Small, expected census inaccuracy cheats Americans

September 16, 2007
By Victor Landa

By its own admission, the U.S. Census Bureau doesn't have an accurate count of every person living in the United States .

Not that it's a bad thing, in the sense that it can't be avoided. It's a monumental task, after all, to count every person in the country. So there's an acceptable margin of error. Some people are going to be missed; that's just the way it is.

The census miscount is acceptable for practical reasons, or rather it's accepted because it's practical to let the number stand. It's too much trouble to correct the mistake, and the error is considered too small to make a difference. When you're considering several hundred millions of people, 1 million or 2 million missed is not that many.

But the miscount is impractical as well, unfair even, because of the consequences implied in getting the number wrong. Every estimate, projection and plan undertaken by the federal government, from budget planning to highway construction, education, health care and voting rights, uses the latest census figures as a starting point. So if the number is wrong at the foundation, the entire structure is flawed.

For instance, it's acceptable that the Latino population in the U.S. will be undercounted by 2.5 percent. In fact, it's expected. The white population is overcounted by the same percentage. The mistake always fluctuates within a few points of an ideal count.

The consequence of the 2.5 percent undercount (a 5 percent aggregate undercount when the white overcount is taken into account) is important to mention. The hardest people to count are those who live in dire poverty or constant motion. It's the same population that has the most to lose in terms of government attention.

The problem is compounded when the people who are most affected by a census undercount (the poor, the homeless, the undocumented) are the ones who are not being counted.

The next census is scheduled for 2010, so why talk about it now, in the middle of September 2007? We're going to need time to get it right; it's as simple as that.

The 2.5 percent undercount of Latinos in the U.S. is anticipated, but there is no adjustment planned to make up for the error. We've got less than 21/2 years to change that fact. A 2.5 percent undercut of services and opportunities is devastating in some areas of the country. And it's just as damaging when the census numbers are used, or misused, for political gain.

After the count in 2010, there will come a traditional realigning of voting districts in 2011. It's not codified to happen that way, but it has been the common practice because the numbers, although imperfect, are fresh. The thinking is that if you trace new voting districts with new numbers, your result will be true - or truer, if that's possible.

There was a protracted controversy when the Republicans in Congress undertook a nontraditional redistricting campaign. They didn't wait for the freshest numbers to come from the Decennial Census; they just up and did it, scoring impressive political gains as a result.

The Democrats were incensed, as would be expected. I imagine part of their indignation was that they didn't think of the ploy first. It was a bold attempt to solidify power for years in the future. Then the table shook under the Republican house of cards.

Time has passed, and now we're watching the Democrats build something on that same table. But while we wait to see what it's made of, we can't let the opportunity of the next census and redistricting pass us by, again.

Eight years ago, the Clinton administration promised to adjust the census results to account for the undercount of Latinos. But the promise was lost in the shuffle of the Bush administration's coming into the White House. This time around, there is no such promise, lost or found. And there's no clear front-runner for next year's presidential election.

That's just the background; the rest of the story still needs to come together. If any progress is to be made to ensure just and fair redistricting in 2011, including adjustments for undercounts, the work should begin now.

The first step is to ensure that the count will be as accurate as possible.