Redistricting Matters: Small changes in the boundary lines mean huge electoral effects

by Peter Wagner, July 26, 2004 Link

Our project, Prisoners of the Census, addresses the political impact of the Census Bureau’s policy of counting prisoners where they are incarcerated, not at home. In some cases, a large population is transferred to prison towns, but the potential impact of prisoner miscount can be large even when the number of prisoners transferred is small.

This example will illustrate how small changes in district boundary lines - even without errors in the underlying Census data — can determine which party or group controls the government. If the underlying Census data is distorted, the potential impact could be far larger.

For this exercise, we’ll use an imaginary territory with 25 residents, 15 “O” voters and 10 “X” voters. (See top image.) (You can assign the X’s and O’s political parties, if that makes it easier.)

In a territory-wide race for Governor or Mayor where the whole territory participates, the candidate backed by the O’s would probably win. But what about elections for a five-seat legislature, where each legislator represents an individual district with one-fifth of the territory’s population? The answer depends on how the districts are divided.

In real life, district boundary lines are redrawn after each federal Census, usually by the political party with majority power. Our example assumes that a Census has just taken place.

The remaining images on this page illustrate three possible ways to divide the districts and their resulting outcomes. Compare the next two images, labeled “Packing” and “Cracking.” The differences in the district boundaries appear subtle, but results are radically different. The final image reflects a fairer way to draw the districts: there is still an “O” majority, but the X’s and O’s are proportionally represented throughout the territory.

In real life, political and ethnic populations are not evenly distributed throughout a territory. In addition, legislators have considerable flexibility in how the boundary lines can be drawn. For example, legislators have some discretion to make districts deviate slightly from population requirements. This deviation from the principle of strict population equality makes additional outcomes possible.

Furthermore, when the underlying Census data is faulty — for example, data that counts tens of thousands of urban prisoners in rural towns - redistricting is even less likely to result in a democratic reflection of the territory.

More information and credits:

For how changes in district size can effect outcomes in other electoral districts, see Hanging in the Balance about New York State Senate district 34, which straddles the border of NYC’s Bronx county and Westchester county in such a way as to use Bronx residents to create a Westchester district.

WXXI (Rochester, NY) and NYCitizens.org has a complex Redistricting Imaginary State Simulation where you can balance racial/ethnic factors against political affiliations for a variety of results. The website is a role playing exercise where you can be an advocate for a certain interest group and experiment with how to best meet “your” objective.

This article is inspired by and draws some text from the Center for Voting and Democracy’s interactive Redistricting Wheel, created by Matthew Pierce. Any mistakes on this page are ours. We welcome feedback.

See our Resources page for more information about redistricting and the census generally.

Texas bill shows states could correct Census counts of prisoners

by Peter Wagner, July 19, 2004 Link

As I wrote in May, there are 3 ways to fix how the U.S. Census counts prisoners:

1. The Census could change its methodology
2. States could adjust the counts after the fact
3. States can ignore the federal Census and bring back their state Censuses.

The first option is the best, but the others are far more practical than they may sound at first glance. One state already does something quite similar: Kansas adjusts how students and the military are counted.

In 2001, Texas Representative Harold Dutton introduced a bill to restore Texas prisoners to their home addresses prior to redistricting. Although the bill was not ultimately successful, the bill was approved by the Elections committee and does illustrate one approach that could be taken by states if the Census Bureau does not change its policy.

The bill would have required the operators of all public and private prisons in Texas to submit to the Texas Controller a report containing the name, age, gender, ethnicity and pre-incarceration address of each person counted in the Census as a resident of the prison. The Controller would then deduct these persons from the Census tracts with the prisons and restore them to their home Census tracts.

The bill is notable for two other facts:

1. The bill puts state and private agencies on notice that they must collect this data. (While most state Departments of Corrections keep extremely detailed demographic records, some, such as Maryland, do not.)
2. Unlike the approach taken in Kansas with students and the military, out-of-state prisoners are retained at their prison addresses rather than being deleted from the redistricting dataset. Representative Dutton’s approach seems to be fairer than the Kansas approach at ensuring that everyone is represented somewhere.

It is certainly possible for states to accept U.S. Census data, modify it to correct for how special populations are counted in the Census, and then rely on that modified data in their redistricting. But the ideal solution would be for the U.S. Census to take into account the huge growth in incarceration as well as the modern requirements of “one person one vote” and start counting people in prison as residents not of the prison town but of their home communities.

Source: Thanks to Will Harrell for drawing this bill to my attention.

Census Bureau Policy Costs Ohio’s Cities Political Power

by Peter Wagner, July 12, 2004 Link

As described in last week’s column, how the Census Bureau counts prisoners undercounts Ohio’s urban areas while boosting the population of the rural areas that host the prisons. Also skewed is how much representation each region receives in the state legislature, because Ohio relies on Census Bureau data to redraw its state legislative boundaries

States are required to redraw their legislative boundaries each decade so that each will contain the same number of people as required by the 14th Amendment’s One Person One Vote principle. Equally sized districts ensure that each resident has an equal access to government regardless of where she or he lives.

The Census counts everyone including people who can’t vote such as prisoners and children. But children are at least a part of the surrounding community and share some common interests with it. Children can with some confidence rely on their neighboring adults to represent their interests. But prison communities are often very closely aligned with the prison industry and are likely to be quite dissimilar to the communities that the prisoners came from.

So while prisoners are barred from voting for or against the legislator that “represents” them while they are incarcerated, Ohio restores a prisoner’s right to vote on the day that he or she is released. But that is also the same day that the former prisoner will be getting on a bus to leave the prison district and return back home.

Including disenfranchised non-resident prisoners as population for purposes of redistricting creates prison districts with substantially fewer constituents than elsewhere. The real residents of the prison district have more access to their legislator than other state residents.

In Ohio, a House district is supposed to contain 114,678 people. But because of how the Census Bureau counts the incarcerated population, the drafters of legislative districts should not assume that everyone is a resident of the place where they are counted. District 85 (in Ross, Pickaway and Fayette counties), currently represented by John Schlichter, is 8.92% prisoners. These disenfranchised prisoners are overwhelmingly from homes outside the district, meaning that the actual population of the district is very small. Every group of 91 residents in District 85 gets as much of a say over state affairs as 100 people in Columbus or Cleveland.

This is precisely the situation the One Person One Vote rule was designed to correct.

Source: Peter Wagner and Rose Heyer Importing Constituents: Prisoners and Political Clout in Ohio