Loss Aversion Motivates Tax Sheltering: Evidence from U.S. Tax Returns

46 Pages Posted: 27 Sep 2013 Last revised: 9 Dec 2014

Alex Rees-Jones

University of Pennsylvania - Operations & Information Management Department

Date Written: December 8, 2014


This paper presents evidence that loss aversion affects taxpayers as they file their annual tax returns. I model the decisions of a loss-averse tax filer who may use tax shelters to manipulate the "balance due" exchanged with the IRS. I use this model to derive distinguishing predictions of loss aversion which facilitate its identification and quantification in the field. Under loss framing, the discretely steeper marginal utility of a dollar motivates greater pursuit of shelters. These motives imply that the post-sheltering distribution of balance due will exhibit a structural shift in the loss domain, due to discretely higher sheltering in this region. Furthermore, the discontinuity in marginal incentives generates excess mass, or "bunching," at the gain/loss threshold. Using the 1979-1990 IRS Panel of Individual Returns, I document the predicted bunching and shifting in the distribution of balance due, and examine the causes and correlates of these features. The observed distribution is consistent with the framing of tax payments as losses and tax refunds as gains, and is difficult to rationalize with plausible alternative theories. Using two complementary structural approaches — identified from the bunching and shifting predictions, respectively — I estimate substantial potential policy impact of this psychological bias. These results have direct implications for tax policy and public finance.

Keywords: Loss aversion, income taxation, tax sheltering, tax avoidance, tax evasion

JEL Classification: D03, H24

Suggested Citation

Rees-Jones, Alex, Loss Aversion Motivates Tax Sheltering: Evidence from U.S. Tax Returns (December 8, 2014). Available at SSRN: https://ssrn.com/abstract=2330980 or http://dx.doi.org/10.2139/ssrn.2330980

Alex Rees-Jones (Contact Author)

University of Pennsylvania - Operations & Information Management Department ( email )

Philadelphia, PA 19104
United States

HOME PAGE: http://https://opimweb.wharton.upenn.edu/profile/26948/

Paper statistics

Abstract Views