On March 18th, Republican Congressman Don Young passed away, leaving Alaska without a representative in the Congress for the first time since 1973. Young, formerly Dean of the House, enjoyed a half-century tenure, making him the longest serving Republican representative in history. Along with that of former Senator Ted Stevens, Young’s legacy has made an indelible impact on the Last Frontier and its inhabitants.
The ensuing special election was the first in state since 1973, when Young himself narrowly won a race to replace Democrat Nick Begich, who had died in a bush plane accident while campaigning for reelection. This year’s contest was particularly-notable because it occurred after the passage of a 2020 Alaskan referendum to adopt ranked-choice voting, making it the first House special in American history to be held under such a system.*
Voters first went to the polls in a June primary to choose which four candidates would advance to the general election ballot. They selected former Governor Sarah Palin (R), businessman Nick Begich III (R), surgeon Al Gross (I), and ex-state representative Mary Peltola (D). Gross, who lost to Republican Dan Sullivan in Alaska’s 2020 Senate race, proceeded to drop out of contention.
On August 16th, voters utilized the new RCV system to indicate their preferences for the three remaining candidates. Members of the electorate had multiple options when it came to their ballots. Most ranked every contender, some only one, while others opted to write-in or leave the ballot blank.
The majority of the returns was reported shortly after election day, but the raw count continued to change for days after the initial vote due to absentee and overseas ballots. Despite the lack of complete numbers, Split Ticket was quickly able to project a (D) Peltola vs. (R) Palin instant-runoff.
After weeks of anticipation, Alaska conducted all of the ranked-choice elimination rounds at once and published the final results for the special election. In a fascinating turn of events, Peltola beat Palin 51.5–48.5%. Her victory marks the first time that Democrats have won Alaska’s At-Large district since 1972, when Don Young lost to Congressman Nick Begich.
Why did Palin lose? That question will be touched on in detail over the coming paragraphs, but the simplest explanation is as follows: excluding exhausted ballots and write-ins, 37% of Begich’s first round voters ranked Peltola second. To win, Palin would have had to receive more than 63% of those voters’ second place preferences, something she could not accomplish because of high unfavorability ratings.
Trivia: Because Peltola won by a competitive, albeit not particularly-close, margin in a double-digit Trump seat, she now holds the record for overperformance relative to presidential partisanship in a special election occurring between 2021-2022.*
*Maine is the only other state that uses RCV for House elections; it has not had a special election since the system’s adoption*
*The TX-06 special election is excluded*
ANALYSIS – DATA BREAKDOWNS
Pundits, by their nature, love heuristics because they fuel their ability to recognize patterns. But empirics also shed light on instances when patterns and long-standing assumptions fail, giving observers the chance to guess where political tailwinds are blowing.
It is for these reasons that pundits are preternaturally drawn to exceptional events and circumstances like the election of an independent to national office or a midterm election in which the party holding the White House defies history to gain seats.
Against the backdrop of Democrats hoping to overcome political gravity and make 2022 a so-called “asterisk election“, former state representative Mary Peltola surprised many after defeating Sarah Palin by a margin of 2.9% – as of September 1st. She is the first Alaska Native to represent her state federally.
The question now is whether Peltola’s victory is destined to become a mere political footnote, or a harbinger for her state, the country, and her next election this November?
To answer that question, one ought to put Peltola’s win in the context of the turnout and persuasion environments that the election took place under. First and foremost, Alaska’s base partisanship has grown inexorably bluer over the past two decades, down to a presidential lean of only 14 points right of the nation as a whole (from a whopping 40 points only two decades ago).
Notably, congressional results have also been converging with presidential partisanship, serving as a sign of the polarized times we live in.
Because the state held its regular primaries at the same time as the special election, the environment is probably reasonably close to that of a primary. But what does that say about the electorate which gave Peltola her win?
Looking at all of the recent historical statewide elections without confounding irregularities, such as the Murkowski coalition and true third-party independents, Peltola’s winning environment was relatively in line with other primaries, simply with increased turnout.
Notably, disregarding the final margin, the conditions of the fall primary look far more like the 2014, 2016, and 2018 elections’ primary environments than they do 2020’s, an important distinction, as 2020 is somewhat of an outlier when it comes to general principles surrounding Alaska primaries.
Reasons for its anomalous status include: the combination of a non-competitive Republican senate primary and COVID-induced absentee ballots, both of which caused Republican margins to grow between the primary and general elections.
Using those cycles as a baseline, a trend-only primary model would predict a base 2022 R vs. D margin of about 22 points; the actual primary had a two-party margin of 22.29, though this number is mostly based on guesswork.
Looking more closely, the primary-general relationship in Alaska has a few key dynamics. The first and most significant factor is differential turnout. Namely, Democratic voters in Alaska tend to make themselves heard to a lesser degree in Democratic Primaries. There are a multitude of factors undergirding this claim.
On the one hand, there are only about half as many registered Democrats as Republicans, meaning that a greater percentage of Democratic general election voters are unaffiliated and are thus represented less in primary elections than partisan Democrats.
Democrats might also vote strategically in Republican primaries in order to have a greater chance to affect election outcomes. Geography plays a large factor as well because the strongest members of Alaska’s democratic constituency are its Natives in the rural parts of the state.
By analyzing the changes between primary and general electorates, one can see that (with the exception of 2020) Republican margins have fallen from primaries to general elections, often precipitously in midterm years.
In the chart below, the left bars represent the primary electorate, and the right the general electorate. Dark colors represent that parties’ respective primary voters. 2020, in this context, is brought into focus because the low margins in the primary gave the Democrats less room to grow. This was partially because GOP primary turnout was also low, but could also be attributed to Democratic primary participation, which suggested Republican voters were taking advantage of the open system to vote in more-competitive primaries.
Looking even more granularly, we see that, except for the 2012 House election (a blowout victory for the now-late Rep. Don Young) and the aforementioned 2020 cycle, there is a fairly consistent marginal shift of between 15 and 20 points in the Democrats’ favor. One caveat: this is the first primary election post-RCV, so exact partisan composition is difficult to determine. There is obviously no way to know for sure that this shift will have an impact this year, but if it were to actually occur, catastrophe would meet Palin or any Republican.
With the exception of 2020, Democrats improved on their primary totals by a greater factor than Republicans did. In both midterm elections in the dataset, Democrats even made gains in the raw vote margin, an impressive feat for a party half the size of the state GOP.
The explanation for this is fairly straightforward. Alaska’s Native population is its most democratic-leaning constituency, but it generally turns out to greater degrees in general elections than in primaries. In AK House Districts 36 through 40 (Bush region), for instance, primary turnout only averaged 20%, whereas for general elections, it more than doubled.
Putting it all together, the environment in which Peltola won her seat was in line with the continuation of long-term trends in the state towards the Democratic party, with a combined Republican vote margin in the low-20% range. That is certainly lower than in past elections, but not by much, and nearly aligns with the year-over-year partisan drift of the state.
Forecasting the November rematch without knowing the particulars of what plays out between now and then is a fool’s errand. After all, according to polling, Begich likely would have soundly defeated Peltola had he advanced to the final round instead of Palin. The overall difficulty that Peltola will face in November is highly dependent on the relative strength of Palin and Begich come election day.
If the data supporting this hypothetical prove accurate, Peltola would end round one with between 42 and 46 percent of the vote. That would be enough to put her over the line, even with a modest exhaustion/defection rate among Palin’s voters. In this scenario, Palin would find herself in an even tougher position than she did during the special, meaning she would almost certainly lose in the instant-runoff if she were to advance.
While the candidates’ November positions are unknown, Split Ticket can look at their respective geographic coalitions to see where each hopeful’s pockets of strength are concentrated. The data also show which parts of the state are over- and underrepresented in the primary electorate, allowing us to predict the environmental conditions which could provide headwinds for specific candidates.
Alaska’s primary-general cycles have predictable, regional trends in primary and general elections. Some occur in areas representing a smaller portion of overall ballots cast and others in regions accounting for a higher portion of total ballots received.
When looking at this relationship, it is important to understand that comparing ballots cast to 2.5% is not empirically sound, largely due to differences in district-level and demographic turnout.
Given those issues, one should instead weight each district by how the ballots cast in this primary election compare to those same house districts’ cast ballots in previous general elections. We are using the 2022 house districts to look at historical data.
Upon completing the weighting process, Split Ticket looks on a precinct-level at where each candidate over- and under- performs his or her statewide proportion of first-choice ballots; this is designed to signify how each contender’s specific geographic coalitions will become more or less powerful going into November.
These factors are small compared to persuasion, with the shift in each house district only responsible for maybe a tenth of a percent change in general election margin, but could nonetheless prove decisive for the pivotal question of whether or not Palin or Begich makes it to the final round.
Peltola’s coalition of first-choice voters was very strong in Anchorage, particularly south Anchorage, which is traditionally more Republican than the city’s northern half. Despite its light-blue tint, this represents quite a significant overperformance.
The Democrat also performed well in Juneau, while falling behind a bit in the Bush – though the North Slope and the Bush in general is likely skewed by the presence of write-in Tara Sweeney, whose voters splintered between the candidates by the final round. Make no mistake, Peltola assembled an impressive coalition for a Democrat in Alaska that is representative of the continued blue encroachment into previously-pink suburban communities.
Peltola’s coalition might benefit less, in terms of regional primary-general shift, than a ‘generic’ Alaska Democrat, because the gains she has made over past Democratic’ primary electorates come significantly from South Anchorage, a place which is relatively less important in November elections than in primaries.
Begich’s coalition did well relative to Palin’s in Anchorage, especially in its southern portion. He also performed admirably in the eastern parts of the Mat-Su Valley and the Bush. Palin, by contrast, squeezed Begich hard in Fairbanks, the western Mat-Su Valley, Ketchikan, and the Kenai Peninsula.
Looking at the geography, it would appear that Begich gets a significant edge over Palin from geographic primary-general change. But this expectation is deceiving for a few reasons.
Firstly, the parts of the Bush where Begich overperformed Palin simply don’t cast a lot of votes for Republicans; the high turnout changes will not really benefit him all that much, and Palin can look forward to Fairbanks and Ketchikan being a more significant part of the electorate in November.
Worse still, Begich is saddled with dominating Palin in regions which will punch below the weight they punched in the special election around Anchorage and the eastern Mat-Su Valley. While Palin did outperform Begich in the Kenai Peninsula (an area which has the most precipitous decrease in impact from primaries to generals), she also did extremely well in the crimson-red areas that stretch from Anchorage east to the Canadian border and north to Fairbanks (Chugach, Copper River, and Southeast Fairbanks). All told, Palin likely has a slightly better coalition going into November than does Begich.
It is also important to point out that another candidate will be appearing on the November ballot, Libertarian Chris Bye. Because of how ranked-choice voting works, this may have larger implications than one might think if he takes votes disproportionately from either Begich or from Palin due to exhaustion, though at this point it is difficult to determine the extent to which this is a factor.
It is, in Split Ticket’s opinion, pundit malpractice to put Alaska into a generic house model with the rest of the country after the state’s adoption of ranked-choice voting. Why? The parties’ are far more geographically based and locally-oriented for a national model to understand.
Based on a primary-shift model of Alaskan elections specific to the state, Peltola would be favored to win re-election if she were to face Palin in the final round in November. Against Begich, she would be an underdog.
Before the special election, Split Ticket rated this race SAFE REPUBLICAN. On election day, the contest shifted to LIKELY REPUBLICAN to better account for vulnerabilities that could be accentuated by Palin’s unpopularity. That move proved to be correct in principle, but wrong in reality. Peltola won despite the fact that we predicted a Republican win, even with Palin.
Mistakes happen, which is why our approach is changing. Following the final results, Split Ticket announced a ratings change for the upcoming general election from LIKELY REPUBLICAN to LEANS REPUBLICAN. Until we know whether or not a Palin vs. Peltola instant-runoff will occur after November, only a modest single-category change is merited. If the November electorate is in fact more beneficial to Peltola, as analyst Galen Metzger argues it could be, Democratic chances against someone like Palin would be very good. In other words, a TOSSUP rating might be on the horizon in Alaska. However, until we know whom Peltola will face in the instant-runoff, our rating will remain the same.