjective well-being and happiness increased by almost 20%, as was the case with the satisfaction with their lives and their neighbourhoods as well as their self-perceived state of health. The project did also include an ethnographical evaluation performed through 75 semi-structured interviews which were completed with three waves of in-depth interviews to 40 participant households. Finally, those assigned to the communitarian participation active policy were also interviewed, revealing an improvement in their attitude towards community, an increase of their associative dynamics, and a better appreciation of public administration.

Consistent with the initial hypothesis, these results confirmed that the B-MINCOME pilot got quite positive impact results both on its quantitative and qualitative dimensions. Simply, as it occurred in similar experiments, granting people cash for a certain period usually brings about a positive impact on their quality of life, happiness and well-being, communitarian and social participation, and – obviously – economic and financial situation. However, experiments’ impact on labour markets are much more controversial, particularly in rich countries. Participation in labour market in general and labour supply dynamics in particular are two dimensions highly sensitive to the very experiment’s design and tightly constricted by macroeconomic and cultural and institutional national patterns (Van Parijs & Vanderborght, 2017). This makes labour impact hard sizeable and almost impossible to extrapolate from one case to another. For example, Finnish participants were only those receiving unemployment benefits. In Utrecht, the experiment was restricted to those already receiving other social benefits, while in Ontario the design established an economic threshold to be eligible. The same occurred in Stockton, California, where the criterion was established at, or just below, the city’s median household income (Castro Baker, West, Samra, & Cusack, 2020). Likewise, the B-MINCOME project was also restricted to former users of municipal social services below a certain economic threshold. Therefore, although some appealing evidence about the supposed labour effects of a basic income might be collected through these experiments, “that evidence can be deceptive” (Widerquist, 2018: 120).

Both due to the beneficiaries’ characteristics and to the same experimental limitations (duration, number of participants, value and conditions of the grant, etc.), an experiment is not very well-equipped to capture the possible effects on labour supply if it is “to be translated into a permanent unconditional scheme” (Groot, 2006). Labour participation comparison between treatment and control groups always “attracts attention because it is a nice, neat, apparently-easy-to-understand number” (Widerquist, 2018: 117). In contrast, qualitative effects such as subjective well-being and happiness, communitarian and social participation, leisure and care work time, or household stress are much more difficult to grasp and require a more detailed and expensive research. Besides most experiments prioritized quantitative results – particularly on labour supply – and that the methodological requirements of qualitative techniques are usually higher, in much of the already conducted experiment’s results related to these qualitative indicators have been quite positive and statistically more robust and significant than those related with labour participation. Consequently, and despite these methodological limitations, collected empirical evidence suggests that basic income would bring about positive results, either in Barcelona or elsewhere. Nevertheless, despite all of these positive effects, none of the countries where these experiments have been conducted has already rolled out – nor seems to be about to do so – a real nationwide basic income scheme (Laín, 2021).

It might be argued that this did not happen yet because profound disagreements on the interpretation of BIE effects remain among politicians, policymakers and the population in general (Virjo, 2006). Of course, experiments cannot answer all questions that basic income does raise. Rather, they can only add some, though important, evidence to the actual corpus of scientific knowledge. But what we suggest here is that, precisely because all BIE have already demonstrated that BI is likely to work by triggering positive results, this “empirical claim” is not consistent enough in justifying (nor denying) the implementation of more basic income experiments. Indeed, most of the time, the debates, doubts and disagreements on BIE results are “more of an ethical debate about the desirability of its effects than an empirical debate about what those effects are”. Simply: basic income experiments cannot respond to “ethical and subjective questions, such as do we want what this does?” (Widerquist, 2018: 87–88). As a result, the “ethical claim” becomes now fundamental, not only because it is necessary to properly explain and discuss the BIE results, but also because the ethical dimension is at the center of the very basic income debate.

4 Ethical Claim When Conducting Basic Income Experiments

Once we have briefly argued that the “empirical claim” is not consistent enough to justify more basic income experiments, the resulting question is whether they might be justified on ethical grounds therefore. Now what we are wondering is not if we are conducting experiments to get empirical evidence, but if we conduct them to prove that basic income is ethically desirable to meet some moral concerns, such as poverty, exclusion, or inequality for example. Properly speaking then, we are not referring to the ethical issues arising when experimental techniques involve human beings – those summarized by the famous Belmont Report and neutralized by its principles of respect for persons, beneficence, and justice –, but to those ethical issues associated with the hypothetical effects of a basic income which basic income experiments would presumably advance.

Of course, although the “empirical” and “ethical” claims must be analytically differentiated, in substantive terms they are always interrelated since a particular experimental result does likely spur people to discuss one or another ethical implication of the basic income. For example, one of the most interesting results of the Barcelona pilot was that its participants became more aware about the municipal and communitarian resources available in their neighbourhoods (such as public libraries, social and cultural centres, communitarian networks, etc.). That impact triggered a discussion about the ethically desirable-positive effect of a basic income on communitarian and social participation of poor people (Blanco, Fernández, Gallego, Marra, & Yanes, 2019; Hill-Dixon, Davis, Patel, & Boelman, 2020). Likewise, many participants were suffering from different health conditions which made them unable to perform any kind of job activity. This led both researchers and participants to discuss whether it is ethically acceptable to make mandatory participating in, for example, a job training program in order to be eligible for a conditional minimum income, or instead if an unconditional program like a basic income would not be more just. Similarly, as an overall effect, women reported some slightly more positive results than men in some indicators, which posed important ethical questions on how social policies (if any) should tackle this kind of gender problematics. Interestingly, the Iranian experiment sharply illustrates such an ethical issue, though in the opposite direction. There the experiment granted all households with a regular cash payment which was assigned to the head of the family – meaning basically men. In this case therefore, the experimental design reinforced the traditional men’s role within the household while perpetuating women’s subsidiary position. There, it was not the result, but the very same experimental design, one with important ethical implications.

As seen, basic income experiments have the capacity to introduce or situate some important debates involving ethical issues, such as the social participation of poor people, the conditionality of social policies, or women’s subaltern role, which are highly strategic for the basic income cause. However, this capacity to generate debate does not necessarily rely on the availability of empirical evidence, but also and paradoxically, on its absence. This is the case of the labour market impact. For example, in the Finnish experiment participants were employed, on average, just 5 more days than those in the control group. Even more, the assessment of this effect cannot be really conclusive due to the introduction of the ‘activation model’ in 2018. Although these results were really poor and not conclusive in statistical terms at all, the experiment was accused of proving itself a failure and basic income, consequently, as ethically undesirable and a “thoroughly wrongheaded idea” (Ezrati, 2019; Tiessalo, 2017).

Beyond biased commentaries, media misunderstandings and (lack of) statistical significance, these cases point out that the sole fact of conducting a BIE is a powerful mechanism to induce debates on the functioning of labour markets and their ethical desirability as a mechanism to distribute social wealth (in the form of wages and salaries) and to protect the worst-off (through unemployment benefits or retirement pensions, for example). Should labour participation be the main mechanism to assure material existence? And if so, under which conditions (how many hours, for which salary, which type of activities, and which kind of relationship between employers and employees, etc.) would it be ethically acceptable? Due to the media impact of basic income experiments, these and other related-questions on work ethics (and potentially on workplace democracy) are now more frequent in our daily life than they used to be a few years ago. Undoubtedly, this a strategic window of opportunity that basic income advocacies can (and should) not pass up.

By leading debates on work ethics, basic income experiments do also contribute to foster other ethics-related discussions about how current social protection systems operate. As it was the case of Barcelona’s experiment, most of the BIE are designed to solve, patch or strengthen some traditional social policies, such as minimum income schemes or social assistance benefits which are “falling short of alleviating poverty”, and some of their associated problematics such as the poverty-traps, stigmatization, non-take-up, bureaucratic inefficiency or administrative arbitrariness “tests their effectiveness still further” (EESC, 2013: 9). BIE are particularly well-placed to discuss these issues by forcing questions such as whether social policies should be conditioned, to which extent welfare regimes should be based on contributory performance or, more broadly, why inclusion and social citizenship should be subject to work and labour markets (Dwyer, 2000; Meade, 1997). In sum, the “ethical claim” (do we conduct BIE to prove that basic income is an ethically desirable measure?) is a powerful and strategical mechanism to advance the basic income cause, not by directly addressing its potential benefits and advantages, but indirectly by addressing the ethical background of the current welfare regimes.

5 Conclusions

The question of what (if anything) can justify basic income experiments requires a twofold response. The “empirical claim” assumes that they are justified as far as they gather empirical evidence of the hypothetical impact of a basic income, which in turn is necessary to justify the very implementation of such a measure. In contrast, the “ethical claim” sustains that basic income experiments should provide some insights into the desirability of basic income. Hence, we would perform experiments to justify it as a just and ethically desirable measure to deal with some moral issues, such as poverty for example. The first response is consistent in methodological and political terms – experiments are necessary to collect data and empirical evidence to estimate if a new policy should be implemented. Nevertheless, the accumulated scientific knowledge after several basic income pilots would permit us to infer that, first, positive results are easily to be expected from the upcoming experiments and, second, that enough evidence would have already been collected to demonstrate that basic income brings about positive results. Though necessary, the “empirical claim” would thus reveal itself not sufficient to justify conducting more basic income experiments.

It has also been argued that, although basic income experiments must face several limitations regarding their methodological design (the length, characteristics of participants, the amount of the grant, etc.) they can inspire, provoke or lay out several ethical discussions which are profoundly strategic for the basic income cause. “Ethical claim” would seem therefore quite persuasive in justifying conducting more basic income experiments. This is not to deny the importance of gathering new data and empirical evidence. Undoubtedly, social experiments are required to inform political proposals such as basic income. What we suggest is that basic income experiments do (and should) also play a central role both to directly foster collective debate on the ethical implications of social policies and welfare regimes, and indirectly to justify basic income.