Association for Public Policy Analysis & Management

Income data are available from many sources, including surveys, tax records, and administrative data from government programs. When used alone, each of these sources has important strengths and major limitations, and none of them independently offers a comprehensive and accurate measure of income. While previous research has combined some sources of income to address these drawbacks, the scope and generalizability of these efforts are limited. To address this gap, we are developing the Comprehensive Income Dataset (CID), a restricted micro-level dataset with income and demographic characteristics that provides an accurate and comprehensive measure of income for the population of U.S. families and households.

The CID focuses on four household surveys: the Current Population Survey, the American Community Survey, the Survey of Income and Program Participation, and the Consumer Expenditure Survey. To these surveys we will initially link tax return data on income and its sources and calculated tax credits from IRS Forms 1040 and W-2, and retirement distributions from Form 1099-R. We also integrate administrative data from key government programs, including (but not limited to) Social Security, SSI, housing assistance, SNAP, Public Assistance, Medicare, Medicaid, WIC, and LIHEAP.

We envision a number of important uses of the CID. The first use is the power of the CID to improve Census Bureau surveys and statistics, through providing additional information on the nature of measurement error, improving imputation methods, and adding new or more accurate variables to surveys. Second, the CID can be used to improve the administration of taxes and by policymakers to forecast and simulate changes in programs and taxes. Finally, the CID would be a valuable tool for program evaluation and research, as it is far and away the best source to analyze poverty, inequality, mobility, and the distributional consequences of government transfers and taxes, as well as numerous additional research topics.