Mitigating loss of health insurance and means tested benefits in an unconditional cash transfer experiment: Implementation lessons from Stockton's guaranteed income pilot.

•Cash transfers, universal basic income, and guaranteed income have re-emerged as potential solutions to income volatility.•Methods used in Stockton's guaranteed income experiment, are testing how GI can exist alongside existing safety net benefits.•A multi-tiered approach to mitigating potential means tested benefits loss is both effective and ethically sound.

Introduction

As the novel coronavirus threatens the health of Americans and the US economy, cash transfers have received broad bipartisan support in both the House and Senate as one method to address the severe financial distress brought on by the pandemic (Grisales, Snell, & Davis, 2020). In March 2020, members of the Republican and Democratic parties introduced cash transfers as part of a coronavirus stimulus package, illustrating that the concept of cash transfers and guaranteed income (GI) programs have moved out of the research space and into the policy conversation (Grunwald, 2020). Nonetheless, to date, neither the proposed stimulus package nor other GI policy proposals discuss how these influxes of cash will or will not interact with means tested benefits and income thresholds, leaving some households at risk of preserving their financial stability at the expense of their access to health insurance and other benefits (Stockton Economic Empowerment Demonstration, 2019) (Stockton Economyic ic). In the midst of an unprecedented crisis and a surge of GI experiments, the field of public health and allied disciplines are ideally situated to lead the policy community in developing models that assist policy makers in ensuring that cash transfer interventions and GI experiments preserve benefits rather than produce new forms of inequality. This paper uses the ethical framework and loss mitigation strategies employed by the Stockton Economic Empowerment Demonstration (SEED) as a pilot and policy proof of concept to demonstrate how the new wave of cash transfer policies and GI experiments can preserve existing health insurance and other benefits.

Background

The concept of combating poverty by providing Americans with a cash transfer or GI based on economic need first entered policy debates in the 1960s (Steensland, 2008). Recent research demonstrating the efficacy of cash transfers in improving health equity outcomes—from reductions in hospitalization rates (Forget, 2011), obesity (Lebihan & Takongmo, 2019), psychological distress (Baird, de Hoop, & Özler, 2013), incidences of psychiatric disorders (Costello, Erkanli, Copeland, & Angold, 2010) and depression (Powell-Jackson et al., 2016; Yoshikawa, Aber, & Beardslee, 2012), to increases in life span (Aizer, Eli, Ferrie, & Lleras-Muney, 2016)—has renewed interest in cash transfers and GI as a means of mitigating the impacts of financial insecurity and income volatility. Given the rise of the gig economy, unemployment risks posed by automation, and the fluctuating nature of waged labor, income volatility—an annual income fluctuation of 25% or more already impacting nearly half of the US population (Pew Charitable Trusts, 2017)—is likely to continue influencing health outcomes, highlighting the inextricable link between public health and economic stability (Standing, 2011). In a call to the public health community, Sanjay Basu (2017) identified rising income volatility as a leading preventable threat to public health. GI may be one component of a larger economic justice structure that could smooth this volatility, and the onus for implementation of ethical and just GI models rests on public health and allied disciplines.

Philosophical and ethical debate about GI and its interaction with the safety net dates back to the original negative income experiments of the 1960s (Harris, 2017). Throughout 2019 and early 2020, Andrew Yang popularized a “freedom dividend”, providing every American citizen with a $1000 cash transfer in his bid for the Democratic presidential nomination that would make existing welfare and social program beneficiaries choose between the cash transfer and benefits (Friends of Andrew Yang, 2019), leading to a renewed public debate on whether Yang's proposal could lead to more, rather than less, inequality by depriving low-income Americans of targeted support (Jarvis, 2019). While most GI experiments operate within the existing benefits structure, little is known about how cash transfers may interact with the social safety net in the US. As demonstrated through the debate around Yang's proposal and concerns around the coronavirus stimulus package (Okun, 2020), it is difficult to anticipate the impact and related ethical concerns of unconditional cash transfer programs on benefit recipients.

First, depending on recipients’ circumstances, GI may contribute to the “benefits cliff,” where a small increase in earnings can lead to a reduction in public benefits, such as health insurance, food assistance, and housing support (Dinan et al., 2007). However, there is no clear way to understand this risk as social safety net programs vary significantly on eligibility requirements, income calculations, and corresponding benefits. While some are open-ended entitlements available to all who qualify, others are limited, in part, to program funding or location (US Government Accountability Office, 2017). As a result, decision-making among low-income welfare recipients involves weighing short-term and long-term costs and benefits within these constraints and the availability of other formal and informal resources (Edin & Lein, 1997; Roll & East, 2014; Zippay, 2002). In particular, social supports provided by informal networks, through both direct (e.g., financial contributions) and indirect (e.g., child care) mechanisms, are used by low-income families to “help alleviate the stress of the cliff effect, with tangible resources and reduced isolation” (Roll & East, 2014, p. 185).

In addition, eligibility and income calculations, divergent as they may be, are interpreted and implemented by “street-level bureaucrats,” who serve as front-line workers in direct service positions within agencies, lower courts, educational settings, and other government entities (Lipsky, 1980). Since street-level bureaucrats are those operationalizing policy on a micro-level, they are typically the first ones to interpret and apply emerging policy during the implementation phase of a new program or intervention (Tummers, Bekkers, Vink, & Musheno, 2015). The legal and ethical gap created when introducing new policies or interventions sharply departing from legacy bureaucratic structures can generate tension and uneven programming during implementation, particularly when workers feel caught between competing auspices and codes of ethics (Keiser, 2010).

In means tested benefit programs, all clients ought to receive equivalent treatment, yet they are also individuals with unique needs at the mercy of workers interpreting complex rules (Lipsky, 1980). In an implementation study of public programs, Roberta Rehner Iverson (2000) found that “invisible barriers” to implementation, such as faulty policy logic and personnel incompetence might prompt implementation delays and programmatic strains that harm recipients while undermining the efficacy of programming. Often absent from research attention, these barriers can generate cascading effects on program outcomes and produce, rather than eliminate, inequalities (Rehner Iversen, 2000).

Policy landscape

Given the recent proliferation of GI pilots in the US (Jain Family Institute, 2019), including recent announcements of pilots in Chicago (Chicago Resilient Families Task Force, 2019), Jackson (Economic Security Project, 2019), Philadelphia (Blumgart, 2020), and Newark (Crowe, 2019), it is essential to bring awareness to what, at this point, is essentially a policy and implementation gray space carrying competing sets of mandates and ethical dilemmas. At present, the public discourse around GI reflects three options: (1) the federal coronavirus cash transfer program which would provide American tax filers earning less than $75,000 ($150,000 joint return) a single payment of $1200 per individual ($2400 joint return) and $500 per qualifying dependent, but does not specify how or if this transfer would impact eligibility for existing means tested benefits (Jamerson, Duehren, & Andrews, 2020); (2) proposals which seek to replace existing means tested benefits with a basic GI for all Americans; and (3) the Stockton Plan, described here, which seeks to provide GI alongside existing means tested benefits.

Option two, espoused by Yang, is predicated on Milton Friedman's ideal GI structure, whereby cash transfers are designed to completely replace all means tested benefits under the auspices of improved government efficiency and enhanced work incentives for recipients (Moffitt, 2003). Friedman's proposal was intended to be tested in the Seattle-Denver Income Maintenance Experiment (SIME/DIME), one of the largest negative income experiments of the 1970s. Members of the experimental group already receiving cash welfare were required to choose monthly whether to receive cash benefits from Aid to Families with Dependent Children (AFDC) or the negative income tax. The intended design testing SIME/DIME against existing safety net benefits was upended, as treatment group members receiving AFDC would not sign over to receive the negative income tax without protection of existing benefits, including health insurance, housing vouchers, and subsidized childcare (Christphersen, 1983). Thus, a GI option that fails to protect against the detrimental health and financial impacts of benefits loss in the current policy landscape is demonstrably infeasible.

Herein, we present option three, the Stockton Plan, which draws on an ethical framework informed by public health and social work codes of ethics, and research informed by Institutional Review Boards (IRBs). This makes SEED the first contemporary pilot to bring the ethical concerns around loss of benefits to the forefront of the GI policy debate, and one that can guide ethical and just implementation of GI. Rather than relying on economic projections, GI pilots are taking place in real time and carry real world impacts for recipients. Thus, there is a clear immediacy for public health and social work professionals to uphold the ethics of the respective professions while producing sound empirical evidence capable of influencing public policy. Moreover, while some developing GI pilots may not be governed by human subject research protections mandated by IRBs, similar ethical measures should be considered to protect GI recipients within the existing policy landscape. This paper demonstrates how SEED used ethical mandates of both public health and social work to attend to ethical conflicts inherent in GI experimentation. We begin by describing the pilot design itself as a rubric and exemplar for discussing how implementation and design decisions such as sampling, recipient onboarding, and democratizing human contact effectively preserved recipient benefits in the existing policy landscape.

The Stockton plan: recruitment and intervention

To understand how GI could exist alongside, and interact with, an existing means tested welfare system, SEED captured a diverse sample of Stockton, California residents recruited from census tracts at or below the City's Area Median Income (AMI) of $46,033 (n = 42). We selected a proportion of the population represented in each census tract at random (n = 4200 households) to receive a SEED invitation mailer. After removing duplicate or invalid responses from the mailed invitation (N = 27), the final sample at recruitment baseline was 475. Respondents meeting limited inclusion criteria (i.e., at least 18 years of age with a current Stockton mailing address) were randomly assigned to one of three groups: treatment (n = 125), active control (i.e., participated in all data collection activities; n = 200), or passive control (i.e., did not participate in data collection activities but consented to use of administrative data; n = 150). The treatment group includes five additional participants to account for medical attrition due to a pre-existing condition or a catastrophic medical event that would preclude ongoing participation in research activities. In February 2019, the treatment group began receiving the intervention, an unconditional cash transfer of $500 per month for 18 months on a pre-paid debit card, without work requirements or spending restrictions. All research methodology was approved by the University of Tennessee, Knoxville IRB.

Ethical framework

The “benefits cliff” is a structural component of many existing social safety net programs that effectively causes resource stagnation or reduction as participants increase income (Dinan et al., 2007). Thus, a critical ethical and practical consideration of implementing GI alongside existing means tested benefits is the risk of existing benefit reduction or disqualification. Because the purpose of SEED is testing GI alongside existing safety net benefits, not in place of them, we implemented several loss mitigation steps including a sampling frame without a household-level designee, onboarding of potential recipients that included benefits counseling as a part of informed consent, and intentional human contact between SEED staff and recipients as they navigate the existing benefits structure.

The ethical underpinning for SEED rests on principles within the public health and social work codes of ethics. Specifically, the principles of the ethical practice of public health promote the right and resources to health, recognize both social interdependence and individual rights, and encourage collaborations that build trust and bolster effectiveness (Public Health Leadership Society, 2002). Similarly, the ethical principles of social work bid social workers to act honestly and responsibly and promote ethical practices. This includes valuing the dignity and worth of every person by promoting socially responsible self-determination, as well as recognizing the importance of human relationships to enhance both community and individual well-being (National Association of Social Workers, 1994). The following loss mitigation steps detail implementation and research design decisions that were specifically formulated with attention to ethical practice.

Loss mitigation steps

Sampling strategy

In traditional survey research design, address based samples select a household member a priori, or request that the eldest member of the household, head of household, or household member with the next upcoming birthday respond to the recruitment ask (Battaglia, Link, Frankel, Osborn, & Mokdad, 2008). SEED invitation mailers were not addressed to any one person in the residence, typically referred to as the “any adult” approach. While this sampling strategy may be less robust for randomization than a two-stage screener sampling strategy (Battaglia et al., 2008), it represented a critical step in preserving the ethical foundations of the project.

Specifically, a key standard of ethical public health practice requires the recognition of both social interdependence and individual rights (Public Health Leadership Society, 2002), and employment of one of the fundamental ethical standards of social work, self-determination, requires that social work professionals regard clients as experts on their own lives (National Association of Social Workers, 1994). By addressing the invitation letter to “Resident”, the individual or family's right to determine who was best to respond, based on existing relationships and networks within the household, financial need, use of benefits, health status, or other factors, was preserved. For example, if a person receiving Supplemental Nutrition Assistance Program (SNAP) benefits received the invitation mailer, she may have had a conversation with her partner about the potential for benefits reduction and the family might have collectively determined her partner as the better fit to participate in the study.

Onboarding process

Similar to sampling procedures, onboarding into experiments typically represents a hands-off approach. Participants receive introductory materials through mail or email, may attend a single orientation, and have limited human engagement with the individual researchers or program staff. Centering both the right to health and resources outlined by public health practices (Public Health Leadership Society, 2002) and the importance of human relationships highlighted in social work principles (National Association of Social Workers, 1994), SEED invited all members of the treatment condition to attend an individualized onboarding session. These individual meetings not only helped recipients navigate a complex benefits structure and preserve access to means tested programs like health insurance provided through MediCal, but also helped establish rapport and trust between the program and recipient from the outset.

During those onboarding meetings, a licensed clinical social worker discussed with the recipient that while GI payments do not exceed the yearly gift tax limit of $15,000 (Department of the Treasury, Internal Revenue Service, 2018), as a pilot GI program, SEED could not guarantee how a front-line caseworker may classify the payment when certifying means tested benefits. Eligibility for benefits that use this same gift criterion carried a low risk of being impacted by GI, but other benefits include gifts as part of eligibility calculations. Where possible, the SEED team pursued waivers excusing GI from income calculations to ensure it would not be counted against recipients' benefits eligibility. There were, however, limits on which benefits waivers could preserve.

As part of informed consent, potential recipients received a comprehensive overview of the potential impact of GI on other benefits. Table 1 describes benefits (i.e., financial aid, food assistance, health insurance, housing support) potentially impacted by GI and recommendations that are part of an ongoing exchange between SEED staff, local partners (e.g., San Joaquin County Human Services Agency, Housing Authority of the County of San Joaquin, legal counsel) and recipients. Those at risk of experiencing any benefit reduction also received a one-page cost-analysis fact sheet that detailed the amount of benefits that could potentially be impacted. At this stage, five recipients opted out of SEED, and other households were randomly selected from the passive control group to take their place.

Table 1

Overview of benefits potentially impacted by GI and recommendations.

Benefit	Description	Potential Impact	Recommendation
CalWorks	Aid and services to eligible families	No risk: Waiver successfully secured in partnership with San Joaquin County Human Services Agency.	Recipient encouraged to participate, assuming no other benefits concerns.
Social Security	Retirement income payable to adults at least 61 years and 9 months old	Low risk: Eligibility is based on previous contributions to Social Security.	Recipient encouraged to participate, assuming no other benefits concerns.
Social Security Disability Income (SSDI)	Aid to adults who can't work due to a medical condition	Low risk: Eligibility is based on previous contributions to Social Security and inability to work.	Recipient encouraged to participate, assuming no other benefits concerns.
Supplemental Security Income (SSI)	Aid to disabled low-income children and adults; also payable to adults at least 65 years old without disabilities who meet financial limits	Impacted: GI payment is considered a gift and counts against eligibility.	Recipient does a cost-analysis of how much they will lose in SSI, and then makes a decision to (not) move forward.
Unemployment	Temporary aid for adults who lose their job to help them meet expenses while seeking new employment	Low risk: Eligibility is based on earned income.	Recipient encouraged to participate, assuming no other benefits concerns.
Women, Infants, and Children (WIC)	Supplemental foods, health care referrals, and nutrition education for low-income mothers, and to infants and children up to age five years old who are at nutritional risk	Low risk: Eligibility is based on income and/or participation in Medi-Cal, CalWorks, and CalFresh.	Confirm potential impact of GI if receiving multiple benefits.
MediCal	Free or low-cost health insurance for low-income children and adults	Low risk: GI is considered a gift from a non-profit organization and therefore is not considered taxable income by the IRS, and should not count against MediCal eligibility.	Recipient does not need to report GI payment on taxes; tax assistance provided to ensure accurate reporting.
Supplemental Nutrition Assistance Program (SNAP)/CalFresh	Benefit enabling low-income adults and families to purchase eligible food in authorized retail food stores	Impacted: GI payment is considered unearned income and may impact SNAP/CalFresh eligibility. The decrease, however, would not be $1:$1 and the process for re-enrolling at the end of the study if the recipient is eligible based on their earned income is straightforward and relatively fast (within 30 days).	Recipient does a cost-analysis of how much they will lose in SNAP, and then makes a decision to (not) move forward.
Public Housing and Section 8 Vouchers	Housing assistance for low-income families, the elderly, and persons with disabilities either through government-owned housing or vouchers to be used in the private market	Low risk: Recipient share-of-cost does increase but is reimbursed by the Hold Harmless Fund through a partnership with the local housing authority.	Recipient encouraged to participate, assuming no other benefits concerns.

Finally, in the event that recipients do experience a partial or total loss of means tested benefits during program participation, SEED established a Hold Harmless Fund to ensure no recipient is worse off financially as a direct result of receiving GI. Modeled after the hold harmless provision included in the statute for the Alaska Permanent Fund Dividend (Social Security Administration, 2019), this fund provides reimbursement to recipients if they become ineligible for benefits due to receipt of GI.

Democratizing human contact

Valuing relationships and building trust through collaborations were key components that allowed SEED to democratize human contact at all points in the pilot. Reducing friction in the social safety net and the threat automation and artificial intelligence pose for technological unemployment are often cited as rationales for introducing a universal basic income or GI (Goodin, 2013; Walker, 2016). These potential shifts are already present in low-wage retail sectors where automation is eliminating human contact in a variety of transactions previously marked with relational interactions (Hegewisch & Hartmann, 2019). Meanwhile, unlike early research around the digital divide citing socio-economic gaps between those with access to technological advances and information technology and those without (Norris, 2001), the drop in the cost of technology coupled with the comparatively expensive nature of human employment is rendering new questions about who is granted access to human contact in their everyday lives and under what terms. As Nellie Bowles (2019) argues, “human contact is now a luxury good,” whereby those with considerably more capital have the latitude to ensure their children's education, their market transactions, their care work, and medical interactions are laced with relational interactions rather than mediated through screens and artificial intelligence. Part of GI's appeal is its potential to address the threat of technological unemployment through a mechanism that simultaneously reduces the friction of human contact required to maintain large bureaucratic structures governed by means testing. Paradoxically, however, the final critical tier of mitigating benefits loss, upholding ethical principles and ensuring implementation efficiency, rests on the human contact and relationship building between SEED's program staff and the research team. Maintaining intentional recipient engagement while navigating a fluid policy landscape requires relational friction in order to preserve benefits while receiving GI.

The intense research activities of SEED, alongside the very human nature of it, have created unique partnerships among the intervention staff, research staff, and recipients. The Stockton-based program officer not only ensures completion of surveys on a monthly basis, but a licensed clinical social worker also checks in on the financial lives and well-being of recipients. As scheduled recertification appointments for housing vouchers or SNAP benefits come about, the program officer and other SEED staff remain in conversation with recipients about how they are choosing to protect limited resources while leveraging the $500.

In keeping with the project's commitment to agency and self-determination, recipients also communicate regularly with the program office as to whether or not they would like staff to communicate with the caseworkers managing their benefits regarding the amount of money provided through the GI. Additionally, when the 18-month pilot ends, staff will assist recipients who may need to re-apply for benefits, leveraging existing relationships with local agencies to ensure that recipients maintain all benefits for which they are eligible once they are off-boarded from the study.

Implications

GI has demonstrated a relationship between cash transfers and improved health outcomes, but the implications of potentially losing benefits when receiving a cash transfer remain under examined. Researchers operate under ethical mandates to ensure that their work does not produce new inequalities, threaten individual agency, or discredit the field by failing to account for feasibility concerns. While IRBs seek to protect human subjects in research, given the lack of standard benefit eligibility requirements and income calculations, they may fail to realize the potential impact of cash transfers on recipients’ overall resources. For pilots that are not governed by these boards, it is even more important that researchers remain proactive to protect recipients by upholding professional practice standards. Both the public health and social work professions are uniquely positioned to conduct research that is empirically and ethically sound. Here, we provide effective practices for mitigating the loss of existing benefits from SEED, one of the first contemporary GI experiments.

It is essential to consider implementation factors alongside policy to protect GI recipients. As researchers seek to advance knowledge related to GI, tensions around benefit loss may impact research activities that are carried out in a traditionally “distanced” way lacking the human contact noted prior. Investigating innovative solutions responsive to income volatility requires a combination of “on the ground” engagement and self-determination where researchers make place-specific program considerations, recognizing the interaction of their work within the existing benefits structure. They must also engage with “street level bureaucrats” to understand how their programs will be interpreted and treated locally (Lipsky, 1980) and, if possible, advocate for waivers that will ensure no harm comes to recipients.

Upholding recipients’ ability to make informed decisions regarding their participation remains crucial to this process and is mandated by the professions of public health and social work (National Association of Social Workers, 1994; Public Health Leadership Society, 2002; Thomas, Sage, Dillenberg, & Guillory, 2002). GI recipients who may have previously experienced long waiting periods for public assistance (Seefeldt, 2017) may be reluctant to participate in GI experiments that may jeopardize their existing benefits and require them to repeat onerous applications and processes. At the same time, recipients should be encouraged to consider the costs and benefits of receiving GI to ensure they do not face a penalty for collecting a cash transfer.

Experiences from SEED can inform other GI pilots being conducted within the context of the US social safety net and in the midst of a pandemic-driven economic crisis where cash transfers are being introduced on an unprecedented scale. While the current coronavirus stimulus package proposes a single cash transfer, the scale of the package creates a considerable policy precedent in an era of GI experimentation. The boundaries determining which benefits are preserved and which are left to chance can inform future public health research while offering public health and allied disciplines an opportunity to shape research and policy.

Funding

This work was funded by the Economic Security Project & the Evidence for Action Program of the Robert Wood Johnson Foundation.

CRediT authorship contribution statement

Amy Castro Baker: Conceptualization, Methodology, Investigation, Writing - original draft. Stacia Martin-West: Conceptualization, Methodology, Investigation, Writing - original draft. Sukhi Samra: Methodology, Writing - review & editing, Project administration. Meagan Cusack: Writing - original draft.

Declaration of competing interest

The authors have no conflicts of interest or financial disclosures.