Do We Need Basic Income Experiments?

1 Introduction

In this essay terms will be defined, the difference between ‘Basic Income’ and ‘Basic Income scheme’ will be clarified, the feasibility of Basic Income schemes will be explored, a variety of past, current and future Basic Income pilot projects and experiments will be discussed, the difficulties of running a pilot project in a country with a more developed economy will be explored, methods for evaluating the financial feasibility and a number of effects of Basic Income schemes will be compared, and the question that constitutes the title of this essay will be answered. The thesis of the article is that microsimulation research can deliver much of what we might normally hope to learn from running pilot projects, so a microsimulation of a Basic Income scheme should be counted as a pilot project: but that there are some effects of a Basic Income scheme that microsimulation cannot test, and for those effects on-the-ground pilot projects are required.

2 ‘Basic Income’

The UK’s Citizen’s Basic Income Trust (CBIT) defines a Citizen’s Basic Income/Basic Income as ‘an unconditional, automatic and nonwithdrawable income for each individual as a right of citizenship.’ (Citizen’s Basic Income Trust, 2018, p. 3).

And the Basic Income Earth Network (BIEN) defines it as ‘a periodic cash payment unconditionally delivered to all on an individual basis, without means-test or work requirement.’ (Basic Income Earth Network)

These definitions represent a consensus among BIEN’s affiliated organisations (Torry, 2017), and they reflect common usage of the term ‘Basic Income’. For the purposes of this essay, I shall employ a working definition based on the two definitions above: A Basic Income is ‘an equal and regular payment of the same amount, varying only by an annual uprating, to every individual of the same age legally resident in the jurisdiction in question.’

(In this working definition, ‘equal’ means that every individual receives the same amount as every other individual when each regular payment is made; and ‘the same amount’ means that all of the payments to each individual are of the same amount, except that once a year the regular amount might increase slightly. The working definition states the practical consequence of the BIEN and CBIT definitions, and makes one particular normal assumption explicit: that different amounts might be paid to different age groups.)

National organisations affiliated to BIEN disagree over whether the definition of ‘Basic Income’ implies payment of Basic Incomes at a ‘subsistence’ level, somehow defined (Torry, 2017): but neither CBIT’s nor BIEN’s definition assumes that a Basic Income will be paid at a particular amount, and our working definition similarly does not assume that it will be. For the purposes of this essay, a regular, unconditional and individual income of any amount counts as a Basic Income.

The definitions do not require the incomes to be permanent in order to be called Basic Incomes. We would expect the social and economic effects of short-term unconditional incomes to be different from those of long-term ones: in particular in relation to employment market behaviour, because if the Basic Income is only short-term then behaviour will be largely determined by the situation that will follow the experiment, whereas with a permanent Basic Income behaviour will be determined by the new Basic Income scheme. However, that difference does not compromise the fact that both short-term unconditional incomes as well as long-term ones can properly be called Basic Incomes; and that consequence means that short-term experiments with unconditional incomes can properly be called Basic Income pilot projects or experiments.

In this essay, the definition of Basic Income will be strictly applied. If an income fails to fulfil any of the elements of the definition, then it is not a Basic Income.

3 ‘Basic Income’ and ‘Basic Income Scheme’

A Basic Income is ‘an equal and regular payment of the same amount … to every individual of the same age …’. A Basic Income scheme is a Basic Income with its levels for different age groups specified, along with the funding method specified in detail, and any changes to existing tax and benefits systems also specified in detail. A Basic Income is always the same: ‘an equal and regular payment of the same amount … to every individual of the same age …’: but for any country, there will be an infinite number of possible Basic Income schemes.

The distinction is important. Arguments against any particular Basic Income scheme are not arguments against Basic Income (Torry, 2018b); and if in any country just one particular Basic Income scheme is feasible, then Basic Income is feasible in that country.

The distinction is also important because it draws our attention to precisely what must be tested if a genuine Basic Income pilot project is to be carried out. The whole idea of a pilot project is to test something that could be done. This means that simply paying a Basic Income to a pilot community does not necessarily constitute a Basic Income pilot project. Only if the Basic Income scheme tested in the pilot community could be rolled out as a national scheme should the project be termed a Basic Income pilot project. This requires a scheme to be tested for national financial and administrative feasibility before the pilot project is carried out, and that a view has been taken that the scheme could potentially meet psychological, political, behavioural and policy process feasibility tests (Torry, 2016).

4 ‘Experiment’ and ‘Pilot Project’

In The Palgrave International Handbook of Basic Income, a distinction is made between Basic Income experiments and Basic Income pilot projects:

Only if a project is testing a genuine Basic Income is it right to call the project a ‘Basic Income pilot project’. A project that tests anything else should be called an ‘experiment’ from which useful lessons might be learnt. (Torry, 2019, p. 10–13)

The title of this essay contains the phrase ‘Basic Income experiments’, which suggests that genuine Basic Incomes are being tested. In order to align this article with the definitions in the Handbook, if an experiment is testing a Basic Income, then it will be called a ‘Basic Income pilot project’, and if it is not then it will be called an ‘experiment’.

5 Basic Income Pilot Projects

In order to understand what constitutes a Basic Income pilot project I shall study a number of on-the-ground experiments.

There have been two genuine Basic Income pilot projects, in Namibia and India. In Namibia, everyone in the chosen community was paid a Basic Income for a period of two years, and in India the residents of a number of villages were paid Basic Incomes for shorter periods. The ‘jurisdictions’ were small, and the periods of payment were short, but those characteristics of the experiments do not invalidate the payments as Basic Incomes. Every individual of the same age in those jurisdictions was paid an income of the same amount, unconditionally. There were differences between the projects: the Namibian project took place in a single community, whereas the larger Indian project took place in a large number of communities of three distinct kinds; and for the Namibian pilot community there was no control community that did not receive the Basic Income, whereas in India the pilot communities were matched with control communities that did not receive the Basic Income so that the effects of the Basic Income could be properly evaluated (Widerquist, 2018, p. 19–25). These differences did not invalidate the Namibian project’s right to be properly termed a Basic Income pilot project.

In both cases, the outcomes were significant. The pilot communities developed democratic structures to control aspects of the local economy; women, people with disabilities, and elderly people, experienced emancipation; economic activity increased, especially among low-earning households; no unwanted economic effects occurred; new businesses were established; money was pooled to enable such infrastructure as toilets and a post office to be built; school attendance and results improved; the community’s health improved; nutrition improved; greater social cohesion was experienced; and crime was reduced (Basic Income Grant Coalition, 2009; Davala, 2019; Davala, Jhabvala, Mehta, & Guy, 2015; Haarmann & Haarmann, 2007, 2012; Haarmann, Haarmann, & Nattrass, 2019; Torry, 2013, p. 69–73, 2018a, p. 134–35; Widerquist, 2018, p. 58–59).

The income recently tested in the aborted Ontario experiment was income-tested and household-based (Ontario, no date; Widerquist, 2018, p. 65), so it was not a Basic Income: so although the experiment was an interesting one, it ought not to have been called a Basic Income experiment or pilot project.

The recent Finland experiment (De Wispelaere, Halmetoja, & Ville-Veikko, 2019; Widerquist, 2018, p. 64–65) selected two thousand unemployed individuals throughout the country and made their unemployment benefit unconditional for a period of two years. The important question is whether we can count widely dispersed individuals as a community or jurisdiction. Because this question can be answered either way, we have to conclude that the Finland experiment fell between being a Basic Income pilot project and not being one.

Experiments during the 1970s in the USA and Canada were Minimum Income Guarantee experiments: that is, they ensured that household annual disposable incomes did not fall below prescribed annual minimum income levels. In order to reach the prescribed minima, income-tested benefits were paid to households: so these were not Basic Income pilot projects, and the extent to which the significantly useful results of the experiments (in terms of improved health and other outcomes) can be used to predict the effects of genuine Basic Incomes must remain an open question (Forget, 2011; Widerquist, 2018, p. 43–55).

As far as I know, no genuine Negative Income Tax (NIT) experiments have ever taken place (that is, experiments that weekly or monthly pay with each individual’s salary an amount of money proportional to the extent to which earned income falls below a tax threshold): but if any such experiments were to take place, then we would have to say the same as we have said in relation to Minimum Income Guarantee experiments: that because of the substantial administrative differences between a Negative Income Tax and a Basic Income, results from the NIT experiments could not necessarily predict the effects of a Basic Income scheme (Widerquist, 2018, p. 15–17).

The major difference between all of these experiments—whether with unconditional individual incomes or not—and a genuine nationwide Basic Income, is that the experiments were all short-term, which means that results must be treated carefully. But that is probably not a problem. Take, for instance, employment market behaviour. In Namibia and India, economic activity increased, and the employment market changed in character, with shifts away from employment and towards own-account economic activity. If the Basic Incomes had become permanent during the pilot project, then we can theorise that because everyone’s economic activity would then have been determined entirely by the new Basic Income scheme, and no longer partially by the situation that would transpire at the end of the two-year experiment, the employment market effects might have been more pronounced than they were, and were unlikely to be less than they were (Basic Income Grant Coalition, 2009; Davala, 2019; Davala et al., 2015; Haarmann & Haarmann, 2007, 2012; Haarmann et al., 2019). Similarly with the Finnish experiment. No employment market effects were noticed during the first year of the experiment, which means that any employment market withdrawal was balanced by new employment market engagement. The second year saw a small increase in employment market activity (Kangas, Jauhiainen, Simanainen, & Ylikännö, 2020, p. 188). With a permanent scheme, in which a permanent Basic Income would not be affected by additional earned income, we can theorise that there would be less incentive to leave employment, and more incentive to engage with it, suggesting that employment market activity would be higher with a permanent Basic Income than during the experiment. Clearly only permanent Basic Incomes can test this: but those would no longer be pilot projects. Similarly, all of the projects delivered enhanced wellbeing, so we can theorise that permanent Basic Incomes would deliver even more of it.

As discussed, an essential test as to whether an experiment is a genuine Basic Income pilot project is only passed if the Basic Income scheme tested would be feasible to roll out nationwide. The small amounts paid in Basic Incomes in the Namibian pilot project if paid to the entire Namibian population would constitute between 2.2 per cent and 3 per cent of the country’s Gross Domestic Product (Haarmann & Haarmann, 2005, p. 2; Haarmann et al., 2019, p. 358), and so would be financially feasible; and in India a nationwide Basic Income at the level of that paid during the pilot project could be funded by abolishing existing corrupt poverty-reduction schemes (Davala, 2019, p. 380–382). Both the Indian and Namibian schemes would clearly be administratively feasible to roll out across their respective countries, and although neither is currently politically feasible, it is possible to envisage situations in which they might become so. The Finnish experiment made unemployment benefits unconditional for two thousand randomly selected individuals, without changing existing tax and benefits systems. The scheme would not be financially feasible if rolled out nationwide, suggesting that what might otherwise have been ambiguously a Basic Income pilot project was in fact not one.

6 The Conditions for Running a Genuine Basic Income Pilot Project in a Developed Country

Now that we are clear as to what might constitute a genuine Basic Income pilot project, we can ask whether a Basic Income pilot project could be conducted in a country with a more developed economy, and in particular one with more developed tax and benefits systems than are to be found in Namibia and India. Those two pilot projects were possible because, even though such projects are never without their difficulties, there were no complex tax and benefits systems that had to be changed just for the pilot communities if the experiments were to mirror Basic Income schemes that could be rolled out nationwide.

The situation is very different in more developed countries with existing complex tax and benefits systems. Take the United Kingdom (UK) as an example. For a nationwide Basic Income scheme to be feasible in the short to medium term, the existing Income Tax Personal Allowance—the amount of someone’s income that is not taxed—would have to be removed or reduced; Income Tax rates might have to rise; National Insurance Contributions (the UK’s social insurance contributions) rates and thresholds might have to change; and means-tested benefits claims would have to be reduced (Torry, 2019, 2021). For a genuine pilot project to take place, these changes would have to take place just for the pilot community or for randomly selected individuals. The problem is that it is exceedingly difficult to alter complex tax and benefits systems just for a single community or individual. A number of Scottish boroughs and English cities have expressed a wish to run Basic Income community-wide pilot projects, but a report published following a Scottish feasibility study makes it clear that only if the UK government were to be willing and able to alter the tax and benefits systems just for the pilot communities would a pilot project be possible (Citizens’ Basic Income Feasibility Study Steering Group, 2020, p. 7–8). If there was a UK government that wished pilot projects to take place, then it could ask the Department for Work and Pensions (which runs means-tested benefits) and Her Majesty’s Revenue and Customs (which runs Income Tax and National Insurance Contributions) to make the necessary changes for pilot communities: but that would still be remarkably difficult to achieve. Equally difficult to achieve would be a pilot project for a Basic Income scheme that redistributed from rich to poor, even if only mildly (Widerquist, 2018, p. 37–42). Because of the complexity of tax and benefits systems in more developed countries, an experiment of which revenue neutrality or something close to it was required (because a nationwide scheme would probably require that) would need to prevent losses to low-income households, and so at least some wealthier households would have to lose money. Given that wealthier households are likely to enjoy political influence, it is easy to see that only experiments that did not require any household to lose money would be likely to be politically feasible, which would of course mean that no genuine Basic Income pilot projects could take place.

But that is not necessarily the end of the story. Two possibilities have been researched and have been found to be feasible were a UK government wishing to run pilot projects (perhaps because policymakers were sympathetic towards Basic Income as an idea but were worried about possible press reactions to a decision to establish one) (Widerquist, 2018, p. 74–75, p. 131–39). One possibility would be a Basic Income for a single age group; and the other would be a Basic Income for an entire community.

1. The working definition of Basic Income employed by this article, with which the CBIT and BIEN definitions are consistent, does not require every age group to be provided with an unconditional income for every individual: what it requires is that everyone of the same age should be paid an equal and unconditional regular individual income. In the UK, it would be feasible to pay Basic Incomes to younger adults—for instance, 18–21 year olds—and to allocate to each of them a ‘BR’ (basic rate) tax rate, which automatically removes the Income Tax Personal Allowance for each of those individuals. Because young adults are usually not already enmeshed in the worst complexities of the UK’s tax and benefits systems, there would be very few anomalies to negotiate. It is of course debatable whether such a nationwide experiment should be regarded as a pilot project. First of all, there would be no control community (Widerquist, 2018, p. 19–25); and secondly it would be difficult to take the Basic Income away once it had been implemented, so the experiment might turn out not to be an experiment at all. The extent to which evaluation would be possible is debatable.

2. The second option assumes that a more ambitious UK government would be able to employ a similar approach for an entire community. It would not be able to make every change that would be required for the pilot project community’s Basic Income and associated tax and benefits changes to match those that would be required for a nationwide Basic Income scheme, but it could get close enough to enable valid conclusions to be reached as to some of the likely effects of a national Basic Income scheme (Torry, 2019).

Every country is different in relation to its tax and benefits systems, so research would have to be undertaken separately for each country to determine whether it would be possible to make the tax and benefits changes for a single community that would enable a Basic Income pilot project to be run that would be close enough in character to a scheme that it would be feasible to roll out nationwide for valid conclusions to be reached as to the likely effects of a nationwide scheme.

We rarely experiment with changes to tax and benefits systems before making the changes across an entire population. First of all, as we can now see, it is always difficult to run such experiments; and secondly, decisions about reforms are taken on the basis of political ideology and a variety of common presuppositions as much as on the basis of the kinds of effects that a pilot project might be able to measure (Widerquist, 2018, p. 87–91). A Basic Income scheme can be viewed in two different ways: as minor administrative changes to current tax and benefits systems (Torry, 2019), or as wholesale reform of the benefits system. If the former assumption is made, then a pilot project would seem to be an unnecessary adjunct to normal policy reform processes; if the latter assumption is made, then a pilot project might be thought appropriate, and if there might be widespread demand for one, or policymakers were hesitant about implementing what they might see as a major reform (a not unlikely eventuality given the problems that successive UK governments have had with the roll-out of the UK’s means-tested Universal Credit), then a carefully planned pilot project could be useful.

7 Methods for Testing a Basic Income Scheme’s Financial Feasibility

If we are to ensure that we only run pilot projects that mirror a Basic Income scheme that could feasibly be rolled out nationwide, then we shall need to test Basic Income schemes for feasibility, and particularly for financial feasibility. This is by no means the only feasibility test that a scheme would have to pass, nor the most difficult. The others would be tests for administrative feasibility (easy to pass), psychological feasibility (whether people understand the idea of Basic Income and understand it to be beneficial), behavioural feasibility (would it have the predicted effects?), political feasibility (do current political ideologies support it?), and policy process feasibility (could the scheme navigate its way through the country’s complex policy process from idea to implementation?): tests for which a variety of research methods, including opinion surveys, would be required—and different countries might find that additional feasibility tests would need to be passed. But however many feasibility tests there might be in a particular context, proving financial feasibility would always be essential, and in two respects: first of all, would it be possible to pay for the Basic Incomes? and secondly, would low-income households be made worse off by the combination of their Basic Incomes and the changes that would have to be made to the tax and benefits system to pay for it? (Torry, 2016) (It is sometimes suggested that if alternative funding mechanisms were to be employed, then this problem would not arise. That is not true. For instance, a carbon tax would increase the costs of heating homes and of transport, and therefore of food and other commodities. The overall effect could be net disposable income losses for low-income households. The same could occur if Basic Incomes were to be financed by increasing consumption taxes.)

So what is required is a method for ensuring that the net cost of the Basic Income scheme as a whole is low to zero, and that no significant losses would be suffered by low income households. We might also wish to ensure that the scheme would reduce all poverty indices, inequality indices, and by a substantial proportion the number of households receiving means-tested benefits.

During the earlier years of the modern Basic Income debate, the question being asked was whether schemes were affordable. If the scheme abolished existing means-tested benefits, then a ‘national accounts’ method could be employed. This used the national accounts, census data, and other national statistics, to calculate the cost of giving to every member of the population a Basic Income, the money saved by abolishing tax allowances and means-tested and other benefits, and the additional revenue that would be collected from increased tax rates. The net cost of the Basic Income scheme was the total cost of the Basic Incomes, less the money saved by abolishing benefits, and less the additional tax revenue collected. If a revenue neutral scheme was required, then by a process of trial and error the net cost could be reduced to zero. This was the method that used to be used in the UK (House of Commons Work and Pensions Committee, 2007: Ev.84–90) and that is still sometimes used by researchers (Citizen’s Income Trust, 2016; Miller, 2017; Painter & Thoung, 2015). The problem with this method is that it cannot tell us about losses that households would experience at the point of implementation: a serious problem, because research in 2012 and 2014 showed that those losses were likely to be substantial for low-income households (Torry, 2014). A further problem was that if means-tested benefits were retained and recalculated rather than abolished, then the national accounts method could not tell us how much additional revenue would be available to fund the Basic Incomes. This is because the reduction in any individual household’s means-tested benefits if a Basic Income scheme were to be implemented would depend on a wide variety of factors relating to household structure, household members’ earned incomes, and so on, and only knowing all of those circumstances for every household, and being able to calculate their effects on means-tested benefits for every household receiving them, would enable the additional funds available for the Basic Income to be known. The national accounts method deals only in country-level statistics and cannot handle household-level information or calculations.

Because the ‘national statistics’ method is inadequate in the context of today’s more demanding debate, in which detailed questions are often asked about the financial feasibility of illustrative Basic Income schemes, and in which the possibility of net disposable income losses at the point of implementation is well understood, we are fortunate to have available an additional research method: microsimulation. This can provide all of the information that we need: the net cost of a scheme; household losses at the point of implementation, including for low-income households; changes in poverty and inequality indices; redistribution patterns; and the numbers of households taken off retained and recalculated means-tested benefits. Two different strategies are available. Some researchers propose illustrative Basic Income schemes and then research their financial effects. Another approach is to set criteria for a financially feasible Basic Income scheme and then to test multiple illustrative schemes until one is found that fits the criteria (Torry, 2019).

Microsimulation is constituted by a) a computer programme into which are coded all of the tax and benefits regulations of a country, and b) financial data obtained from a survey of a statistically significant sample of the population. The data is run through the programme to discover a wide variety of statistics; new benefits are then written into the programme (for instance, Basic Incomes at different levels for different age groups), along with changes to existing taxes and benefits, and the programme is run again. The new set of statistics generated can then be compared to those generated by the current tax and benefits system, and conclusions can be drawn about changes in poverty and inequality indices, numbers of households that would be taken off means-tested benefits by the Basic Income scheme, household net disposable income losses, and the net cost of the scheme (Morgan, Reed, & Torry, 2019).

8 Does Microsimulation of an Illustrative Basic Income Scheme Constitute a Basic Income Pilot Project?

Because the data that a microsimulation programme uses is real world data from a statistically significant sample of the country’s population, microsimulation is effectively a pilot project run on the entire population of the country. This means that in order to discover a large number of the financial effects of an illustrative Basic Income scheme, there is no need to run a pilot project: microsimulation can do the job for us; and it can also, by a trial and error method, find an illustrative Basic Income scheme that fits prechosen criteria for a feasible scheme, if such a scheme exists; and it can generate a lot of information about the effects that that scheme would have. So not only can microsimulation function as a Basic Income pilot project, but it can also set the parameters for a Basic Income scheme to be further tested by an on-the-ground pilot project. Only such a two-stage method—microsimulation, followed by an on-the-ground pilot project—will ensure that what is being tested by the on-the-ground pilot project will be a Basic Income scheme that could be rolled out nationwide, so it is the only way to ensure that a genuine pilot project will take place.

However useful microsimulation might be—and it is very useful—there are of course things that it cannot do on its own. For instance, it can calculate marginal deduction rates (MDRs)—that is, the proportion of additional earned income that is withdrawn by income taxes and the withdrawal of means-tested benefits—but it cannot tell us what effect MDR changes would have on employment incentives; and neither can it tell us the ways in which household members might rearrange their employment patterns if they all had Basic Incomes, and how such pattern changes might alter employment incentives. For all of that, a real on-the-ground pilot project is required. Similarly, microsimulation offers only a static ‘morning after’ evaluation of the implementation of a Basic Income scheme. It cannot tell us about the dynamic effects: for instance, if employment patterns were to change, then earned incomes and therefore tax revenues would change, thus changing the Basic Income scheme. Nor can microsimulation tell us about levels of wellbeing or social cohesion, and those are factors in which we might be interested. The Centre for Microsimulation and Policy Analysis at the University of Essex is currently working on an extension of its microsimulation programme that will model employment market and other dynamic effects in order to predict macro-level as well as micro-level effects of policy changes over time. Such an extension will be extremely useful as it will enable us to model some of the longer-term employment and economic effects of illustrative Basic Income schemes: but because labour market models are based on assumptions about how labour markets operate, there will always be a level of doubt as to how accurately such models can predict what would happen if the tested Basic Income scheme were to be implemented. With current static microsimulation, we can be fairly sure that the morning after the implementation of a Basic Income scheme we would see the effects that the microsimulation research tells us that it would have. We cannot be quite so sure that computer modelling can predict longer term labour market and other effects with such accuracy, so on-the-ground pilot projects will still be required, and preferably pilot projects that involve entire communities.

Widerquist suggests that it is important for those running Basic Income experiments to be clear about the questions that the project needs to answer (Widerquist, 2018, p. 1–12, p. 93–98), and also to be clear about the questions that experiments cannot answer or can only answer in a partial fashion (Widerquist, 2018, p. 99–129). We are now clearer about the financial feasibility questions that an experiment in a more developed country would not have to answer, and the questions to which answers might still be required.

9 Do We Need Basic Income Experiments?

We are now in a position to answer the question set by the title of this essay.

In countries with less developed economies, and therefore without existing complex tax and benefits systems, pilot projects that test genuine Basic Income schemes can reveal some of the effects that a nationwide Basic Income scheme might generate, and they can facilitate active public debate about Basic Income: although we must recognise that the limited length of the pilot project will mean that results must be treated carefully, and that detailed reasoning will be required to work out the extent to which the pilot project results might or might not reflect the effects that the scheme tested would deliver if it were to be rolled out nationwide and permanently.

In more developed countries, microsimulation itself functions as a pilot project, and because it can discover an illustrative Basic Income scheme that it would be feasible to roll out nationwide, we can be sure that an on-the-ground pilot project for that scheme would be a genuine Basic Income pilot project. However, there are some things that microsimulation cannot tell us, and to discover the likely effects that a microsimulation programme cannot tell us about would require an on-the-ground pilot project; and although it might not be easy to establish a pilot project in a more developed country, compromise might be possible. In the UK, and presumably in other countries as well, it would be possible to construct a pilot project for a Basic Income scheme that would be sufficiently like a scheme that could be rolled out nationwide that the results from the pilot project would give us useful information about the effects that a national version of the scheme might deliver. Such an experiment would provide useful information and would stimulate public debate about Basic Income.

This leaves two issues to be addressed:

First of all, how useful to the Basic Income debate are experiments that test things other than Basic Income schemes?—for instance, projects that test Minimum Income Guarantee schemes that establish household-based minimum income levels and then pay income-tested benefits to enable household incomes to reach the guaranteed levels, such as the 1970s experiments in Canada more recently researched by Evelyn Forget (Forget, 2011). Generalisation is impossible, but what we might be able to say is that the closer an experiment gets to being a Basic Income pilot project, the more useful the results might be as predictors of what would happen if a Basic Income were to be rolled out nationwide.

Secondly: Precisely what is being requested by the general public needs to be carefully evaluated. If there is public demand for a Basic Income pilot project rather than for a Basic Income, and there are questions in the minds of the general public that a pilot project might be able to answer, then holding a Basic Income pilot project would be fully justified, and the project should be held. But if policymakers are motivated to hold a pilot project because demand for a Basic Income is growing, but they fear how the media might react to a proposal to establish a nationwide Basic Income scheme, then if they are convinced of the desirability and feasibility of Basic Income, it might be better for them to make the argument for establishing a Basic Income and then to do that (Widerquist, 2018, p. 141–44). Under these circumstances the justification for holding a Basic Income pilot project is reduced because the project could be either a distraction from the policy process (because it might mean a nationwide Basic Income scheme missing a political opportunity) or detrimental to it (because the results might be poorly communicated or misinterpreted, whether purposely or not) (Widerquist, 2018, p. 43–59, p. 77–85, p. 145–49).

10 Conclusion

The outcome of our discussion is that in any country a genuine Basic Income pilot project might be useful, at least to some extent, in telling us what some of the effects of a nationwide Basic Income scheme might be, but that if microsimulation were to be available then the questions that would need to be addressed by a pilot project would only be those that could not be addressed by microsimulation. And if an on-the-ground pilot project is being considered, then only a Basic Income scheme that microsimulation has found to be feasible to roll out nationwide should be tested.

We have seen some excellent pilot projects in the developing world. If in a more developed country the factors discussed in this essay were to suggest that a Basic Income pilot project might be justified, then if at all possible it should be carried out. Holding a carefully planned Basic Income pilot project in a more developed country could be of enormous benefit to the global Basic Income debate.