Congressional Apportionment Watch

A man with one watch always knows what time it is. A man with two watches is never quite sure.

With that in mind, here is some good news for New Yorkers who may have recently come to believe that they will lose a second congressional seat to Florida in the 2010 apportionment of congressional seats among the states, and some bad news for Texans who thought they would gain four seats.

First, a little back story.

Every ten years, the congressional seats are reallocated among the states based on their population revealed by the national census. The Census Bureau is scheduled to report the April 1, 2010, state population numbers this December. From this, an apportionment formula is used to assign congressional districts among the states. While the exact numbers are a mystery until the Census Bureau announces the official numbers, we can make educated guesses using various government and privately produced population estimates. All signs point towards a continued shift in population from the northeast to the Sun Belt states.

Election Data Services is an influential organization that produces apportionment predictions. Using data solely from the Census Bureau, New York is already expected to lose one seat. A recently released EDS report analyzing data from ESRI -- a demographic research company that mashes up Census Bureau population estimates with housing starts and U.S. Post Office change of address lists to produce their population estimates -- generated a number of news stories and created heartburn for New Yorkers. This new report predicts New York losing a second seat, apparently to Florida, which now gains two. One other noteworthy change in the fight between Minnesota and Missouri to retain a seat, Minnesota now would keep all their seats and Missouri would lose one.

There is an important flaw to the EDS study that a close read of the report reveals, and that affects its conclusions.

The EDS report uses ESRI's July 1, 2010, population estimates, while the population numbers that will be used to conduct congressional apportionment are from April 1, 2010. Three months may not seem like a big deal; however, populations in states like Florida and Texas are growing very quickly. We can turn the clock back on the ESRI data by linearly interpolating the July 1, 2009, and July 1, 2010, estimates to produce an April 1, 2010, estimate. When we do, New York no longer loses a second seat. In fact, these reconfigured numbers suggest that Florida still gains a second seat at the expense of Texas, which was expected to gain four seats but now only gains three. (This adjustment also continues to have Missouri losing one seat and Minnesota losing none.)

If there is one thing we know about estimates, it is that they are almost always wrong. Further, states on the knife's edge will gain or lose a congressional seat by thousands of persons -- well within the uncertainty of these population estimates. So, take this analysis and the EDS analysis with the appropriate large grain of salt. We will not know with certainty how the next congressional apportionment will shift political power until December, when the Census Bureau officially announces the states' April 1, 2010, population counts. Until then, a number of states will anxiously await to discover how well they did in convincing their hard-to-count populations to respond to the census.