While I wait for the LRC to post the text of the truckload of new bills whose titles they published yesterday, I’m musing over Legislative district demographics.
As of 11 a.m. this morning, the average number of registered voters in each of our 35 Legislative districts is 15,116. Actual registered voters range from 11,077 in District 15 (central Sioux Falls) to 19,038 in District 33 (north and west edges of Rapid City, Black Hawk, Summerset, then west out into the Hills around Johnson Siding and Pactola).
Districts are supposed to include a roughly equal population. In 2010, that ideal population would have been 23,262. Based on the 2016 Census estimate, fair Legislative districts would each include 24,727 South Dakotans.
According to the Census, 75.4% of South Dakota’s population is 18 or older. That means that, in an ideally drawn district, we should find 18,644 voting-age residents. Some small percentage of those folks would be unable to vote due to felony convictions (although once you’ve served your full sentence, you can register to vote again!) or lack of citizenship, but we won’t get that deep into the numbers.
Comparing the 2016 population estimate with current vote registration, we find that 81.07% of voting-age South Dakotans appear to be registered to vote. Yet in four districts (15, 7, 26, and 17), the ratio of registered voters to ideal voting-age population is less than 70%. In six districts (8, 12, 14, 11, 30, and 33) the ratio of registered voters to ideal voting-age population is greater than 90%.
| District | Active Registered Voters 2018.01.03 | Active/ Ideal2016 | Difference from average participation (% pts) |
| 1 | 15,246 | 81.77% | 0.70 |
| 2 | 15,077 | 80.87% | -0.21 |
| 3 | 14,908 | 79.96% | -1.11 |
| 4 | 15,038 | 80.66% | -0.42 |
| 5 | 13,981 | 74.99% | -6.09 |
| 6 | 16,615 | 89.12% | 8.04 |
| 7 | 12,259 | 65.75% | -15.32 |
| 8 | 16,786 | 90.03% | 8.96 |
| 9 | 16,337 | 87.62% | 6.55 |
| 10 | 15,662 | 84.00% | 2.93 |
| 11 | 17,393 | 93.29% | 12.22 |
| 12 | 16,932 | 90.82% | 9.74 |
| 13 | 16,497 | 88.48% | 7.41 |
| 14 | 17,213 | 92.32% | 11.25 |
| 15 | 11,077 | 59.41% | -21.66 |
| 16 | 16,039 | 86.03% | 4.95 |
| 17 | 12,833 | 68.83% | -12.24 |
| 18 | 13,257 | 71.10% | -9.97 |
| 19 | 15,902 | 85.29% | 4.22 |
| 20 | 14,009 | 75.14% | -5.93 |
| 21 | 13,281 | 71.23% | -9.84 |
| 22 | 13,164 | 70.61% | -10.47 |
| 23 | 15,281 | 81.96% | 0.89 |
| 24 | 14,997 | 80.44% | -0.64 |
| 25 | 16,528 | 88.65% | 7.58 |
| 26 | 12,458 | 66.82% | -14.25 |
| 27 | 13,149 | 70.53% | -10.55 |
| 28 | 14,871 | 79.76% | -1.31 |
| 29 | 14,963 | 80.25% | -0.82 |
| 30 | 17,854 | 95.76% | 14.69 |
| 31 | 16,193 | 86.85% | 5.78 |
| 32 | 14,968 | 80.28% | -0.79 |
| 33 | 19,038 | 102.11% | 21.04 |
| 34 | 15,828 | 84.89% | 3.82 |
| 35 | 13,409 | 71.92% | -9.15 |
| SD | 529,043 | 81.07% |
The seeming voter deficit in District 15 could be accounted for entirely by the penitentiary population. Subtract those 1,500-some inmates from District 15’s estimated voting-age population, and District 15’s active/ideal ratio jumps a hair above the statewide average.
The deficits in 7 and 17—Brookings and Vermillion/Clay and Turner counties) may related to the university populations boosting Census count but depressing voter registration, as students either don’t vote or vote back home. The smaller university populations of Aberdeen, Spearfish, and Madison don’t appear to have the same impact on their district ratios: NSU’s District 3 is just one percentage point below the state average, while BHSU’s 31 and DSU’s 8 are both several points above the average.
Overall, the numbers could reflect shifting populations. A lower-than-average active/ideal participation ratios could signal a decline the district’s population since 2010. A higher-than-average active/ideal participation ratio could mean lots of people have moved into the district since 2010. If that’s the case, low-participation districts like 35 (East Rapid) need to be drawn bigger when we redistrict in 2021, while districts like neighboring 30 and 33 need to be drawn smaller to restore population balance.
If these voter registration patterns remain steady in the 2018 election, and if we assume equal turnout rates in all districts, then one voter in District 15 voting to reëlect Senator Reynold Nesiba has 36% more power than a voter in District 2 voting to reëlect Senator Brock Greenfield and 72% more power than one voter in District 33 voting to reëlect Senator Phil Jensen.
So here’s something to think about: should we draw Legislative district boundaries with an eye toward actual voting populations? Or would such math disadvantage districts with large populations who need representations (children and inmates are people, too) but who can’t vote?
Don’t give kookburgers in the legislature even more ideas on how to gerrymander for fun and self serving reasons.
I’m pretty sure if this had a chance of passing constitutional muster it would have been tried, but the “one man one vote” rule established by Reynolds v. Sims thankfully prevents it.
https://en.wikipedia.org/wiki/Reynolds_v._Sims
Here is a link to the case.
Cory
Dist 15 is interesting. Do you have the numbers from 2010? I wonder if this is a population shift (or lack of growth) in an inner city district? would be interesting to see the change over the eight years
Lee, I don’t have 2010 voter registration by district. I would love to have a slider on the SOS map that would allow us to roll the map back and compare years. Perhaps we could check election results and compare turnout….
Speaking of voting, Trump just disbanded his voter fraud commission.
Why did the skunk do that?
I am liking the numbers. Room to dump 1,000 Democrats into District 15 from every other metro district is a good thing if we base redistricting on voter participation, which is what we should be doing. We should be looking at feedlots in the Texas panhandle for inspiration for 2020 when it comes to housing low-turnout liberals in District 15. I hope some creative minds take this post to heart. Interestingly enough, if one strictly looks at voter participation numbers, one can gerrymander rather effectively without using any other identifying characteristic. This makes it easier to weed out liberal Independents.
I think districts should be of equal population, regardless of actual voting ability. I vote for the best interest of those in my family who cannot vote yet, and those representatives still enact laws that affect the non-voters.
According to Ballotpedia, 23,262 was the ideal number and the maximum deviation was 9.6% in the last redistricting. It doesn’t ID what district was off by that much, about 2300 people. Assuming ballotpedia is close to accurate (they have the ND majority leader’s death as a SD legislator!), the population out-migration in Dist 15 must have occurred in the last 8 years. FYI
Spencer’s desire to manipulate redistricting for his partisan purposes is duly noted.
Lee, the trick is, we don’t know from the outside whether the low voter numbers in District 15 are due to out-migration, including the Pen, high non-citizen population, low interest in voting, a combination of all of those factors, or something else.
Joe, I also hesitate to move away from straight population count. Basing district maps on voter participation could make Republican voter suppression efforts a vicious cycle: Republicans make it harder for poor minority populations to vote, fewer poor and minority voters come out to vote, Republicans like Spencer draw the boundaries to crowd even more poor and minority voters into low-turnout districts….
Roger! See this morning’s post on Kobach!