Public opinion on government funding of the arts in the United States: demographic and political factors

ABSTRACT

I examine associations between demographic and political factors and attitudes regarding government funding of the arts in the United States. I also generate dynamic, state-level estimates of public opinion on governmental support the arts from 1990 to 2016. While elite rhetoric and policy proposals sometimes call for major reductions in government funding for the arts, I find no evidence of majority support for cutting arts spending at the national level during this period. At the state level, majorities supported reducing government funding for the arts only rarely. I find that partisanship and ideology are associated with support for government funding to a modest extent. While ideologically conservative members of the public were more likely than others to want to cut arts spending, majorities of Republicans favoured maintaining or increasing government spending on the arts in three of the four survey years analyzed. Overall, proposals to substantially reduce public funding of the arts in the United States were not well-aligned with public opinion.

KEYWORDS:

• Government funding
• arts spending
• public opinion
• partisanship
• ideology

Acknowledgments

I would like to thank Kyu Chul Shin for research assistance.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Similarly, opposition to public arts funding in the U.S. is not entirely restricted to conservatives or Republicans. Liberal commentator Jonathan Chait, for example, has argued against public arts subsidies (Chait 1997).

2. Total per capita government spending on the arts in the United States is still far lower than typical per capita spending levels in OECD-member counties (Feist et al. 1998; National Endowment for the Arts 2000).

3. Beyond analyzing responses to survey questions that ask directly whether respondents support government funding of the arts, a number of studies have used contingent valuation methods, or “willingness-to-pay” approaches, to estimate support for government arts spending in the U.S. and other nations (Hansen 1997; Martin 1994; Thompson et al. 2002; Santagata and Signorello 2000). While some of these studies analyze socioeconomic correlates of support for government arts spending, they have not to date examined the effects of political attitudes.

4. In 2006 and 2016, only some respondents were asked these questions.

5. Means are based only on respondents who answered all seven questions.

6. Appropriations data are drawn from American Academy of Arts and Sciences, Humanities Indicators (2018).

7. Descriptive statistics for all variables used in the regression analyses are provided in Table A1 in the appendix.

8. In 1998, respondents were asked whether they attended a dance performance, attended a classical music performance, attended a live drama, or visited an art museum in the past year. Respondents who did one or more of those things were coded as having attended an arts event.

9. While the subsamples of black respondents and respondents identifying with other racial categories are not extremely small, their relatively small sizes does likely contribute to many of the coefficients on the racial identification variables being statistically insignificant while in some cases being larger in magnitude that coefficients on variables such as the partisan identification variables.

10. Ideally, an indicator of identification with evangelical Christianity and/or the Christian right would be included in this analysis. Unfortunately, the GSS has not consistently included variables that allow scholars to capture such identities. Nevertheless, frequency of church attendance is regularly associated with more conservative policy preferences in the United States (Wald and Calhoun-Brown. 2018), and as such it is somewhat surprising that the associations between religious service attendance and opinions on arts spending reported here are not stronger.

11. The demographic characteristics used in the procedure were gender, age, education level, and race. Demographics from the 5% sample of the most recently administered U.S. Census were used in generating estimates for 1990, 1996, and 2006. Because the necessary education variable was not collected during the 2010 U.S. Census, demographics from the 2010 American Community Survey (ACS) were used to generate the estimates for 2016. IPUMS (Ruggles et al. 2018) was used to facilitate the retrieval of Census and ACS data. The state-level indicators employed were the percentage of a state’s population that were evangelical Christians or Mormons (from the most recent Religious Congregations and Membership Study by the Association of Statisticians of American Religious Bodies, hosted by the American Religion Data Archives at www.thearda.com) and the state-level share of the two-party vote in most recent presidential election (Leip 2019).

12. In a different context, Toshkov (2015) does not find this suggestion to be well-supported.

13. Data on the total state arts appropriation per capita was drawn from the Public Funding Sourcebook published by National Assembly of State Agencies (2017).

14. Alaska and Hawaii are included in this calculation, but are often omitted in state politics research. The correlation between appropriations and opinion is even weaker if Alaska and Hawaii are omitted.

15. Based on the estimates from this simple model, increasing a state’s general ideology by one standard deviation would be expected to result in $0.55 more per capita in arts funding, holding state-level opinion on arts spending constant. Given that a large majority of states spend less than $2.00 per capita on the arts, this is a substantial effect.

Additional information

Funding

This project was supported in part by a West Virginia Humanities Council Fellowship Grant.

Notes on contributors

Matthew L. Jacobsmeier

Matthew L. Jacobsmeier is an associate professor of political science at West Virginia University.  His research interests include public opinion and political behavior, race, ethnicity and politics, and religion and politics.