The Supreme Court agreed to hear Evenwel v. Abbott, a challenge to Texas's state legislative redistricting plan, which draws districts on the basis of total population and not on a voter-based metrics (such as citizen voting-age population). There are important, fundamental constitutional problems with this case, which I've raised briefly elsewhere. But I want to focus on the impact of the case if petitioners succeed--because, of course, sadly, few want to spend time on what the Constitution says, or how theories of representative government change, and most just want to figure out what's in it for which groups.

Ostensibly, the case, narrowly viewed, would shift power away from voters residing in districts with a large number of non-citizens and toward districts with a large number of citizens. That's, at least, the superficial reason for the challenge in Texas. Given that ostensible fact that disproportionately more Latino voters reside in districts with non-citizens, that would theoretically mean a dilution of their political power if other voters were added to their districts. And given that Latino voters tend to favor Democratic candidates, this would tend to advantage Republicans.

This is probably, at best, partially right, and, at worst, mostly wrong.

Recall that Latino voters often qualify for majority-minority districts under Section 2 of the Voting Rights Act. The baseline there is voters, not simply population. To the extent that Latino voters have secured districts under Section 2, they will not suddenly be disempowered if districts must be drawn to include solely voters; instead, it is likely that, at least in regard to Section 2 districts, the effects would be less pronounced than immediately reported, and perhaps even marginal. (This, of course, assumes that Latino voters are residing in Section 2-drawn districts, and that the Supreme Court continues to interpret Section 2 as it has.) It might be the case that Latino voters would lose a marginal district or so if they were unable to muster sufficient voters in newly-redrawn districts. But they could also add previously-marginalized Latino voters in non-majority-minority districts to shore up a district that might otherwise have been lost under a new rule. And that means the partisan impact might be less than otherwise projected, too. (I'm sure very careful political scientists will have more to say about the more nuanced impact of such policies in the months ahead.)

But there are other alterations that such a lawsuit might bring, all depending on the fashioning of the remedy. To name a few.

• Districts with a felon prison population would lose political power. While a few states count prisoners as inhabitants of their last place of residence before prison, most include them as members of the population where they currently reside. That, obviously, can create districts, especially in the state's lower chamber of the legislature, with a significant number of ineligible voters. And these prisons tend to be in fairly rural areas (PDF)--or, areas that tend to have more Republican voters. A couple of million felons reside in these extremely concentrated areas--prisons, after all, are probably amount the most concentrated non-voting demographic you can find--and the practice of "prison gerrymandering" might be found unconstitutional.
• Districts with a disenfranchised ex-felon population would lose political power. To the extent a decision gets so granular as to exclude even ex-felons in redistricting in states that disenfranchise them, we'd have to look at where the couple of million ex-felons reside. And if most ex-felons are in largely concentrated in urban areas, then we'd expect a loss of political power for those urban dwellers as voters are added to their districts--which might benefit Republicans more than Democrats.
• Districts with a large number of children would lose political power. Children under the age of 18 are ineligible to vote in every state, with a couple of very marginal exceptions. But they are currently included in redistricting schemes. Localities that have more children would lose out; localities with aging populations would likely gain power. Whether this benefits any particular group is debatable. But to the extent there's a "fertility gap," in that Republicans tend to have more children than Democrats, it likely would also diminish Republican power. (Further complicating this analysis, of course, is that minorities tend to have children at higher rates than whites.) Taking California congressional districts as an example (assuming the case would be extended to congressional districts, but it's a useful data point because of the ease of obtaining Census data), the results are stark. There are just 91,000 under-18 residence in CD-12 (held by Nancy Pelosi, Democrat), but almost 227,000 children in CD-21 (David Valadao, Republican). Thousands of voters would be shifted out of districts like Pelosi's and into districts like Valadao's. (Further, it might also be the case that the shift from child-filled districts to elderly or childless districts would result in more substantive changes in political outcomes, including shifting of spending away from education toward end-of-life care or sustaining pensions.)
• Robust expansion of the Census Bureau's duties. If such granular data is required before states can draw districts, the Census is going to have a much bigger job. As it is, there have been instances where conservatives have opposed the existing breadth of the Census survey.

These are, of course, projections and guesses and possible results. But they are much deeper than the original "Latino loss, Republican gain" picture originally suggested. (I've written about the "invisible federalism" that underlies the political structure of the Electoral College, for example, as a result of our system of apportionment of representatives, includes some meaningful choice in jurisdictions about voters and non-voters.)

These potential ramifications come with an extremely important caveat (which, of course, I save until the end). It's not entirely clear what a system of redistricting based on "voters" would necessarily look like. The appellants in Evenwel seem to offer at least three possibilities before the three-judge panel: citizen voting-age population, total voter registration, and "non-suspense" voter registration. If the Court did decide that voters was the proper metric, but allowed flexibility within that regime, then redistricting would change based on citizenship status and age, but not felon prison population; however, a more narrow holding, or a state choice to use something narrow, would impact prison gerrymandering.

But even under the most generous Supreme Court interpretation in favor of appellants, I find it hard to believe that states would be allowed to continue to include vast numbers of imprisoned ineligible voters contained in a very tiny geographic area. And (regardless of the merits of this or any other impact on redistricting) that, I think, would be a fairly significant impact of Evenwel--and certainly not the one that either appellants or most commentators have really identified.