Demographic microsimulation integration

Individual Data

For individual income and payroll taxes, the backbone of the Tax Module is the IRS’s Statistics of Income (SOI) Public Use File (PUF). This dataset is an anonymized representative sample of tax returns with more than one hundred variables on income sources, deductions, credits, and demographic characteristics. We start with the 2012 PUF and updates all variables in accordance with the latest publicly available aggregate data through a process called historical aging.

The PUF is then statistically matched with population data from the CPS. This process imputes non-tax demographic data to each tax return, and provides data on the population of individuals and families that do not file taxes. The result is an augmented PUF with a rich variety of economic and demographic variables. For a full technical discussion, see this report.

Demographic Microsimulation Integration

We then use demographic projections from PWBMsim to forecast the tax microdata forward into future years. Differential growth rates in subpopulations imply that today’s tax-filing population will look substantially different in the future. Sample weights are adjusted to reflect these changes.

In addition to demographic projections, PWBMsim produces detailed aggregate income account projections. We use these series to project baseline growth rates of all variables found on tax returns, which vary for individual taxpayers by income and demographic characteristics. When the Tax Module evaluates a tax policy change, some of these variables respond endogenously to account for micro-dynamic behavior (as discussed here).

Overall, this integration process takes full advantage of PWBM’s proprietary economic and demographic projections to create a consistent and unified framework for making revenue projections.