Free Groceries for All: A Realistic Path Towards Universal Basic Income

1 Economic Insecurity

It is well-known that economic insecurity is rampant and devastating in most parts of the world. Among the rich countries, this is especially true of the Anglo-American nations. For instance, in the US, 770,000 people experience homelessness every night (HUD 2024); 21 million people spend more than 30 % of their income on rent (USCensusBureau 2024); 40 % of households are only three paycheques away from falling below the poverty line (Wiedrich and Newville 2019); and 35 million people are food insecure (Gundersen 2021). Across the 62 urban areas of Canada, a minimum wage worker, working full-time, cannot afford a one-bedroom apartment in 53 of the 62 areas (Tranjan 2024). In the UK, 74 % of parents find it difficult to meet childcare costs (2024) and 10 % are food insecure (GovUK 2024).

A UBI is one important approach (though not the only possibility) for improving economic security. Pilot projects have typically shown robust improvements in economic security, education, and health outcomes (Bastagli et al. 2016; Gibson et al. 2018). Likewise, models of full-blown projects predict substantive improvements in poverty and inequality (Clark 2013; Widerquist 2017). For instance, Widerquist shows that a UBI scheme set at the official poverty rate would reduce poverty from 13.5 % of the population to approximately 0 %, lifting 43.1 million people (including 14.5 million children) out of poverty (at a cost of $539 billion, or about 2.95 % of GDP) (2017, p. 9).

Nevertheless, the hard fact is that UBI remains quite unpopular across the OECD. In the US, support for UBI ranges from about 38 % to 45 % of the population (Gilberstadt 2020; Ipsos 2017). In Canada, approximately 44 % support (Ipsos 2017). In the UK, polls have found support ranging from 41 % to 54 % (Baranowski and Jabkowski 2021; Laenen 2023, p. 62). Across Europe, the most rigorous survey to date is probably round 8 of the European Social Survey (ESS), held in 2016–2017, which polled 44,000 respondents from 23 different countries. Roughly 55 % of those interviewed by the ESS supported the introduction of a basic income. The highest levels of support – above 65 % – were observed in Central and Eastern Europe (Lithuania, Russia, Hungary and Slovenia), and Israel. The lowest levels of support – below 45 % – were found in Switzerland, Sweden and Norway (Baranowski and Jabkowski 2021).

The most extensive examination of the polling evidence to date comes from Tijs Laenen’s book, The Popularity of Basic Income: Evidence from the Polls (2023). He points out that these numbers likely overestimate the real level of support for UBI, since asking a person if they “support” something in the abstract doesn’t necessarily mean that, when push comes to shove, they would actually vote for it (This is the problem of so-called “cheap support” [De Wispelaere 2015]). Laenen (2023) shows that when respondents are not just presented with an abstract question about whether they support UBI in general, but are also presented with arguments for and against UBI, their support typically drops.^[1]

Perhaps the most compelling reason to suspect that the survey support numbers are inflated comes from the only country that has actually held a concrete vote on UBI. On June 5, 2016, the Swiss had a referendum on whether to adopt a UBI (loosely defined as a scheme that would provide a monthly cash payment of around €2,330 per month, to all Swiss adult residents, without any means test or work requirement). The proposal was overwhelmingly rejected – 23 % voted “yes,” 77 % voted “no” (Laenen 2023, p. 50). It’s instructive to compare this result with the ESS survey, where 36 % of Swiss respondents said that they theoretically “supported” UBI; notice that when it actually came down to it, 13 % fewer actually did (Baranowski and Jabkowski 2021). Of course, the Swiss referendum is only one instance; nevertheless, to date it is the most direct measurement of genuine, concrete political support for UBI that exists (as opposed to the abstract “cheap support” elicited in surveys).

Why the ambivalence towards UBI? The survey data consistently show that large swathes of the population are concerned with one or more of the following four main objections.

1. The most prominent objection is that a UBI will disincentivize work, leading to “lazy” people watching TV or surfing while everyone else still has to work. Surveys typically find this to be the most prominent worry (Laenen 2023). Concerns of this sort are shared by 63 % of Americans, 55 % of British people, and 54 % of Canadians (Ipsos 2017).

2. A second objection is that UBI is ethically problematic because it constitutes “something for nothing,” and thereby violates an ethic of reciprocity.

3. A third objection is that UBI will be spent on drugs or alcohol or other “temptation goods.” For instance, a recent survey of Americans predicted that if homeless people were to be given a UBI, their spending on drugs and alcohol would be 81 % higher than for non-homeless people (Dwyer et al. 2023).

4. A final objection is that UBI is too expensive – requiring a significant increase in taxes. For example, on average, 54 % of Canadians (up to 66 % in Alberta) say that a UBI would be “too expensive” (2020).

Over the last 30 years, academic philosophers and sociologists have developed detailed responses to all of these objections.^[2] Nevertheless, the populations of the rich countries remain highly ambivalent. For those of us who are supportive of UBI, it is uncomfortable but undeniable that no country anywhere in the world has implemented UBI, only one country has had a national referendum on its introduction (which was roundly rejected), and almost no major political party in any country officially supports a full-blown UBI in their platform.^[3]

So what can be done? The long-term approach is to try to convince people that the merits of UBI really do outweigh the disadvantages. This is important work, but it is slow. The short-term approach is to take the preferences of the population as fixed, and then try to design a UBI-type program that can avoid (or at least significantly dampen) these four objections so that it could receive wide support, and actually get implemented in the near term by a mainstream social-democratic government (e.g., the Democratic Party in the US, the Labour Party in the UK, the New Democratic Party in Canada, the Socialist Party in France or Spain, etc.).

2 A Proposal for Free Groceries for All (FGA)

Here is the proposal: All citizens^[4] should be provided with a relatively small amount of money unconditionally, say $50 per month ($600 per year) for adults and $25 per month ($300 per year) for children. The money will be automatically transferred to a Grocery Electronic Card registered to each adult or parent. The cards may only be used in registered establishments for the purchase of groceries (any other shopping items including restaurant food are prohibited).

The total cost of the FGA program in the US would be approximately $177 billion, or 0.6 % of GDP. In Canada it would cost CAD$24 billion (1 % of GDP), and in the UK it would cost £24 billion (0.9 % of GDP).^[5]

The main strength of the FGA proposal is its achievability, since it is specifically designed to mitigate the main objections to UBI.

In terms of concern (i), about disincentivizing work, the proposal avoids this problem because the amounts at stake are so modest. No conservative politician can claim with a straight face that $600 per year is sufficient for anyone to quit their job and sit on the beach all day. It is possible that an additional $600, on top of pre-existing benefits, may encourage a small fraction of people on the margin to leave their low-paying job, but even if one thinks this is a bad thing, the numbers are sure to be quite small. The key point is that it’s hard to whip up moral outrage about $600, and even harder to do so for free food, especially for children.

In terms of concern (ii), about people receiving “something for nothing,” survey data suggests that most people have quite different moral intuitions about “free food” than they do about “free cash.” The Populus (2018) survey on UBI is the closest example that I am aware of directly asking people to state their preferences for free money versus free food. The survey asked 2,070 British adults their opinion on the following statement: “Rather than cash, the state should provide citizens with basic food supplies and social housing to meet their needs”. In response, 43 % agreed, 27 % disagreed (20 % were neutral, and 9 % didn’t know) (Populus 2018, p. 5).

Additional evidence comes from the fact that although many Americans are opposed to “welfare,” they tend to be significantly more supportive of the Supplemental Nutrition Assistance Program (SNAP, colloquially known as “food stamps”) which provides poor people with money to buy food (and only food). A Pew study found that significant numbers of people (44 % of the population) believe “government aid to the poor does more harm than good by making people too dependent on government” (Pew 2014), whereas a survey from the John Hopkins Center for a Liveable Future found that 61 % of people opposed reducing funding for SNAP (HopkinsCenter 2018, p. 14). Likewise, a Rasmussen (2019) poll finds that 54 % of American adults believe that too many people are dependent on the government for financial aid, whereas far fewer are critical of food stamps – only 36 % say they’re too easy to get. Indeed, when Americans are presented with the objective facts of how much money for groceries SNAP recipients truly receive (an average of about $5.70 per family member per day), 66 % of registered voters believe that it should be increased, and only 4 % believe it should be decreased. Strikingly, 53 % of Republicans also believe that SNAP benefits should be increased (SavetheChildren 2023) – which speaks to the broad bipartisan appeal of free food for the poor, even in a highly polarized political climate.

This widely shared sentiment of support for food security is common in most places. For another example, in Canada, a recent survey of 2,002 people found that 85 % agreed that the government should ensure that no child in Canada goes hungry, and 82 % agreed with the statement that “people going hungry in Canada goes against our values” (MLCentre 2024).

Taken together, this evidence suggests that many people have significantly different moral intuitions about the government providing free cash versus free food. Whereas free cash is widely perceived as an undeserved luxury, free food is not; it is more often seen as a basic human right. Providing free cash strikes many as morally questionable, whereas providing people with food seems like basic human decency.

By design, the FGA proposal directly addresses concern (iii) about spending on drugs and alcohol by explicitly banning it. The money can be spent on groceries and nothing else.^[6]

In terms of the final concern – (iv) about the cost – we saw above that the cost for the three Anglo countries ranges from 0.6 % to 1 % of GDP. These are significant costs, it’s true, but they are not extreme. To put these costs in perspective, the $177 billion cost for FGA in the US is similar to the amount currently spent on Veterans Benefits or on Federal Student Aid. It is about 15 % of what the government currently spends on the military (Cernadas 2023).

A program of this size could be paid for by raising taxes on the total income of the richest 10 % of Americans by approximately 2.86 %. Or, alternatively, by instituting a wealth tax on only the richest 1 % at a rate of 0.41 %.^[7] Rich families would thereby pay more in taxes than they receive from FGA, making them net contributors to the program. Hence the program would be overall redistributive from the rich to the poor.

3 Further Advantages

An important strength of the FGA policy is its universal nature. Every single citizen would receive it. This means that truly huge numbers of people would receive the benefit – roughly 330 million Americans, 38 million Canadians, or 68 million British citizens. These gigantic numbers mean that once instituted, FGA would quickly become normal and familiar to everyone. And as soon as it becomes normal, it will also likely become normative: viewed as good and natural, with the result that any attempt to remove it will be widely resisted as an unwelcome threat. Of course, we should not overstate the case – $600 will not radically change anyone’s life. Nevertheless, millions and millions of people will appreciate the help. Life will become a bit easier and a little less scary.

The underlying rationale for setting the grant so low is that it can be very difficult to get universal programs established. But once they are established, they tend to be sticky. This social phenomenon was first discussed by Titmuss (1975) and Skocpol (1995). But it has recently been given a major elaboration by the brilliant work of David Calnitsky (forthcoming). Using a wealth of data from the last 100 years, Calnitsky demonstrates that across the OECD there is a strong tendency for universal social programs to move in one direction: upwards (he names this phenomenon the “Stickiness of Universal Programs” [SUP]). One striking example is the case of pensions. Its stickiness is such that even 40 years of neoliberalism in America has not succeeded in markedly reducing pension generosity (Calnitsky 2022). It is not exactly clear what drives the SUP tendency. Calnitsky argues that it is mainly due to risk aversion (the well-known fact that, psychologically speaking, most people dislike losses more than they value gains [Tversky and Kahneman 1991]), plus the power of the democratic masses to maintain the benefits they currently enjoy. Whatever the explanation, there is little doubt that this social phenomenon exists and is powerful.

Taking the SUP seriously implies that the most sensible strategy for achieving a full-blown UBI is to start small. Getting a universal policy on the books in the first place is the really hard part. But if this can be accomplished, there are good reasons for thinking that, over time, the populace will come to enjoy the benefit, take it for granted, push for it to be expanded, and refuse to tolerate its reduction.

A final strength of the FGA proposal is that there already exists a working model in the SNAP program, so we can be quite confident about its institutional viability. Originally set up in 1974, SNAP is now a well-established program. By 2022, SNAP benefits were being claimed by 41.2 million people in 21.6 million households. The national average benefit per month was $181.72 per person and $343.00 per household. SNAP works by providing registrants with an Electronic Benefits Transfer (EBT) card. Most groceries are eligible, but alcohol, tobacco, gasoline, nutritional supplements, medicines, pet food, household goods like paper towels or cleaning supplies, and foods that are sold hot are not. Grocery stores must register with the government before they can accept EBT payment, but it is quick and easy to do so (the process is free and online), and almost all are.^[8]

Compared to the FGA proposal, SNAP is more generous but also much narrower in terms of who gets it (only the poorest 10–15 % of the population qualify, as opposed to the 100 % proposed by FGA). In 2022, SNAP cost roughly $119 billion (Desilver 2023). SNAP has operated for decades and remains quite popular. As we saw above, it is consistently more popular than general cash “welfare” programs, and is supported even by a majority of Republicans (recall that even though 65 % of Republicans say they’re opposed to welfare because it does more harm than good, 53 % actually want SNAP to be increased [Pew 2014; SavetheChildren 2023]). Its popularity is likely due to the fact that providing food to poor people (as long as there’s no worry about money being spent on drugs or alcohol) is widely seen as acceptable and even commendable.

The evidence shows that SNAP is quite successful at reducing food insecurity, and consequently leads to improved health outcomes (Bitler and Seifoddini 2019; Gundersen and Ziliak 2015). For instance, Gundersen and Ziliak find that “food-insecure children are at least twice as likely to report being in fair or poor health and at least 1.4 times more likely to have asthma, compared to food-secure children” (2015, p. 1830).^[9]

All-in-all, SNAP provides good evidence that a program like FGA is institutionally workable. We know how to provide people with electronic cards; we know how to register grocery stores; we know that when people get extra money for food, it helps their lives and improves their health.

As it presently exists, a major problem with SNAP is that it remains stigmatized. Celhay and colleagues (2025) find that massive numbers of SNAP recipients (up to 25 %) hide the fact that they receive it. The authors examine a range of possible explanations for this fact and conclude that “these results provide robust evidence that welfare participation is associated with stigma” (p. 3).

The FGA program would likely do significantly better on this score because it would be universal (received by 100 % of the population, not 10–15 %). Indeed, it is hard to imagine that any policy would long be seen as stigmatizing when it’s received by literally everyone. After all, this means that it’s no longer a small group of “them” who are receiving the benefit. It’s everyone; recipients are no longer “them,” but “us.” Not only would FGA be received by 330 million Americans, but this includes tens of millions of white men. Given their dominant social status in the US, it is hard to believe that a program in which every prototypical American middle-class white man participated would have any stigma attached to it. It is far more likely that it would quickly come to be seen as normal and natural and “American.” (Analogously, recall the evolution of pensions from beings seen as a socialist abomination in the early 20th century when they were rare, to a source of patriotic pride now that almost everyone has them). In their study, Celhay et al. (2025) find, as we would suspect, that social stigma drops when the level of participation of one’s peers increases (a 10-percentage point increase in local participation [in SNAP] leads to a 0.9-percentage point decline in the conditional probability of misreporting). This suggests that a universal program like FGA would be far less stigmatizing than SNAP currently is.

The FGA proposal has been designed with Anglo countries in mind. However, in the US case, the proposal would clearly need to be modified because SNAP already exists and operates quite differently in that it is more generous but serves far fewer people. So what should be done in the specifically American context?

I agree with Gundersen (2021) that the goal here should be to slowly expand SNAP into a kind of UBI. Gundersen proposed two models for this. One is a straightforward UBI model where every single household, 100 %, would receive the maximum SNAP benefit, irrespective of income. This would reduce food insecurity in the US by 88.8 % (presuming the costs are paid for by richer people), but comes with a hefty price tag of $730 billion.^[10]

Gundersen also examines a second proposal whereby benefits are increased from current SNAP levels by roughly 25 % and eligibility is also expanded, not universally, but to everyone below 400 % of the poverty line, which is roughly 58 % of the adult population (2023). He estimates that this second policy would reduce food insecurity by 98.2 % at a cost of $565 billion.

These policies are ethically attractive. The difficulty, of course, is that expanding SNAP in either of these ways would result in the total cost increasingly hugely, by 5–6 times its current size. In terms of achievability, the major problem is that since SNAP currently benefits only a relatively small number of relatively powerless people, it is unlikely to have a particularly powerful coalition pushing for the program to be expanded. Hence, while Gundersen’s proposals are commendable goals, it seems doubtful that they will be achieved anytime soon.

All to say that the US faces unique obstacles to the implementation of FGA because of the prior existence of SNAP. In other parts of the world, the path forward seems clearer. Even if FGA starts out as a less generous grant than SNAP (and hence more affordable for the government), the fact that it would benefit a much larger constituency (including not just poor people but the much-more-powerful middle and some of the upper classes), means that there would naturally be a much larger and more powerful political coalition interested in its expansion and resistant to its retraction.

4 Conclusions

Economic insecurity is a major problem of the contemporary world, even in the rich countries. UBI has been widely advocated as a potentially powerful antidote. However, it remains relatively unpopular because much of public believes that it disincentivizes work, immorally provides something for nothing, will be spent on drugs and alcohol, and/or is too expensive.

Rather than developing another academic response to these worries, this paper attempts to design a UBI-type policy that could be highly popular right now, and therefore realistically implemented in the short-term. The policy of Free Groceries for All would entitle all citizens to a small amount of money every month via an electronic card, that could be used for purchasing food and only food. The evidence from the survey data suggests that this policy would be significantly more appealing to the public at large than a conventional UBI, since it is able to substantially mitigate the major objections.

A policy of $50 per month is, of course, insufficient to rectify the scale of the insecurity problems. But the growing social scientific evidence regarding the tendency of existing universal social programs to ratchet upwards implies that it is strategic to start small in order to get such a program on the books. Once millions of people are tangibly enjoying the benefit, it is much more likely to expand over time. In this way, Free Groceries for All may well serve as a practical stepping-stone towards the achievement of a more robust and radical UBI.