Inclusion of “Undesirable” Groups and Support for Universalistic Welfare Policy

• In Event: Health and Public Opinion

Abstract

Does immigration reduce support for universalistic welfare policy? A long literature has examined how racial and ethno-religious heterogeneity undermines support for the welfare state. The “anti-solidarity hypothesis,” suggests that people dislike redistributing towards outgroups, especially where outgroups may be concentrated below the median income. Less research has looked specifically at how inclusion of outgroups affects support for universalistic social insurance programs that make exclusion difficult. We undertake a survey experiment with a panel of 1,000 respondents through YouGov in the U.S. to test whether providing information about who will be included (immigrants & transgender youth) in universalistic programs (Medicare for All & Universal Pre-K) decreases support for those same programs. The experiment was preregistered with Open Science Framework. We find that nearly 70% of the sample supports Medicare for All in the control arm. Introducing information that immigrants will receive benefits reduces support by 6.4 percentage points (p=0.05) and providing information that gender affirmative care for transgender youth is covered in the benefit package reduces support by 4 percentage points (p=0.26). The inclusion of information about immigrants and transgender youth benefits has no effect on subsequent questions about support for more universalistic versus means-tested programs (Universal Pre-K vs means-tested subsidies). Differences within parties across study arms are not significant. However, among Democrats introducing information about immigrant benefits reduces support for Medicare for All by 4 percentage points and by 10 percentage points for transgender benefits. For Republicans, providing information about immigrant benefits is more powerful than information about transgender benefits reducing support by 10 percentage points versus 1 percentage point. Overall, party identification is by far the strongest predictor of support for Medicare for All, followed by age and education. We conclude that specific messaging about who will benefit from universalistic policies can reduce support for such policies in line with the anti-solidarity hypothesis, but the magnitude of the effect is relatively small compared with more stable differences such a party identification. Emphasizing that specific marginalized groups will benefit also does not increase support for universalistic programs, including among Democrats who we might think would view inclusion of such groups favorably. Future studies may want to examine whether specifying that certain out-groups will be excluded could increase support for universalistic welfare, particularly among conservatives. Overall, messages emphasizing universal shared benefits for all without reference to specific identity group appears to enjoy the most popularity in line with the anti-solidarity hypothesis.

Authors

• Ashley M Fox, SUNY at Albany

• Hyesang Noh, Rockefeller College of Public Affairs & Policy, University at Albany, State University of New York