Does Frequency or Amount Matter? An Exploratory Analysis the Perceptions of Four Universal Basic Income Proposals

Abstract

Advocates for a Universal Basic Income (UBI) argue that it would provide citizens with a basic foundation for financial security, boost the economy, alleviate poverty, encourage entrepreneurship, reduce crime, and insulate the employment sector against job losses due to automation. Still, the idea lags in popularity in the United States compared to existing cash policies such as the annual Earned Income Tax Credit and one-time COVID-19 relief packages. We hypothesize that this disparity is related to predicted uses of a UBI in comparison to annual or lump sum cash programs. In this survey of 836 Americans, we explore whether predicted behavioral responses to four randomly assigned hypothetical cash transfer scenarios vary across the domains of amount and frequency. Respondents are more likely to associate monthly payments with work disincentives and lump-sum transfers with debt repayment. Implications for UBI advocates include the need to continue educating the public on the empirical associations between UBI, employment, and expenditures.

1. Introduction

Economic volatility since the 2008 Great Recession and, more recently, the COVID-19 pandemic has renewed American public interest to an old idea: universal basic income (UBI). Andrew Yang, a technology entrepreneur who ran for the Democratic Party’s nomination for presidential candidate in 2020, is credited with raising awareness of UBI as a current policy option, proposing that all Americans receive $1000 monthly. Several cities and some states have recently proposed or enacted basic income pilots. Most modern UBI proposals include three conditions: the payment should be distributed directly to individuals (rather than households), should not be means-tested, and hold no mandatory conditions (i.e., workforce or educational participation) to receive the payment (Johnson and Roberto 2020). Put simply, UBI is “a periodic cash payment unconditionally delivered to all on an individual basis, without means-test or work requirement” (Basic Income Earth Network 2020). However, basic income programs currently being piloted in cities such as Stockton, CA, and St. Paul, MN, are sometimes referred to as “guaranteed income” rather than basic income because they are means-tested to target low-income households. Whether universal or means-tested, basic income programs have no requirements for how the money is spent. That is, they may be conditional on income or not (universal), yet are unrestricted concerning benefit usage.

Though basic income proposals are gaining more consideration and attention at the local and state levels, recent polls indicate that support for a UBI among the American public is split, with roughly half in support and half opposed (Reinhart 2018; Freeland 2019; Gilberstadt 2020). Meanwhile, other existent “lump sum” cash transfer policies in the United States, such as the Earned Income Tax Credit (EITC), the more recent 2021 COVID relief package and enhanced Child Tax Credit, and targeted tax relief to low- and moderate-income families enjoy much higher approval ratings (Data for Progress n.d.; Lake Research Partners 2022; Pew Research Center 2021; Rendleman and Yoder 2019). One factor underlying these differing positions may be the expectations concerning how Americans would use a UBI benefit and the cultural meaning ascribed to these uses. While there is extensive research on how individuals themselves respond to different payment structures and amounts in the context of public benefits such as tax credits (Mendenhall et al. 2012; Smeeding et al. 2000) and social welfare programs such as unemployment insurance (Ganong and Noel 2016), little research exists concerning how individuals perceive that other people will use their payments and, to our knowledge, there is no research concerning the perceived uses of a UBI.

Because voters tend to make decisions based on heuristics or mental shortcuts (Lau and Redlawsk 2001), the ongoing UBI policy debate would benefit from a better understanding of American assumptions about the hypothetical impacts of a UBI. To that end, in this study, we surveyed more than 800 adult Americans regarding their expectations concerning how recipients would respond to various models of unconditional cash transfers (including various work and spending behaviors) and the degree to which these expectations vary based on the generosity and allocation schedule. We also examined whether respondent demographics influenced their perceptions of recipient behavior. Ultimately, we find that the amount of transfer did little to influence behavioral expectations. Still, monthly disbursements were more likely to elicit predictions of work disincentives, while lump sum transfers were more often associated with debt repayment. These findings may help to inform differing support between tax credits, pandemic relief, and UBI policy proposals.

2. Background

Advocates argue that implementinga basic income would provide every citizen with a basic level of financial security, boost the economy, alleviate poverty, encourage entrepreneurship, reduce crime, and help compensate for jobs lost due to technological advances. A Stanford University meta-analysis of 16 other systematic reviews of worldwide UBI pilot test data with thousands of participants, dating from the early 1970s to today, reveals several consistent themes. UBIs, a type of unconditional cash transfer (UCT), decrease poverty, increase consumption, have minimal effects on labor force participation, improve school attendance and achievement of the children of recipients, and improve physical and mental health (Hasdell 2020).

Meanwhile, critics of the existent American welfare state, which comprises a patchwork of conditional and/or means-tested fungible and non-fungible benefits, argue that it disincentivizes work and savings through restrictive income and asset limits and stigmatizes poor and minority women (Hamilton 2020). Further, welfare “sanctions” and lost benefits are associated with comparatively poor school performance, increased child abuse, and decreased family preservation (Kortenkamp et al. 2004; Larson et al. 2011; Slack et al. 2007). Therefore, supplementing or replacing means-tested assistance with UCTs can better support low-income families and improve the well-being of their children. A UBI also has potential policy advantages over more traditional means-tested assistance, as it minimizes administrative costs and increases efficiency since there is no need to determine eligibility (Fouksman and Klein 2019). Finally, advocates argue that UBI would be less vulnerable to future budget cuts, as more universal programs such as Social Security for retired persons and tax credits enjoy broad political support (Kasy 2018).

However, the national UBI discourse also includes several ethical, practical, and political objections. For example, ethicists argue that via the principle of reciprocity, people who receive benefits should make societal contributions if able (Richards and Steiger 2021). Practical objections include the high cost of implementation, efficiency distortions from tax increases, and benefits to non-poor people. Political objections to UBI include the possibility that it would require the elimination of other safety-net programs and would lead to incentives for immigration (Richards and Steiger 2021).

While versions of the concept first appeared in the 16th century, UBI did not truly gain traction in the United States until the civil rights and welfare rights movements of the 1960s. With proponents ranging from Martin Luther King Jr. to conservative economist Milton Friedman, federally funded Negative Income Tax (NIT) pilots launched at multiple sites throughout the United States with thousands of participants (Hamilton 2020). NIT has several similarities to a UBI (there are no work requirements or expectations on how the money is spent), but instead of being dispersed universally, it is a means-tested refundable tax credit that brings low-income households up to a designated income floor. However, because most UBI proposals include a “taxing back” of the benefit for higher-income individuals, the distributional effects of a UBI and NIT are functionally similar (Groot 2004).

In 1972, President Nixon proposed the Family Assistance Plan (FAP). This policyhad some similarities to an NIT, except that it did not cover unemployed households and was to be disbursed at a household level. The plan was ultimately defeated in Congress by forces on both the political left and right. Welfare rights organizers viewed it as insufficient (Withorn 2006) and as an effort to reinforce male dominance over women, as the FAP explicitly sought to support men, particularly Black men, as household heads through job-training programs and wage supplements to support low-wage employment (Quadagno 1990). Southern Congressmen and business elites also opposed it because they believed it threatened the local low-wage labor market that depended upon the labor of economically marginalized Black people, particularly Black women, who were largely excluded from locally controlled public assistance programs (Ventry 2000; Quadagno 1990).

Ultimately, this political impasse led to a compromise that launched the Earned Income Tax Credit (EITC) and transformed the former Aid to the Blind into the modern-day Supplemental Security Income (SSI) program (Social Security Administration 2018), but interest in UBI and NIT fizzled as national politics moved to the right. Therefore, while the EITC is conditional on work, its political history provides important insight into the political challenges of adopting a UBI in the U.S. Currently, the EITC serves as one of the primary anti-poverty mechanisms for families with children in the United States. The credit is delivered via an annual lump sum as part of a taxpayer’s tax refund and research suggests families use their EITC to catch up on bills, save for emergencies, and reduce debt (Halpern-Meekin et al. 2014; Mendenhall et al. 2012; Shaefer et al. 2013). However, while the lump-sum delivery of the refundable credit gives households a sizable income boost, waiting until tax filing to receive it may induce families to accumulate unsecured debt and defer necessary health care in anticipation of their refund (Farrell et al. 2018; Michelmore and Jones 2015; Weber 2016). Furthermore, unlike other stigmatizing means-tested programs such as Temporary Assistance to Needy Families (TANF), Sykes et al. (2015) provide evidence that EITC recipients imbue the EITC with social meaning as a reward for work, an opportunity for mobility and to treat their children as “ordinary kids” and therefore promotes social inclusion and offers access to social rights of citizenship. Therefore, while still conditional on work, the EITC may at least be perceived by recipients as offering some aspirational citizenship rights many believe a UBI may offer.

The defining difference between UBI/NIT, the EITC, and SSI is that the EITC and SSI require a recipient to either work or have a documented inability to work. Widerquist (2005) argues that the media and public misunderstanding of an NIT’s effects on labor force participation were the primary reasons for its defeat in the 1970s. While overall work did slightly decline among recipients of the four NIT pilots, several important nuances were not captured in the public discussion. First, there was no measurement of whether external workforce demand declined during the period under observation. Further, the pilots only included participants with low incomes. Moffitt (1979) estimates that NIT would reduce work among low-income recipients by 4.5% but only 1.6% for mid and high-income earners (presumably because NIT makes low-wage jobs less attractive), an overall effect which might be offset by the other benefits and efficiencies described above. Additionally, the fact that unconditional cash transfers give workers greater bargaining power against poverty-level wages is now seen as a benefit among many basic income proponents (Lowrey 2018a).

The NIT’s effects on workforce participation differed for men and women. While men worked 20–130 h less per year, women (mostly mothers) worked zero to 166 h less per year (Widerquist 2005). One might argue that an NIT giving mothers more choice in whether to stay home with young children or work outside the home is positive. Further, while women contribute significantly more unpaid care to household and national economies (Johnson and Wiener 2016), this labor is rarely counted as “work” in American discourse. This lack of a nuanced conversation around a UBI or NIT’s effects on the workforce soured early supporters of the idea, such as Senator Daniel Patrick Moynihan, who wrote, “But were we wrong about a guaranteed income! Seemingly it is calamitous” (Widerquist 2005, p. 24).

3. Perceptions

Still, growing inequality and economic instability since the Great Recession of 2008 and the COVID-19 pandemic have brought renewed attention to UBI as a legitimate policy alternative, with at least 11 pilots launched in cities across the country in 2021 (Holder 2021). Public opinion surveys conducted in recent years find that roughly half of Americans favor the idea (Reinhart 2018; Freeland 2019; Gilberstadt 2020). However, demographics seem to play an important role in UBI support. Two-thirds (67%) of young people under 30 support a UBI (Gilberstadt 2020). Race also seems to play a factor in UBI favorability, as 45% and 35% of Black and Hispanic respondents strongly favor a UBI, compared to only 16% of White respondents in a survey by the Pew Research Center (Gilberstadt 2020). Further, only 5% of self-described Conservatives strongly support a UBI compared to 38% of Liberals (Gilberstadt 2020).

Meanwhile, existing cash transfer programs such as the EITC and CARES Act economic impact payments enjoy much broader public support. The EITC began as a small tax credit to help offset payroll taxes of low-income parents (Hotz and Scholz 2003), and program eligibility and generosity have expanded under both Democratic and Republican administrations (Mendenhall 2006). Governors implementing state-level EITCs have higher approval ratings and vote shares (Rendleman and Yoder 2019). Similarly, President Biden’s 2021 $1.9 trillion COVID-19 relief package enjoyed a 70% approval rating, although Republicans are still less likely to support it than Democrats (41% vs. 94%) (Pew Research Center 2021). Within this package, most Americans received a $1400 one-time cash payment, phasing out after the first $75,000 in income for individuals and $150,000 for married households. The Child Tax Credit (CTC) was also temporarily expanded to $3000 for school-age children and $3600 for children under six, removing previous earnings requirements and making the credit fully refundable so that even those who are unemployed can receive the full benefit. Half of the child credit was issued monthly between July and December 2021 (Taylor 2021).

The expanded CTC is argued by some to be a form of guaranteed income for families with children (DeParle 2021). It contains no work requirements and begins phasing out at $150,000 in annual income for two-parent families. While the EITC does include a work requirement, it and the 2021 CTC both lack other more intrusive eligibility requirements present in traditional welfare programs such as TANF, such as asset limits, drug testing, paternity establishment rules, and extremely low income limits. They also lack restrictions on how the money is spent, arguably making them more akin to an unconditional cash transfer than in-kind benefits such as food stamps and housing assistance.

Disparities in support for UBI and existent cash transfer programs such as the EITC and CTC may partially stem from long-standing welfare narratives suggesting that generous social policy breeds dependence. Furthermore, these attitudes towards welfare policy often have racist undertones, playing into prejudicial stereotypes (Gilens 2000). Indeed, instead of adopting principles of universality and equity as in some European countries, the American social safety net has been deeply altered by the racist political rhetoric of “welfare queens” and intergenerational dependence (Hamilton 2020). Conservative and neoliberal skepticism of the welfare state reached its zenith in 1996 when the 60-year-old Aid to Families with Dependent Children program was replaced by Temporary Assistance to Needy Families (TANF), ushering in new work requirements and lifetime limits on assistance. Furthermore, critical race theorists argue that the neoliberal establishment of TANF was simply a replication of Reconstruction-Era efforts to control the lives and reproduction of newly freed Black women (Kandaswamy 2021). Many Freedman’s Bureau policies after the American Civil War were concerned with ensuring that Black women worked outside the home, often in domestic service, caring for the children of White families while their own children were cared for by others (Kandaswamy 2021).

Similarly, “welfare queen” rhetoric amidst the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, which created TANF, was concerned with ensuring that recipient mothers return to work as quickly as possible. Kandaswamy pointedly argues that “the implicit message of [TANF] is that impoverished children do not need care as much as they need role models in labor discipline (Kandaswamy 2021, p. 3).” Today, this racialized context and the state-level control afforded by policy devolution means that African Americans are more likely to live in states with harsher eligibility criteria and lower benefits than White Americans (Soss et al. 2008). This has created a patchwork of welfare policies with greater interstate variation than between many European countries (Bruch et al. 2018).

Much evidence suggests that racist, sexist, and classist heuristics such as these play an important role in voter perceptions of policy proposals (Bernhard and Freeder 2020; Kahneman 2013). This is especially true when examining public attitudes towards “welfare” policy. When asked to evaluate policies aimed at low income individuals, respondents often rely on the age-old heuristic of “deservingness”, with those willing to work but unable labeled as “deserving” and those able to but unwilling to work labeled as “undeserving” (Petersen et al. 2012). When asked what they might do with a basic income, Americans state that they would use the money to pay down debt or save for education and homeownership, but when asked what they think others would do, respondents fall back on debunked stereotypes of work disincentives and negative consumption (Evans and Popova 2016; Hamilton et al. 2021). When asked to explain these reactions, qualitative responses include statements such as “people nowadays [sic] will take everything they can get and then some like they are owed” and “Americans can be lazy, so they’ll do anything to reduce time at work” (Hamilton et al. 2021).

Still, little public opinion polling thus far has examined whether the design and expected use of unconditional cash transfers might influence these perceptions. For example, is the amount or frequency influential? While UBI is potentially a more effective and efficient method of remediating economic insecurity, public support lags behind other annual or one-time cash transfer programs. A UBI may be perceived as frivolous if Americans use their benefit on non-essential goods and services and decrease labor force participation, while annual or one-off benefits may be perceived as addressing important financial challenges and goals.

A monthly allocation that represents a consistent cash inflow may be perceived as having a substitution effect for earned income, which may trigger concerns (which can also be racialized) about work disincentives. The expected impact of UBI on labor force participation, especially among economically and racially marginalized communities, may also be an important factor that shapes these differing positions. Indeed, several politicians and lobbying entities helped derail past efforts to establish a guaranteed income through a negative income tax (NIT) by invoking fears of lazy and undeserving individuals, particularly Black women, that will choose not to work and become a drain on society (Ventry 2000; Soss et al. 2011).

Conversely, a lump-sum allocation may be perceived as “extra” money that enhances agency and choice or helps meet family needs for which income is insufficient. A lump-sum allocation also represents deferred consumption and “forced” savings (Barrow and McGranahan 2000), which may allay concerns about social welfare benefits discouraging thrift and personal responsibility. Additionally, these expectations may depend on the generosity of the benefit.

These hypothetical suppositions would be consistent with behavioral economics research, which has found that individuals tend to treat money viewed as windfalls differently than money viewed as regular income (Epley and Gneezy 2007) and are more likely to consume income distributed through regular payments than an equivalent amount of income distributed in a single lump sum (e.g., Shefrin and Thaler 2004). Similarly, Friedman’s (1957) permanent income hypothesis argues that consumer behavior is more influenced by permanent changes in income than one-time infusions. Alternatively, sociological research suggests that individuals imbue different sources of money with social meaning, shaping how they spend and allocate the money (Zelizer 1989, 2017). Indeed, several scholars have illustrated how individuals who receive the EITC regard their tax refund as “family money”, and this guides their decisions to spend the money in ways to support their and their children’s general well-being (Bellisle 2022; Sykes et al. 2015).

To our knowledge, there is little to no existing literature on the perceived uses of unconditional cash transfers. The following exploratory analysis examines whether the perceived uses of unconditional cash transfers differ across various frequencies and amounts and therefore helps to explain its uneven public support. We then seek to learn whether these behavioral expectations differ by demographic characteristics of the respondent to add to previous research findings that respondent’s positionality can influence UBI support. Specifically, we examine age, gender, and employment status as young people, women, and persons with low incomes have been found to have higher favorability towards UBI proposals (Freeland 2019; Gilberstadt 2020).

4. Methods

4.1. Study Design

Amazon’s Prime Panels platform is an increasingly popular recruitment tool in social science research, one which can provide diverse respondent samples and high-quality data (Chandler et al. 2019). Prime Panels was used to recruit adult Americans randomly assigned to one of four hypothetical UBI scenarios. The four scenarios reflected two dimensions of an unconditional cash transfer: amount (a $3000 or $6000 total benefit) and allocation frequency (monthly vs. lump-sum). These amounts were chosen as they are similar to existing cash transfer programs and modern UBI pilots. For example, the 2021 CTC provides $250 per month for children over age 5, and the EITC includes a maximum credit of $6660 for 2020. Meanwhile, UBI pilots such as those in Stockton, CA, and Hudson, NY, provide recipients with $500 per month. Respondents in each condition were then asked how they thought most people would spend these funds. A university Institutional Review Board approved the study in August 2018, as it involved minimal risk to participants.

4.2. Sample

Initially, 877 participants were recruited; however, 41 cases were excluded because participants either did not either (a) provide consent (n = 5), (b) report US citizenship (n = 35), or (c) respond to the full survey (n = 1). The final sample of 836 participants was limited to adults (18+) who were also US citizens. As a randomization check, Chi-squared tests of independence were conducted to ensure that demographic variables did not differ significantly based on the UBI plan condition. As expected, group assignment was not significantly associated with participant age (p = 0.696), gender (p = 0.563), race/ethnicity (p = 0.716), Hispanic/Latino status (p = 0.681), marriage status (p = 0.212), education (p = 0.609), employment status (p = 0.729), or annual household income (p = 0.643). In addition, minimum detectable effect size (MDES) calculations indicated that we were able to detect effect sizes of 0.097, 0.146, and 0.181 with the final sample of N = 836 for chi-square tests with 1, 13, and 39 degrees of freedom, respectively, setting alpha at 0.05 and power at 0.80.

Table 1. Sample demographic characteristics.

	n	%
Age
18–24	142	17.0
25–29	140	16.7
30–39	251	30.0
40–49	154	18.4
50–64	102	12.2
65+	47	5.6
Gender
Male	322	38.5
Female	507	60.6
Non-binary/third gender	6	0.7
Prefer not to say	1	0.1
Race/Ethnicity
White	619	74.0
Black or African American	108	12.9
American Indian or Alaska Native	5	0.6
Asian	45	5.4
Native Hawaiian or Pacific Islander	3	0.4
Other	27	3.2
Multiracial	29	3.5
Hispanic or Latino
Yes	122	14.6
No	714	85.4
Marriage Status
Married	330	39.5
Widowed	22	2.6
Divorced	80	9.6
Separated	22	2.6
Never married	382	45.7
Education
Less than a high school degree	14	1.7
High school degree or equivalent	192	23.0
Some college but no degree	207	24.8
Associate degree	124	14.8
Bachelor’s degree	212	25.4
Graduate degree	87	10.4
Employment
Employed, 1–39 h/week	261	31.2
Employed, 40 h/week or more	264	31.6
Not employed, looking for work	110	13.2
Not employed, not looking for work	103	12.3
Retired	51	6.1
Disabled, not able to work	47	5.6
Annual Household Income
$0–$9999	61	7.3
$10,000–$19,999	76	9.1
$20,000–$29,999	104	12.4
$30,000–$39,999	87	10.4
$40,000–$49,999	80	9.6
$50,000–$59,999	108	12.9
$60,000–$69,999	54	6.5
$70,000–$79,999	57	6.8
$80,000–$89,999	34	4.1
$90,000–$99,999	37	4.4
$100,000 or more	138	16.5
Total	836	100

reports the sample’s demographic characteristics. Survey participants represented a broad age range, with 17.0% college-age (18–24 years), nearly three-quarters of working age (16.7% 25–29, 30.0% 30–39, 18.4% 40–49, and 12.2% 50–64) and 5.6% of retirement age (65+). Approximately 60.6% of participants were female, 38.5% were male, and 0.7% were non-binary or third gender; one participant elected not to disclose their gender. Compared to the U.S. population, participants were younger and more likely to be female (US Census Bureau 2019b). Respondents also represented a range of familial and living arrangements, with 39.5% married, 45.7% never married, 2.6% widowed, and 12.2% either divorced or separated. Of those reporting that they were not married, 29.5% lived with a partner.

The sample was racially and ethnically comparable to the total US population (US Census Bureau 2019a), with 74.0% reporting that they were White, 12.9% Black, 0.6% American Indian or Alaska Native, 5.4% Asian, 0.4% Native Hawaiian or Pacific Islander, and 6.7% either multiracial or some other race. Additionally, 14.6% of participants identified as either Hispanic or Latino. The sample also varied in terms of education, employment, and income. Only 1.7% of the sample had less than a high school diploma, 23.0% had a high school diploma or GED, 24.8% had some college, 14.8% had an Associate’s degree, 25.4% had a Bachelor’s degree, and 10.4% held a graduate degree. Educational attainment was similar to that for the U.S. population, in which 32.1% have a Bachelor’s degree or higher compared to 35.8% in the study sample (US Census Bureau 2019b).

Just under a third (31.2%) of the sample worked part-time, 31.6% worked 40+ hours per week, 13.2% were unemployed and looking for work, 12.3% were unemployed and not seeking work, 6.1% were retired, and 5.6% were unable to work due to a disability. By comparison, in 2019, 70.2% of Americans aged 16+ worked full time (Bureau of Labor Statistics 2019). Finally, 16.4% of respondents had an annual household income (in 2019) of less than $20,000, 32.4% earned $20,000–$49,999, 34.7% earned $50,000-$99,000, and 16.5% earned $100,000 or more. In 2019, the median US household income was $65,712 (Guzman 2020).

4.3. Data Collection

In addition to several demographic questions, participants were randomly assigned to receive one of four questions related to cash transfer benefit usage. The average response time was 2.8 min and participants were reimbursed approximately $1 for their time.

The four randomized questions were as follows:

• If every American received a payment of $500 per month with no strings attached and no matter their situation, how do you think most people would spend this money?
• If every American received a payment of $250 per month with no strings attached and no matter their situation, how do you think most people would spend this money?
• If every American received a one-time payment of $6000 with no strings attached and no matter their situation, how do you think most people would spend this money?
• If every American received a one-time payment of $3000 with no strings attached and no matter their situation, how do you think most people would spend this money?

Use of the phrases “no strings attached” and “no matter their situation” were meant to characterize the benefit as unrestricted and unconditional, respectively. These questions were designed to assess variation in opinions concerning how participants thought recipients might use the benefit relative to the amount and allocation frequency. After a random assignment in Qualtrics, 227 participants responded to Q1 ($500 per month), 198 to Q2 ($250 per month), 200 to Q3 (one-time payment of $6000), and 211 to Q4 (one-time payment of $3000). For each of the questions, respondents were allowed to choose one of the following responses: (A) Quit working or seeking work, (B) Reduce working hours, (C) Continue working as they do now, (D) Put the money in savings, (E) Pay down debt, (F) Apply the money towards education or student loans, (G) Apply the money towards homeownership, (H) Apply the money towards small business development, (I) Apply the money towards regular expenses (housing, groceries, utilities, etc.), (J) Apply the money towards childcare, (K) Apply the money towards healthcare expenses, (L) Apply the money towards a major consumer purchase, such as a vehicle, television, or appliance, (M) Spend it on small luxuries or non-essentials (e.g., eating out, travel, gifts, alcohol, clothes), and (N) Other (please explain). These categories were informed by existent research on expenditures and usages of other cash transfer and UBI programs (West et al. 2021; Halpern-Meekin et al. 2014; Mendenhall et al. 2012; Shaefer et al. 2013). While these categories are not necessarily mutually exclusive, we asked respondents to choose only one to gauge their first reaction to the questions. As discussed above, these “knee-jerk” heuristics are important in voter perceptions and decision-making.

4.4. Data Analysis

Chi-squared tests were used to examine whether differences in the amount and allocation frequency of hypothetical UCTs might influence perceptions of how the general public might use those benefits. Associations between hypothetical benefit and expected uses were tested across all four benefit configurations and between amount ($3000 vs. $6000) and allocation frequency (monthly vs. lump-sum allocation) groups. Finally, z-tests were used to directly compare the proportions of expected benefit uses by hypothetical scenario groups using Bonferroni corrected p-values to adjust for the large number of comparisons made across analyses. The Bonferroni approach is a conservative one allowing us to minimize the chance of Type I errors (VanderWeele and Mathur 2019).

5. Results

Across all hypothetical conditions, the majority of respondents believed that the average person receiving unconditional benefits of any amount or allocation frequency would most likely use benefits on regular expenses (27.5%), paying down debt (22.5%), and purchasing small luxuries or non-essentials (14.7%). Nearly twice as many respondents believed that the average benefit recipient would continue to work as they do now (10.2%) than reduce their working hours (5.1%), and even fewer believed that the average recipient would quit working or seeking work (2.5%).

Chi-squared tests comparing perceived uses of the four types of benefits are reported in

Table 2. Cross-tabulation of perceived outcomes of the adoption of UBI plans by plan type.

	Basic Income Plan
	$250/mth		$500/mth		$3000 Once		$6000 Once		Total
Outcome	n	%	n	%	n	%	n	%	n	%	χ2	p	V
Quit working or seeking work	6	3.0	5	2.2	4	1.9	6	3.0	21	2.5	0.83	0.843	0.031
Reduce working hours	13	6.6	20	8.8 a	7	3.3	3	1.5 b	43	5.1	14.0	0.003	0.129
Continue working as they do now	29	14.6	22	9.7	20	9.5	14	7.0	85	10.2	6.71	0.082	0.090
Put the money in savings	15	7.6	16	7.0	17	8.1	16	8.0	64	7.7	0.20	0.977	0.016
Pay down debt	39	19.7	34	15.0 b	60	28.4 a	55	27.5 a	188	22.5	15.4	0.002	0.136
Apply the money towards education or student loans	2	1.0	2	0.9	3	1.4	3	1.5	10	1.2	0.50	0.920	0.024
Apply the money towards homeownership	1	0.5	2	0.9	1	0.5	3	1.5	7	0.8	1.66	0.645	0.045
Apply the money towards small business development	0	0.0	2	0.9	1	0.5	3	1.5	6	0.7	3.41	0.333	0.064
Apply the money towards regular expenses	53	26.8	63	27.8	63	29.9	51	25.5	230	27.5	1.05	0.789	0.035
Apply the money towards childcare	1	0.5	0	0.0	1	0.5	1	0.5	3	0.4	1.13	0.771	0.037
Apply the money towards healthcare expenses	2	1.0	5	2.2	0	0.0	3	1.5	10	1.2	4.71	0.194	0.075
Apply the money towards a major consumer purchase	1	0.5 b	13	5.7 a	9	4.3	15	7.5 a	38	4.5	12.2	0.007	0.121
Spend it on small luxuries or non-essentials	35	17.7	42	18.5	24	11.4	22	11.0	123	14.7	8.10	0.045	0.098
Other (please explain)	1	0.5	1	0.4	1	0.5	5	2.5	8	1.0	6.61	0.085	0.089
Total	198	100	227	100	211	100	200	100	836	100	69.4	0.002	0.166

Note. Percentages are reported by column. Significant (p < 0.05) differences in outcomes by plan type are in bold type and denoted by subscripted letters “a” (higher values than other plan types) and “b” (lower values than other plan types). Chi-squared tests of independence are reported for each outcome as χ² (3, 836) at the end of each Outcome row and for the total table of 2 × 4 comparisons as χ² (39, 836) at the end of the “Total” row.

. A statistically significant association between plan type and the outcome was observed, X² (39, 836) = 69.4, p = 0.002, though the effect size for this association was small (Cramer’s V = 0.166). Individual comparisons of plan types by outcome revealed statistically significant differences in perceptions that UBI benefits would lead to a reduction in working hours, Χ² (3, 836) = 14.0, p = 0.003, debt repayment, Χ² (3, 836) = 15.4, p = 0.002, major consumer purchases, Χ² (3, 836) = 12.2, p = 0.007, and spending on small luxuries, Χ² (3, 836) = 8.1, p = 0.045. In addition to examining differences across all treatment conditions, we also examined the extent to which individual treatment conditions led to significantly different responses compared to other conditions. This analysis shows that respondents believed that the average person would reduce their working hours more when provided with a $500/month (8.8%) than when provided with a one-time payment of $6000 (1.5%), pay down their debts more when given $3000 (28.4%) or $6000 (27.5%) one-time payments than when given $500/month (15.0%) and spend more on major purchases with $500/month (5.7%) or $6000 (7.5%) one-time payments than the $250/month plan (0.5%). No other statistically significant differences in outcomes by plan types emerged.

Another set of Chi-squared tests were conducted between outcome and plan type, collapsed by frequency of benefits (i.e., monthly vs. one-time payments). These tests are reported in

Table 3. Cross-tabulation of perceived outcomes of the adoption of UBI plans by benefit frequency.

	Benefit Frequency
	Monthly		One-Time		Total
Outcome	n	%	n	%	n	%	χ2	p	V
Quit working or seeking work	11	2.6	10	2.4	21	2.5	0.02	0.886	0.005
Reduce working hours	33	7.8 a	10	2.4 b	43	5.1	12.2	<0.001	0.121
Continue working as they do now	51	12.0	34	8.3	85	1.2	3.18	0.075	0.062
Put the money in savings	31	7.3	33	8.0	64	7.7	0.16	0.689	0.014
Pay down debt	73	17.2 b	115	28.0 a	188	22.5	14.0	<0.001	0.129
Apply the money towards education or student loans	4	0.9	6	1.5	10	1.2	0.48	0.490	0.024
Apply the money towards homeownership	3	0.7	4	1.0	7	0.8	0.18	0.671	0.015
Apply the money towards small business development	2	0.5	4	1.0	6	0.7	0.74	0.389	0.030
Apply the money towards regular expenses	116	27.3	114	27.7	230	27.5	0.02	0.886	0.005
Apply the money towards childcare	1	0.2	2	0.5	3	0.4	0.37	0.543	0.021
Apply the money towards healthcare expenses	7	1.6	3	0.7	10	1.2	1.49	0.223	0.042
Apply the money towards a major consumer purchase	14	3.3	24	5.8	38	4.5	3.12	0.077	0.061
Spend it on small luxuries or non-essentials	77	18.1 a	46	11.2 b	123	14.7	7.99	0.005	0.098
Other (please explain)	2	0.5	6	1.5	8	1.0	2.16	0.142	0.051
Total	425	100	411	100	836	100	40.6	<0.001	0.220

Note. Percentages are reported by column. Significant (p < 0.05) differences in outcomes by plan type are in bold type and denoted by subscripted letters “a” (higher values than other plan types) and “b” (lower values than other plan types). Chi-squared tests of independence are reported for each outcome as χ² (1, 836) at the end of each outcome row and for the total table of 2 × 2 comparisons as χ² (13, 836) at the end of the “Total” row.

, and responses to each outcome by benefit frequency are displayed in Figure 1. The association between benefit frequency (monthly vs. one-time) and outcomes was statistically significant, Χ² (13, 836) = 40.6, p < 0.001, with a small effect size (Cramer’s V = 0.220). Consistent with previous analyses, respondents believed that most people would reduce their working hours more on monthly plans (7.8%) than one-time plans (2.4%), Χ² (1, 836) = 12.2, p < 0.001, and pay down debt more on one-time plans (28.0%) than on monthly plans (17.2%), Χ² (1, 836) = 14.0, p < 0.001. Additionally, a statistically significant difference in smaller purchases (e.g., luxury items, non-essentials) emerged, Χ² (1, 836) = 7.99, p = 0.005, such that respondents thought that people would use monthly benefits (18.1%) more than one-time benefits (11.2%) to make these purchases.

We also examined differences in responses by the yearly amount offered through the plans (i.e., $3000 vs. $6000), but we did not observe any significant differences based on the payment amount. This likely indicates that the bulk of the differences we observe in this study is due to the frequency of the payments rather than the amount.

6. Exploratory Analyses

Since some response options were selected less frequently than others regardless of cash transfer plan, outcomes were collapsed into theoretically similar categories and compared once more with benefit frequency. Specifically, new categories were created for quitting or reducing work hours (outcomes A and B), saving or investing the money (outcomes D, F, G, and H), paying off debt (outcome E), spending on necessities (outcomes I, J, and K), and spending on major or minor consumer purchases (outcomes L and M). If participants had responded in the affirmative to any of these outcomes, they were added to the associated response class.

Once more, significant differences emerged in comparisons of outcome frequency by benefit frequency, Χ² (4, 713) = 24.6, p < 0.001. Compared to one-time plans, monthly plans were significantly associated with perceptions that the average person would use benefits to quit or reduce work (10.4% vs. 4.9%; p = 0.003) and that the average person would spend on major or minor consumer purchases (18.6% vs. 12.7%; p = 0.018); conversely, one-time plans were significantly associated (p < 0.001) with perceptions that the average person would pay off debt with their benefits (28.0%) relative to monthly plans (17.2%). No statistically significant differences were observed by benefit frequency in perceptions that people would save/invest benefits (p = 0.338) or spend benefits on regular/necessary expenses (p = 0.943).

Chi-squared tests were also used to compare these collapsed outcome groups with dichotomized age (under 30 vs. 30 or older), gender (male vs. female), and employment (working part-/full-time vs. not working) variables, regardless of plan assignment. While outcomes did not statistically significantly differ by gender (p = 0.273) or employment (p = 0.083), significant differences in outcomes were observed between participants over/under 30 years of age, Χ² (4, 713) = 30.6, p < 0.001. Compared to respondents older than 30, younger respondents more frequently believed that the average person would use benefits for quitting or reducing work (13.9% vs. 6.4%) and saving or investing (18.0% vs. 9.2%), but not for necessary/regular expenditures (25.0% vs. 38.8%; p’s < 0.05). No other statistically significant differences appeared between these demographic variables and outcomes (p’s > 0.05)

7. Discussion

Overall, we did not discover much variation in how respondents projected that the money would be spent. Across all four questions, a plurality (25–30%) of respondents assumed that recipients would apply the money towards regular expenses (housing, groceries, utilities, etc.), the intensive consumption margin. It is also notable that the second most frequent behavioral expectation was debt reduction. Far smaller proportions of respondents expected a response on the extensive margin of consumption concerning non-essentials and an income effect on labor supply; only 5.1% said recipients would reduce their work hours. An even smaller proportion expected movement on the extensive margin of labor supply, perhaps because the magnitude of the benefit was too low to facilitate labor market exits. Thus, the overall picture that emerges is that respondents mostly expect recipients to use the additional money from a UBI to smooth consumption (inclusive of debt) rather than adjust labor force participation. Clearly, the idea that a UBI might discourage work—a chief concern among many policymakers when considering income support programs—was not a widespread issue among respondents. Instead, they considered how recipients could use a UBI to optimize household resource management.

Two areas that did create significant variations with respect to benefit configuration were the reduction of work hours and repayment of debt. While 8.8% and 6.6% of those predicting the effects of $500 and $250 per month (respectively) believed that recipients would reduce working hours, this was only the case for 1.5% and 3.3% of those responding to questions regarding lump-sum payments ($6000 and $3000, respectively). Thus, the overall expectation of an income effect on labor supply was low, but greater when considering a monthly versus lump-sum benefit. This may be because a monthly benefit makes it easier for individuals to consider reducing work hours based on X fewer hours and X hourly pay relative to a $250 or $500 monthly payment.

This relatively small prediction of labor force reduction adds nuance to similar previous studies. For example, Richards and Steiger (2021) presented a basic income-like scenario and then asked respondents for their agreement with the statement “(T)he number of people not working would increase”, to which 47.8% agreed. Similarly, in an IPSOS poll, 63% of Americans agreed that “Basic income will discourage people from being in or seeking paid employment” (Colledge and Martyn 2017). However, the difference between these findings and ours mostly likely lies in the question-framing. In these previous studies, respondents were asked about overall labor force participation. They believed it would decline but were not asked by how much it would decline. We asked what “most people would do.” Taken together, it is reasonable to surmise that American respondents believe that some people would work less (resulting in an overall decline in labor force participation) but that most people would continue working.

Those considering the effects of lump-sum payments thought that the repayment of debt would be a likely outcome more often than those considering a monthly benefit. A lump-sum allocation could make it easier for individuals to pay down if not eliminate debt—an important consideration given record levels of student debt, that 17% of student loan borrowers are having trouble repaying their loans, and that over a quarter of U.S. households carry credit card balances most or all of the time (Board of Governors 2020). Conversely, expectations concerning using the benefit to put money in savings did not differ with respect to benefit amount or allocation method. Because so many Americans struggle to save—less than half have the recommended amount of emergency savings, and only 58% have retirement accounts (FINRA Investor Education Foundation 2019), any additional income in any form may be expected to make a positive difference in household balance sheets.

While informative, the research design includes several important limitations. First, because we asked about perceived uses of an unconditional cash transfer rather than the strict favorability of the various proposals, inferences regarding favorability cannot be made and require follow-up research. However, because previous research has linked UBI proposals with welfare narratives of work disincentives and dependence (Hamilton et al. 2021), one could reasonably hypothesize that expectations of reduced work hours in the context of monthly benefit proposals were seen as a negative outcome by participants. At the same time, because no more than 8.8% of respondents predicted reduced work in any scenario, it may be that these attitudes are changing. Interestingly, those under 30 were more likely to predict reduced work hours, and previous research has found young people are more supportive of UBI proposals (Gilberstadt 2020), suggesting that the conception of work and UBI among young people requires further study. For example, could Generation Z be less likely to intertwine the concepts of work and deservedness than Baby Boomers? The fact that our sample was younger and more heavily female than the general population both complicates this dynamic and provides fodder for further study. Similarly, we did not ask about the respondents’ political leanings, which could arguably influence their perceptions. Levay et al. (2016) recommend using MTurk’s available statistical controls to ensure demographically representative samples.

Further, we did not ask respondents about annual payment scenarios in addition to lump-sum and monthly payments. In the lump-sum condition, we referred to the UBI allocation as one-time rather than specifying it as a recurring annual payment. In contrast, the monthly allocation condition might have been construed as a recurring payment over an indefinite period. Thus, differently inferred time horizons between the lump-sum and monthly conditions might have influenced responses. We recommend this for future research in this area. Future research might also investigate a wider array of disbursement amounts beyond the four examined here. Finally, because we surveyed American respondents, all of whom operate in the uniquely American capitalist, neoliberal paradigm, we expect our findings to be primarily useful to researchers and advocates in the same context.

Despite these limitations, our results provide insights into the continued disparities in the popularity of UBI proposals and existent cash transfer programs such as the EITC and CTC. Monthly payments (rather than a lump sum) still seem, for at least some respondents, to evoke decades-old imagery of work disincentives and dependence. While we did not ask exactly to what degree work might decrease (for example, a conservative estimate of 1–2% might be in line with some NIT pilot research), most UBI pilots find this notion to be inaccurate. Indeed, the recent Stockton, CA UBI pilot found a 40% increase in full-time employment among participants, compared to only 5% in the control group (West et al. 2021). We did not ask respondents whether they might predict an increase in employment, which might pose a limitation. However, there is no research to our knowledge suggesting that this perception would predict a political roadblock. Interestingly, the Stockon pilot does indicate that respondents in the current survey were correct about predicted expenditures on regular expenses, as 80% of participant funds were spent on food (37% alone), merchandise, utilities, and auto expenses. Less than 1% was spent on alcohol and tobacco (West et al. 2021).

8. Conclusions

What is interesting about our findings here in the context of existent cash transfer policies such as the EITC and 2021 CTC expansion is that the split in behavioral expectations echoes the experiences of UBI advocates of the 1960s and 1970s. While proposals for an NIT were unsuccessful in the 1970s, the resulting programs were those that required work (EITC) or a documented inability to work (SSI). Similarly, in our study, the more basic income-like proposals (monthly benefits) were significantly more likely to evoke expected work disincentives than lump-sum payments even when the amounts were equivalent. Indeed, popular modern-day national programs either require work (EITC, again) or the inability to work (SSI), and Congress failed to extend the expanded CTC arguably due to opposition from politicians who will only support cash support programs conditional on work (Spotlight on Poverty and Opportunity 2022; Rubio 2021). Yet, recent research on the enhanced CTC suggests the policy had negligible impacts on employment (Ananat et al. 2021; Roll et al. 2022). Instead, a growing line of research suggests that ongoing income support throughout the year is related to various positive outcomes, including reduced income poverty, food insecurity, and financial and mental stress (Andrade et al. 2019; Bellisle et al. 2021; Hamilton et al. 2022; Kramer et al. 2019).

Our results might be useful for basic income advocates in a few ways. First, there is a hypothetical argument that building a UBI into the tax code rather than as a stand-alone program could increase its favorability since progressive tax policies such as the EITC garner relatively bipartisan support (Lowrey 2018b). Indeed, this appears to be the logic behind President Biden’s expansion of the existing Child Tax Credit rather than other less-popular programs such as TANF. Several scholars, advocates, and members of Congress have proposed transforming existing tax-based programs such as the EITC with a more UBI-like delivery, including one-time advances, quarterly periodic payments, and monthly payments (Holt et al. 2020; Halpern-Meekin et al. 2018; Vallas et al. 2014).

Secondly, an incrementalist policy advocate might acknowledge that the notions of work and deservedness date to the Elizabethan Poor Laws of 1601 (Jansson 2014) and are unlikely to be unraveled in the short policy window created by the COVID-19 pandemic and Democratically-controlled Executive and Congressional branches. The more pragmatic approach might be to advance basic income for children and persons with disabilities as an incomplete but definitive step forward. In addition, as Widerquist (2005) argues that a “failure to communicate” the 1970s NIT pilots’ effects on labor was a primary factor in the movement’s failure, there appears significant potential for a similar scenario today. Advocates hoping to advance a basic income agenda must aggressively educate the general public about the true relationship between work and UBI.