Answer: When he or she is counted as part of a legislative district, despite not technically being a resident and not being able to vote at all.
Let me explain. Due to a little-known rule of the U.S. Census, persons incarcerated in state or federal prisons as of April 1 of a Census year are counted as living in the legislative district where the prison is located, rather than, say, the legislative district where they lived prior to incarceration and to which they presumably will return upon release from prison.
While this method of counting has minimal impact on federal-level Congressional redistricting, it’s not difficult to see how it could affect state-level legislative redistricting, both in terms of district sizes and in vote dilution.
Case in point, House District 11 in Jefferson County contains the Varner, Cummins, Maximum Security, and Tucker facilities of the Arkansas Department of Correction. When “guests” of those cross-bar hotels are counted as residents of District 11, they make up 16% of the population of that district. Meaning, then, that every 84 actual, non-incarcerated residents of District 11 have the same voting influence as 100 people in a district that does not contain a prison facility.
On top of this problem, certain prisonless counties, due to higher-than-average rates of incarceration, are even more under-counted in the Census figures than other counties that do not have prisons. For example, Sebastian County is home to 125,744 people (4.3% of the population), yet 8.3% of the inmates in the ADC (560 prisoners, based on 2010 numbers) come from Sebastian County. Likewise, Crittenden County has 1.8% of the state population, but it accounts for 3.9% of the prison population. In both counties, the failure to count those inmates as residents of those counties results in state legislative districts that are larger than they would otherwise have to be and they receive an incorrect amount of monies that are distributed based on population.
The impact on vote dilution is even more pronounced at the the local level. Jackson County bases its Justice of the Peace districts on the Census numbers that include the Grimes-McPherson Correctional Facility. As a result, nearly two-thirds (64%) of the JP District 5 is composed of incarcerated persons, and every 36 actual residents of that district have as much influence as 100 Jackson County residents in the other districts.
Most troubling, perhaps, the 1,551 inmates at the Federal Correctional Center in St. Francis County (Forrest City) will account for 5.3% of the population of whatever House district they are placed in through redistricting, despite the fact that many of the inmates are not even Arkansas residents. That is, over 1,500 people, the vast majority of whom have never voted in Arkansas, will never vote in Arkansas, and will leave Arkansas as soon as they are released, will be counted as “constituents” of an Arkansas legislator.
What can be done? Ideally, the time would be taken to account for where the inmates are actually domiciled and maintain a residence. After all, Arkansas Code Annotated § 7-5-201(b), entitled Voter Qualification, provides:
(b) “Voting residence” shall be a voter’s domicile and shall be governed by the following provisions:
(1) The domicile of a person is that place in which his or her habitation is fixed to which he or she has the intention to return whenever he or she is absent;
(2) A change of domicile is made only by the act of abandonment, joined with the intent to remain in another place. A person can have only one (1) domicile at any given time;
(3) A person does not lose his or her domicile if he or she temporarily leaves his or her home and goes to another country, state, or place in this state with the intent of returning;
(4) The place where a person’s family resides is presumed to be his or her place of domicile, but a person may acquire a separate residence if he or she takes another abode with the intention of remaining there;
Using these rules, temporary incarceration with the intent to return to one’s home upon release would seemingly not cause forfeiture of legal residence. (The equation would change slightly for prisoners who were in for life.) Of course, because those in prison are not eligible to vote, it could reasonably be argued that the section on voter qualification is irrelevant.
Under Article 19, section 7, of the Arkansas Constitution, it might not matter if § 7-5-201 applies. That section states:
Absence on business of the State, or of the United States, or on a visit, or on necessary private business, shall not cause a forfeiture of residence once obtained.
While this provision does not specifically mention incarceration, one could argue that being locked up in state (or federal) prison is “business of the state” (or “of the United States”). More generally, the general intent of this provision — that being away from your residence because of something you are required to do won’t change your legal residence — could be applied to incarceration.
Just for the sake of argument, let’s assume that neither of these laws is dispositive of the issue. What can be done in that case? Well, if local solutions are any indication, the best approach might just be to subtract the prison populations from the Census numbers for each county. Lincoln County was more or less forced to do this when drawing their JP districts; otherwise, they would have had two districts for which there were no actual voters. Similarly, Lee County had to ignore the prisoner count when drawing districts for its county board of supervisors in order to avoid having a district with no voters. Finally, Forrest City, when dealing with the population at the federal prison, decided to ignore the prison population when drawing city wards; had they counted the inmates, one of the wards would have been over 60% incarcerated persons.
At the end of the day, whether the answer is counting the inmates at their county of residence or simply ignoring them (because, after all, they can’t actually vote while they are in), it’s pretty clear that something needs to be done here. There was a rule change by the Census for 2010 that allowed Census takers to count inmates at their county of residence, but nothing I’ve seen thus far suggests that this option was followed in Arkansas. In fact, while I would LOVE for someone to show me otherwise, given how readily Arkansas has accepted the population counts that include prison populations, I would literally be shocked if they were not currently included in all the numbers you’ve seen.
Unfortunately, I haven’t heard or read too many people talking about this issue, so the odds of anything being done in this redistricting are somewhere between slim and none.