Microsimulation Model of the U'S' Federal Individual Income Tax

1
Microsimulation Model of the U.S. Federal
Individual Income Tax

• Ralph A. Rector, Ph.D.
• Tracy L. Foertsch, Ph.D.
• The Heritage Foundation
• August 22, 2007

The analysis and conclusions presented here are
strictly those of the authors. They should not be
interpreted as reflecting the views of The
Heritage Foundation.
2
Overview

• ObjectiveAnalyze changes in Federal Income Tax
• Account for details of tax provisions
• Include economy-wide effects
• Target AudienceLegislative Decision-makers
• Method and PlatformFile Types and Languages
• Databases SPSS / SQL
• Parameter and Output Files Excel / VBA / AREMOS
• Microsimulation Model C / C
• Policy Environment
• Produce revenue estimates consistent with
  official forecasts
• Estimate dynamic revenue and economic effects
• Provide estimates of marginal tax rate effects

3
Primary Components of the Microsimulation Tax
Model

• Statistically-matched public use base year
  micro-level records
• Tax return data Internal Revenue Service
  (IRS)/Statistics of Income Division (SOI) Public
  Use Tax File
• Household survey data Census March Current
  Population Survey (CPS)/Annual Income and
  Demographic File
• Federal individual income and payroll tax
  calculator
• Linear-programming data aging routines
• Age microsimulation data by adjusting sample
  weights and dollar amounts to hit target values

4
Aging the Microsimulation Tax Data

• Historical data (post-base year)
• Official biannual baseline economic and budgetary
  projections
• Assumes current law

5
Aging the Microsimulation Tax Data

• Historical data (post-base year)
• Official biannual baseline economic and budgetary
  projections
• Assumes current law
• Serves as a common baseline for scoring proposed
  changes in taxes and spending

6
Aging the Microsimulation Tax Data

• Historical data (post-base year)
• Official biannual baseline economic and budgetary
  projections
• Assumes current law
• Serves as a common baseline for scoring proposed
  changes in taxes and spending
• Used to calibrate US macroeconomic model (Global
  Insight forecasting model)

7
Aging the Microsimulation Tax Data

• Historical data (post-base year)
• Official biannual baseline economic and budgetary
  projections
• Macroeconomic model projections
• Baseline forecast from macroeconomic model used
  to calibrate microsimulation model

8
Aging the Microsimulation Tax Data

• Historical data (post-base year)
• Official biannual baseline economic and budgetary
  projections
• Macroeconomic model projections
• Baseline forecast from macroeconomic model used
  to calibrate microsimulation model
• Policy simulation forecast from macroeconomic
  model used to calibrate microsimulation model

9
Inputs From the Macroeconomic Model

• Components of personal income
• Wages and salaries
• Investment income (interest and dividend income)
• Proprietor income (farm and non-farm)
• Other business income (rental income)
• Corporate profits
• Price levels
• CPI for urban consumers
• GDP deflator for consumer medical goods and
  services
• Other variables used to compute tax

10
Outputs From the Microsimulation Model

• Revenue estimates
• Individual income tax
• Marginal Tax Rates
• Ordinary income
• Business income
• Capital income

11
Calibrating the Microsimulation and Macroeconomic
Models for Policy Simulations

• Step 1. Use the microsimulation model to estimate
  the change in individual income tax revenues
  under conventional assumptions
• Step 2. Use the macroeconomic model to estimate
  first-round economic and dynamic revenue
  effects
• Step 3. Use the alternative macroeconomic
  forecast to update incomes in the microsimulation
  model

12
Produced by Microsimulation and Macroeconomic
Models
Example of Iterative Differences Between
Individual Income Tax Revenue Estimates
13
Ongoing Model Development

• Increase the number and quality of links between
  the microsimulation and macroeconomic models
• Use the microsimulation model to improve
  estimates of behavioral effects
• Improve distributional analysis

14
END