Finding 30,000 disenfranchised voters in Houston

In 48 states, a felony conviction means that you lose the right to vote. This can be temporary or permanent, depending on your state. During the 2018 election, Florida pushed this issue into the national spotlight with a state constitutional amendment to restore voting rights to felons. It passed with two-thirds of the vote.

Texas, like 21 other states and Florida, suspends a convicted felon’s right to vote during incarceration, parole, and probation. Once they go “off paper,” their voting rights get restored automatically, although they still have to register to vote.

This approach is more progressive than states like Wyoming (where an ex-offender must apply to restore their voting rights) or Delaware (which may require a formal pardon from the Governor).

But fourteen states and the District of Columbia restore voting rights automatically upon release from prison. And two states — Vermont and Maine — don’t suspend voting rights at all.

How do these attitudes shape our national conversation? Already, there are seeds of a proposal to end the suspension of voting rights at the Federal level. All presidential candidates may have to take a stand.

If this issue becomes a pillar of criminal justice reform in the next presidential election, what would that mean for Houston? Over 30,000 people living in Harris County would gain the right to vote in the upcoming 2020 contest.

Specifically, I looked at people convicted of felony crimes in Harris County who are still completing their sentence of incarceration or probation, and will not be “off paper” until after October 4, 2020. That’s the last day to register to vote in the 2020 presidential election.

I wanted to understand where these people live, so I did the math and built the maps:


Heatmap showing the number of disenfranchised ex-offenders per voting precinct.
Sources: Harris County District Clerk, Texas Secretary of State.

What would this mean for Houston elections?

An additional 30,000 registered voters isn’t much in a city as big as Houston, but it can make a difference. Generally, Houston has low voter turnout and extremely close elections. An additional 30,000 registered voters might lead to:

  • 3,800+ additional voters in a municipal election (based on 20.51% voter turnout from 2015 and adjusted for addresses inside the Houston city limits)
  • 18,000+ additional voters in a presidential election (based on 61.33% voter turnout from 2016)

Since we don’t know the political party of these new voters, this analysis stops short of making a case that ex-offender voting rights would alter the outcomes of any election.

But the 2015 municipal election for Mayor came down to 586 votes.

What would happen if another 3,800 people voted? Could that influence how our elected officials make policies that affect ex-offenders?

Disenfranchising Houston’s neighborhoods by precinct and Census block group

The data show that Houston’s current and ex-offender home addresses cluster in particular voting precincts. In fact, the top 25 precincts represent fewer than 1% of all voting precincts, yet make up 10% of the current and ex-offender population, even after controlling for people who list the jail or a homeless shelter as their home address.

PrecinctCurrent or Ex-OffendersRegistered VotersBallots CastVoter TurnoutBeto Margin
036921578915519.64%70.59%
0031150270599636.82%96.14%
0794147308866721.59%91.06%
07571455511227541.28%86.57%
01971412972113338.12%96.21%
01091392934116039.53%96.84%
00681373224140643.61%96.73%
03591373407100529.49%88.79%
05201295520190734.54%86.47%
04541272485103841.77%96.76%
02831237413287038.71%86.79%
01521213211112835.12%80.51%
05761192854126344.25%94.17%
01691142596101439.06%95.99%
03751122968101934.33%66.5%
0253109214982338.29%96.65%
02401082599107641.4%96.11%
06491084886214443.88%75.62%
01801072967117039.43%94.91%
0237106216582237.96%96.63%
00451056701295944.15%83.66%
0611104250599839.84%95.46%
04311034140162139.15%81.29%
05591025242230443.95%80.83%
0284994007181045.17%67.86%

Since all of these precincts swing left, the “Beto Margin” helps me understand the severity of the swing. It reflects the percentage of votes that Beto O’Rourke earned over Ted Cruz.

Almost all of the addresses in the most popular precinct for current and ex-offenders, Precinct 0369, correspond to the Harris County Jail or the homeless shelter Star of Hope. Located just north of Downtown Houston, this is in the heart of Harris County’s criminal justice system.

Finally, it’s worth nothing that the average voter turnout in Harris County in the 2018 election was 52.86%, and all 25 of these precincts had a below average voter turnout.

Through the lens of race and income, the picture is even more stark:


Heatmap showing the number of disenfranchised ex-offenders per Census block group.
Sources: Harris County District Clerk, US Census Bureau.

There are some trends that emerge in the 15 most affected block groups:

  • Nine of the block groups are majority Black, three are majority Hispanic, three have no majority. No block groups have a White majority.
  • Seven out of fifteen block groups are at least 66% Black, with five of them at least 80% Black.
  • Ten out of fifteen block groups are below the 2017 Texas threshold of 200% FPL ($49,200).

My approach to the data

I followed these steps to complete this analysis:

  • Subset all felony cases from the Harris County criminal court dispositions dataset.
  • Standardized sentence length to a numeric field containing number of days. In the data, it is a text field such as “1 YEAR” or “6 MONTHS.”
  • Added the sentence length to the disposition date to determine when the person goes “off paper.”
  • Subset the dataset to October 4, 2020 and later (the last day of voter registration in Texas for the 2020 presidential election).
  • Removed duplicate cases by case number and defendant name to get a count of people, not records.
  • Geocoded the home addresses of the remaining people and subset the points to Harris County.
  • Aggregated home addresses by voter precinct and census block group.

The result is a dataset of 30,308 individuals who have been convicted of a felony and will not have their voting rights restored by October 4, 2020.

There were an additional 1,274 people with unknown addresses or marked as homeless. They are not included in this analysis, although it is likely that many of them live in Harris County.

If you’d like to talk about this analysis, please tweet @januaryadvisors.

Jeff Reichman

Jeff is passionate about data. He founded January Advisors and Sketch City, and serves on the board of the League of Women Voters of the Houston Area. Read his full bio on LinkedIn.

Into civic tech, maps, data, and building awesome stuff?

Then you'll love our monthly newsletter.