What do people want from a welfare system? Conjoint survey evidence from UK adults

Abstract

Current debate about the appropriate level and form of social safety net leads us to ask the question: what do people want from a welfare system? We conducted a conjoint survey experiment with 800 UK-resident adults. We presented them with welfare schemes that varied generosity; in their fiscal implications (rates of personal income tax as well as other funding mechanisms); in their population consequences (effects on the rate of poverty, on inequality, and on physical and mental health); and in their institutional design features (means testing, conditionality, and universality). The strongest driver of preference for a welfare scheme was its effect on poverty: people liked schemes that reduced poverty, and disliked schemes that increased it. Respondents were prepared to trade off their dispreference for higher personal income taxes against poverty: even for Conservative voters, substantial income tax rises were acceptable in exchange for sufficiently large reductions in the poverty rate. Taxes on wealth and carbon emissions were positively valued relative to increasing government borrowing. Respondents paid some attention to the effects of schemes on inequality and physical and mental health as well as poverty. Preferences over institutional design features, such as means testing, conditionality and universality, were weak. We discuss the findings with respect to the envelope of welfare systems that would be politically feasible in the UK context.

The Design of Welfare: Unraveling Taxpayers' Preferences

“…Taken altogether, these results show that the preferences expressed in our survey and selfreported voting behavior are fairly consistent, which makes us confident that our respondents answered our survey seriously and truthfully. These findings are also in line with those of Lara & Shores (2022), who find that preferences for income equality and income mobility are related to partisanship, and with those of Nettle et al (2023) on the relationship between partisanship and preferences for conditional welfare vs. universal benefits.…”

“…Respondents assign eight percent fewer points to a policy package that costs one hundred euros more on an annual basis. Nettle et al (2023) also find that UK respondents have a preference for more generous benefit levels and lower taxes, all other things equal.…”

“…Similarly, Rincón et al (2022) find in a choice experiment that the Finnish public is more favorable to income support that is conditional upon looking for work or not being able to work than of unconditional income support. However, Rincón (2023) finds no important role for job search requirements in a similar study in Spain, and Nettle et al (2023) find that job search requirements do not significantly influence preferences for income support in the UK.…”

“…Since our research design was not incentive-compatible, it is interesting to relate respondents' distributed points in our choice experiment to actual choices they made during the Parliamentary election that took place in the Netherlands just a couple of months before our survey, in March 2021. In doing so, we follow Hainmueller et al (2015), De and Nettle et al (2023), who compare choices made in a choice experiment with voting behavior in order to check the validity of the choice experiment. Respondents of the LISS panel were asked to report for which party they voted.…”

“…Third, our paper relates to an emerging literature in political science that measures multidimensional policy preferences using choice experiments to let respondents make trade-offs among dimensions of policy packages. In particular, a number of studies examine the preferences of the public for a universal basic income vs. conditional welfare benefits (Nettle et al 2023;Rincón 2023;Rincón, Vlandas, and Hiilamo 2022;Stadelmann-Steffen and Dermont 2020). We focus on the preferences of the public for the organization of means-tested welfare.…”