Running Elections with RCV Does Not Increase “Undervote”

Posted by Theodore Landsman on November 08, 2016

A review of undervotes in Bay Area elections suggest that the percentage of voters at the polls for state and national elections are no more likely to skip local elections with ranked choice voting than those held without ranked choice voting. 

Election Run with RCV

Average Undervotes (%)

No

17.7%

Yes

14.6%

A common anxiety among political scientists about ranked choice voting (RCV) is that voters may be confused by a different voting system, and skip RCV elections when at the polls for other elections -- leaving what is referred to as an “undervote”. However, new research conducted by FairVote on Bay Area elections from 2000 to 2014 suggests that RCV does not increase undervoting in local elections, and in fact, generates less undervoting than other election systems that are common in the region.

We examined results in 107 ranked choice voting races in four cities that have adopted RCV: Berkeley, Oakland, San Francisco and San Leandro. We also examined results in another 115 elections in the  following cities  without RCV: Anaheim, Richmond, Santa Anna, Santa Clara, San Jose, and Stockton.   These cities were chosen because of their proximity to cities with RCV, and their demographic similarity, which made them good control cases.
Overall, voters in the Bay Area undervoted in an average of 16.22% of races. In RCV elections, 14.6% undervoted, while elections conducted using other systems, 17.7% of voters undervoted. 
The two factors that drove undervoting up the most (see the table below) were uncontested races and races held in multi-member districts using block voting or at-large numbered posts. One solution for Bay Area cities struggling with undervoting in multi-winner elections might be to adopt multi-winner RCV, since multi-winner RCV ballots are no more complicated than single winner RCV ballots and would create more proportional outcomes.

Variable

Effect On Undervotes

Effect on UnderVotes for RCV Elections

Election for City Executive Position? (0-1)

Races for city executive positions saw 3.8 percentage points less undervoting.

No Significant Effect

District Magnitude (0-6)

Races in multi-winner districts saw a 7.8 percentage point increase in undervoting for every additional winner.

No RCV elections were held in multi-winner districts

Total Candidates (0-28)

Races with more than one candidate saw a .6 percentage point decrease in undervoting for every additional candidate.

RCV races with more than 1 candidate saw a .5 percentage point decrease in undervoting for every additional candidate.

Unopposed (0-1)

Races where one candidate ran unopposed saw a 13.3 percentage point increase in undervoting.

RCV races where 1 candidate ran unopposed saw a 11.6 percentage point increase in undervoting.

Open Seat (0-1)

Races where the incumbent was not seeking re-election saw a 2.4 percentage point increase in undervoting.

RCV races where the incumbent was not seeking re-election saw a 2.7 percentage point increase in undervoting.

Competitive? (0-1)

No significant effect

No significant effect

City Partisanship (0-100)

For every 1 point increase in democratic partisanship, undervoting went down by .1 percentage points.

No significant effect

Turnout (CVAP) (0-100)

No significant effect

No significant effect

Did Election use RCV? (0-1)

No significant effect

Invalid criterion for this regression

At-large election? (0-1)

No significant effect

No significant effect

African American % of District (CVAP) (0-100)

No significant effect

No significant effect

Hispanic % of Disctrict CVAP (0-100)

No significant effect

No significant effect

Asian % of District CVAP (0-100)

For every 1 point increase in Asian turnout as a percentage of civilian aged population, undervoting decreased by .2 percentage points.

For every 1 point increase in Asian turnout as a percentage of civilian aged population, undervoting decreased by .1 percentage points.

Population 25 and over with a HS degree (0-100)

For every 1 point increase in percentage of the population 25 and over with a highschool degree undervoting increased by .1 percentage points

No significant effect

 

Three factors that drove undervoting down significantly were high competition, whether a race was for an executive position (e.g. mayor, treasurer, district attorney) and city partisanship. It makes intuitive sense that voters would participate more in competitive elections, and executive positions likely saw less under-voting because they are the top of the ballot for odd-year elections. More research is needed on why city partisanship might drive down undervoting.

Another major conclusion is that the racial breakdown of a district has little to do with its level of undervoting either in RCV elections or the total election pool. This suggests that the fear of RCV skeptics that more complicated ballots would lead to disproportionate drop off among minority voters is unfounded.

Undervoting, exhausted ballots, and more will be explored in an upcoming report on voter behavior and ballot irregularities in the Bay Area.

Show Comments
comments powered by Disqus

Join Us Today to Help Create a More Perfect Union