Tax-Information Avoidance

Abstract: 
Many of our most consequential decisions are influenced by the taxes we face. When comparing different options for retirement savings, for example, individuals must account for potentially different tax consequences of different savings options. When considering two employment opportunities in different states, workers must compare not only the wage and benefits packages, but also the different tax rates. When deciding whether or not to work an extra shift, workers must consider how much of the extra wage will make it to their pockets post-tax. When deciding whether to donate to charity, buy a house, or to pursue any other tax-incentivized behavior, knowledge of the relevant deductions or credits is necessary for calculating the after-tax price of these activities.
While information about tax law is clearly necessary for optimal decision making, a growing body of evidence suggests that taxpayers are imperfectly informed about key tax parameters (e.g., Blaufus et al. 2013; Gideon 2015; Rees-Jones and Taubinsky 2016). Recent research confirms an intuition many have developed through personal experience: doing taxes is highly aversive, and many people leave money on the table to avoid filling out additional tax forms (Benzarti 2016).
In this project, we seek to connect these phenomenon to a well-developed literature in economics and psychology on information avoidance (for a recent review, see Golman, Hagmann, and Loewenstein forthcoming). The key insight of this literature is acknowledging the psychological effects of information. When modeling Bayesian decision makers, economists often treat information as a purely instrumental construct: it helps individuals make better forecasts a decisions, but has no direct effects on utility. This approximation abstracts from the possibility that bad news might be psychologically aversive, might generate dread, and might lead to new paths of behavior that a time-inconsistent decision maker might wish to avoid in the short term (even if optimal from a time-consistent perspective). To take an archetypal example from this literature that illustrates these trade-offs, Ganguly and Tassoff (forthcoming) demonstrate that undergraduate experimental participants are often eager to avoid learning the results of common STD tests, and some are willing to pay to stop from learning the results. While it is clear that a perfect decision-maker would want the results of these tests, one can easily rationalize why a less-than-perfect decision-making might behave in this manner.
If similar information-avoidance occurs in the tax setting, it has substantial implications for our modeling of tax systems. In our project, we will undertake a series of experiments meant to cleanly isolate the presence of tax-information avoidance and to highlight several of its consequences. Conditional on these experiments demonstrating the presence of this phenomenon, we will then work to integrate this type of behavior into common models of tax analysis and of response to tax incentives, with a specific focus on it’s implications for uptake of tax-preferred retirement savings.
Funded By: 
Boettner
Award Dates: 
July 1, 2017 - March 30, 2018