CBI Pilot Project vs CBI Experiment
I shall employ the same distinction between a CBI pilot project and CBI experiment that Charlie Young employed in a paper presented at the 2018 BIEN congress:
- ‘Pilot: A full basic income pilot adopts all of its principles […] Pilots will, however, be temporally limited and may be applied to only a subset of the wider population eg a town, city or region.
- Experiment: Experiments are not full pilots as they may not be universal (eg they will target a particular cohort), have elements of conditionality, or do not meet the criteria of other principles laid out in the following sections. This is often due to financial, political and legal constraints.’ (Charlie Young, Realising Basic Income Experiments in the UK, RSA, 2018)
This is important to note as some of the ‘pilot projects’ described will be CBI experiments, as they will focus on a particular section of the population, whilst others will be fully fledged pilot projects where the results can be taken as more representative of a CBI’s effects.
Principles of a Basic Income
Before evaluating the state of each of the pilot projects, the principles of a Basic Income need to be established so that each pilot project can be compared. Charlie Young suggests the following principles:
- Basic Payments
- Regular Payments (at least once per month)
- Unconditional Payments
- Universal Payments
- Non-withdrawable Payments
- Equal and Individual Payments (exc. Children)
- Payments do not leave anyone worse off
- Ceteris Paribus – other things held constant
Do the pilot projects fit these principles?
Using Charlie Young’s principles, the following can be concluded:
(Table from Charlie Young, Realising Basic Income Experiments in the UK, RSA, 2018, used with permission)
(In the context of this table, ‘unconditional’ has a restricted meaning. It means ‘not work tested’. It does not mean that the payment is not withdrawn as earnings rise.)
Reviewing Historical and Current Projects
- Seattle-Denver, Dauphin (Manitoba) and Gary (Indiana): The 1960s/70s
These experiments were large-scale trials of a Minimum Income Guarantee in various places in Canada and the United States. Here, the structure of each experiment and the findings will be described.
All of the experiments focused on a Minimum Income Guarantee (MIG) system, and thus were experiments rather than pilot projects. A minimum guaranteed income was set for each household type, below which benefits would be paid to bring people up to this guaranteed income, and then withdrawn at a specified rate, so that a ‘breakeven point’ was reached, when the family would stop receiving benefits and start paying income tax.
All of the families involved in the Seattle-Denver experiment had at least one dependent family member and each single-parent family had an income of no more than $11,000 or $13,000 for 2-parent families.
In the Gary, Indiana, experiment, the subjects were predominantly black, single-parent families living in the local area. Similarly, urban and rural participants with incomes below C$13,000 from Winnipeg and Dauphin formed the 1,300 families that were involved in the Dauphin ‘Mincome’ experiment in Canada.
Labour Market Effects:
|Experiment||Work reduction amongst the following groups (% of hours p.a.):|
|Gary, Indiana||3% – 6%||0%||26% – 30%|
Clearly, work reduction took place in all of the experiments, bar ‘Wives’ in the Gary, Indiana experiment. However, Karl Widerquist highlights four key areas where these experiments may not be representative of the national impact if a CBI were to be introduced:
Firstly, the payments were offered to families on low incomes: In both the Seattle-Denver and Dauphin experiment, all of the families involved had a relatively low income. Therefore, the effect of the payments on their net incomes, and their ability to reduce their working hours, would have been disproportionately high.
Secondly, due to the relatively small size of the sample groups, and thus the relatively small reduction in labour market supply, the increase in wages would have been negligible. Yet, on a national level, the increase in wages would have been much larger, thus attracting workers, and so lessening the reduction in average hours worked.
Thirdly, Widerquist highlights how the large decreases in hours worked suggests people may have completely dropped out of the labour force; however, he finds that this drop was actually caused by workers taking longer to find their next job, thus having the positive effect of reducing underemployment.
Finally, the rise and falls in the macroeconomic labour market appeared to contribute to the changes in hours worked. For example, in some experiments, the hours worked actually increased as the macroeconomic labour market grew whilst the experiment took place. In the table above, the values are calculated with reference to a control group who were not provided with the Basic Income, and so it may be the case that those in the control group were forced to take less appealing jobs in order to earn the same level of income as those receiving the MIG payments. This would have the effect of decreasing worker satisfaction and so the figures here can be seen to support a Basic Income, as it allows workers to choose more satisfying and stimulating work.
Non-labour Market Effects:
Although not the focus of the experiments, the non-labour market effects of the experiments provided promising reading. For example, quality of life indicators such as school attendance, teacher ratings, and test scores, all increased during the experiment. Moreover, children were found to stay in school longer as a result of the experiments, and more adults re-entered education. Moreover, there were reduced incidences of low-birth-weight babies and both food consumption and nutritional content of the participants’ diet increased.
The effect of the payments on divorce rates (increasing them in the Seattle-Denver experiment) was widely reported. Some claimed this to be a negative consequence of a Basic Income, as it allowed family break-up to occur, but others claimed that the Basic Income freed women from financial dependence on their husbands and thus gave them the ability to live a dignified life.
Changes in the poverty indices highlighted the benefits of the payments, and the experiments demonstrated that the decrease in labour market participation was within an acceptable and sustainable range. However, the reports at the time focused on the work disincentive – taking the figures from the experiments and assuming that they would be representative of a national Basic Income (which they were not suggested to be) – and also the divorce rate increase. Thus, the experiments were not able to support Richard Nixon’s efforts to implement a MIG in his Family Assistance Plan sufficiently for it to pass into law, so the idea decreased in popularity until the 2000s.
- Otjivero, Namibia: 2007-2009
Namibia was a particularly poignant location for the next of the large Basic Income experiments, given that two-thirds of the population live below the poverty line and that it has one of the most unequal income distributions in the world.
All residents below the age of 60 years involved in the programme received an unconditional Basic Income of NAD100 per person per month. Given this universality, it can be seen as being the first ever universal cash transfer programme, and thus the first of the experiments to resemble a true Basic Income.
Firstly, the community responded surprisingly by establishing an 18-person panel that advised local residents on how they could best use their cash transfer. This highlights the importance of Basic Income in both bringing the community closer together and mobilizing it, so that it can have a more promising future.
Secondly, in the first 6 months of the trial, the number of underweight children fell from 42% to 17%, a remarkable statistic given the time frame. Similarly, the number of parents paying school fees and buying school uniforms doubled during the first 6 months, whilst the school drop-out rate fell from 30-40% to just 5%.
Thirdly, the majority of people in the experiment were able to increase their working hours, and the total income of the Otjivero community increased by a greater amount than the grants given, thanks to a number of new enterprises beginning as a result of the Basic Income.
Fourthly, health clinic attendance increased by five times more than at the beginning of the trial, as more people could afford to pay the NAD4 fee for an appointment. A similar trend appeared when looking at the quantity of ARV treatment for HIV/AIDS, thanks to Basic Income enabling residents to access both government health services and good nutrition.
In addition, the Basic Income resulted in the formation of a committee aimed at shutting alcohol taverns on the day of the grant, thus lowering the chances of alcoholism. The number of transactional sex cases also fell as a result of the Basic Income, whilst economic and poverty-related crime fell by 20%.
The funding for the Basic Income was provided by the Evangelical Lutheran Church, but the funding ran out after two years. Responsibility for the failure to implement a national Basic Income following the experiment lay with the Namibian government, which claimed that a Basic Income might ‘make people lazy and dependent on hand-outs’. The Namibian government also claimed that ‘we can’t dish out money for free to people who do nothing’. It seems that again a powerful group has focused on preconceived ideas surrounding a Basic Income, rather than looking at the research itself. Unfortunately, this seems to be a hard pattern to change.
- Madhya-Pradesh, India: 2010
India trialled an ‘unconditional cash transfer’, or Basic Income, among 22 villages in Madhya-Pradesh. The universality of the transfers, as Guy Standing has highlighted, starkly contrasts with the social assistance targeted at the poor by the Indian government. Cost saving and effective distribution were seen to be two of the potential benefits of a universal cash benefit, given its simpler administration.
Two pilot projects took place in the rural areas of Madhya-Pradesh, largely thanks to the Self Employed Women’s Association (SEWA) in partnership with UNICEF. The experiment was split, with the Madhya Pradesh Unconditional Cash Transfer (MPUCT) involving 8 villages who received the grant, and 12 who did not (and so acted as a control); whilst the Tribal Village Unconditional Cash Transfer (TVUCT) involved 1 tribal village receiving the transfer and 1 as a control. Thus, for 18 months, over 12,000 individuals were tested as part of the trial, 6,000 of whom received the cash transfer.
It is interesting to note that, unlike the American and Canadian experiments of the 1970s, which were focused on labour market effects, the Indian pilot project was designed to ‘identify the effects of basic income on individual and family behaviour and attitudes, and on community development’.
The table below shows the different amounts received by participants in the pilot projects:
|Pilot Project||Amount an adult received (Rupees per month)||Amount a child received (Rupees per month)|
|MPUCT (first year)||200||100|
|MPUCT (after one year)||300||150|
Conversion of these figures shows that the average family received the equivalent of $24 or £15 per month. This was a quarter of the income of median-income families in India, just above the official poverty line, and enough ‘to make a difference for basic needs’.
Any income destined for under-18-year-olds was paid to the mother, in a move to further emancipate and empower women under this scheme. Similarly, as with a Citizen’s Basic Income, this pilot paid the income individually in order to increase personal freedom and bargaining power, particularly for women, the elderly and the disabled in the household.
- Implementation and fiscal inclusion
93% of participants received the cash grant in the first month of the pilot. This highlights the administrative feasibility of a CBI, as a majority of participants in the pilot successfully received the income. Although paying the grant into a bank account presented SEWA officials with challenges during the first few months, meaning that cash was still being distributed, the transfer from cash to bank account payment was successfully achieved within a few months, by which time almost all participants had a bank account set up.
- Housing, Sanitation and diet
The Final Evaluation Survey (FES), conducted shortly before the end of the pilot, highlighted the fact that recipients of Basic Income grants were significantly more likely to make improvements to their dwellings. Specifically, in tribal villages, the grant was used to repair old houses, switch to better drinking water sources, and shift to better lighting, as well as 10% of the grant being used to build entirely new houses.
Diet similarly benefitted, as the weight-for-age score for children improved significantly, thus supporting similar findings in the Otjivero study. The cash grants also caused an increase in spending on fruit and vegetables, whilst those receiving the cash grants were no more likely to increase spending on goods such as alcohol and tobacco, again supporting the findings of the Otjivero study.
Many health benefits as a result of the cash grant were found, for example:
- Lower incidences of common illnesses
- More regular medical treatment and taking of medicines, particularly in the TVUCT pilot
- Higher levels of immunization
- Significantly more cash grant households acquiring health insurance compared with those who did not receive the grant
The improvement in health was mostly attributed to the increased affordability of medicines, thanks to the grant; however, many households also mentioned the higher consumption of better food, and lower anxiety resulting from the grants.
The two key statistics in terms of schooling were the 12% increase in enrolment rate in villages that received the cash grant, and the 29% increase in school attendance, also in these villages, compared with a 13% increase in the villages that did not receive the grant.
- Economic Activity
Cash grants, contrary to the preconception, were associated with an increase in labour and work, and also resulted in households that received the grant being twice as likely to increase their production work than those not receiving it. This also resulted in a switch from waged work to own-account farming and small-scale business, thus elucidating the entrepreneurial benefits of a CBI. All of this culminated in the households in receipt of the grants being more likely to increase their income from work than those not in receipt.
Furthermore, the cash grants also enabled some people with disabilities to become economically active. For the entire disabled population who participated, the grants also provided greater access to food and medical assistance, as well as a greater voice in how household income was spent.
- Debt and Savings
Finally, the grants improved a household’s chances of escaping from the poverty trap by providing an income to pay off existing debt or accrue savings. Those receiving the grant were twice as likely to reduce their debt, whilst the accounts the grants were paid in to were often utilised as savings accounts, thus providing more liquidity in the local economy.
The findings of both the MPUCT and TVUCT proved especially supportive of the introduction of a CBI, highlighting improvements in housing, sanitation, diet, health, schooling, economic activity, and savings. Guy Standing has highlighted how an unconditional income allowed people to meet their own needs by choosing their own spending priorities. Similarly, he suggests how beneficial a Basic Income would be for Indian families, as it would enable them to gain greater control over their lives. Although there are challenges in implementing a larger CBI in India, such as the large rural expanse and the cost of the project, the government is looking into such a proposal.
- Current Pilots
a) Ontario, Canada
The Ontario government, in Canada, has decided to stop an experiment involving 4,000 low income people aged 18 to 64 at three sites within Ontario: Hamilton, Thunder Bay, and Lindsay. Withdrawing the payments as earned income rose, and basing the payments on households rather than individuals, made the experiment a trial of a (non-work-tested) means-tested benefit. However, the results from this experiment, that paid C$16,989 p.a. less 50% of earned income to singles, and C$24,027 p.a. less 50% of earned income to couples, could still inform the debate surrounding a Basic Income’s impact on low income households. With a withdrawal rate of 50% of earned income, and payments differing in relation to household size, the usefulness of the experiment’s results will be limited if they are used to predict the effect of an individualised Basic Income.
Unlike an individualised, non-withdrawable CBI, the experiment provided relatively less for couples living together, thus creating an incentive to live apart. The potential motivation behind this, perhaps to provide a Basic Income Grant at low cost, is understandable, although the effects must be considered when evaluating the results.
KELA, the Finnish social insurance body, is conducting a 2-year study involving 2,000 participants aged between 25 and 58. The participants have been randomly selected from a nationwide group all receiving unemployment benefits. The experiment replaces the €560 per month unemployment benefit with a Basic Income of the same amount. Interestingly, the Finnish government made participation in the CBI trial mandatory for those selected. The reason for imposing the unemployment criterion is a lack of funding for a larger and more expansive CBI.
Such a low CBI has the advantage of being directly comparable with the unemployment benefit that it is replacing. Given this match, the effects of a CBI in one of its most hotly debated areas, the unemployed section of society, can be usefully tested. The focus of the experiment is stated to be ‘the effects of a Basic Income on employment’, a focus very similar to the studies in the United States and Canada 40-years prior. However, whilst the preconceptions of the 1960’s and 1970’s experiments were that a CBI would decrease employment, Finland is approaching these experiments with a long-term unemployment problem which they are looking to solve. Finland’s long-term unemployment insurance eligibility criteria are said to be causing a disincentive to work, and so the results of the CBI experiment will hopefully point towards CBI being a partial solution to the problem, as opposed to the pessimistic attitude adopted following the first CBI trials in America and Canada.
GiveDirectly, a US not-for-profit organization, is establishing an experiment in Kenya involving tens of thousands of local participants across numerous villages for several years. As a charitable organization, it has raised a lot of money through celebrity endorsement and is particularly keen to analyse the effect of cash, as opposed to other infrastructural aid programmes, on developing economies. The proposed CBI amounts range from US$0.50 to over US$1 per day, which may seem small, but given that the average income in some of the villages involved in the experiment is less than US$1 per day, this is a significant amount. The study will show how a CBI will impact less developed nations.
d) Other potential studies
Y-Combinator Research (YCR), a non-profit section of a technology company, is planning a CBI project in the USA involving payments of US$1,000 per month to 1,000 individuals for 3-5 years. Again, the participants will be households with below-median incomes for their local communities, and will have individuals aged between 21 and 40 years-old. Like the Finnish pilot, there is no withdrawal rate, and so a true CBI is being tested.
The Netherlands, concerned with the harsh sanctions imposed by ‘active labour market’ policies, are trialling a welfare scheme in which some ‘job requirement’ conditions are removed, and some benefits remain in place when a participant, previously unemployed, finds work. The local experiments involved in the trial were made possible by a 2015-ruling that allowed experimentation of such a scheme at municipal level.
Stockton, a city in California, has, like YCR, secured private funds to provide a CBI of US$500 per month for 100 people over 18 months. The impetus behind the trial is like YCR’s: the threat of technology to jobs within the California area, given the close location of Silicon Valley.
Other areas such as Scotland (working with the Royal Society of Arts), Barcelona (employing a Minimum Guarantee model with differing conditions), and British Columbia, are all planning future CBI projects.
In a different vein, filmmakers have raised enough money to provide 20 US citizens with a CBI of US$231 per adult and US$77 per child, and are planning to follow the participants across 2 years, with the goal of producing a feature film entitled ‘Bootstraps’.
This article draws on Charlie Young, Realising Basic Income Experiments in the UK, RSA, 2018
Sources for pilot projects
- Seattle-Denver, Dauphin (Manitoba) and Gary (Indiana): The 1960s/70s
For an overview of the 1960s and 1970s experiments, see: https://basicincome.org/news/2017/12/basic-income-guarantee-experiments-1970s-quick-summary-results/
For a detailed analysis of the Seattle-Denver trial, see:
- Otjivero, Namibia: 2007-2009
For a brief overview of the Otjivero (Namibian) study, see:
Alternatively, for an in-depth analysis of the outcomes of the Namibian study, see:
- Madhya-Pradesh, India: 2010
For a description of the Madhya-Pradesh study, see Guy Standing’s analysis:
Alternatively, please see the official website for the trial:
- Current Pilots
For a discussion of the current CBI experiments, please refer to Karl Widerquist’s blog post: