Ranked Vote Choice (RVC) or Instant Runoff Voting (IRV)

In Austin, Texas, two important changes in voting for Austin City Council members were implemented in November 2014. First, Austin moved from an at-large-by-place election to a single-member district election for its council members, and at the same time, increased the council from six members and a mayor to ten council members and a mayor. Second, the election date was moved from the first Saturday in May to the first Tuesday after the first Monday in November, which coincides with the general election to elect national, state, and county officials.

In the first election under this system, seventy candidates were competing in the ten single-member districts. In the election on November 4, 2014, only two of the ten council candidates won a majority of the votes in the district; so eight districts were scheduled to hold runoff elections between the two highest vote getters on December 16, 2014. Subsequently, one candidate withdrew, leaving seven districts to conduct runoff elections.

Is there a better method of deciding elections than with a runoff? One method of eliminating the necessity of a runoff election would be to abandon the majority vote requirement and institute a plurality vote requirement. That is, the candidate with the largest number of votes in the election wins, regardless of whether that is a majority or not. The problem with this system is, of course, that the winner may have garnered a small percentage of the vote. For example, in Austin’s District 8, the candidate who received the largest number of votes won only 26.38 percent of the vote, leading the candidate with the second highest vote total by only 179 votes out of 21,538 votes cast in the district. Other districts had similar results. Had the plurality system been in place, the District 8 candidate who was not the first choice of nearly 74 percent of the voters would have won the council seat in the district. This does not seem right.

So, the question remains: Is there a better method of deciding elections without a runoff? The answer involves ranked vote choice (RVC) or instant runoff voting (IRV), which mean essentially the same thing. Here’s how the system works: each voter ranks the candidates for the office according to his/her first, second, and third choice. When the votes are tallied, if one candidate receives a majority of the first- choice votes, that candidate is elected. However, if no candidate receives a majority of first-choice votes, then the candidate who received the fewest first-choice votes is eliminated, and the voters who selected that candidate as their first choice have their second-choice votes assigned to one of the remaining candidates. If, as a result of those second-choice votes, one of the candidates now has a majority of votes, the candidate is elected. However, if no candidate still has a majority of the votes, then the remaining candidate with the fewest votes is eliminated, and his/her second-place votes are assigned to one of the remaining candidates. This process is repeated until one candidate receives a majority of the votes in the contest.

The RVC or IRV system is used in several cities in the United States to conduct their mayoral and/or city council elections. Examples include Berkeley, California; Oakland, California; Portland, Maine; San Francisco, California; San Leandro, California; St. Paul, Minnesota; and Telluride, Colorado.

The advantages of the RVC or IRV system include lower election costs by eliminating runoff elections, higher voter turnout, and, according to Fair Vote, RVC or IRV:

• Gives voters the option to rank as many or as few candidates as they wish without fear that ranking less favored candidates will harm the chances of their most preferred candidate.
• Empowers voters with more meaningful choice.
• Minimizes strategic voting.
• Creates a positive atmosphere where candidates campaign to the voters rather than against each other.

Source: Fair Vote: The Center for Voting and Democracy, http://www.fairvote.org/reforms/instant-runoff-voting/

Thus, Austin should amend its charter to substitute instant runoff voting for the current runoff election system.

The Battle for Texas (Part I)

The results on November 4th were not what Democrats had hoped for, and as a result, there are many assessments of Wendy Davis’ campaign and the efforts of Battleground Texas that assign blame for the poor showing (see this , this , this or this ). That is not the purpose of this post. I want to move forward—to indicate what might be different about 2018. In Part I, I will assess what happened in 2014.

First, what was the composition of the electorate and which gubernatorial candidate did each category favor? From the exit poll, we get a good idea of who voted and for whom they voted. So here are the salient data.

Females were more prominent among voters than males, but the expectation that Wendy Davis would do better than Republican Greg Abbott was incorrect: Abbott received 54 percent of the female vote, and Davis received 45 percent. A closer examination of the vote by gender reveals Davis’ strengths as well as her weaknesses. First, Davis did best among African American women (7 percent of the electorate), winning 94 percent to Abbott’s 5 percent. Among Hispanic women (9 percent of the electorate), Davis bested Abbott 66 percent to Abbott’s 39 percent. Among Anglo women (33 percent of the electorate), Abbott won 66 percent to Davis’ 31 percent. Second, Davis won women who were not married (19 percent of the electorate) with 57 percent to Abbott’s 42 percent. However, Davis lost married women (33 percent of the electorate) with 36 percent of the vote to Abbott’s 62 percent. Third, Abbott won Republican women (19 percent of the electorate) with 95 percent of the vote as well as independent women (14 percent of the electorate) with 57 percent of the vote. Clearly, Anglo women and married women presented a challenge to Davis’ hope of victory. Furthermore, whereas Davis expected to do well among independent women, she lost them badly.

Next, consider how Davis performed among various ethnic groups. As expected, she won the African American vote (12 percent of the electorate) handily, receiving 92 percent of the vote to Abbott’s 7 percent. She also won the Hispanic vote, but the margin was not what was needed to result in a Davis victory. Davis won 55 percent of the Hispanic vote (17 percent of the electorate) to Abbott’s 44 percent. Among Anglos (66 percent of the electorate), Davis was trounced—Abbott’s 72 percent to Davis’ 25 percent. Previously, I maintained that Davis needed 70 percent of the Hispanic vote and their percentage of the electorate needed to be 22 percent as well as 38 percent of the Anglo vote to win. Only among African Americans was Davis’ support sufficient and their percentage of the electorate large enough to support a Davis victory.

Davis also fell short among the age categories that she had to win in order to win the governorship. According to the exit polls, Davis tied with Abbott among the 18-29 year olds who voted, barely lost the 30-44 year olds, and lost the 45-65 year olds (32 percent to 67 percent) and voters over 65 (29 percent to 69 percent). Voters who were 18-44 years old constituted 41 percent of the electorate—the same percentage as voters who were 45-65 years old. The only age category won by Davis was the 30-39 year olds (53 percent to 45 percent).

Finally, among family income categories, Davis won only those families who earned less than $30,000 annually (51 percent to 47 percent). As family income increased, Davis’ percentage of the vote decreased as Abbott’s percentage of the vote increased.

 What do the exit polls from the 2014 general election indicate for future Democratic Party efforts in Texas? There are several groups that are not voting in the numbers necessary for Democratic candidates to win, and several categories do not support Democratic candidates sufficiently to result in Democratic victories. Consequently, Democratic appeals to voters must be sharpened and groups that are likely to support Democratic candidates need to be motivated to vote. Elections are about who participates and what choices those who do participate make. In Part II, I'll consider how the electorate might change for 2018.

How Texans Did (or Did not) Vote in 2014 Elections

There are 16,679,393 eligible voters (VEP) in Texas. This number excludes people under 18 years of age as well as those disenfranchised because of a felony or who are not citizens of the United States. Here is what those people did on Election Day (only 28.4 percent of VEP voted). If you're a Texan who is eligible to vote but didn't, I hope that you're embarrassed. If you're not registered to vote, you need to do so!

Predicting Voter Turnout from Early Voting in the SOS Counties: An Update

First, I predicted the total early vote in the state based on an estimation of the percentage of the total early vote that the early votel in the SOS counties (the 15 most populous counties in Texas) would represent. I predicted that it could be 67.16 percent. In actuality, it was 67.32 percent, which made my estimate of the total early vote 5,950 votes more than the actual early vote.

Year	Day 12	Total EV	D12/TEV
2006	1,074,824	1,728,030	62.20%
2008	3,556,156	5,351,660	66.45%
2010	1,731,589	2,640,918	65.57%
2012	3,407,497	5,020,901	67.87%
2014	1,715,731	2,554,692	67.16%	Prediction
2014	1,715,731	2,548,742	67.32%	Actual

Then I used my prediction of the total early vote in the state to predict the total vote (early vote plus Election Day vote) for Texas. I predicted that the early vote percentage of the total vote would be the same as the percentage in 2010 (53.03 percent). It was actually 54.13 percent, which made my prediction of the total vote (4,817,446) 108,890 votes more than the actual total vote (4,708,556). The error was 2.31 percent.

Year	EV State	Total Vote	EV/TV
2006	1,728,030	4,399,068	39.28%
2008	5,351,660	8,077,795	66.25%
2010	2,640,918	4,979,870	53.03%
2012	5,020,901	7,993,851	62.81%
2014	2,554,692	4,817,446	53.03%	Prediction
2014	2,548,742	4,708,556	54.13%	Actural

Then I used my prediction of the total vote to calculate voter turnout (total vote/registered voters). My prediction was 34.35 percent. The actual turnout rate was 33.57 percent. The error was 2.3 percent. That's not bad!

Year	Total  Vote	Reg Voters	Voter T/O
2006	4,399,068	13,074,279	33.65%
2008	8,077,795	13,575,062	59.50%
2010	4,979,870	13,269,233	37.53%
2012	7,993,851	13,646,226	58.58%
2014	4,817,446	14,025,441	34.35%	Prediction
2014	4,708,556	14,025,441	33.57%	Actual