Existing social safety nets play an important role in reducing poverty and inequality, especially in the rich countries of the Global North. These safety nets, however, also have numerous shortcomings and limitations. How should these safety nets be reformed or renewed? Two proposals for such reform are universal basic income (UBI) and job guarantee (JG) programs.
We have participated in the evaluation of recent policy experiments piloting such programs, including the MAGMA job guarantee in Marienthal, Austria (Kasy & Lehner, 2026), and the Pilotprojekt Grundeinkommen in Germany (Bernhard et al., 2025; Bohmann et al., 2025). In the following, we review the empirical findings from these experiments, as well as from the wider literature. But before turning to our empirical findings, we first discuss a conceptual framework for assessing the merits and potential drawbacks of UBI and JG policies, relative to existing alternatives.
A UBI provides unconditional cash payments made regularly to all citizens or residents of a country, regardless of their employment status or other circumstances. When UBI is financed by suitably progressive income taxes, only individuals with low market incomes are net recipients. JG programs, by contrast, offer employment to anyone willing and able to work, typically at a living wage with basic benefits, with a focus on socially useful work.
UBI and JG proposals have much in common: They are unconditional, in the sense that eligibility is not based on prior contributions, work history, demonstrated willingness to seek paid employment, or the exhaustion of all personal assets. Because they are unconditional, neither UBI nor JG require surveillance and control by welfare administrators, in the way that many existing welfare or social insurance programs do.
Relatedly, UBI and JG are intended to be universal, so that everyone is entitled to receive them – though the boundaries of that universality, that is, of who is part of the eligible polity, are certainly contested. (Residency requirements, for example, might in practice exclude more people than the work history requirements of traditional safety nets.)
In these ways, both UBI and JG differ from many existing safety nets that are conditional, not universal, involve surveillance, involve work requirements, provide distortionary incentives, and that might benefit (low-wage) employers more than nominal recipients.
UBI and JG also have some important differences: UBI only involves a cash transfer, while a JG additionally involves labor, of some sort or other, by recipients. In the policy debate, UBI and JG are sometimes pitted against each other, based on seemingly very different visions of a better society (see Lehner 2026). We will argue, however, that these policies should instead be seen as complements, serving different purposes and solving different problems, for different parts of the population.
Conceptual considerations
To evaluate any UBI or JG program, relative to traditional welfare policies, we need to ask three questions.
We first need to consider how such a program impacts the welfare of recipients directly. Here the question of individual autonomy will be central, where autonomy in turn depends on the set of choices effectively available to recipients.
We second need to consider how the program impacts both recipients and others via labor market responses. Here the recipients’ outside options and bargaining power are key, which are affected by the backstop offered by public guarantees such as a UBI or JG.
We third need to assess how the costs for public coffers stack up against the benefits for recipients. Here the effect of program expansion on the tax base is the most important consideration, which depends in particular on labor supply responses.
Autonomy and welfare
Economists typically assess individual welfare in terms of utility. Utility determines what people choose: among the choices available to them, they pick the one with the highest utility. Utility also determines how we assess people’s wellbeing: the higher their utility, the better.
Defining welfare as utility has important implications for evaluating social policy. It implies that individual welfare can only be increased by giving individuals more choice, rather than less. This insight is especially relevant in light of the work requirements or other conditionalities that are often associated with welfare benefits. One might erroneously reason that (i) employment is a key determinant of welfare, that (ii) work requirements can increase employment, and therefore (iii) work requirements increase welfare. This is wrong if we understand welfare as utility: Imposing conditionalities such as work requirements for benefit eligibility can only ever make recipients worse off, relative to the alternative of unconditional benefits.
A second, more subtle implication of defining welfare as utility is given by the envelope theorem (Milgrom & Segal, 2002): Consider a small policy change, such as an expansion of transfer generosity. According to the envelope theorem, the effect of such a small policy change on recipient welfare does not depend on how recipients adjust their behavior.
To get some intuition why this is the case, note that recipients could leave their behavior unchanged, following the policy change, and simply enjoy the more generous benefit. Behavior change can only make the recipients even better off, in terms of utility. Since (by assumption) recipients’ original behavior was already optimal in terms of utility, any change of behavior after the policy change has only negligible effects on utility, however. Changing recipient behavior might, by contrast, impact the tax base and public revenue. We will get back to this point when comparing the costs and benefits of UBI or JG programs.
The utilitarian argument for autonomy is an instrumental one. By allowing people to make their own choices, rather than imposing choices via conditionalities, they can achieve higher utility. There is a related but distinct argument that autonomy is intrinsically valuable, whether or not it allows individuals to make choices that achieve higher utility. This argument can be made on normative grounds; see e.g., Sen (1995), who argues that capabilities are a better measure of welfare rather than utility, since the latter depends on idiosyncratic individual preferences. The argument for the intrinsic value of autonomy can also be made on psychological grounds: Plenty of evidence suggests that subjective wellbeing depends on perceived agency and autonomy. Our own empirical findings confirm this (Bohmann et al., 2025).
Autonomy is one dimension of subjective wellbeing, but a range of others are important, as well, in the context of welfare provision. This includes questions of meaning, social inclusion, social status, time-structure, activity, etc. These dimensions have been discussed by sociologists as the “latent and manifest benefits” of work (Jahoda, 1984). Here, then, is a key difference between UBI and JG. UBI provides recipients with cash, and with a sense of autonomy and safety, but it does not by itself promote meaning, social inclusion, status, or time-structure. This matters more for some recipients than for others. Those who are currently long-term unemployed, for a wide range of individual reasons (such as the program participants in Kasy & Lehner, 2026) might benefit much more from a JG, while those in the wider population who do not suffer from cumulative individual disadvantages in the same way (such as the participants in (Bernhard et al., 2025; Bohmann et al., 2025)) might benefit more from the autonomy and safety provided by UBI.
A final point to note regarding the subjective advantages of both UBI and JG is the question of stigma. Given existing social norms about the desirability of wage labor and individual deservingness, benefit receipt is often associated with stigma, due to a perceived individual failing. Universal, guaranteed benefits avoid such stigma by construction: When eligibility is not based on individual circumstance, it also does not signal individual failure.
Outside options and incidence
The preceding considerations are all based on the direct effect of welfare programs on recipients. Equally important are indirect effects, due to shifting power and market prices.
The most important ingredient determining an individual’s bargaining power in all kinds of social relationships are their outside options: What is the best available alternative, should the relationship end? This applies to labor relations with employers, but it also applies to private (romantic) relationships, and to relationships with welfare administrators.
Outside options set a lower bar. No one needs to accept worse conditions than their outside option, if they can walk away. Crucially, this applies even if the outside option is never actually taken up. A public JG, for instance, sets a standard of employment quality and remuneration below which private employment cannot fall, since otherwise employees would switch to the public option; similarly for UBI.
Beyond this lower bar, outside options also impact how the surplus of a relationship is shared. Different models of bargaining (such as for instance Nash bargaining) make different predictions about how the difference is split, but they consistently predict that those with better outside options receive a bigger share of the pie.
This prediction is key when we assess the impact of unconditional benefits (such as a UBI financed by progressive income taxes) relative to welfare benefits that condition on labor market participation (such as for instance the Earned Income Tax Credit in the United States): Unconditional benefits financed by taxes improve the (potential) recipients’ outside option, relative to a given employment relationship. This implies that employees can more easily stand up for themselves, whether to demand higher wages or better working conditions or to resist abuse. Benefits that are conditional on employment (or on demonstrated willingness to accept employment), by contrast, weaken the outside option of potential benefit recipients. In the presence of such conditional benefits, a loss of employment additionally entails a loss of benefits, and thus makes it harder for employees to demand their rights.
Bargaining, as just described, takes place in bilateral relationships in the presence of match surplus, when both parties are better off by maintaining their match, relative to their outside options. This is a typical situation both in private relationships and in the labor market, where match surplus exists in the presence of job search frictions or employer market power.
Related considerations also apply in (hypothetical) competitive labor markets, where there are many equivalent alternatives to any given relationship. In such markets, economists typically assume that wages and the conditions of work are determined by the interplay of supply and demand. A policy change can lead to higher wages and better working conditions if it reduces labor supply. A policy change can worsen wages and working conditions if it increases labor supply. The extent to which either happens depends on the relative slopes of labor demand and labor supply.
Advocates of JG who are skeptical of UBI sometimes voice the concern that a UBI would allow employers to lower wages. Both models of bilateral bargaining and of competitive labor markets suggest that this is not likely; for a UBI to lower wages would require that it worsens outside options or induces an increase of labor supply, in these models. It is theoretically possible that the income effects associated with a UBI make recipients less responsive to changes in wages, which would lead to an increase of employer market power. Empirically, we find no evidence of this.
Spillovers and program costs
Programs such as UBI or JG potentially impact not only recipients and their employers or direct relationships; these programs might also generate spillovers on the local community or the wider economy. Such spillovers could be both positive and negative, and might operate through a variety of channels: By increasing purchasing power, increased local consumption demand could stimulate the economy. More broadly, guaranteed benefits such as UBI or JG provide individual safety and stabilize expectations for the future, which serves to stabilize consumption, and thereby the economy at large. On the other hand, if the work provided through a JG substitutes for activities previously performed on the market, then a JG might decrease market labor demand.
There is finally the question of program costs. Beyond the direct (mechanical) cost of transfer payments, labor supply responses might induce an excess cost by reducing the tax base. If, hypothetically, a UBI induced many people to stop working, then this would induce a loss of income taxes for the state. To assess this excess cost, we need to determine the causal effect of a program on the tax base; this is the type of effect that the sufficient statistics literature in public finance aims to determine.
Recall the envelope theorem discussed above: An expansion of program generosity has a direct mechanical effect on recipients, but any behavioral responses can be ignored when assessing recipient welfare. By contrast, the impact on public revenue entails both the direct mechanical cost and the behavioral effect on the tax base. The desirability of increased transfer generosity then depends on the tradeoff between the benefit of added redistribution to transfer recipients versus the excess cost of (potential) behavioral responses.
Evidence
We now turn to the empirical evidence, using the preceding conceptual framework to guide both empirical questions and normative interpretation.
Job guarantee
The most direct evidence on job guarantees in a rich country context comes from the MAGMA pilot in Gramatneusiedl, Austria (Kasy and Lehner, 2026). The program offered a guaranteed job, after a short preparatory phase, to residents in long-term unemployment. Take up was close to 100% among those offered a job, an important first stage result if we think of a guarantee as a credible outside option.
The headline impacts line up closely with the conceptual distinction between cash, autonomy, and the latent benefits of work. On the one hand, MAGMA increased employment and income security. On the other hand, it produced large improvements in time structure, activity, social contact, collective purpose, and social recognition (Kasy and Lehner, 2026). These outcomes mirror long standing descriptive and quasi experimental evidence that unemployment is associated with substantial losses in wellbeing (Clark and Oswald, 1994; Clark, 2003; Kassenboehmer and Haisken DeNew, 2009; Hetschko et al., 2014; Pohlan, 2019). They also connect to more recent work that treats job amenities and meaning as objects of demand, estimating willingness to pay for schedule flexibility, working from home, and related non-wage attributes (Mas and Pallais, 2017; Maestas et al., 2023).
A key policy concern is unintended labor market consequences. If eligibility begins after some unemployment duration, job search could in principle slow down before eligibility. MAGMA offers a useful test because eligibility starts after nine months. In hazard rate comparisons against synthetic control municipalities, transition rates out of short-term unemployment are higher, not lower, in the treated municipality (Kasy and Lehner, 2026). This is consistent with the mechanism that the promise of a backstop can stabilize expectations and reduce the demoralization that often accompanies repeated rejection, raising effective search rather than lowering it.
A second concern is displacement. Public employment might crowd out private employment, which is an arguable reason why evaluations of traditional public job creation programs often find weak post program earnings effects (Card et al., 2010; Card et al., 2018). MAGMA finds no evidence of displacement. Municipality level unemployment falls sharply, driven by near elimination of long-term unemployment, without a rise in short-term unemployment (Kasy and Lehner, 2026). Total employment increases by about 555 days per eligible worker over the program horizon, and the increase is not only due to direct job provision but also a rise in unsubsidized employment and self employment (Kasy and Lehner, 2026). This pattern is closer to evidence from public works programs in lower income settings that find positive wage and employment spillovers, such as Muralidharan et al. (2023) and Imbert and Papp (2015) for India, and Franklin et al. (2024) for Ethiopia, as well as to evidence on the psychosocial value of employment in refugee settings (Hussam et al., 2022).
Finally, empirical evidence on costs has grown. MAGMA’s net costs to the public sector are estimated at about 220 euros per participant and month, while participant income increased by 153 Euro, after accounting for reduced benefit payments, revenues of the social enterprise, and tax and social insurance contributions (Kasy and Lehner, 2026). This is a reminder that public employment should be evaluated against its welfare objectives: If the goal is to raise utility and capabilities, then the relevant outputs include social inclusion and dignity, not only later unsubsidized earnings.
Basic income
Evidence on basic income is often framed around a single fiscal worry: reduced labor supply and hence a smaller tax base. The German basic income experiment provides unusually clean evidence on this margin. In a preregistered randomized trial, recipients received 1200 euros per month for three years. In the study, we find essentially zero effects on employment and on job transitions into and out of work (Bernhard et al., 2025). This result is consistent with the Finnish basic income experiment, which found modest employment responses and emphasized the role of removing welfare traps rather than creating large income effects (Verho et al., 2022).
There is some evidence of intensive margin adjustment. Survey and administrative data suggest a small shift toward part time work and modest reductions in social insurance contributions and income taxes, adding up to about 90 euros per month in reduced public revenues, an excess burden around 7.5 percent of the transfer (Bernhard et al., 2025). This corresponds to the tradeoff highlighted by the conceptual discussion: Recipient welfare gains are largely mechanical in the sense of the envelope theorem, while fiscal costs depend on behavioral responses. In this setting, those responses appear limited.
The incidence concern runs in the opposite direction: could basic income enable employers to lower wages? The search and bargaining framework in Bernhard et al. (2025) shows that theory is ambiguous once we allow for amenities and bargaining over surplus, so evidence is essential. Empirically, there is no sign of a fall in employment or of deteriorating outside options, and stated willingness to accept wage cuts before quitting changes little (Bernhard et al., 2025). This aligns with related evidence from permanent universal cash transfers, such as the Alaska Permanent Fund, which finds limited adverse labor market impacts (Jones and Marinescu, 2022).
The German experiment is very informative on the autonomy channel: In Bohmann et al. (2025), we show large improvements in mental health, purpose in life, and life satisfaction, with a substantial share of these gains persisting after payments end. We link these improvements to increased perceived autonomy and to concrete life changes, including greater saving, more prosocial transfers, more time with friends, and more sleep. This speaks directly to the conceptual claim that autonomy is both instrumentally welfare improving and intrinsically valued. It also helps reconcile mixed evidence in high income settings: Several recent guaranteed income trials in the United States and Canada find no or weaker mental health effects (West and Castro, 2023; Dwyer et al., 2023; Miller et al., 2024; Vivalt et al., 2024). We argue that generosity and predictability matter because they determine whether recipients can implement meaningful change (Bohmann et al. 2025), an interpretation that fits with earlier evidence from low and middle income cash transfer studies and meta analyses (Haushofer and Shapiro, 2016; McGuire et al., 2022).
Taken together, these findings suggest a complementarity of benefits, rather than a contest between basic income and job guarantees: Basic income primarily expands autonomy and financial security, with limited labor supply reductions (Bernhard et al., 2025; Bohmann et al, 2025). Job guarantees, by contrast, directly provide the latent benefits of work that cash alone does not reliably produce, especially for disadvantaged groups and long-term unemployed (Kasy and Lehner, 2026). If the policy objective is welfare in the broad sense emphasized in our conceptual discussion, the evidence points to complementarity: basic income as a general floor that strengthens outside options and reduces surveillance and stigma, and job guarantees as a targeted instrument that restores social inclusion, time structure, and recognition where these are missing.
References
Bernhard, S., Bohmann, S., Fiedler, S., Kasy, M., Schupp, J. and Schwerter, F. (2025). Basic income and labor supply: Evidence from an RCT in Germany, Working Paper.
Bohmann, S., Fiedler, S., Kasy, M., Schupp, J. and Schwerter, F. (2025). Cash Transfers, Mental Health and Agency: Evidence from an RCT in Germany. Working Paper.
Card, D., Kluve, J. and Weber, A. (2010). Active labour market policy evaluations: a meta-analysis, The Economic Journal, 120(548), pp. F452–F477.
Card, D., Kluve, J. and Weber, A. (2018). What works? A meta-analysis of recent active labour market programme evaluations, Journal of the European Economic Association, 16(3), pp. 894–931.
Clark, A.E. (2003). Unemployment as a social norm: psychological evidence from panel data, Journal of Labor Economics, 21(2), pp. 323–351.
Clark, A.E. and Oswald, A.J. (1994). Unhappiness and unemployment, The Economic Journal, 104(424), pp. 648–659. Hetschko, C., Knabe, A. and Schöb, R. (2014). Changing identity: retiring from unemployment, The Economic Journal, 124(575), pp. 149–166.
Hussam, R., Kelley, E. M., Lane, G., and Zahra, F. (2022). The Psychosocial Value of Employment: Evidence from a Refugee Camp. American Economic Review, 112(11):3694–3724.
Imbert, C. and Papp, J. (2015). Labour market effects of social programmes: evidence from India’s employment guarantee, American Economic Journal: Applied Economics, 7(2), pp. 233–263.
Jones, D. and Marinescu, I. (2022). The labour market impacts of universal and permanent cash transfers: evidence from the Alaska Permanent Fund, American Economic Journal: Economic Policy, 14(2), pp. 315–340.
Kassenboehmer, S.C. and Haisken-DeNew, J.P. (2009). You’re fired! The causal negative effect of entry unemployment on life satisfaction, The Economic Journal, 119(536), pp. 448–462.
Kasy, M. and Lehner, L. (2026). Employing the unemployed of Marienthal: evaluation of a guaranteed job programme, Working paper.
Lehner, L. (2026). What Do Unemployed Workers Want: Guaranteed Work or Guaranteed Income? Working Paper. Mas, A. and Pallais, A. (2017). Valuing alternative work arrangements, American Economic Review, 107(12), pp. 3722–3759.
Maestas, N., Mullen, K.J., Powell, D., von Wachter, T. and Wenger, J.B. (2023). The value of working conditions in the United States and implications for the structure of wages, American Economic Review, 113(7), pp. 2007–2047. McGuire, Joel, Caspar Kaiser, and Anders M Bach-Mortensen, “A systematic review and meta-analysis of the impact of cash transfers on subjective well-being and mental health in low-and middle-income countries,” Nature
Human Behaviour, 2022, 6 (3), 359–370.
Pohlan, L. (2019). Unemployment and social exclusion, Journal of Economic Behavior & Organization, 164, pp. 273–299. Verho, Jouko, Kari Hämäläinen, and Ohto Kanninen. 2022. “Removing Welfare Traps: Employment Responses in the
Finnish Basic Income Experiment.” American Economic Journal: Economic Policy 14 (1): 501–22.
West, S., and Castro, A. (2023). Impact of Guaranteed Income on Health, Finances, and Agency: Findings from the Stockton Randomized Controlled Trial. Journal of Urban Health, 100, 227-244.
Vivalt, E., Rhodes, E., Bartik, A. W., Broockman, D. E., Krause, P., and Miller, S. (2024). The employment effects of a guaranteed income: Experimental evidence from two U.S. states. Working Paper 32719, National Bureau of Economic Research.
Share:
