Universal basic income and the Finnish experiment
The preliminary results on the Universal Basic Income (UBI) experiment in Finland, and what they mean for the long-standing questions over the potential impact of UBI in developed countries.
“How can sufficient goods be available when anyone can withdraw from work – when people are no longer motivated by the need to earn their own living, or when reliance on the efforts of others makes them lazy? If only you’d been with me in Utopia (…)”
Utopia, Thomas More, 1516.
The conversation between Rafael Contrasenso and Thomas More has been rehashed this week, following the publication of preliminary results on the Universal Basic Income (UBI) experiment in Finland. There may have been more hashtags, but no fewer mentions of the word ‘utopia’.
In this review, we look at how the results of the Finnish UBI experiment were viewed in the blogosphere and social media, and how they fit with longstanding questions over the impact of UBI in developed countries.
Effects on well-being and long-term labour supply
Opponents of UBI, such as Nathan Keeble at Mises, argue that a substantial reduction in labour supply necessarily follows the introduction of UBI, through income effects. Hoynes and Rothstein (2019), in ‘Universal Basic Income in the US and Advance Countries’, point to the Blundell and Macurdy (1999) literature review as evidence of negative income elasticities of labour. Similar estimates are obtained when looking at lottery estimates, a possible for UBI (Cesarini et al. (2017)). Adam Perkin goes further and argues anti-work incentives breed employment-resistant personalities.
The most extreme-left position is the work of anarchist Pëtr Kropotkin who, to the contrary, believes “idleness” would be minimal after work is disassociated from livelihood. Kropotkin argues the true lazy man is rare. Intellectually stimulating work, aligned with interests and skills and improvements in child welfare, would prevent reductions in labour market participation (in the Conquest of Bread (1892)). Hoynes and Rothstein (2019) write that any UBI would be expected to decrease labour supply in the short run. However, they show how modern pro-UBI theory (and some empirical work) support these offsetting mechanisms.
Hoyes and Rothstein (2019) mention extensive evidence on how credit constraints are binding on education and job decisions. A UBI, some argue, would allow for higher educational investment and moves to jobs with training. Furthermore, “higher skilled individuals tend to work more”. Increases in wages would follow productivity gains. UBI second-order effects on child development “may be an important and until recently largely overlooked part of the benefits”. Finally, Jones and Marinescu (2018), in their analysis of the Alaska Permanent Fund, conclude that the demand from increased income might be enough to compensate for the initial labour supply reduction.
An important discussion around UBI instead focuses on how it compares to current welfare schemes. Arguably the most important pro-UBI argument is the elimination of welfare traps. The UBI does away with the opportunity cost of finding work in unemployment insurance and the steep marginal tax rates of means-tested transfers. Such an argument has been advanced by Milton Friedman in defense of a negative income tax. Compared to current welfare schemes, a UBI could thus increase labour supply.
On the other hand, a UBI does away with mandated job search. Such unconditionality can disincentivise work, as pointed out in the University of Bath’s IPR Policy Brief on ‘Assessing the Case for a Universal Basic Income in the UK’. Friedman, as well as libertarians such as Matthew Feeney at Cato Institute and Ed Dolan, however, argue that even if effective in increasing the labour supply, such conditions might wrongly shape individuals’ decisions and ultimately be harmful to society.
Other points in favour of a UBI include improvements in subjective well-being, health status and network/community spillovers. Examples include stress reduction, reduction in hospitalisation rates and in crime. An increase in wellbeing can simultaneously increase productivity, decrease financial costs, and increase social welfare. Universality also eliminates the stigma of welfare programmes. Stigma is nonetheless an intended part of welfare schemes and, some say, necessary as an incentive to find work.
A more pragmatic perspective to support UBI over current welfare schemes is the reduction of bureaucracy (see libertarian author Ed Dolan or, yet again, Friedman) and how it would be harder to cheat the system (e.g. John Davis Davidson at The Federalist). Iain Murray at the National Review suggests UBI might boost mutual aid (private charity), alleviating one of the main libertarian criticisms of the welfare state. Yet, the affordability/redistribution trade-off is a concern for many. If income from current welfare schemes was simply repurposed as a UBI, there would be an income shift to middle-class families, non-disabled families (as shown in OECD simulations and in IZA simulations). Increased taxation to ensure progressiveness threatens affordability and faces opposition on the same grounds as other redistributive policies.
Effects on labour allocation
Staunch defenders of UBI, such as political economist Van Parijs, argue that UBI could maximise welfare by allowing job uptake aligned with personal preferences. Such arguments rest on the existence of substantial market failures in labour markets, which others such as Nathan Keeble at Mises dispute. On entrepreneurship, while Brittany Hunter at FEE defends the view that it would be stifled, UBI proponents highlight that it would instead encourage risk-taking – a view with some empirical (non-experimental) support.
Brittany Hunter at FEE criticises UBI on the grounds that “there is no greater incentive than financial security and holding a job is essential to that end”. Yet, the idea that people currently work from need instead of want is seen by some (namely Groot in his work Basic Income, Unemployment and Compensatory Justice, building on Hamminga (1995)) as a market failure that a UBI could correct.
What the Finnish experiment tells us (not) about UBI
Two thousand people aged 25-58 years who received an unemployment benefit from Kela, the Finnish Social Security Institution, in November 2016, were selected as the treatment group for a Randomised Control Trial. They received a partial basic income of €560 per month, the equivalent to the basic unemployment allowance and the labour market subsidy usually provided in Finland. The first results of the basic income experiment were published on February 8th 2019 in a Report.
Media and social media reception
The experiment and now its results have been represented by media outlets and on social media in inconsistent and, at times, contradictory ways. In April 2018, several media outlets (e.g. Sky News, The Guardian, Business Insider) announced that Finland was cancelling the experiment, which prompted a statement from Kela: “Contrary to reports, the Basic Income Experiment in Finland will continue until the end of 2018.”
Today, regarding the results of the experiment, the University of Bath IPR scholar [Ville-Veikko Pukka] draws attention to “reckless interpretation of the results by media, policy-makers, and the public” which “can have far-reaching consequences on the future of policymaking”. Focusing on partial results has resulted in very contradictory assessments of the experiment.
Effects on labour supply and the welfare trap
The Report states that “through the basic income experiment, the government wishes to investigate whether a social security model based on basic income could promote more active participation and provide a stronger incentive to work than the present system”. On the primary outcome – number of days in open employment – the treatment group did not find it statistically significantly easier to find employment than the control group. Most outlets which consider the experiment a failure give prominence to this result.
Many factors might have muted the positive effects of the UBI on labour supply. Heikki Hiilamo highlights: “The Finnish experiment was about partial basic income targeting able-bodied people without work, it was not about universal basic income. That has been a source of major confusion around the experiment and a source of critique of it.” Moreover, even for the unemployed targeted, the short time span of the trial (and particularly the published first-year results) might not allow for long-term effects to arise (see financial journalist Anni Lassila and Ville-Veikko Pukka).
Olli Kärkkäinen, economist at Nordea, raises the point, echoed by many (e.g. Juhana Vartiainen), that it is impossible to disentangle effects from the elimination of the welfare trap – which would increase labour supply – and the elimination of the need to search for a job – which would reduce labour supply.
It must also be highlighted the welfare trap was not entirely eliminated. The majority of the treatment group received social insurance benefits on top of the UBI, unlike originally planned by the research group. As confirmed to Bruegel by Miska Simanaien, researcher at Kela, such unemployment benefits would be taken away in the case of full employment. The treatment group and control group were arguably not sufficiently different from each other. Finally, maybe the welfare trap was not the binding constraint on the sampled population. The report concedes that the sampled population was in “a difficult labour market situation”.
Other factors would instead bias labour uptake upwards. The report states a “majority of individuals in the treatment group did not benefit from non-compulsory active labour market measures” due to legislation changes. Finally, the beneficial tax treatment (income from the UBI not being) created a higher incentive to take on work than would realistically exist. Olli Kärkkäinen says he is surprised at the absence of positive effects, given such a “high incentive to work”.
Suomen Kuvalehti gives a detailed piece showing how these design issues ultimately translated into a lost opportunity for clear results.
Effects on wellbeing
Other outcomes analysed in the Report are trust and satisfaction with life, health and financial well-being. Individuals receiving basic income were more likely to assess their health status positively, to rate positively their ability to concentrate, and to remain interested in things previously considered enjoyable, with such a difference being statistically significant. Most voices considering the experiment a success highlight these results.
Basic income recipients were also more trusting in other people, in the legal system and in politicians. They were more positive about the future, confident in their future financial situation and on their ability to influence societal matters.
Such results support theories on well-being gains and network spillovers. Yet, the benefits are not quantified and are based on self-assessment. Heikki Hiilamo argues well-being variables are not reliable since there was no baseline survey. The direction of the associated bias in unknowable. He adds that the positive effect could be due to attention created by the study, a concern raised by Kela itself, which would inflate the real effects of UBI on well-being.
Effects on bureaucracy
The report evaluates experiences of bureaucracy. Individuals in the treatment group were more likely to consider that basic income would reduce the bureaucracy involved when accepting a job offer, with the difference between the groups being statistically significant. This is a second result highlighted by those considering the experiment a success. The FT, for instance, highlights “World’s largest trial cuts bureaucracy but is a long way from viable system”; Van Parijs mentions a “lower administrative cost”.
The report highlights that a “majority of the individuals in the treatment group did not benefit from the lower bureaucracy”, since the basic income was not truly unconditional. People were forced to apply for benefits whenever these were superior to the level of the basic income. This would produce a downward bias on the difference in perceptions, but also jeopardises the value of the result to guide policy, since many individuals had an experience with bureaucracy comparable to the control group. The absence of a baseline survey remains potentially problematic.
Effects on job distribution and entrepreneurship
The test group was 8 percentage points more likely to be in part-time employment (not statistically significant), 0.85 percentage points more likely to receive income from self-employment, and received on average €21 less in earnings from self-employment. Per the report, statistical significance is not reported to control for multiple hypothesis testing.
Santens, basic income advocate, points out that entrepreneurship effects are muted because there was no demand stimulation, only initial capital stimulation. Yet, he argues the treatment group “earned a bit less, which supports how people would prefer to be their own boss” and states findings include “increased ability to be an entrepreneur regardless of earnings”. Such information cannot be derived from small differences whose statistical significance is unknown.
More compelling positive signs of UBI effects on entrepreneurship come through in a Financial Times piece. Looking into the experience of three recipients, two state their decision to start their own business was ‘unlocked’ by the income transfer.
Analogous to communication on negative income tax experiments (as argued by Karl Widerquist) the conclusiveness of the Finnish UBI experiment might be overstated in the media, “though some important conclusions can be drawn, if they are drawn carefully”.
Regarding the main outcome of interest – labour supply – the experiment is unfortunately not very informative. Being a partial income, applied only to a subset of the unemployed, in such a short span, means that income effects, human capital accumulation and demand factors do not come through. Long-term effects on labour supply cannot be inferred.
Though the experiment was intended to test short-term labour supply effects as compared to current welfare schemes, design issues led to unclear results. The zero difference between treatment and control group in employment tells us little about the relative effects from abolishing the welfare trap and eliminating conditions to access welfare The treatment and control group were also not sufficiently different, which might have biased results towards zero. Moreover, the fact the policy tested is not revenue-neutral,undermines its value as policy guidance.
Results on other measures are threatened by the absence of a baseline survey and quantitative measures. The strong increase in subjective well-being has been widely celebrated by UBI proponents, yet estimates might be upward-biased. No clear effects on entrepreneurship arise, though case studies on recipients hint at positive effects. The evidence on the reduction of bureaucracy is arguably the most robust, since it comes through despite a downward bias. Design issues unfortunately undercut its value for policy advice.
One aspect does find broad support in the blogosphere, media and social media: the desire to experiment and carefully design public policy is welcome. Nonetheless, careful experiment design is needed. Isolating the experiments from political influence is fundamental for useful results and policy guidance.
Republishing and referencing
Bruegel considers itself a public good and takes no institutional standpoint. Anyone is free to republish and/or quote this post without prior consent. Please provide a full reference, clearly stating Bruegel and the relevant author as the source, and include a prominent hyperlink to the original post.