Along the shores of Lake Victoria in western Kenya, mobile phones in several hundred villages ding in unison on the first of every month. For more than 21,000 adults, the sound means one thing: 2,250 Kenyan shillings appearing in their bank accounts. The cash equals one-quarter to half of the average income for a two-adult household in Bomet County, one of the poorest in Kenya.
The money (roughly US$22.50) arrives courtesy of the US-based charity GiveDirectly, which is studying the effects of handing people lumps of cash with no strings attached — an idea known as a universal basic income (UBI). The mobile phones in these villages will ding every month for the next 12 years, making this UBI trial the longest and largest ever conducted.
“It’s a poverty-alleviation tool. Participants can invest in riskier things because they have their basic needs taken care of,” says Tavneet Suri, an economist at the Abdul Latif Jameel Poverty Action Lab at the Massachusetts Institute of Technology in Cambridge and one of the lead investigators on the Kenya trial.
The Kenya experiment is one of a handful of UBI trials in various stages of development around the world. Finland has already begun a trial, as has Ontario in Canada. Stockton, California, is planning to roll out its own experiment later this year. Although the concept isn’t new — it was first proposed by Enlightenment philosophers — it remained a fringe idea until the past few years, and governments are now starting to take it more seriously. Interest in the idea grew in the aftermath of the 2008 economic crisis and because of endorsements from Silicon Valley tech gurus such as Elon Musk.
Proponents of guaranteed income schemes argue that poor people will benefit more from unrestricted funds than from current welfare systems, which tend to have stringent requirements that often leave recipients trapped in poverty. “Universal basic income is about giving people cash without question, and trusting that they know how to use it in the most-effective way they can,” says Luke Martinelli, an economist at the University of Bath, UK.
For economists and public-policy scholars, the current interest in UBI provides an opportunity to conduct rigorous trials to determine whether it will produce measurable benefits. But translating a grand economic theory into workable policy is far from easy. Almost all trials have involved a small number of people or lasted just a few years, which limits their power. And there is no clear definition of success; researchers try to balance measuring potential gains in one area, such as health care, with potential offsets in another, including education and labour-force participation.
But for the growing chorus of voices calling for data-driven policy, trials such as the one in Kenya are the only way to see whether UBI actually works. “This is one of the first rigorous randomized control trials of UBI,” says Suri. “This is our chance to understand UBI and its impacts.”
The modern welfare state emerged out of the ashes of the Great Depression and the Second World War. As governments across the Americas, Europe and Commonwealth tried to rebuild their economies, they began taking an active role in providing for the well-being of their poorer citizens through grants, services and money earmarked for purposes such as housing and food. Although such welfare systems have improved standards of living, most require an immense bureaucracy to administer benefits and to ensure that recipients meet strict qualification standards.
Welfare critics have long argued that the administrative costs are huge and provide limited positive results; in some cases, they discourage people from finding jobs. In response, leaders across the political spectrum have latched onto the idea of UBI — which has been promoted over the centuries by luminaries such as Thomas More (in his 1516 novel Utopia), philosopher Thomas Paine, the liberal US President Franklin Delano Roosevelt and economist Milton Friedman, a favourite of conservatives including US President Ronald Reagan and UK Prime Minister Margaret Thatcher. Progressive politicians and thinkers have seen the idea as a way to end poverty; conservatives have viewed it as a streamlined welfare system that is easier and cheaper to run.
In the 1960s and 1970s, a handful of sites across the United States tested a scheme related to UBI called a negative income tax. In this kind of programme, individuals making below a certain amount receive supplemental money from the government. But after early results from one of the trial sites revealed an increase in divorce rates, politicians nixed the idea as being toxic to the American family.
Another early test happened across the border in the small prairie town of Dauphin, Canada. In a trial called Mincome, supported by the federal and provincial governments, the town’s poorest residents received monthly cheques from 1974 to 1978 with no constraints on how the money should be spent. Researchers tracked changes in the proportion of people working full- and part-time, as well as in nutrition, education and basic health outcomes. But before the trial could be analysed, waning funds and political change scrapped the idea, and all the data were packed in more than 1,800 boxes and stored in a warehouse. They sat there until economist Evelyn Forget at the University of Manitoba in Winnipeg brushed off layers of dust and opened the boxes.
The documents Forget uncovered revealed that teenage children in MINCOME families completed an extra year of schooling compared with teens in similar small Manitoba towns. Hospitalizations decreased by 8.5%, with the largest drops in admissions for accidents and injuries and mental-health diagnoses. Importantly for economists, who worried that the programme might encourage people to quit their jobs, Forget found that employment rates stayed the same throughout the trial (E. L. Forget Can. Public Policy 37, 283–305; 2011).
Now, supporters of UBI in several countries are trying to build on the results from those earlier studies and develop trials to decide whether governments should give UBI a chance.
The Kenyan experiment grew out of smaller trials that the charity GiveDirectly had done in sub-Saharan Africa. Starting in 2009, the group tried to alleviate poverty by providing relatively modest direct cash transfers. These shorter and smaller injections of cash created ripple effects through the communities involved. In a trial in Zimbabwe, one year of cash transfers improved childhood vaccination rates and school attendance (L. Robertson et al. Lancet 381, 1283–1292; 2013). Because the transfers were only short-term and too small to cover living expenses, they weren’t full-blown UBI trials. But that early work gave the group a leg up on planning a UBI trial, says Michael Faye, co-founder and president of GiveDirectly.
Experts say that full-blown experiments are particularly difficult to design because they require a great deal of advance planning. “With a lot of these projects, the devil is in the details, and the design of research depends on a fine-grained knowledge of its impact,” says Rob Reich, a political scientist at California’s Stanford University, who is not part of these trials.
Guaranteed-income experiments are different from many clinical trials because researchers are looking for improvements in a wide variety of areas and doing so across communities rather than in individuals. Suri describes a cycle of improvements that UBI might create: well-fed pregnant mothers should have healthier children than would undernourished women. Longer education would create better job opportunities and delay marriage and childbirth, resulting in healthier mothers and babies. Suri says that her team plans to track everything from entrepreneurship to health, education and nutritional status, with the help of a platoon of locals who will go door-to-door, and do a series of short phone check-ins and some in-depth interviews with village elders to get a big-picture view of the intervention’s effects.
Because the trial will be so long and expensive, Suri helped to design four different arms to answer as many questions as possible. The largest arm provides 2,250 Kenyan shillings every month for 2 years to each adult in 80 villages. A second arm provides the same amount of money each month for 12 years. A third arm provides a total equivalent to US$505 — 2 years’ of basic income — in 2 payments, separated by 2 months. The fourth arm serves as a control group that gets nothing.
“We can run a horse race between different types of UBI,” Suri says.
Participants in a pilot project that began in 2016 are enthusiastic about their prospects. “This has made me believe that I can commit and be able to pay school fees for my children and I am also confident of saving money to improve my business,” says Jael.
A UBI experiment in Finland has been struggling. The project grew out of concerns that the country’s complicated unemployment benefit system keeps some people from returning to full-time work because that would curtail their support. In March 2016, the government welfare agency Kela teamed up with a non-profit research organization Tänk to announce a UBI trial that would provide €560 (US$658) per month to a group of 2,000 adults currently receiving unemployment benefits around the country. The extra income would not be taxed at the same rate as normal unemployment benefits, nor would participants be required to actively search for work during the two-year trial. They also wouldn’t lose the UBI income if they found work.
Although global media initially praised the programme, popular opinion later soured on the scheme, which cost €20 million. The monthly payments were nowhere near enough to cover an adult’s basic living expenses, and the trial addressed only adults who were unemployed at the time. Plus, there was no adequate control group. In late April 2018, the Finnish parliament refused requests from Kela for another year of funding and instead expressed preferences towards developing other welfare schemes. To UBI advocates, the problems with the Finnish programme have proved a serious barrier to getting other trials up and running.
“People’s expectations were far higher than the trial could deliver,” says Karl Widerquist, a political economist at Georgetown University in Qatar and co-chair of the Basic Income Earth Network, which promotes UBI.
Other researchers have started a range of UBI trials and precursor projects (See ‘Money for nothing’). The city of Stockton has begun an experiment that is underwritten by philanthropic organizations. The trial will be smaller than the Kenya trial — including only 100 people for 12–18 months — owing to funding limitations, says Taylor Jo Isenberg, managing director at the Economic Security Project, who is helping to fund the experiment. But the effort will provide good preliminary data for later trials, she says, because Stockton is a microcosm of the United States in terms of diversity, poverty levels and job loss from automation and outsourcing.
“Updating previous research on unconditional cash transfers is really important but really expensive. We hope to open the door for other stakeholders to step in later,” Isenberg says.
The Canadian province of Ontario is also hopping on the UBI bandwagon. Its trial began late last year, enrolling more than 4,000 individuals across the province. As spectators watch for signs of success and failure in these trials, the researchers involved need to define what, exactly, ‘success’ or ‘failure’ means. Given that some of the impacts of UBI won’t be felt for 5–10 years, onlookers might be waiting for quite some time.
The announcement in April that the Finnish UBI trial wouldn’t be funded beyond this year provided a sobering reminder that politics — more than data — will determine the fate of such programmes. The government pulled the plug before Markus Kanerva, managing director at Tänk, and his colleagues had examined the data to see how well the trial worked, something Kanerva says his team is waiting to do until late 2018 so as not to bias the results.
The outcome of all these trials is far from clear, not least because the Kenya project — the most ambitious — only just began.
Problems of scale
Over time, the trials could generate data on the costs and benefits of UBI schemes, such as whether the initiatives reduce health-care expenditures. But Martinelli thinks that the data will show that it will cost too much to make a programme effective. “An affordable UBI is inadequate, and an adequate UBI is unaffordable,” he says.
But even a clear win in these trials won’t necessarily indicate that UBI would work in practice, says economist Damon Jones at the University of Chicago in Illinois. Because they are relatively small and most of the funding comes from private sources, the trials won’t provide a sense of whether governments could afford a big public programme or whether citizens would be willing to fork out extra taxes to fund them. “Medicine can be scaled up, but this isn’t as easy,” says Jones. A new cancer drug might extend lifespan by 3 months, which stays true whether 10 people take the drug or 10,000. In a UBI trial, 10 people receiving cash will have a very different impact on a community compared with 10,000.
Jones cautions that this doesn’t mean the UBI trials shouldn’t be done or that they will produce meaningless data, just that even the best-designed studies have inherent limitations.
Regardless of the outcomes, the trials will have an ongoing impact because they can identify potential flaws in the process, help researchers refine the questions they ask and give policymakers some of the answers they crave. If the trials succeed, “it wouldn’t just be an outlier in social policy, it would be a minor miracle”, Reich says.
For the participants of the Kenya trial, that minor miracle has already arrived. The knowledge that GiveDirectly will deposit funds into their accounts every month for more than a decade has already begun to shift how some of them think about money. Each text alert means a chance to invest in their own lives or their businesses with the security that they can still put food on the table. And that, they say, is priceless.
Nature 557, 626-628 (2018)