How to measure local incomes

American Demographics, March, 1998 by Josh Galper

You need accurate estimates of a market's economic potential. Heres what you need to know to make an intelligent choice.

Take a good look at your organization's strategic plan, or your product's marketing plan. The chances are good that those plans use local-area income estimates. Gauging the economic power of neighborhoods and counties is essential to understanding any market's ability to buy, invest, donate, or pay taxes. Accurate income estimates are crucial to the success of new stores, new schools, and new products. Inaccurate estimates can lead to economic disaster at worst, and missed opportunities at best.

Income estimates in the U.S. come in many forms. Public and private data firms produce annual income estimates for many geographic areas, from individual city blocks to the nation as a whole. Some income estimates are for individuals; others are for households. Some give the average or mean income; others give a median. Some measure total pre-tax income; others give income post-tax, or money income excluding non-cash goods and services.

Moreover, different income estimates are based on different assumptions and methodologies. All income estimates are definitely not created equal, so proceed with caution.

Every income estimate depends on a model. Every model rests on a set of assumptions that specify how the model is built and for what purpose. Understanding these assumptions is the key to evaluating the validity and proper use of any estimate. For example, a model that incorporates all possible income before taxes will show a different mean income than a model that removes taxes from the equation. Also, per capita income estimates are derived from different sources than estimates of household income, so the two cannot be compared. If you fail to recognize differences like these, the result may be a costly mistake.

This article reviews several sources of county-level household income estimates. County-level data are often the smallest area available that covers the whole U.S., and household income is probably the most widely used type of estimate. Households are almost always based on the total population minus the "institutional population," a term that generally includes military personnel, college students, retirees in homes, prisoners, homeless persons, and any others who live in an area but not in households. Hence, it is not possible to divide total household income by population and get a per capita income figure.

WHO ESTIMATES, AND HOW

Although private data companies estimate U.S. income in different ways, all of them start with government data. The federal government is the only entity in the U.S. with the mandate, funding, and reach to manage the process of data collection in all 50 states over long periods of time. Several agencies contribute to this process, including the Census Bureau, the Internal Revenue Service (IRS), the Bureau of Labor Statistics (BLS), and the Bureau of Economic Analysis (BEA).

Each federal data source has its strengths and weaknesses. For example, many BLS data are based on the ES-202 form, a federally mandated quarterly report of employment submitted by non-exempt businesses and government agencies. These forms produce the best data available on the wages and salaries of workers by place of employment. Yet the ES-202 does not go to self-employed workers. It also skips 14 million other workers in a wide range of industries, including some types of agricultural workers, domestic workers, some government employees, and members of the Armed Forces.

The Internal Revenue Service calculates taxpayer income from annual tax returns, which may be the most truthful source of income data available. Unfortunately, a line-by-line accounting of returns is only available to the public at the state level, for reasons described by the IRS as political. A less complete accounting, known as "total money income," is available at the county level. Unfortunately, total money income is neither the "total income" taxpayers report on line 22 of Form 1040, nor the "adjusted gross income" they write on line 31. Rather, it is the sum of wages; total taxable interest income; alimony received; business income or loss; total pensions and annuities; profit or loss from rent; royalties; farm income or loss; unemployment compensation; and total Social Security benefits.

This may sound like a good way to describe the average person's economic resources, but it misses a lot. The current IRS definition of total money income does not include tax-exempt interest; dividends; taxable refunds of state and local income taxes; capital gains and losses; other gains and losses; the sale of business property; and taxable IRA distributions. These are such significant exclusions that the IRS total income measure cannot be viewed as a useful definition of income.

In non-census years, the most popular income estimates come from the Census Bureau's Current Population Survey (CPS). This survey measures both houshold and individual income at the state level, using questions about 18 kinds of income. The definition used for these surveys is called "money income received" or "total money income," although it should not be confused with the IRS definition. While the CPS is popular, the bureau's own research has shown that CPS data underreport certain kinds of income. The decennial census uses a similar methodology on smaller units of geography and contains similar errors.