How do you design a universal basic income experiment? Tips for policymakers

Posted in: Basic income, Economics, European politics, Evidence and policymaking, Global politics, UK politics, US politics, Welfare and social security

Dr Joe Chrisp is a Research Associate in the Institute for Policy Research (IPR) at the University of Bath. Laura Smyth is a Research Assistant in the IPR at the University of Bath. They are co-authors of a rapid evidence review, ‘Basic income experiments in OECD countries’, published by the International Public Policy Observatory (IPPO).

Since 2015, there has been a surge in political and media interest in a universal basic income (UBI) in OECD countries.

For example, on 16 February 2022, the Welsh government announced plans to roll-out a basic income pilot for care leavers. It will provide a benefit of around £1600 a month for two years to roughly 500 people. An amount that would be one of the highest levels of benefit offered in such an experiment and roughly equivalent to the statutory minimum wage. Although it will be taxed as income by the UK government.

A UBI is often defined as a “regular income to all individuals within a political community, irrespective of working status or income from other sources, with no strings attached” and as such marks a radical departure from existing social security systems that have job-seeking requirements and are either means-tested or based on a contribution record.

The most tangible development in response to this growing interest has been the mushrooming of social policy experiments, either loosely or directly tied to the idea of a basic income, instigated by governments of various levels.

Working in collaboration with the IPPO, we sought to examine these experiments, and co-authored a first-of-its-kind evidence-based summary of existing basic income experiments for policymakers in OECD countries. We identified 38 relevant experiments, 21 of which had been completed by November 2021, and compared their core characteristics, key results, and policy outcomes.

What did we find?

  • The majority were ‘bottom-up’ and not led by national governments, with a growing trend for the involvement of NGOs.
  • Most had a small number of participants.
  • Most focused on low-income households or benefit recipients rather than on a sample from a universal population.
  • Nearly all were targeted and dispersed rather than universal within saturated sites.


Key takeaways

  • Evidence on employment outcomes is weak or not statistically significant in most cases. Except for certain sub-groups in specific contexts.
  • Most experiments show positive wellbeing effects, although the evidence is also often limited and subjective.
  • No experiments to date have ended with the implementation of a basic income
  • Most have not led to any clear policy reform according to available data.


Key points for policymakers

Using these key takeaways, we developed a ‘toolkit’ for policymakers looking to implement universal basic income schemes in the future:

How to meet research goals:

  • Pay attention to sample size and design simplicity to enhance evidence robustness.
  • Under-researched UBI elements include: effect on non-benefit recipients or members of low-income households; effect over a longer period of time; and its effect on a small community when it is universally provided.


How to meet pilot goals:

  • Experiments should test schemes and interventions that would be fiscally and legally feasible for that level of government to implement.


How to meet political goals:

  • More work should be done to build coalitions within civil society and political groups with outreach done before, during and after the experiment itself.


We also grouped policymakers’ goals into three distinct categories:


‘Research’ goals relate to the advancement of (global) knowledge about the effects of a basic income on a variety of outcomes. On this front, broadly speaking, we suggest careful consideration should be paid to the sample size and the simplicity of the design to enhance the robustness of the evidence.

Past experiments have often included multiple interventions the effects of which are difficult to disentangle, particularly with a small sample size. Policymakers could also consider the gaps in current knowledge and design experiments so that aspects of a basic income that we do not have clear evidence on are examined. This may include, for example, (1) the effect of the benefit on those that are not already either benefit recipients or members of low-income households, (2) the effect of the benefit over a longer period of time or (3) the effect of the benefit on a (small) community when it is universally provided.


‘Piloting’ goals relate to the desire to road-test a policy that could feasibly be implemented by the government pursuing the experiment.

With such a goal in mind, it is essential to use a design that would be fiscally and legally feasible for that level of government to implement. This has not been the case in many past experiments, which inevitably limits the extent to which policy reform can be initiated that directly builds upon lessons from the experiment.


‘Politics’ goals relate to the enhancement of public understanding and support for basic income and its principles.

The insights from academic literature are highly limited here given existing research does not focus as much on the political and policy achievements or failures of UBI experiments.

Yet, clearly political goals are at the heart of many of these experiments given they often test schemes that they could not implement.

Thus, we can infer that a key aim of the experiment is to persuade others that may have the power to implement such a policy. Designers of experiments here must tread a fine line between wanting to convey the reliability of their findings, precisely to strengthen the message that their policy intervention works, or to involve themselves more directly in outreach efforts and political campaigning alongside the experiment.

This is a trade-off because the latter risks falling foul of RCT guidelines and avoiding any outreach or campaigning risks entirely losing the public relations battle in framing the purpose and findings of the experiment.

The announcement from the Welsh government should be commended on the grounds that it meets our recommendations on ‘piloting’ goals. While a Senedd committee has criticised the lack of ambition, and particularly the sample size, the scheme is designed so that it is possible for it to be rolled out following a successful experiment. The fact that the basic income will be treated as taxable income is also likely to mean that recipients will also pay tax from their first £ of income, as other modelled basic income schemes tend to be.

Of course, it is still too early to say what its political strategy will be, and we will be observing with interest how this more pragmatic course ends up in terms of future policy development.

All articles posted on this blog give the views of the author(s), and not the position of the IPR, nor of the University of Bath. Read more of our blogs on basic income, and learn more about our research in this area.

Posted in: Basic income, Economics, European politics, Evidence and policymaking, Global politics, UK politics, US politics, Welfare and social security


  • (we won't publish this)

Write a response