U.S. Individual Income Tax Return Simulated Data: A Methodology

Abstract

ABSTRACTThis paper proposes and tests a methodology for the development of simulated data for individual income tax returns in the U.S., enabling students of taxation and accounting, and other interested parties, to examine changes to the tax code, the effects of tax planning alternatives, and conduct repeated experimental testing on tax return data. The simulation draws insight from statistical and economic literature into the stochastic nature of certain line items from the tax returns, using the information provided by the Internal Revenue Service's (IRS) Statistics of Income (SOI) program. The simulation produced data for 147,000 tax returns, representing ∼1 percent of the population of filed tax returns, and was then validated against the results of the summary data provided by IRS SOI. Our simulated data closely matched the number and combined dollar value of the IRS SOI summary data at the AGI, state, and filing status levels.