On June 12, 2018, Maine held primary elections to determine the Democratic and Republican nominees for Governor, the Democratic nominee for Maine’s 2nd Congressional District, and nominees for several state legislative districts. These were the first elections to be held under ranked-choice voting (RCV) since it had been approved in a referendum in 2016 (more details on the 2016 referendum can be found here), and there was an additional referendum on the ballot to decide whether or not to continue using RCV for Maine elections.
As my previous articles on Maine have shown, Maine is an extremely territorial state. Each candidate’s hometown is printed on the ballot under their name, and most serious candidates win their hometowns by wide margins. This turned out to be the case for all four serious Democratic gubernatorial candidates, both of the top two Republican gubernatorial candidates, and all three Democratic candidates for ME-02.
Democratic gubernatorial primary
The Democratic gubernatorial primary was ground zero for RCV usage in Maine. The election featured seven candidates, four of whom were strong, so it was a foregone conclusion that none of the candidates would receive an absolute majority of votes. And indeed, the candidate who received the greatest number of first-choice votes, Attorney General Janet Mills (D-Farmington), received only about 33 percent of those votes.
Mills performed very strongly in her hometown of Farmington (where she received 72 percent of the vote) and all across Franklin County, of which Farmington is the county seat. She also performed very well in the neighboring rural counties of Oxford and Somerset, as well as in the adjacent areas of Androscoggin and Kennebec counties. Her numbers were weaker in the Portland area and especially in York County, where two of the other candidates had their political bases. Mills’ performance was generally in line with her statewide percentage in the remainder of the state. Mills did not have much of an ideological base of support, but her endorsement by EMILY’s List, as well as her being the best-known candidate in the race (she served as state AG for eight years, frequently coming into public conflict with Paul LePage, and was a state legislator for six years before that), was enough to give Mills both a plurality of the vote, and, due to the second-, third-, and fourth-choice ballots she received, the Democratic nomination.
Coming in second place in the Democratic gubernatorial primary was Adam Cote (D-Sanford). Despite never having previously held elected office, Cote ran a strong campaign and ultimately received 28 percent of first-choice votes.
Cote performed the strongest in the vicinity of his hometown of Sanford, in York County (where he received 63 percent of the vote), as well as in the Portland area and along the coast between Kennebunk and Bath. He was particularly strong in some of Portland’s wealthy suburbs, such as Cape Elizabeth, Falmouth, and Yarmouth, in which his first-choice votes exceeded Mills’ by more than 10 percent. Those towns have lots of socially-liberal voters who swung toward Hillary Clinton in 2016. However, Cote performed rather poorly in most of Maine’s rural areas, and as Maine is a mostly rural state, his lack of support there cost him dearly. The same socially liberal campaign that earned him support in the Portland suburbs had little appeal in the rural areas, where many people voted Obama/Trump. Cote, however, did perform well in Aroostook County, where the large Franco-American population noticed that of the four major candidates, he was the only one with a French last name. Cote’s relatively strong first-round support allowed him to advance to the final round of RCV, in which he received 46 percent of the vote, losing to Mills’ 54 percent.
The third-place finisher in the Democratic gubernatorial primary was lobbyist Elizabeth “Betsy” Sweet (D-Hallowell). Sweet received 16.5 percent of first-choice votes.
Sweet did not have as strong of a base of support as either Mills or Cote, but her voters were nonetheless disproportionately located in the three coastal counties of Knox, Waldo, and Hancock. Her campaign seemed to have a strong appeal to small-town, coastal, and back-to-the-land liberals (many of whom live in those three counties), rather than to either urban, suburban, or working-class voters (who gave Sweet lower levels of support). Those creative-class liberals who voted for Sweet are more numerous in places like Vermont and western Massachusetts, but in Maine they are just one part of the Democratic base, and they were not numerous enough to give Sweet a victory.
The final strong candidate in the Democratic gubernatorial primary was former state House speaker Mark Eves (D-North Berwick). Eves received 14 percent of first-choice votes.
In contrast to Sweet’s lack of a strong geographic base, Eves had a very strong geographic base – the southwestern tip of Maine, which includes his hometown of North Berwick (where he received 68 percent of the vote). Unfortunately for Eves, the seven towns in SW Maine colored dark blue on the above map contain only about 48,000 people, or less than five percent of the state’s population. One major factor preventing Eves’ support from spreading north was that it met up with Adam Cote’s base in Sanford and the socially-liberal coastal towns nearer to Portland. Eves also performed quite poorly in most of the rural areas, as the votes there were being taken by Mills (Oxford, Franklin, Somerset, and Washington), Cote (Aroostook), and Sweet (Knox, Waldo, and Hancock). Eves’ inability to substantially expand his support beyond his small geographic base is ultimately what doomed his candidacy.
Republican gubernatorial primary
In contrast to the Democrats’ gubernatorial primary, the Republicans did not need RCV to determine the winner. Shawn Moody (R-Gorham) won the primary with a solid 56 percent majority of first-choice votes. Moody had previously run for Governor in 2010 as a right-leaning independent, garnering only 5 percent of the vote, but this time he had the support of much of outgoing governor Paul LePage’s political network.
Moody had a very clear geographic base of support – the southwestern corner of the state, including all of York and Cumberland counties and part of Oxford County. Moody’s hometown of Gorham (a suburb of Portland) is located in the heart of this area. In addition, the rural areas of York, Cumberland, and Oxford have all trended substantially Republican during LePage’s governorship (the massive swings toward LePage in 2014 were a major factor in his win that year), so it makes sense that LePage’s tacit endorsement would have some sway here. Moody won most towns in Maine outside of this area as well, but the totals were less one-sided in the remainder of the state.
Coming in a distant second to Moody was state Senate Majority Leader Garrett Mason (R-Lisbon).
Mason was strongest in his home county of Androscoggin as well as in nearby Franklin County and in some of the rural areas in Penobsot and Washington counties. It should be noted, however, that, even in those areas, Mason lost most of the towns there to Moody. Moody even beat Mason in several towns in Mason’s own state senate district. Surprising no one, Mason’s best town was his hometown of Lisbon, where he received 72 percent of the vote, however in only one other town in his state senate district – Turner – did he receive an absolute majority of the vote. The inroads that Moody made in his opponents’ home turf show how sweeping his win was.
The third-place finisher in the Republican gubernatorial primary was Mary Mayhew (R-China), a former LePage Administration official.
Mayhew, more than any other major candidate in either party’s gubernatorial primaries, lacked any semblance of a geographic base of support. She lost her hometown of China to Moody by a margin of 45-28, and did even worse in the towns around China. She exceeded 35 percent of the vote in only three towns across the state, and received more than 30 percent in only seven more. Furthermore, these towns were all small and scattered around the state. Her support was in the single digits in both Gorham (Moody’s hometown) and Lisbon (Mason’s hometown).
ME-02 Democratic Primary
The ME-02 Democratic primary was largely fought between two candidates – state representative Jared Golden of Lewiston, and conservationist Lucas St. Clair of Hampden (who happens to be the son of controversial businesswoman and environmentalist Roxanne Quimby). Golden ended up defeating St. Clair by a margin of 8 percentage points among the first-choice votes, and ended up winning the nomination on the second round of RCV by an almost identical margin.
As with most Maine elections, both candidates had clear geographical bases of support. Golden performed very strongly in his home county of Androscoggin – he won it with 4,887 votes to St. Clair’s 1,331. Androscoggin provided Golden’s entire margin of victory – the election was pretty much a tie in the remainder of the district. St. Clair won his home county of Penobscot, as well as the neighboring counties of Piscataquis and Aroostook, by solid margins, however they were unable to overcome Golden’s margin in Androscoggin. There are two reasons for this: Firstly, the counties (and areas within them) that St. Clair won by larger margins are generally Republican-leaning in general elections, meaning that there weren’t enough Democratic votes in those towns to give St. Clair the victory. In addition, Golden won his home region by a larger margin than St. Clair won his. The largest town in the counties that compose St. Clair’s political base is Bangor, and St. Clair won it by a fairly narrow 50-45 margin. Golden, by contrast, won Lewiston by an 82-14 margin.
Golden now moves on to face Republican incumbent Bruce Poliquin in the general election. ME-02 voted for Trump by a margin of 10 percent, and Poliquin won his 2016 election by a 55-45 margin, so he won’t be easy to beat. However, Obama did win ME-02 in both 2008 and 2012, and that should give Democrats some hope. While Poliquin is slightly favored as of right now, this will probably end up being a very close race.
RCV Referendum
At the same time that ranked-choice voting was being used to determine the Democratic nominees for Governor and ME-02, its future in Maine was in doubt. In that same election, Mainers voted in a referendum on whether or not to continue to use RCV for all elections for which RCV is constitutional. Ultimately, the voters of Maine reaffirmed the use of RCV in Maine elections by a 54-46 margin, slightly larger than the 52-48 margin that Maine implemented RCV by in 2016.
The coalition that reaffirmed RCV usage in Maine is largely the same as the one that implemented it in the first place in 2016. The results indicate that few voters changed their minds, and the slightly larger margin this time is mostly due to the fact that Democratic turnout was higher than Republican turnout this year, since Democrats have always been more supportive of RCV in Maine than Republicans (which itself is largely due to the 2010 gubernatorial election, which Paul LePage won with 39 percent of the vote and would have lost to Eliot Cutler under RCV).
Other Odds and Ends
I have made a few more maps that I will share. This first map shows the percentage of 2016 voters who voted in the 2018 primary.
The most noticeable pattern to be found here is that the entire coast of Maine from Portland to Bar Harbor saw turnout that was substantially higher than the state average. Some of this should not be surprising – the wealthy coastal suburbs of Portland such as Falmouth, Yarmouth, and Cape Elizabeth always have high turnout. The more interesting part is that this zone of higher turnout extends well up the coast to also include liberal towns such as Camden, Belfast, Blue Hill, and Bar Harbor. This is a sign that coastal liberals are more fired up to vote this year, which is good news for Democratic state senate candidates Erin Herbig (Waldo County) and Louis Luchini (Hancock County). However, inland, neither working-class Democrats nor Republicans seem to be especially fired up for the elections this year. But there is some good news for Democrats from the results in the inland working-class towns, and my next couple maps will show it.
Before I present my final map, this next map below shows the partisan composition of the electorate of each town in the 2018 primary. To create the map, I took the number of votes cast in each party’s gubernatorial primary in each town, added them together, and then calculated the percentage of the total votes that were cast in each primary. Statewide, approximately 56 percent of votes were cast in the Democratic primary, while 44 percent were cast in the Republican primary.
The map looks fairly similar to most Democratic vs. Republican maps of Maine, but there are a few interesting differences with this map compared to the Trump vs. Clinton map in 2016. To demonstrate the most important difference, here is my final map. This map compares Trump’s performance in the 2016 election to the Republican percentage of the partisan composition of the 2018 primary electorate.
At first glance, it may be difficult to find any major patterns here. But there is one. In northern Maine, there’s a series of white working-class towns that swung very heavily from Obama to Trump in 2016. These towns include Rumford, Mexico, Jay, Livermore Falls, Millinocket, and the entire St. John Valley. And it turns out that those towns are all colored dark blue on the map.
What does this mean? It means that many of the voters who swung from Obama to Trump still consider themselves Democrats enough to vote in Democratic primaries. Now, this undoubtedly partly because Maine has closed primaries, but it also suggests that Clinton’s poor performance in those towns is an aberration, and that those Obama/Trump voters are still very winnable for Democrats in general elections as well. After all, Maine does not have the long history of registered Democrats voting for Republicans that states like West Virginia and Oklahoma have. The fact that those Obama/Trump voters have not changed their party registration is a very good sign for Democrats. Additionally, this could also mean that Republicans in those towns are unenthusiastic about voting this year. Either way, Democrats should be very happy with the partisan compositions of those towns for the 2018 primary.
I will be writing similar analyses of the Connecticut primary in August and the Massachusetts and New Hampshire primaries in September – be on the lookout for them!
Gregory Naigles 