Off the Kuff Rotating Header Image

What is “safe”?

Saw this on Twitter, and it got me thinking:

AOC isn’t the only person I’ve observed referring to CD28 as “safe” Democratic. This WaPo story from 2019, reprinted in the Trib, calls CD28 “a strongly Democratic district…which gave the president just 38.5 percent of the vote in 2016”. This DMN story has a subhed that calls CD28 “Vast and overwhelmingly Democratic district”, and notes that “Trump lost here by 20 percentage points”. The American Prospect is a bit more circumspect, saying CD28 is “a safely (though not extremely) blue district, with a +9 Democratic lean”, and also noting the 20-point margin for Clinton over Trump in 2016.

But 2016 isn’t the only election we’ve ever had, and the Clinton-Trump matchup isn’t the only data point available. Here’s a broader look at the recent electoral history in CD28:

Year  Candidate    Votes    Pct
2012  Obama      101,843  60.2%
2012  Romney      65,372  38.6%
2012  Sadler      90,481  55.1%
2012  Cruz        68,096  41.5%
2012  Hampton     93,996  58.5%
2012  Keller      61,954  38.6%

2014  Alameel     41,901  46.6%
2014  Cornyn      42,010  46.7%
2014  Davis       48,451  52.7%
2014  Abbott      41,335  45.0%
2014  Granberg    45,658  51.7%
2014  Richardson  38,775  43.9%

2016  Clinton    109,973  57.8%
2016  Trump       72,479  38.1%
2016  Robinson    95,348  52.6%
2016  Guzman      77,590  42.8%
2016  Burns      102,778  57.1%
2016  Keasler     69,501  38.6%

2018  Beto        97,728  58.7%
2018  Cruz        67,483  40.5%
2018  Valdez      87,007  52.7%
2018  Abbott      75,939  46.0%
2018  Jackson     94,479  58.3%
2018  Keller      63,559  39.2%

Yes, in 2014, John Cornyn topped David Alameel in CD28. To my mind, if it is possible for a candidate of the other party to beat a candidate of your own party in a given district, that district is by definition not “safe”. It’s true that in Presidential years, most Democrats win CD28 comfortably, with the closest call being a win by just under 10 points. But in off years, even factoring out the crapshow that was the Alameel campaign, Dems generally win CD28 by smaller margins.

None of this is to say that CD28 is a swing district. It’s not, and I have no reason to be concerned about it in 2020. But if Trump-versus-Clinton-in-2016 is the gold standard here, I’ll point out that of the six districts Dems are targeting this year, four of them (CDs 02, 10, 22, and 31) were won by Trump by larger margins than Wendy Davis won CD28 by in 2014 and Lupe Valdez won it by in 2018. Different years, different conditions, and different candidates may provide a different perspective.

Another way of looking at this is to see how Democratic CD28 is compared to other Congressional districts represented by Democrats:

Dist  Clinton    Beto
CD07    48.2%   53.3%
CD32    48.4%   54.9%
CD15    56.2%   57.4%
CD28    57.8%   58.7%
CD34    59.1%   57.7%
CD20    60.2%   66.2%

All other Dem-held districts were at least 63% for Clinton and 70% for Beto. Again, none of this is to say that CD28 is vulnerable. Whoever wins the CD28 primary will be the strong favorite, like 99%+, to win it in November. This is not a comment on that race, but on public perception and objective reality. It’s why I generally try not to make blanket statements like “safe district” but try instead to put a number or two on it, so you have some context to my evaluation. I doubt anyone will adopt this as their style guide, but it’s very much how I prefer to operate.

And I have to say, I might have let this go by if I hadn’t also seen this little gem in the Chronicle story on the announced resignation of State Sen. Kirk Watson:

Abbott soon will have to schedule a special election for the remainder of Watson’s four-year term, which ends in 2022. Watson’s District 14, which mostly lies in Austin, leans Democratic.

“Leans Democratic”??? Here’s that same set of numbers for Watson’s SD14:

Year  Candidate    Votes    Pct
2012  Obama      193,112  60.2%
2012  Romney     116,001  36.1%
2012  Sadler     187,717  59.4%
2012  Cruz       109,877  34.7%
2012  Hampton    181,614  59.1%
2012  Keller     106,581  34.7%

2014  Alameel    123,058  56.2%
2014  Cornyn      80,818  36.9%
2014  Davis      140,602  63.3%
2014  Abbott      75,206  33.9%
2014  Granberg   127,108  59.7%
2014  Richardson  73,267  34.4%

2016  Clinton    249,999  65.3%
2016  Trump      106,050  27.7%
2016  Robinson   218,449  58.8%
2016  Guzman     124,165  33.4%
2016  Burns      223,599  60.8%
2016  Keasler    120,727  32.8%

2018  Beto       289,357  73.8%
2018  Cruz        98,589  25.1%
2018  Valdez     257,708  66.3%
2018  Abbott     119,889  30.9%
2018  Jackson    264,575  69.4%
2018  Keller     104,375  27.4%

LOL. SD14 “leans” Democratic in the same way that a wrecking ball leans against the side of a building. Data is your friend, people. Use the data. I know I’m tilting against windmills here, but at least you can see why I noticed that tweet.

Related Posts:


  1. Richard says:

    I’ve been in the valley for the last couple of days and Cuellar’s opponent has been running ads that “Cuellar is Trumps favorite Democrat.” That could leave a mark.

  2. blank says:

    I once tried to find a mapping of probabilities winning to race ratings, and the closest thing I could find was a 538’s Atlas of Redistricting project, but is has only 3 categories, Usually Democratic, Highly Competitive, and Usually Republican. It looks like that if both parties have a better than 17% (5 out of 6) chance of winning, then it’s considered a highly competitive seat. I don’t subscript to CPR, so they may say what those numbers are. If you forced me to guess, it’s something like this:

    If the favorite has less than a 60% chance to win, then it’s a toss up.
    Else if the favorite has less than a 75% chance to win, then it’s lean.
    Else if the favorite has less than a 90% chance to win, then it’s likely.
    Else it’s safe.