An Information Model to Present Proposals to Increase Taxes: Two Examples in the Context of American Values

Abstract

This article offers an information model for educating the public on the rationale for a tax increase. It argues that citizens exposed to two types of information—a need narrative and an anchor value—are more willing to support a tax increase. The applicability of the model is illustrated with two examples in which the model was used to structure the presentation of information. In both situations, exposure to the specific types of information suggested by the model is associated with support for a tax increase. The conclusion relates the model to the findings of previous research on the attributes of successful referenda campaigns for tax increases.

Keywords:

• information model
• public deliberation
• tax increase
• narrative paradigm
• anchor heuristic
• transit funding
• motor fuel tax
• transportation finance

Notes

¹The public has traditionally been strongly opposed to increases in transportation-related taxes or fees. In the United States, polls by Quinnipiac University found that voters in the Northeast were strongly opposed to raising gasoline taxes. A poll of New Jersey voters found that 69 percent were opposed to raising the state gasoline tax (Quinnipiac University, 2006). A similar poll found that 61 percent of Connecticut

voters were opposed to gasoline tax hikes proposed in 2005 (Quinnipiac University, 2005). There are other incidences of overwhelming taxpayer opposition to raising the gas tax and other fees. In 2002, Missouri voters soundly defeated Proposition B—which would have increased sales and gasoline taxes to pay for $483 million worth of transportation infrastructure projects—by an almost 3-to-1 margin.

²This non-parametric statistical test was chosen because the variables are ordinal, the analysis involved small samples sizes (10 for the treatment focus group with information and 37 for the control focus group without information), and we cannot assume that the responses are normally distributed. We could, however, assume that the population distributions are continuous, making it possible to perform the Mann-Whitney test.