(see Poster Gallery) Understanding American Welfare State Retrenchment through the Lens of Public Opinion Toward Welfare Spending

Background/Purpose: Although fickle, measures of public opinion have been used by policymakers for centuries to determine the course of public policy. Set in a context of decades of welfare state retrenchment, this study investigates contemporary public opinion toward social welfare spending through the lens of policy feedback theory. In this light, variation in support for welfare spending and influence on policymakers may partially explain trends toward welfare state retrenchment.

I seek to answer whether there an association between various demographic characteristics and opinion toward social welfare spending and, if so, what demographic characteristics are predictive of one’s opinion toward social welfare spending.

Methods: I test the bivariate and multivariate relationships between participants’ response to the question of “. . . are we [the United States] spending too much, too little or about the right amount on welfare?” and theoretically relevant demographic variables—including work status, marital status, parental status, age, education, gender, race and ethnicity, household income, region of residence, political party, and religious status—using multivariate logistic regression with data from the 2014 General Social Survey. This sample used probability sampling, making it nationally representative. I limit the data to responses from 2014, and of the 2538 participants in that cross-section, my final model has an N of 1106 participants.

Results: I find that work status, gender, political party, and religious status all had statistically significant relationships with opinion toward welfare spending using Wald and likelihood ratio tests when controlled for all other variables. The predicted probability of saying welfare spending was too much was lower for students and those with a work status of “other” when compared with those who worked fulltime and for Black participants when compared to white participants. On the other hand, the predicted probability of saying welfare spending was too high was higher for men when compared to women, for Republicans when compared to Independents and Democrats, and for those who were religious when compared to those who were nonreligious.

Conclusion/Limitations: These findings are important for explaining welfare retrenchment through the lens of policy feedback theory because they delineate what social identities are associated with sympathy for and concern with levels welfare spending.

The groups represented by each of the tested social identities have varying influence on how governments and other stakeholders shape the American welfare state. It follows that as the American welfare state has and continues to see broad retrenchment, the interest of the constituencies with the concern that the United States is spending too much on the welfare state (i.e., those who are employed, White, men, religious, or Republican) has been exercised by policymakers.