Finland's hugely exciting experiment in basic income, explained

The headlines are shocking: "Finland plans to pay everyone in the country $876 a month"; "Finland plans to give every citizen a basic income of 800 euros a month"; "Finland plans to give every citizen 800 euros a month and scrap benefits."

The reality is a little less wild. Finland is definitely not close to paying "everyone in the country" or "every citizen" — not yet, anyway. What it is doing is weighing a proposal for a basic income: an approach to welfare in which residents would get a flat amount of money every month, regardless of how rich they are.

Finland is, however, conducting a pilot project, with a tiny fraction of the Finnish population participating, according to Olli Kangas, director of research at Kela, the Finnish Social Insurance Institution that runs basic unemployment, health, family, and many other benefit programs. Kangas and his research team have been tasked by the Finnish government with presenting proposals for testing out a basic income.

So, no, a European country is not on the verge of adopting a basic income. If the trial is a success, they could go all in, but that could be years off. Nonetheless, Finland is on the verge of conducting the most methodologically rigorous and comprehensive test of basic income to date. And that alone is a big deal.

How Finland is going to test out a basic income

The reason Finland is doing this now is thanks to Juha Sipilä — Finland's center-right prime minister — who took office in May after his Centre Party won the largest bloc of seats in parliamentary elections. The party's platform promised to bring a "new political culture of bold experiments," of which the basic income trial is to date by far the boldest. Sipilä has been publicly supportive of the idea of a basic income, and Kangas says the government has set aside €20 million (a little under $22 million) to fund the trial. The basics are laid out in the following PowerPoint presentation, which Kangas forwarded to me:

The consortium running the experiment — headed by Kangas but including researchers at various research centers, think tanks, and universities — plans to file an interim report on March 30 and a final report on November 15 that will lay out how exactly the experiment is to proceed. It will begin in 2017 and last for two years. So far, the consortium has already polled the Finnish population and found widespread support for the idea: 69 percent of Finns reported support for a basic income, with the median respondent calling for a minimum of €1,000 a month ($1,083).

The €800-a-month figure reported in the international press is just illustrative, Kangas says: "€800 is just one sum, nothing more," he wrote in an email. "Nothing is fixed yet. It can be any other sum or many other sums." Indeed, Kangas is looking to test more than one model. He emphasizes that nothing is fixed yet. The research group only makes suggestions, and political decision-makers will then determine which models, which research settings, which target groups, and what benefits levels will be experimented with in 2017 and 2018.

There are four options he says ought to be considered and evaluated:

1. A full basic income, which would replace most means-tested benefits. "It should be a rather high sum," Kangas writes. "Our basic pension is about €750 a month." However, he emphasizes that this can't replace all safety net programs, especially because unions in Finland help run the unemployment insurance system, and displacing them could cause the unions to hemorrhage members and lose political influence.
2. A partial basic income, about €550 ($596) a month, which Kangas says is on the level of basic government unemployment benefits (as distinct from the parallel union system). This option would preserve many existing benefits.
3. A negative income tax, in which benefits would phase out as people earn more money. For example, under a $10,000-a-year negative income tax with 50 percent phase-out, a household with no income would get $10,000, a household with $10,000 in income would get $5,000, and so on, until the household is making $20,000 and getting no benefits.
4. Miscellaneous other approaches. One possibility Kangas suggests is simply merging some existing benefits — basic unemployment allowance, minimum sick leave and maternity benefits, etc. — and then adding additional payments to reward desirable behavior, like caregiving and volunteering. This would create a kind of participatory income.

This degree of comprehensiveness puts the Finnish experiment far ahead of basic income trials in the past. US experiments from the 1970s not only focused exclusively on the negative income tax model, but they also only targeted existing welfare recipients. Utrecht, Netherlands, is conducting an experiment, but it's also only giving money to people already getting benefits. By trying a bunch of different models and not merely limiting the checks to existing welfare recipients, Finland's trial could give us much more meaningful results than past experiments.

The difficulties of doing a national experiment

Roni Rekomaa/AFP/Getty Images

Ideally, Kangas told me, he'd like to take several different kinds of samples. In the scientifically ideal research setting, there would be a national lottery so he gets a representative random sample of Finns across the country who'll receive a basic income. But he also wants to do regional lotteries that are regionally representative, and then lotteries confined to large towns. He also wants there to be some smaller municipalities that have a large portion of their populations (30 percent, say) get checks. In the PowerPoint, he suggests that in a couple of districts 100 percent of households could get checks.

The idea is to see what happens to a community under a basic income, rather than just to individual people. Having a whole town get benefits could have cascading effects as households escape poverty, as some people use the income guarantee as insurance so they can take risks and form companies, as universities see increased enrollment from people better able to afford supplies, etc. "If people in a smaller area are getting the benefits, their behavior vis-a-vis other people will change, employers and employees will change their behavior, encounters between clients and their street-level bureaucrats (social workers, employment offices, etc.) will change, and the interplay between different bureaucracies will change," Kangas says.

These externalities are not possible to catch in a national random sample, but the national random sampling makes for more reliable results on a household level. Thus, the combination of random nationwide sampling with local experiments would yield the most reliable results. But before any experiments start, all the models, their costs and distributional effects — who wins and who loses — will be evaluated by rigorous microsimulation, Kangas says.

A community-level trial has only really been tried once before, in Dauphin, Canada, in the 1970s, and the results were quite positive. While the more poorly designed US studies suggested that a negative income tax could deter people from work, in Dauphin only mothers of young children and teenagers appeared to work less; it's not as clear working less is a bad thing when it takes the form of maternity leave, or when it leaves young adults more time for studying and volunteering. Dauphin was instructive, but a more contemporary trial in Finland could provide still better evidence about what happens when an entire town comes under a basic income regime.

There are some ways in which the trial will necessarily be limited. It only lasts two years, and people's behavior after getting a benefit they know will be taken away shortly could differ from their behavior under a permanent basic income system.

The experiment also has a special difficulty due to the Finnish constitution's guarantee of equality, a guarantee that possibly conflicts with a policy randomly giving some people checks with no strings attached while giving others nothing. The PowerPoint suggests that the researchers will seek to get around this through the local experiments (where everyone in one area can get benefits, so people at least on the local level are equal) and perhaps by making participation voluntary, which could lead to some selection bias.

But all told, Kangas is overseeing what is by far the most important and informative test of a basic income to date. It's the first time a non-negative income tax basic income plan will be experimentally tested using rigorous methods in a developed country. That's a huge, huge deal, and we should all be paying attention in 2019 when the evaluations begin to trickle in. It's still very up in the air what the trial will show, and how the government will decide to use its findings.