Skip Navigation

Effective Federal Tax Rates, 1979-1997 October 2001

On March 31, 2000, the Tax Analysis Division of the Congressional Budget Office (CBO) brought together 15 experts on distributional issues in a one-day conference on conducting distributional analyses. The wide-ranging discussion reached substantial agreement on how best to classify people within the income distribution and assess their effective tax rates. Rather than concentrate on the ideal way to carry out distributional analysis, many of the participants focused instead on how to do so in the face of limitations on the available data that preclude that ideal. All attendees agreed that compromise was required to produce meaningful results. Those compromises demand that findings be qualified but do not render them invalid.

Three major messages came out of the conference. First, the distributional statistics that CBO produces are both important and useful, and CBO should continue to produce them regularly. Second, CBO should investigate changes in and extensions of its current measures to determine the impact of altering its basic methodology. Third, rather than replace the measures it currently uses with a single alternative, CBO should produce statistics using a variety of measures so that users can choose the measures they feel are most appropriate.

The conference posed four specific questions:

• What is the appropriate unit of analysis for distributional studies--the family, the household, the tax unit, or some other group?
  
  
• What should count as income? Should it be cash only or a broader measure?
  
  
• What is the appropriate period over which to measure tax burdens? Should it be one year or a longer period?
  
  
• Who bears the burden of various federal taxes?

Decisions made in carrying out the analysis reported in this study reflect the discussion of those four questions at the conference. Much of CBO's analysis compares the effects of different answers to the questions to evaluate the importance of particular decisions.
 

Unit of Analysis

Participants agreed that which unit was chosen for analysis was probably less important than making the unit consistent throughout a study. Participants expressed a preference for using households--groups of people, not necessarily related to each other, living together in a single housing unit--on the grounds that households probably better represent consumption units than do tax filers (generally a single taxpayer or a couple plus dependents) or families (related people living together).
 

Definition of Income

There was general agreement that CBO should investigate the effects of using a broader measure of income than one comprising only cash earnings in order to rank people in the income distribution. Participants nearly unanimously approved including noncash items such as food stamps, school meals, and housing and energy assistance, all measured at their market value. No consensus was reached about counting the value of health insurance and health programs such as Medicare and Medicaid. The major objection was that the market value of such benefits almost certainly overstates their worth to low-income households. One proposed solution called for a dichotomous classification of households based on whether they had any health coverage, either through employment, through a government program, or through privately purchased insurance. The obvious difficulty with that approach is that it creates two distributions that cannot be readily combined. Conference attendees expressed no interest in including imputed income from durable property, such as owner-occupied housing.

There was little discussion of how to adjust incomes to account for differences among households. Participants seemed to agree that CBO's adjustment for differences in family size improved distributional rankings. No strong feelings surfaced about the different work-related prices or costs that households face.

Two complaints were raised about CBO's distributional rankings. First, some participants noted that Members of Congress and other users of the analyses have difficulty thinking about quintile rankings and incomes adjusted for differences in family size. Those kinds of users apparently prefer stories about individual families rather than statistical averages for amorphous groups of people. A second and related issue was that people seeing CBO's tables cannot easily find their place in the income distribution. Their desire to do so argues for CBO's providing tables not only for adjusted income rankings but also for distributions based on dollar income. At the same time, participants cautioned, any analysis including such tables should clearly explain the important differences between rankings based on cash income and rankings based on adjusted income.
 

Measurement Period

Attendees generally favored measuring the incidence, or burden, of taxes over a person's lifetime rather than annually because year-to-year variations in income--both taxable and tax-exempt--can result in misleading statistics about effective tax rates across the income distribution. Participants agreed, however, that limitations in the available data preclude a lifetime analysis and that analyzing shorter periods is the only current option. Furthermore, because people think of their tax liabilities in annual terms, CBO's focusing on yearly data makes it simpler both to compute incidence and to explain it to readers.
 

Tax Incidence

The assumptions about incidence underlying CBO's previous tax analyses evoked no serious complaints from participants. Their major suggestion was that CBO include more sources of revenue--particularly estate and gift taxes and customs duties--in its analyses. Attendees acknowledged, however, that limitations in the data make including more sources difficult.
 

Conference Participants

Individuals from a variety of organizations and government agencies attended CBO's distributional conference, including:

Tom Barthold, Joint Committee on Taxation
Julie-Anne Cronin, Office of Tax Analysis, Department of the Treasury
Al Davis, House Committee on Ways and Means
Jim Horney, Senate Committee on the Budget (at the time of the conference, Center for Budget and Policy Priorities)
Rick Kasten, CBO analyst, now retired
Arthur Kenneckell, Federal Reserve Board
Tom Koerner, Joint Committee on Taxation
Marvin Kosters, American Enterprise Institute
Jim Nunns, Office of Tax Analysis, Department of the Treasury
Lindy Paull, Joint Committee on Taxation
Wendell Primus, Center for Budget and Policy Priorities
Ralph Rector, Heritage Foundation
Frank Sammartino, Joint Economic Committee (at the time of the conference, Urban Institute)
Bernie Schmitt, Joint Committee on Taxation
Joyce Zickler, Federal Reserve Board

Participants from the Congressional Budget Office included:

Dan Crippen, Director
Barry Anderson, Deputy Director
Steve Lieberman, Executive Associate Director
Tom Woodward, Assistant Director, Tax Analysis Division
Roberton Williams, Deputy Assistant Director, Tax Analysis Division
Mark Booth, Senior Analyst--Revenue Estimation, Tax Analysis Division
David Weiner, Senior Analyst--Modeling, Tax Analysis Division

---