An instant runoff ballot from the Takoma Park 2007 Ward 5 special election

An extract from the sample ballot for the instant runoff (single winner ranked choice) special election held on January 30, 2007, to fill the city council seat for Ward 5 in Takoma Park, Maryland. (Click for a higher-resolution version.) Image adapted from the Ward 5 special election sample ballot and reproduced for purposes of commentary.

tl;dr: I answer questions about ranked choice voting raised by the Howard County Charter Review Commission and others.

I was able to make it to the Howard County Charter Review Commission’s public hearing last night and request that the commission recommend the adoption of ranked choice voting for the Howard County Council, with all members elected county-wide. (See my prior post for a copy of my testimony.)

The members of the commission had various questions about ranked choice voting, as did people on Facebook when I posted a link to my testimony. My answers were off-the-cuff and could be improved in many cases, while one answer (to a question about examples of ranked choice voting in Maryland) was outright incorrect.

Here’s the complete list of questions along with my improved answers.

Wouldn’t this make running for Howard County Council more expensive, since candidates would have to campaign county-wide?

I think running under a ranked choice voting election could be somewhat more expensive, but I think this fear is overblown.

The key point is that running in a ranked choice election for one of five or seven council seats is not like running for county executive. In the county executive race a candidate must aim for more than 50% of the vote, but in a ranked choice election a candidate need only get more than 16.7% (for a five member council) or 12.5% (for a seven person council) of the vote.

That means a candidate could (and should) concentrate expensive and/or time-intensive activities like canvassing or direct mail on their natural base of voters, which might be concentrated in a particular geographical area or based on shared interests. The rest of the voters could be addressed through less expensive means, for example using social media.

If county council members no longer have districts, how would a person know which council member to call for constituent services?

I’m not really the person to ask about this. The people to ask about this are Liz Bobo, Ed Cochran, C. Vernon Gray, Lloyd Knowles, or Ginny Thomas, all of whom were elected to the Howard County Council during a time when council members were elected at large, before council districts were adopted in 1984.

But if you ask me, I’d answer as follows: Even though council members would be elected county-wide, they’d still likely have natural constituencies, based on the area of the county in which they live, particular interest groups they represent, and so on. For example, someone living in Ellicott City concerned about, say, Route 40 development would likely contact whichever council member happened to live in Ellicott City and seemed interested in that issue.

Since there would be no districts, it wouldn’t be possible for redistricting to split up those natural constituencies. If you find a particular council member is responsive to your concerns, you can continue to treat them as “your council member” for as long as they serve. And if that council member ceased to be responsive you’d have the choice of several others to ask instead.

Question: What about retaining council districts and electing members for each district using ranked choice voting?

Answer: This is a perfectly possible scheme to implement, and in fact is implemented in at least one nearby jurisdiction (see below). The technical term for it is “single winner ranked choice voting” or “instant runoff voting”. Voters express their preferences among candidates as before, but only a single winner is chosen.

One problem with single winner ranked choice voting is that it makes sense only if there’s more than two candidates. With two candidates the counting of votes works just like a regular election: whoever has the most first preference votes wins, with the voters’ second preferences not mattering.

So how do we get more than two candidates? There are two possibilities (which are not mutually exclusive). The first is that the two major parties continue to put forward one candidate apiece, but those two candidates are joined by additional candidates who represent smaller parties or who are running as independents.

With ranked choice voting, voters unhappy with the two major parties are free to express their displeasure by giving their first preference votes to someone other than a major party candidate. Their votes would not be wasted, since if their favored candidate doesn’t win their second preferences can help elect someone else, such as the major party candidate they consider the lesser of evils.

Another possibility is that the two major parties put forth multiple candidates of their own, for example, as a substitute for their traditional primaries. Alternatively, they could retain the current primaries but let two or three winners advance to the general election rather than just one. This might help when it’s not clear which of the candidates is more suited for general election success.

However, using single winner ranked choice voting by district does not help with the problem of ensuring that minority groups are represented on the county council. Only if members of such groups were concentrated in one or two districts would they have enough critical mass to help get their favored candidates elected.

What about keeping the five council districts as is, with election by district as done today, and adding two at-large council members elected county-wide?

If the at-large elections were conducted as they usually are, this would almost certainly just result in adding two Democratic council members to the council.

Each party would nominate two candidates for the at-large seats. Given that Democrats have a significant edge in both party registration and votes county-wide (based on the 2018 election), and given that most people (even supposedly unaffiliated voters) would likely vote a straight party ticket, the two Democratic candidates would almost certainly end up with the most votes and be elected.

We know this because this is exactly what happened on a larger scale during the period when Howard County elected all five council members at large. During that era Republicans had more than one member on the county council only once, at a time when Democrats and Republicans were more evenly matched in voting strength than they are today. In 1974 the effect was so pronounced that all five Democratic council candidates were elected and Republicans had no member on the council.

What about using ranked choice voting for the two at-large seats? This would almost certainly always lead to one Democrat and one Republican being elected at-large. In a ranked choice election with two candidates, a candidate can be elected if just around 33.4% of voters select them as their first preference. It’s a simple matter of arithmetic: if two candidates each get 33.4% of first preference votes, or 66.8% between them, the best that any other candidate can do is to get 33.2% of first preference votes and come in third. The first two candidates would then be elected to the two seats.

Given the number of Republican voters in the county and past voting patterns for county council elections (see below), if the Republican party ran a single candidate then that candidate would likely clear the bar of getting 33.4% of first preferences. The same is true of the Democratic party. The two parties would then split the seats between each other.

If ranked choice voting were implemented with a seven-member council and county-wide elections, what would be the likely split between Democratic and Republican council members?

If there were no other parties were involved, my best guess is that we’d see at least two Republicans on the county council (i.e., a 5–2 split), and possibly three (a 4–3 split).

The math again is fairly simple: Using a similar argument as above, with ranked choice voting for a seven-member council a candidate would win if they got just over 12.5% of the first preference votes (100% divided by 8).

In the 2018 election Republican candidates for county council collectively got 43,772 votes out of a total of 136,524 votes in all county council races combined, or about 32%. However, there was no Republican candidate in District 3; if there had been one then the total vote for Republican county council candidates might have been a few points higher, say around 36–38%.

If Republicans ran two candidates in a ranked choice election for seven council seats, and the two GOP candidates each got an equal number of first preference votes, a 32% share of the total vote would translate into 16% of the first preference votes for each candidate, more than the 12.5% needed to get elected.

Could three Republicans win election to a seven-member county council with ranked choice voting? It’s possible, but it would be tight: With three candidates and an equal share of first preference votes, the three candidates collectively would have to get more than 37.5% of the vote (3 times 12.5%), significantly more than the 32% GOP share in the 2018 council election.

However, even if there were not enough first preference votes to elect three Republicans it’s possible that a third Republican might be elected if they could secure a fair number of second preference votes from voters who selected Democratic candidates for their first preferences. The type of candidate best positioned to do this would be a moderate Republican who appealed to many Democrats, for example someone like Allan Kittleman.

Is any jurisdiction in Maryland currently using ranked choice voting?

I answered “no” to this during the hearing, but I was wrong. It turns out that the city of Takoma Park has been using ranked choice voting since 2006 for both its mayoral elections and elections to its city council. All of these are single winner ranked choice (“instant runoff”) elections. Six members of the city council are elected by ward, and the mayor is elected at-large.

Takoma Park’s experience is not fully applicable to Howard County, for at least two reasons. First, Takoma Park does not use a multiple winner ranked choice system like I’ve been discussing above. Second, the total number of votes cast is relatively small. This makes it feasible to count the votes by hand, something that would be relatively time-consuming in Howard County.

However the use of ranked choice voting in Takoma Park does indicate that the concept is not foreign to Maryland, and it seems to be both well-understood and popular among voters in the city.

Would the Maryland legislature have to pass special legislation in order for Howard County to use ranked choice voting?

I’m not a lawyer and can’t give a definitive opinion on this. However to my knowledge the only restriction imposed by the Constitution of Maryland on Howard County is as follows:

The charter for the government of any county governed by the provisions of this Article may provide for the election of members of the county council by the voters of councilmanic districts therein established, or by the voters of the entire county, or by a combination of these methods of election. (Article XI-A, Section 3A.)

A ranked choice election conducted on a county-wide basis would appear to satisfy this requirement.

That’s all the questions I can remember. If you have further questions please feel free to ping me on Facebook or Twitter (@hecker).

For further exploration

For more information on ranked choice voting in Takoma Park, see the following:

Numbers for Republican vote share in the 2018 county council elections are from the official 2018 gubernatorial election results for Howard County. See also my 2012 post “How to win an STV election for Howard County Council, part 1” for an example of how a ranked choice election might have played out based on the 2010 election results.

Finally, Section 3A of Article XI-A of the Constitution of Maryland appears to be the governing language on how Maryland charter counties (like Howard) can conduct council elections.