UConn Advance - February 6, 2006 - Census income stats misleading, study finds

The data used to determine how much aid towns receive from state government is misleading, because it fails to include capital gains income, according to a study just released by the University’s Center for Population Research (CPR).

The new study, “How Census Income Estimates Provide Misleading Statistics on Personal Income for Connecticut Towns,” comes as increasing numbers of local officials are complaining that state government is not providing enough money to cities and towns to pay for services, forcing them to raise local property taxes on homeowners and businesses.

“Our federal and state governments lack an all-inclusive measurement of personal income that is used by all,” says Wayne Villemez, director of the Center.

“There is no agreement on what constitutes personal income,” adds Villemez, a professor of sociology. “This becomes a tangible concern when we see the magnitude of capital gains income that goes uncounted.”

In Connecticut, as in other states with personal income extremes, capital gains have become a large component of income in several towns. Yet currently, the only source of information on personal income for Connecticut towns is the U.S. Census Bureau’s Decennial Census.

Census data, updated every 10 years, relies on estimates from self-reported data to calculate household income for towns. The other main sources of personal income data – such as the Bureau of Economic Analysis, the Internal Revenue Service, and the state Department of Revenue Services – do not break down the data by town, even though such data is vitally important for determining town funding.

The CPR study attempts to fill this void. It analyzes and compares personal income data by town from various sources, synthesizes these findings, and raises important concerns for public policy decision-making.

“The strength of our study,” says Villemez, “is that it reveals just how misleading self-reported income data can be for extreme low- and high-income geographic areas.”

For example, the study shows that per capita income in Connecticut for the 1999 tax year was $28,766 based on Census 2000 estimates, but increased to $31,125 based on 1999 federal tax returns (IRS Federal Adjusted Gross Income).

According to Census 2000 estimates, the per capita income gap between the lowest and highest income towns in Connecticut in 1999 was $69,687 per person. Using IRS Federal Adjusted Gross Income data, however, the study shows that the per capita income gap between the lowest- and highest-income towns was $145,982 in 1999 – more than double the Census 2000 estimates.

The study also points out that Census 2000 estimates have no towns with an income of $100,000 or more per person, but the IRS/CPR data show six Connecticut towns with incomes of $100,000 or more per person. Similarly, Census 2000 estimates show 12 towns with incomes less than $20,000 per person, whereas the IRS/CPR data shows 28 towns with incomes less than $20,000 per person.

For the poorest urban areas, the discrepancy between personal income statistics is not as extreme as for the richer towns, Villemez says: “We knew that Bridgeport and Hartford were the poorest towns in Connecticut.”

The study shows, however, that “the bottom of the income scale is lower than previously thought,” he says. “Basically, it looks like either the poor are poorer, or there could be more people in poverty than what is being counted. What our report suggests is the larger extent of low-income households in rural areas.”

The study also points out that income data from the 2000 Census was not incorporated into education funding formulas until the 2005 fiscal year. This postponed an accounting of the drastically uneven growth in town personal income during the 1990s.

The study concludes that in Connecticut, Census 2000 income estimates produce highs that are too low and lows that are too high. In effect, the Census estimates artificially understate the gap between the wealthiest and poorest Connecticut towns. Census 2000 did not count, for example, $11.6 billion in capital gains income in Fairfield County in 1999.

“Some may argue that the concentration of capital gains in Fairfield County is an anomaly,” Villemez says. “But is it? In reality, we don’t know.”

He maintains that issues raised by the CPR study are of concern beyond Connecticut and apply to all areas of the country where low-income or high-income households are geographically concentrated. The study also includes examples of the disparity in reported personal income between low- and high-income communities in Massachusetts, New York, New Jersey, Texas, Florida, California, Virginia, Maryland, Illinois, and the District of Columbia.

“We may be seeing an increasing reliance on capital gains as a source of personal income due to the aging of the Baby Boomer generation and its dependence on IRAs, 401Ks, etc. for retirement income,” Villemez says.

“We do not know for certain just how widespread these problems are, because we do not have a tangible and unquestionable comprehensive measurement of personal income. If we had such a measurement then we could, with some certainty, identify geographic areas where personal income estimates are misleading.”

Census income stats misleading, study finds

by David Bauman - February 6, 2006