The Most Vulnerable Georgia Districts — Part One
Last week I discussed the lack of competitiveness in the General Assembly races across the state of Georgia. I’d encourage you to go back to that post and start there, but here’s the quick synopsis: The Georgia Democrats are not competing in enough seats to take a majority in either chamber of the General Assembly.
In order to compete in future elections, we have to start planning now. Immediately following the 2014 election we should have started a targeting analysis for the 2016 election. We will need to do the same in preparation for the 2018 election. Without having access to any existing analysis, I decided to do one of my own.
For this part of the analysis, I’m ignoring what candidates are currently running. Examining the individual candidates is an interesting task, but not one that I will enter into here. This will be the first of two parts identifying districts with a potential to flip. Part One will use a vulnerability analysis based on prior election results. Part Two will use current population data to identify demographic advantages.
The Georgia Legislative Vulnerability Index
Why use a vulnerability index?
The easiest way to understand which districts should be competitive this year is simple: How did Democrats perform on the ballot in previous years?
I cannot take credit for the system I’m using here. Developed by Daily Kos, the index identifies vulnerable House seats. They call it the House Vulnerability Index (HVI), but it is easily transferable to the Georgia legislative districts.
The idea of this index is to balance out the incumbency advantage with the natural partisan tilt of the district. This is not a perfect science and assumptions are made about what previous votes mean in today’s elections, but it is quite effective at narrowing the list for further analysis.
How is the index calculated?
The first step is to take the margin in the previous presidential race within each district. The presidential race sees the greatest turnout allowing for this to act as a proxy for the partisan tilt of the district itself. For the current Republican seats, we will use the difference of the 2012 vote for Obama and the vote for Romney within each legislative district. The higher the margin, the more vulnerable the seat. Positive margins indicate the opposing party’s presidential candidate performed better. For example, on a Republican-held seat, a positive margin would indicate that Obama had more votes in the district than Romney. These numbers are then ranked with one being the largest tilt to the opposing party.
Next, Daily Kos takes the margin from the most recent race within the legislative district. For example, for this year’s election, they used the results from each 2014 House race. This will help weight the current incumbent advantage present in the district. For this half of the index, I made a slight adjustment for the Georgia legislative districts. I averaged the results from 2012 and 2014 instead of just using the most recent. This change helps to account for the number of races that had a single candidate on the ballot. This made it more likely to include at least one competitive race resulting in a more accurate measurement of the incumbency advantage.
Open races receive a zero ranking and the rest are then ranked. This margin indicates the incumbency advantage, so if there is no incumbent, then the advantage doesn’t exist. Historically, open seats are more vulnerable, and a zero ranking helps account for that.
Finally, these two rankings (Presidential margin within the district and the average of the two previous legislative races) are combined for the total HVI. The lowest score possible is a two. This would indicate the district has the most partisan tilt to the opposing party and the closest margin in the previous two legislative races.
The index is calculated separately for the seats held by each party. The lowest ranking on a Republican seat indicates that it is the most likely to flip to the Democrats and vice versa.
A few notes for this analysis:
- Some seats are held by incumbents who won in a special election. These special elections were used in place of the 2014 numbers if they were competitive to give weight to the incumbency advantage. The results from the recent special election were still averaged with the results of the 2012 election.
- Any independents/non-partisan seats were included in the analysis for the party with which the representative caucuses.
- The index was calculated separately for each the Senate districts and the House districts. All numbers are for results only within each Senate or House district.
- The election results for each district were compiled by Daily Kos. You can find the interactive maps here and the spreadsheets here.
How is the index used to identify potential targets?
“2016 needs to be a foundation election with the goal of a majority following the 2020 election.”
Once the index is in rank order, we have to determine which districts need to be examined in further detail. Given the unique nature of state legislative races, the index is only a starting point. With such a small amount of voters compared to Congressional Districts, candidates of either party can drum up enough support to win unexpected districts (ex. Gerald Greene in HD-151 and Valerie Clark in HD-101). This index should give us a starting list of districts before clarifying based on a more detailed analysis of the demographics, registration numbers, and turnout expectations.
We also need to look at our goals for this election. As much as we need to have enough candidates on the ballot to give us an opportunity to take a majority, the reality is that taking a majority in 2016 is an extreme long-shot. It would take a massive shift like the one seen in the 2004 election. In that election, the Republicans flipped 27 Georgia House seats. Georgia legislative elections have become less competitive over that time, so the goal is not a majority in 2016. 2016 needs to be a foundation election with the goal of a majority following the 2020 election.
We should be targeting districts in three categories:
- Republican districts that are vulnerable today,
- Republican districts with a potential vulnerability in future elections, and
- Democratic districts that need to be protected.
To start building our list of targets, we can develop some criteria within this index to categorize the districts.
For Category 1, we can automatically include the top five seats ranked by the HVI. Regardless of the numbers, we need to target districts that show vulnerability in both the incumbent measure and the partisan measure. Also, anything with either margin a -2.5% or higher should be included. Anything this close shows either a weak incumbent or a toss-up partisan split. For open seats, I will only look at the presidential margin after taking the top five seats by HVI. I will take a deeper look at open seats when I look at the demographics.
Beyond this, I will include anything between a -7.5% margin and -2.5% margin as a potential for Category 1. I will examine these further in Part Two. These districts will almost certainly end up in either Category 1 or Category 2. Anything that close has a potential to flip this election or next depending on the demographics.
In this first sweep, Category 2 will include anything with a margin between -15% and -7.5%. These are likely too far away to flip this year, but should still be given attention for future elections. These districts could still have opportunities depending on demographics, registration, and voter turnout numbers. After looking at the numbers, we can determine if they need to be in Category 2 or removed. Anything in Category 2 needs to be challenged in order to lay the groundwork for 2018 and 2020. If we don’t have candidates, then we won’t have the opportunity to bring our message to those districts.
Category 3 is simple. Districts in Category 3 are the same as districts that would be included in Category 1 for the Republicans. The cutoffs will be applied to currently held Democratic seats and included as the seats that need to be protected.
Without further ado, let’s move to the numbers.
Senate Targets
The takeaway for the Senate races: There is a lot of work to do to make districts competitive, but the districts we have are solid Democratic districts.
Category 1 —
Top Five by HVI:
- SD-43
- SD-6
- SD-23
- SD-8
- SD-40
No districts meet the alternative criteria for Category 1.
Category 2 —
No districts meet the criteria for Category 2.
Category 3 —
Top Five by HVI:
- SD-26
- SD-42
- SD-15
- SD-33
- SD-35
No districts meet the alternative criteria for Category 3.
Senate Summary —
The Senate side of the race shows the weakness in the vulnerability analysis. These seats are uncompetitive, so even the top five in each of the categories do not appear vulnerable.
SD-43 is the most obvious potential for a flip. It is a solidly Democratic district, but it was lost in a special election. Obama won this district by 43.58%. It should have never been in question, but now a Republican will be running as an incumbent.
SD-6 is another good target based on this analysis. Obama lost the district by 7.34%, but the incumbent Republican has averaged a 13.53% margin of victory. We can take a further look at the demographics, but this should be a target.
Even though it is fifth on the list, SD-40 could be another strong target. Obama lost the district by just over 10%, but the Senate seat was not challenged in 2012 and was lost by over 30% in 2014. When the final list is compiled, this is likely a Category 2 target, but we will have to look.
Beyond these seats, there are no good targets at least on the face of the numbers. We will have to look at the demographics to get any idea of the vulnerability.
There is a consolation prize for the Democrats here. The Republicans have almost no chance of flipping any seats. The closest margin in the 2012 Presidential race was in SD-33 and Obama won the district by 16.86%. The next closest was 29.68% and it only gets worse for the Republicans.
The takeaway for the Senate races: There is a lot of work to do to make districts competitive, but the districts we have are solid Democratic districts.
House Targets
The House looks better than the Senate with a promising set of districts to target.
Category 1 —
Top Five by HVI:
- HD-105
- HD-138
- HD-145
- HD-111
- HD-95
>-2.5% in either margin:
- HD-151
>-7.5% and <-2.5% in either margin:
- HD-101
Category 2 —
>-15% and <-7.5% in either margin:
- HD-107
- HD-40
- HD-117
- HD-37
- HD-173
- HD-147
- HD-140
- HD-106
- HD-51
- HD-79
Category 3 —
Top Five in HVI:
- HD-80
- HD-66
- HD-81
- HD-132
- HD-96
No districts meet the alternative criteria for Category 3.
House Summary —
This is good starting point for the House. There are seven total districts currently in Category 1 and ten more districts in Category 2. There are too many here to get into much detail, but this a good start before we head into the demographic analysis in Part Two.
As far as districts to defend, this list is more worrisome. We were the recipients of an unusual special election victory in HD-80. Romney won this district by 13.15%. In 2012, this Senate seat was not challenged, but in the special election, the Democrat won by 54.5%. The incumbency advantage is present, and a 13.15% partisan tilt is not insurmountable. We will need to look further at this district to see how it can be defended.
Three other districts are definite Republican targets: HD-81, HD-132, and HD-66. Obama won these districts by 5.31%, 5.54%, and 7.82% respectively. These will be the main targets for the Georgia Republicans, so we will have to take a look at what the demographics tell us.
The House looks better than the Senate with a promising set of districts to target. If we can flip the seven Category 1 districts this election, flip the ten Category 2 districts in 2018, and keep all currently held seats, then the House split will be 102R/78D. That leaves us 13 seats to flip in 2020. That is a decent trajectory assuming we have candidates competing in those districts. (Spoiler Alert: We don’t…)
We have a starting list. Now what?
Demographics are shifting all the time, so the results from two and four years ago may not be reflective of the current population. In the next post, I will add to the districts from this index through a different analysis. Ignoring past results, this analysis will be based on the current demographics, registration numbers as of September 1st, and previous turnout percentages specific to each district. Using those numbers, combined with demographic vote splits, I will calculate an estimated voting split for each House and Senate district. This should give us a baseline for each district’s current demographic advantage. I can then assign those districts to the three categories. Based on that analysis, I will reorganize the districts into the three targeting categories.
From the target list, I can use the data compiled to find appropriate strategies within each district (registration goals, GOTV initiatives, etc.), predict outcomes in 2016 based on a variety of turnout ratios, and identify deficiencies to overcome for future elections. In a series of later posts, I will take apart each district individually to understand its inherent potential and the current state of the race. This will include looking at the individual candidates and a simple projection of the outcome based on current nationwide and statewide polling, demographic data, registration information, and historical voter turnout.
By the end, we should have a good idea of what could happen on November 8th and have a baseline to build upon after the election.
— Is there anything you’d like to see in addition to this analysis?
— Do you think the Democrats have a chance at taking a majority in either chamber of the General Assembly by 2020?
Here is Part Two of the Most Vulnerable Georgia Districts:
— Following the POLITICS of today, for the generations of TOMORROW —
Politics for Tomorrow is a publication focused on progressive politics both nationally and in the State of Georgia.
Follow Politics for Tomorrow on Twitter here.
Follow Austin Wagner on Twitter here.
