In discussing income inequality, understand the nature of the data

Is there anything wrong with the richest 5 percent of families earning six times as much as the poorest 20 percent? What if the disparity were only half that? If there is something wrong with unequal incomes, should the government do something to make things more equal?

People frequently hold very strong opinions about these issues. Two studies released last week may add fuel to the flames in an election year.

This is a complicated and important issue. Moreover, it is one where innocent misinterpretation as well as fraudulent misrepresentation are common. Understanding the nuts and bolts of how the data are collected and tabulated is important for interpreting its implications or lack of them for public policy.

Given this complexity, I will examine income distribution in more than one column. This week I want to examine the nitty-gritty of income distribution statistics. In subsequent columns, I will discuss more fundamental philosophical issues underlying the debate.

The Economic Policy Institute, a Washington D.C.-based think tank closely linked to labor unions, conducted the first study. It looked at changes in the incomes and share of total national income received by different groups in society.

The data came from the Current Population Survey conducted by the Bureau of the Census. This survey, conducted every month, includes 50,000 households across the United States and is used to determine unemployment rates, among other things.

In March of each year, a supplemental set of questions is added to the CPS questionnaire, which ask about the household’s income during the previous year. These questions break income quite minutely by source. The responses to this survey are tabulated to provide the best annual information we have about household income.

The 10-year Census of population, which will be carried out this year, is intended to reach every household. It also asks about household income and serves as a benchmark to which the annual survey can be compared.

The second report released this week is the Federal Reserve’s 1998 Survey of Consumer Finances. The Fed carries out this survey only once every three years by conducting detailed interviews with a sample of about 4,300 families across the country.

Tax records supplement the survey for the families involved. While the Census Bureau’s CPS measures income, the Fed survey focuses on wealth, a distinction not understood by many people. Not surprisingly, both surveys found that Americans with higher incomes are seeing both their incomes and net worth increase at faster rates than are poorer ones. Here are some things for which to look out when reading them or other information about income distribution.

  • Is the unit of analysis individuals, families or households? These are not the same, and the rules for determining who makes up a separate household are often arcane. The Census Bureau publishes a 24-page set of Definitions and Explanations to support CPS reports. Why does this matter? Average income per household for the lowest income 20 percent of the population has not grown much in the past 20 years, but per-capita income for the group has grown somewhat more. Why? Because the number of U.S. households has grown faster that the population. Average household size has dropped as more single generation Xers live on their own and as the population ages with more one or two-person retiree households.
  • Be sure whether the subject is wealth or income. Income refers to money coming in during a specific period such as a year. Wealth is net worth, or the value of all assets minus the value of all liabilities. The two terms are not interchangeable. Wealth is distributed even less equally than income. If the subject is income, what income is excluded and excluded? Is only earned income–wages and salaries–included or are interest, dividends, rent, pension benefits and so forth all considered? Is the income before tax or after tax? Are government “transfer payments” such as Social Security included? If the subject is wealth, are nonfinancial assets such as homes and other real estate considered? Some studies of wealth exclude residential housing, which is the largest single asset owned by many households. They then conclude that U.S. wealth distribution is extremely skewed. What about the present value of whole life policies and defined-benefit pension plans in which household members are vested? The Fed study includes pretty much all assets, but does not impute an expected value to vested pensions or Social Security.
  • In studies of ownership of financial assets, find out what assets are included and excluded. If one looks only at stocks and bonds held directly by households, few households own any, and a very few control the bulk of the total value. But if one includes holdings through intermediaries such as 401k plans, mutual funds and insurance policies, many more households have a stake in financial markets and the value thereof is more equally distributed.
  • Recognize that not all households are the same. The poorest 20 percent of society has very little earned income because most households in it are either old or young–retirees or students. The retirees in this group have virtually no taxable income, but some have enough assets, such as homes, and nontaxable income to avoid privation. Students, on the other hand, have some earned income, but virtually none own housing or have other wealth. Most will spend only a few years in this low-income group.

© 2000 Edward Lotterman
Chanarambie Consulting, Inc.