This election is from the information available, going to be a coin flip. I hope to do better than a coin flip in these predictions I will make. First thing’s first, however, a history of RealClearPolitics average polling for major races (President, Senator). Gubernatorial races are not figuring prominently this year, and the only close one to speak of worth attention is New Hampshire, where Republican Kelly Ayotte is slightly leading in polling at the moment. The endorsement of Governor Sununu may be sufficient to put her over the top. Before we carry on here, I want to go over some polling history.
Presidency
| Presidential Election | 2016 | Net Bias | 2020 | |||
| State | RCP Average | Actual | RCP Average | Actual | Net Bias | |
| AZ | R +4 | R +3.5 | R +0.5 | D +0.9 | D +0.3 | D +0.6 |
| GA | R +4.8 | R +5.1 | D +0.3 | R +1 | D +0.3 | R +1.3 |
| MI | D +3.4 | R +0.3 | D +3.7 | D +4.2 | D +2.8 | D +1.4 |
| NV | R +0.8 | D +2.4 | R +3.2 | D +2.4 | D +2.4 | None |
| NC | R +1 | R +3.7 | D +2.7 | R +0.2 | R +1.3 | D +1.1 |
| PA | D +1.9 | R +0.7 | D +2.6 | D +1.2 | D +1.2 | None |
| WI | D +6.5 | R +0.7 | D +7.2 | D +6.7 | D +0.7 | D +6 |
This is indicative of largely Democratic bias over the last two cycles, but 2012 had Republican bias, and as you will see, the 2022 midterms had a considerable Republican bias, but there is a caveat to this one.
Let’s look at some indicators:
Good for Trump:
. Republicans outnumber Democrats nationwide, a new development (Archacki). The independent vote may turn out to be the deciding factor, however.
On November 1st, 538 put Trump at a slight advantage with a 53% chance to win, but greater odds have been overcome, most notably by Trump in 2016.
The betting markets are favoring a Trump victory. The predictive performance of the betting markets since 1916 has only failed thrice: in 1916 itself, 1948 (polling ended two weeks before the election in that one), and the tremendous upset of 2016.
Per Nate Cohn, there isn’t evidence yet that the pollsters’ Trump voter counting problems are fixed.
Nate Silver, formerly of 538, predicted a Trump win up until final projection.
The polling of the election may suffer from a non-response bias that favors Trump (Grover).
Defections from the Democrats in Michigan from Arab American voters, with a possibility that Trump gets more of them than Harris does, with left-wing Arab Americans potentially voting for Green Party candidate Jill Stein.
The Trump campaign, according to Jack Herrera of Politico, has consistently had a stronger ground game in Pennsylvania than the Harris campaign, which could spell victory in the state despite the “garbage island” incident. As someone who has worked in campaigns, I strongly believe in ground game. In 2020, I knew the Democrats were not going to do as well as polls predicted because they were behind on ground game, namely door knocking, and particularly so in Florida. However, although Trump himself survived COVID, his reelection did not.
Kamala Harris is not an inspiring candidate aside from her being a woman and most importantly, not Trump, at least from a standpoint of those not ideologically motivated. Her record in the Senate was one of the most liberal, and she campaigned as a hard California liberal in the 2020 Democratic primary.
Republicans have higher favorability on the economy per Gallup polling. The last time a party’s candidate for president lost when the party had an edge on the top issue, which is the economy this year, it was in 1948. And I’ve mentioned the issue of polling in 1948.
Good for Harris:
Historian Allan Lichtman’s keys to the presidency system has predicted since 1980 the outcomes of all presidential elections except 2000.
The Des Moines Register poll of Iowa actually puts her three points ahead. This is most likely an outlier, but this poll has had some good history behind it, but an Emerson poll also came out that had Trump up in Iowa by ten points and Emerson is a highly reputable firm. If the Iowa poll is indicative of movement towards the Democrats in Iowa, then that doesn’t speak well for Trump’s ability to win less solid states. But one must remember of course that this time their off.
Per 538’s Nate Cohn, pollsters may be overly weary of underestimating the Trump vote, making an undercounting again seemingly unlikely.
Nate Silver’s final projection gives Harris the slightest edge over Trump.
The subject of abortion may be a sleeper factor here that increases Harris’s share considerably with moderate and independent women and wins her the election.
The “island of garbage” line at Trump’s Madison Square Garden rally, which may tip some needed Latino voters in Pennsylvania.
Donald Trump has a lot of baggage to put it mildly, and some older voters may be nervous about him being the more chaotic candidate, again putting it mildly.
Michigan Democrats have a stronger ground game than Republicans. Although Republicans have undergone some course correction after the catastrophic tenure of Kristina Karamo as party chairman and this will help them some, early voting is indicating a trend towards an over 50% turnout from Detroit (Massey & Guillen). A 50%+ turnout is an indicator of a Democratic win.
Now…let’s look at the RCP poll averages from Monday evening for the seven swing states. I know Minnesota and New Hampshire were recently added to tossups, but I doubt this is going to happen.
Arizona – Trump +2.8
Republicans have been making gains in Arizona in registration since 2020, and there is no recent poll that has Trump in a polling deficit. The closest one is Morning Consult, which has a tie. In this final stretch, Arizona is the sick man of the swing states for the Harris campaign and the champion for Trump’s. Trump wins the state.
Georgia – Trump +1.7
Most polls have Trump leading in Georgia, with New York Times/Siena being the only one to indicate a Harris lead. NYT/Siena is a highly credible polling firm, but so is Emerson, which indicates a one-point lead for Trump. I think Trump pulls through in Georgia.
Michigan – Harris +0.5
The numbers in Detroit, as I mentioned in the positives for Harris, make me think that Michigan, despite Arab American defections, will stay blue this time. The polling momentum appears going slightly to Trump, but a flurry of polls late in the game pointed to a GOP wave in 2022 that didn’t materialize in most of the nation. What’s more, the Michigan GOP is still recovering from the disastrous tenure of Kristina Karamo as its party chair. I think ground game matters a lot, and Democrats have a better ground game in Michigan. Much of the ground game right now is through Elon Musk’s America PAC, which is a major push to reach people with a low propensity for voting (LaHut). Harris wins Michigan, the sick man of the swing states for Republicans.
Nevada – Trump +0.6
This one is rather interesting. Although Trump is up, Nevada has had some history of over-polling Republicans. However, it was one of the states in 2020 in which the pollster average was the actual margin of victory for Biden. Nevada political expert Jon Ralston has predicted that Harris will win but by the narrowest of margins with mail-in ballots. Ralston is a credible source, as he predicted the outcome of 2020 in Nevada as well as Republican Joe Lombardo winning in the gubernatorial election in 2022 and Democrat Catherine Cortez Masto winning reelection to the Senate. However, Ralston has also predicted margins of victory, and he has overpredicted Democratic margins in the past. He predicts for this one that Harris will win by 0.3% (Schwartz). In 2020, Ralston predicted a Biden victory in Nevada by four points, but Biden actually won by 2.4%, a difference of 1.6% in favor of Democrats and in 2022 he predicted a 2 point victory for Masto but Masto won with 0.8%. If his being off by 1.2 to 1.6% in favor of Democrats holds up, Trump wins the state. The state’s Democratic registration edge also has declined considerably from 2020 and many of the state’s new residents are from California and may be against their old state’s political leadership. Not good for Harris if true. Ben Margiott of Las Vegas Channel 3 News reports that “Republicans continue to hold a roughly 4% turnout advantage”. I predict a Trump win by the skin of his teeth.
Ralston’s 2020 and 2022 predictions:
https://www.reddit.com/r/YAPms/comments/1gjqsc6/jon_ralstons_2020_presidential_and_2022_senate/#lightbox
North Carolina – Trump +1.2
Trump has been leading in most polls in North Carolina, and the early vote looks good for Republicans and not so good for Democrats. That being said, early voting is not necessarily predictive. It is difficult to see how Harris outperforms 2020 Biden, save a considerable enough migration of college-educated whites from four years ago. I predict a Trump win.
Pennsylvania – Trump +0.4
Most agree that this is the must-win of the swing states. The interesting thing about Pennsylvania is it is one of the two states in which the poll average matched the outcome in 2020! Furthermore, the ground game of the GOP in this state has been outperforming that of the Democrats for months (Herrera). However, there has been more elderly Democratic early voting, which may not bode well for the GOP. Democrats, however, have an advantage in early voting shaved by 600,000 votes. With more Republican votes in, this gives them more room for GOTV on Election Day. On balance, I call this one for Trump.
Wisconsin – Harris +0.4
Harris’s numbers look the second best in Wisconsin, and I think the days of polling discrepancies in the state may be over. I think that Harris is favored to win the state based on Democratic ground game, which is stronger than the Republican game in the state.
Senate
Something to bear in mind about Senate polls historically is that there has always been at least one upset every year. An upset is defined in this post as one in which the opposite result occurs from the RCP polling average.
Races in which upsets occurred over the past ten years:
North Carolina, 2014: Tillis vs. Hagan
RCP Projection: D +1.2
Actual Result: R +1.7
Democrats spent the most money on this race to protect Senator Kay Hagan of North Carolina and she led in most polls, but the midterm went the Republicans’ way and Tillis pulled off an upset.
New Hampshire, 2016: Ayotte vs. Hassan
RCP Projection: R +1.5
Actual Result: D +0.2
Senator Kelly Ayotte was leading in most polls up to the election, but as the state voted for Clinton, Governor Maggie Hassan got enough of the vote to pull a squeaker.
Pennsylvania, 2016: McGinty vs. Toomey
RCP Projection: D +2
Actual Result: R +1.6
Senator Pat Toomey of Pennsylvania won a narrow reelection victory even though only poll put him ahead, this tracked with Trump winning the state, although Toomey ran ahead of Trump.
Wisconsin, 2016: Johnson vs. Feingold
RCP Projection: D +2.7
Actual Result: R +3.4
Senator Ron Johnson of Wisconsin was thought at the beginning of 2016 to have a DOA campaign. However, despite grim prognostications throughout the year and most polls of the Senate race putting former Senator Russ Feingold at an advantage, Johnson won by 3.4. Feingold’s RCP average was 2.7 up.
Arizona, 2018: McSally vs. Sinema
RCP Projection: R +1
Actual Result: D +2.3
Republican Senator Martha McSally was facing popularity problems and ran a lackluster campaign. Despite late polls seeming to go her way, Kyrsten Sinema defeated her.
Florida, 2018: Nelson vs. Scott
RCP Projection: D +2.4
Actual Result: R +0.2
Democratic Senator Bill Nelson of Florida after three terms was defeated by Republican Rick Scott. Nelson had led in most polls, but there was an increase in the number of polls that put Scott over the top.
Indiana, 2018: Braun vs. Donnelly
RCP Projection: D +13.
Actual Result: R +5.9
Democratic Senator Joe Donnelly led in the majority of polls in the final stretch, but Republican Mike Braun pulled off a victory in Republican Indiana.
Nevada, 2018: Heller vs. Rosen
RCP Projection: Tie
Actual Result: D +5
This may not technically be considered an upset, but Heller had been leading in two of the three last polls, and the result was one point higher than Emerson’s poll figuring Rosen at +4.
Maine, 2020: Collins vs. Gideon
RCP Projection: N/A, but all polls had Democrat Sara Gideon up, although RCP considered the race a “toss up”.
Actual Result: R +8.6.
Although polling was sparse, this one is scandalous, as the margin of victory for Collins wasn’t even within 5 points.
North Carolina, 2020: Cunningham vs. Tillis
RCP Projection: D +2.6
Actual Result: R +1.8
Chuck Cunningham looked good to defeat Republican incumbent Thom Tillis, who was seen as lackluster. I thought at the time Tillis would pull through a win anyway as polls had underestimated him before and Cunningham had an extramarital affair scandal.
2022 saw four upsets, all upsets were polling that favored Republicans but the seats went to the Democrats. The Republicans had candidate quality issues with the uncharismatic and weird Blake Masters, the personally troubled and questionably coherent Herschel Walker, and the comically out-of-touch Dr. Mehmet Oz. Nevada’s Democratic turnout machine managed to secure a victory for Cortez Masto over Laxalt, who made the poor decision to be loud in support of Trump’s election denial in 2020. I have my doubts the poll bias will be this bad this time around towards Republicans, and much of this bias actually happened in the last week of the campaign as a flurry of bad polls came out, and contrary to popular belief it was more widespread than just Republican pollsters.
Arizona, 2022: Kelly vs. Masters
RCP Projection: R +0.3
Actual Result: D +4.9
Georgia, 2022: Walker vs. Warnock
RCP Projection: R +1.4
Actual Result: D +0.9
Nevada, 2022: Cortez Masto vs. Laxalt
RCP Projection: R +3.4
Actual Result: D +0.9
Pennsylvania, 2022: Fetterman vs. Oz
RCP Projection: R +0.4
Actual Result: D +4.9
2022 is an example of why the polls may be biased towards Republicans this time around rather than the Democrats, as they were from 2014 to 2020. This is a possibility, but I don’t see the bias being this much, perhaps a little on the side of the Democrats, and maybe even a bit to the Republicans once again. Now for the Senate races!
Michigan – Rogers vs. Slotkin, Slotkin +2.3
Democratic Congresswoman Elissa Slotkin has led in all but one poll, and that is from Rasmussen Reports, which although did very well in the 2004 election predictions, it has had a bit of a spottier record since 2012. Republican Mike Rogers was a decent pick, but Democrats are on their game once again in Michigan. Like Harris wins Michigan, Slotkin wins. If she loses, Democrats have had an awful election.
Montana – Sheehy vs. Tester, Sheehy +7.7
Senator Tester will be defeated in the election despite late-game optimism from Democrats. There has been no time in the history of Senate RCP polls in which a candidate was this far ahead in polling and lost. It is a testament to Senator Tester that he managed to serve three terms from a state that has repeatedly voted for Republican presidents.
Nevada – Brown vs. Rosen, Rosen +4.9
Nevada is figuring to be a tight race presidentially, but the polling has put Democrat Jacky Rosen repeatedly up, with Republican Sam Brown only leading by one in a Susquehanna poll, without doubt an outlier. Rosen wins another term.
Ohio – Brown vs. Moreno, Moreno +1.7
Senator Brown has been banking on a number of people voting for Trump and him, but Ohio’s status as a red state and Trump’s coattails look like they are going to push Moreno over the top. Brown was leading up until very recent polls, and it looks like the momentum is on Moreno’s side. I also reason this because split-ticket voting is a less common phenomenon than it used to be. I also don’t see this as a last-minute bad poll flurry as this turn isn’t widespread.
Pennsylvania – Casey vs. McCormick, Casey +1.8
Democratic Senator Bob Casey has been in office since 2007, and this is his toughest race. However, McCormick has been behind in most recent polls although the margins are close. Since every Senate election has had at least one upset, I will boldly offer the prediction that it is this race. Winner: McCormick.
Texas – Allred vs. Cruz, Cruz +4.4
Democrats try again to take down Ted Cruz with Congressman Colin Allred. They have arguably picked better this time than Beto O’Rourke, but Texas’s Republican status is not going to change with this presidential election, and there won’t likely be enough Trump-Allred voters to elect him. Cruz wins another term.
Wisconsin – Baldwin vs. Hovde, Baldwin +1.8
I was torn between this and Pennsylvania as being the shocker, but the ground game is stronger for the GOP in Pennsylvania and they just seem to be polling better. The one hesitancy I have in this is that Wisconsin polls have from 2016 to 2020 underestimated Republican strength. Indeed, Biden won by less than a point in 2020 when he was polled to win by more than six. This being said, the ground game is stronger for Democrats in Wisconsin this year, and I believe in ground game. The days of poll bias for Democrats may just be over, indeed Republican poll bias manifested in 2022, with Senator Ron Johnson of Wisconsin winning by a narrower than expected margin and Governor Tony Evers winning reelection despite RCP poll averages putting Republican Tim Michels, who denied that Trump had actually lost the 2020 election, narrowly on top. Again, candidate quality matters! Although Hovde is a better candidate and he does have the advantage of being able to self-fund, he also didn’t get the endorsement of the Wisconsin Farm Bureau Federation based on his lack of knowledge on farm issues, so I think Baldwin does get another term.
Overall, I predict Republicans end this race with 53 seats in the Senate. As for the House, it is, like the presidency, anyone’s game, but I’m going to make a little bit of a wild prediction here. Democrats narrowly take the House because of Harris coattails in blue state House races. Yes, I’m one of the few who is predicting a mixed result in which Trump does not get unified government. There are a fair number of vulnerable Republican incumbents and the legislative chaos caused by a stubborn minority in the GOP certainly didn’t help their image to govern.
References
Archaki, L. (2024, September 28). For First Time Ever, More Americans Are Republican Than Democrat. The Daily Beast.
Retrieved from
https://www.thedailybeast.com/more-americans-now-identify-as-republicans-than-democrats/
Herrera, J. (2024, November 4). Trump’s Gains With Pennsylvania Latinos Are Real. Maybe Enough to Withstand ‘Island of Garbage.’ Politico.
Retrieved from
https://www.politico.com/news/magazine/2024/11/04/latinos-decide-election-pennsylvania-00186534
Massey, D. & Guillen, J. (2024, November 2). Black turnout in cities like Detroit is make-or-break for Harris. Axios.
Retrieved from
https://www.axios.com/2024/11/02/harris-black-turnout-detroit-atlanta-philadelphia
Schwartz, I. (2024, November 4). Jon Ralston: Nevada is Going To Be Close, Mail-In Ballots That Come In Late Will Put Kamala Harris Over The Top. Real Clear Politics.
Retrieved from
Mad Politics: The Bizarre, Fascinating, and Unknown of American Political History 