(Mar 08, 2011) T110000 Explanation of Tax Model
The Tax Policy Center Microsimulation Model
March 8, 2011
The Urban-Brookings Tax Policy Center’s large-scale microsimulation model produces revenue and distribution estimates of the U.S. federal tax system. The model is similar to those used by the Congressional Budget Office (CBO), the Joint Committee on Taxation (JCT), and the Treasury's Office of Tax Analysis (OTA).
The model is based on data from the 2004 public-use file (PUF) produced by the Statistics of Income (SOI) Division of the Internal Revenue Service (IRS). The PUF contains 150,047 records with detailed information from federal individual income tax returns filed in the 2004 calendar year. We attach additional information on demographics and sources of income that are not reported on tax returns through a constrained statistical match of the public-use file with the March 2005 Current Population Survey (CPS) of the U.S. Census Bureau. That match also generates a sample of individuals who do not file income tax returns (“nonfilers”). The dataset combining filers from the PUF (augmented by demographic and other information from the CPS) and non-filers from the CPS allows us to carry out distribution analysis on the entire population rather than just the segment that files individual income tax returns.
The tax model consists of two components: a statistical routine that “ages” or extrapolates the 2004 data to create a representative sample of both filers and non-filers for future years; and a detailed tax calculator that computes the individual income tax liability for all filers in the sample under current law and under alternative policy proposals. The calculator also computes the employee and employer shares of payroll taxes for Social Security and Medicare.
Aging and Extrapolation Process
For the years from 2005 to 2020, we “age” the data based on CBO forecasts and projections for growth in various types of income, IRS figures on the growth in the number of tax returns, and Bureau of the Census data on the composition of the population. We use actual 2005 through 2007 data when they are available. A two-step process produces a representative sample of the filing and non-filing population in years beyond 2004. We first inflate the dollar amounts for income, adjustments, deductions, and credits on each record by their appropriate forecasted per capita growth rates. We use the CBO’s forecast for per capita growth in major income sources such as wages, capital gains, and non-wage income (interest, dividends, social security income and others). We assume that most other items grow at CBO’s projected per capita personal income growth rate. In the second stage of the extrapolation, we adjust the weights on each record using a linear programming algorithm to ensure that the major income items, adjustments, and deductions match aggregate targets. For years beyond 2004, we do not target distributions for any item; wages and salaries, for example, grow at the same per capita rate regardless of income.
Individual Income Tax Calculator
Based on the extrapolated data set, we can simulate policy options using a detailed tax calculator that captures most features of the federal individual income tax system, including the alternative minimum tax (AMT). The model reflects major income tax legislation enacted through the end of 2010, including the tax compromise legislation signed into law in December of 2010 that temporarily extended provisions in the 2001 and 2003 tax acts, EGTRRA and JGTRRA. We model the following major provisions of those acts: changes in marginal tax rates, the new 10 percent tax bracket, credits for children and dependent care, itemized deduction limitations, personal exemption phase outs, the AMT, education incentives, retirement and pension measures, and the marriage penalty provisions, which increased the standard deduction, 15 percent bracket, and earned income tax credit for married couples. We also model JGTRRA’s changes to the taxation of qualified dividends and capital gains.
In our distribution tables, we assume that the burden of the individual income tax falls on the payer. CBO, JCT, and Treasury use the same assumption.
Payroll Tax Calculator
Using the extrapolated dataset we also calculate federal payroll taxes for Social Security and Medicare. One complication is that for married couples, our tax return data only provide information on combined earnings. Payroll taxes are based on individual earnings; this is important because the amount of earnings subject to the Social Security portion of payroll taxes is capped at about $100,000 (indexed annually). For married couples, we therefore rely on the split in wages observed on the CPS record to which the PUF record is matched in order to assign individual earnings.
In our distribution tables, we assume that the employee bears the burden of both the employer and employee portions of payroll taxes. This premise is widely accepted among economists; it is the same assumption made by CBO and JCT in its distribution tables. Treasury used the same assumption when they produced distribution tables that included all federal taxes.
Assigning Corporate Tax Burden to Individuals
The tax model distributes the burden of the corporate income tax to individuals so that it can be included in our distribution tables. The incidence of the corporate tax, however, is an unsettled theoretical issue. Although it only directly taxes corporate income, the corporate tax could be passed on to labor in the form of lower wages, to consumers in the form of higher prices, or to the owners of all types of capital since the after-tax rate of return on corporate equity will affect the after-tax rate of return on other types of capital.
We assume that the burden of the corporate tax falls on all capital income. CBO uses the same assumption in its distributional analysis; Treasury formerly used this assumption when they provided complete distribution tables. The JCT does not distribute the corporate tax because of the uncertainty surrounding its incidence. Our methodology most closely resembles that of CBO.
We first determine each return’s share of aggregate capital income, defined as taxable and tax-exempt interest, dividends, realized capital gains, and net income from rents, royalties, and estates or trusts. Each record’s share of the corporate tax burden is then calculated as total corporate tax liability multiplied by its share of aggregate capital income. Thus a return with 0.05 percent of aggregate capital income is assigned 0.05 percent of aggregate corporate tax liability. We rely on CBO for our projections of corporate tax liability and on JCT for changes in baseline corporate tax liability due to tax proposals.
Because the income tax data in our model contain no direct information about wealth holding, we rely on information from the Survey of Consumer Finances (SCF) to develop imputations of assets and liabilities. Specifically, we impute asset items and liabilities to each record in the income-tax file based on regressions of those wealth components against explanatory variables that exist on both the SCF and SOI datasets. To mitigate the problem of the SCF’s small sample size —it contains fewer than 5,000 observations—we pool data from the 2001 and 2004 surveys. In addition to roughly doubling the sample size, combining data from the two years smoothes out some of the temporal variation in asset values. We then calibrate the imputed number of individuals owning each type of asset (and liability) and their aggregate values to match SCF totals, augmented by the net worth of the Forbes 400. We further adjust the imputed distribution of each asset and liability by income class to more closely resemble those reported in the SCF.
We assign values for most estate tax deductions and credits based on averages calculated on the SOI estate tax data. Our estate tax calculator then determines estate tax liability for each record in the database, based on the values for gross estate, deductions, and credits and the relevant estate tax rates and brackets. Finally we calculate each record's expected value of gross estate and net estate tax liability by multiplying by appropriate mortality rates. We employ a linear programming algorithm to reweight the records to ensure that our baseline estimates of the distribution and aggregate values for gross estate and its components match the most recent published estate tax data from SOI.
In our distribution tables, we assume the burden of the estate tax is borne by decedents, the same assumption used in the past by Treasury. Neither CBO nor JCT include the estate tax in their distribution tables.
Recent Model Enhancements
Within the last few years, the Tax Policy Center completed a major update of its microsimulation model of the federal tax system and enhanced the model in several ways. We incorporated more recent data by updating the tax model database from the 2001 to the 2004 public-use file (PUF) of tax returns produced by the IRS. We also performed a new statistical match of the PUF with the March 2005 Current Population Survey. We updated the tax model's estate tax module to incorporate the latest data on estate tax filers from the Statistics of Income Division of the IRS and the retirement savings module to be consistent with the new 2004 data. We also expanded the retirement module to allow us to model the revenue and distributional implications of implementing automatic enrollment in IRAs and 401(k) retirement plans. This latest version of the tax model also includes refined imputations of itemized deductions, such as charitable contributions and home mortgage interest, for “nonitemizers” —people who claim only the standard deduction on their tax return. These imputations allow us to model the distribution and revenue implications of proposals to replace certain credits with deductions.
The latest version of our microsimulation model also includes a completely overhauled and expanded education module. We use data from the October 2003 and October 2004 CPS, as well as the National Postsecondary Student Aid Study (NPSAS) to impute student status, characteristics, and education expenditures onto the tax model database. This allows us to model current tax incentives for education—such as the HOPE and Lifetime Learning Credits and the deduction for higher education expenses—as well as to examine the revenue and distributional implications of combining or modifying these tax programs. We can also model current spending programs such as Pell Grants, and examine the revenue and distributional effects of changes to the Pell Grant rules.
We have also made several improvements that allow us to model a variety of indirect taxes, including excise taxes, broad-based consumption taxes (e.g. a value-added tax or VAT), and environmental taxes. Using data from the Consumer Expenditure Survey (CEX), the Medical Expenditure Panel Survey (MEPS), and the American Housing Survey (AHS), we produce detailed estimates of the consumption expenditures of individuals in the tax model database. We also use the Urban Institute’s DYNASIM model to estimate the amount of future consumption financed out of current wealth, which allows us to analyze transitional issues for options that move the tax system from an income base to a consumption base. This work gives the TPC the ability to estimate the distributional impact of hybrid income-consumption tax systems and other comprehensive reform options, such as the plans endorsed by the President’s Advisory Panel for Federal Tax Reform and the Bipartisan Policy Center’s Debt Reduction Task Force.
Finally, using data from the Medical Expenditure Panel Survey (MEPS), we collaborated with the Health Policy Center to impute details of health insurance eligibility, coverage, and medical expenses in the tax model database. With a modification to the tax calculator, the imputed information allows us to analyze policies that change the tax treatment of health insurance such as repealing or limiting the currently unlimited exclusion of employer-provided health insurance. We are in the process of revising the imputations and the tax calculator to account for provisions in the Patient Protection and Affordable Care Act and modifications in subsequent legislation such as the Medicare and Medicaid Extenders Act of 2010.
Planned Model Enhancements
During 2011, TPC will undertake another major overhaul of our microsimulation model. In addition to incorporating the latest economic and demographic forecasts from CBO, IRS, Census, and other sources, we will update the tax model database to the 2006 public-use file of tax returns produced by the IRS. We will also refine and improve our statistical matching techniques and perform a constrained statistical match of the PUF with the March 2007 Current Population Survey. We also plan to improve our estate tax module using data from SOI that links estate tax returns with the prior-year income tax returns filed by decedents. We will also begin including excise taxes in our distribution tables based on the consumption imputations described above. The addition of excise taxes and customs duties to our distribution tables means that we will incorporate virtually all federal taxes in our distribution tables. Finally, as part of the update process, we will integrate our long-run microsimulation tax model with the TPC model, which will give us the capability to carry out revenue and distribution analysis through the year 2080.