Pants on Fire!
"An audit of the last census found that over 370,000 people living in Texas were not counted, and our state lost an estimated $1 billion in federal funds as a result."

Silvestre Reyes on Thursday, March 11th, 2010 in an e-mail blast

Rep. Silvestre Reyes says census audit found more than 370,000 Texans were missed by 2000 census, costing $1 billion in federal aid

Urging constituents to return their filled-out 2010 census surveys to the federal government, U.S Rep. Silvestre Reyes refers to a dramatic blast from the past.

The El Paso Democrat states in a March 11 e-mail: “An audit of the last census found that over 370,000 people living in Texas were not counted, and our state lost an estimated $1 billion in federal funds as a result.”

Goodness. (And where did that census form disappear to?)

Does Reyes get his big numbers right? And if there was an undercount, did it cause Texas to miss out on a big chunk of federal aid?

For starters, we found no such audit, and Reyes’ office didn’t provide back-up for his statement. However, a 2001 PriceWaterhouseCoopers study prepared for the U.S. Census Bureau Monitoring Board, an entity that no longer exists, stated that nationally, the 2000 census failed to count some 3.3 million residents, including 373,567 Texas residents.

The undercount figures were attributed to the Census Bureau and Eugene Ericksen, a Temple University sociologist.

Some perspective: Reyes’ recent e-mail harkens to concerns aired in 2000, especially by Democrats, that without an adjustment the census would fail to account for many residents, mainly racial minorities and renters. The bureau did not adjust its count of the nation’s population.

The 2001 report said the undercount could result in a federal funding loss of $4 billion across 31 states (including Texas) and the District of Columbia. It raised the prospect of California and Texas losing the most, $1.5 billion and $1 billion, respectively, from fiscal 2002 to 2012.

A press release accompanying the report included this warning from Peter Merrill of PricewaterhouseCoopers' National Economic Consulting group: “Inaccuracies in the census can cause federal funds to be distributed in a way that is not fully consistent with congressional intent. The census undercount not only misallocates funds among jurisdictions, it also causes a net loss to the states of funds from federal entitlement programs. Compounding the problem, many state-funded grant programs to localities also rely on census counts.”

Reyes isn't alone in using figures from the 2001 report to argue the importance of participation in the 2010 census. Since last summer, Texas newspapers including the Austin American-Statesman, Dallas Morning News, Fort Worth Star-Telegram and San Antonio Express-News have mentioned figures from the report. 

Does the report still hold water?

Our finding: Like a sunk boat.

About two months after the PriceWaterhouseCoopers report came out, the Census Bureau said it would reassess how accurately it had counted the nation's residents in 2000. And in 2003, the bureau announced new estimates for what the 2000 counts for each state should have been.

And there was another twist: Instead of undercounting the nation's population, the bureau concluded it had over-counted by about 1.3 million people.

David Whitford, chief of the bureau’s Decennial Statistical Studies Division, told us that the bureau ultimately realized that undercounts nationally were more than offset by double-counting of many individuals and other errors. One example: counting students away at college while also recording them as part of their parents' household back home.

Some population groups were affected more than others. The bureau said it had overcounted whites, Asians, American Indians on reservations and young children, the Associated Press reported, while many blacks and Hispanics were missed.

There was still a small net undercount (by 10,245 people, or .05 percent) in Texas, the bureau said. The census count of Texans should have been 20,862,065, the bureau estimated.

Where does this leave the PriceWaterhouseCoopers study? Professor Ericksen told us that by repudiating its initial undercount estimates, the bureau effectively invalidated both the population figures and federal aid estimates.

And Merrill of PriceWaterhouseCoopers cautioned against using its 2001 report for another reason: "It is certainly out of date... Our firm is not telling anyone that the analysis done in 2001 is relevant for today."

Vincent Perez, Reyes' press secretary, said he's been unable to get the Census Bureau to confirm that it backed off the estimate that more than 370,000 Texas residents weren't counted in the 2000 census.

How does Rep. Reyes’ statement stand up?

Reyes incorrectly refers to the consultant’s projections as an audit of the census; it was a report.

More importantly, Reyes relies -- as have other entities -- on an outdated report based on numbers the Census Bureau has said were flawed. The bureau long ago yanked the statistical rug out from under the claim that more than 370,000 Texas residents were not counted and that Texas stood to miss out on $1 billion in aid as a result.

This isn't a close call. We rate Reyes' statement as Pants on Fire.