Captain K Debunks More Rigged Polling with Facts - Iowa Edition

Topic: 2024 Election Forecast

In June, I wrote about a poll that didn’t get much fanfare since it comes from a state that will vote for Donald Trump by a large margin and has done so in the past two elections.  It was Ann Selzer’s poll of Iowa with Donald Trump leading Joe Biden by 18 points, 50-32%.  Selzer traditionally releases a September poll and then a final about a week before the election.

In 2020, she released her September poll showing Trump and Biden tied at 47% each, before showing Trump +7% at the end.  Trump would go on to win the Hawkeye State by 8.2% in a quasi-election flooded with mail-in voting and every cheat known to man in play.  She has a track record for accuracy in her final polls, and an equal record for portraying false nail biters when the chips aren’t on the table.  In this case, her Trump vs. Harris poll was greatly anticipated, because as my aforementioned article outlines, there is no precedent since 1960 for Wisconsin and Iowa to be more than 9.8% apart from one another in any direction, meaning Trump plus double digits in Iowa presents a relative certainty that he will pull along Wisconsin and its overlapping demographics for another ten electoral votes, and crucial electoral votes at that.

Now, 18 points is a massive beating, especially in a moderate Midwestern state that voted for Barack Obama twice, and 32% for the Democrat nominee is unrealistically low for a final result, as is Trump being stuck at just 50%, given that Trump scored 51.2% and 53.1% in each of the past two elections, respectively.  I think it is fair to expect Harris to improve upon an 18% deficit… but by how much?

Well, yesterday, Selzer’s poll ran for the Des Moines Register.  The result, surprisingly, showed Trump leading just 47-43%.  Optimists will note that Trump is four points ahead of where Selzer had him in terms of victory margin four years ago at this point, but if you think a poll is moving 14 points in three months when swapping members of the same administration, you need to see a shrink.  You should also ask yourself why Trump continues to poll ahead in Pennsylvania and Michigan frequently and hasn’t lost double digits in margin in those states.  She has Trump leading with men by a margin of 27% (59-32%), but Harris up with women 53-36 (17%), a number far surpassing the national polling averages and very unlikely in a state like Iowa, which also has women living in the torch-red boonies and working-class counties.

Nevertheless, I’m sick and tired of dealing with people who produce results as they are ordered to produce them, and for no other reason than to cast doubt on the electoral situation in Wisconsin, which stands to impact the Electoral College winner.  Selzer will most likely return to form and produce a Trump +9 or more poll for the release just before the General Election, but where exactly do I think Iowa sits today?

Simple – let’s consult the ever-reliable party registration indicator:

Let me teach you how to read this chart.  In 2004, the Republican Party of Iowa had 4,402 more registrations than the Democrats had, or a lead of 0.4% (expressed as R+0.4% in political speak).  That year, George W. Bush won the election there by 10,059 votes over John Kerry.  10,059 in margin is 5,657 more than the Republican registration lead (10,059 minus 4,402).

In 2008, Democrats took the lead in registrations, and by a country mile.  They led by 106,442, or 5.4% (D+5.4%), thanks to a swing of 110,844 registrations to the left.  That was telling the electorate that Obama was not only going to flip Iowa, but flip it with ease.  He won it by 146,541 over John McCain en route to a national blowout.  Obama overperformed the Democrat registration advantage by 40,099.

Obama lost more than three million votes in reelection and began undoing the Democrat death grip on the white working class in reelection, and by 2012, the GOP had regained the registration advantage in Iowa by a hair.  That suggested a move to the right, and it proved correct.  Obama held the state, but by 91,927 – a reduction in margin of 54,614.  Obama overperformed registration by 93,327 as the incumbent, with the GOP nominee incapable of speaking to the working class.

Things changed substantially when Donald Trump came on the scene, though registrations didn’t move a lot. The state moved to the right by about a point in party registration, suggesting an electoral move right, and it was a blowout, with Trump taking Iowa by 147,314, or 114,228 over the registration advantage.  Thanks to the Democrat caucus in 2020, an important early indicator in Iowa, compelling party registration, the state dipped slightly to the left, and the result showed a slight dip to the left, with Trump winning by slightly less than he did in 2016, but still outperforming the GOP advantage by 118,021.

What can I deduce from the five elections above with regard to party registration?

·      Obama won independent voters in Iowa by a wide margin, as evidenced by overperforming a Democrat registration advantage in 2008, and vanquishing a small GOP advantage in 2012.

·      Trump dramatically overperformed the GOP registration advantage in 2016 and 2020, showing a swing of independents to his side.  Joe Biden had fewer votes in 2020 than Barack Obama did in 2008 or 2012, and I still think Biden’s total is somewhat inflated.

·      There are 232,587 fewer active Democrats in Iowa today than in 2008, and Iowa continues to gain modestly in population.  The Democrat Party in Iowa is on life support, and all 99 counties in Iowa are more Republican than they were in 2020 by party registration.

That leads me to introduce 2024 to the data set:

Well, would you look at that.  The GOP has expanded its voter registration edge by a mere 147,415, making the 2012 shift look like child’s play, when in fact its primary driver was the same – movement away from Democrats. 

The GOP, on average and including the disastrous 2012 campaign, outperforms its registration edge by 20,896 votes, and with Trump, the number moves to an incredible 116,125, capturing his support among working-class Democrats and independent voters.

The GOP registration lead as of this month = 168,005 (R+10.7%!)

168,005 + GOP average overperformance of 20,896

EQUALS

Trump +188,901 (roughly 12%)

---

168,005 + TRUMP average overperformance of 116,125

EQUALS

Trump +284,130 (roughly 18%)

Take note that party registration is a lagging indicator of vote expression, so a lot of this registration volatility likely represents the formal movement of two-time Obama voters who switched to Trump officially changing sides; however, Iowa’s voter registration indicator is perfect in predicting trajectory and swing this century, and there is no reason to believe it won’t be again – even the major MSM aggregators have Trump winning Iowa in double digits, and if that happens, there is no way Harris will win Wisconsin.  See below in the two columns furthest right:

Iowa and Wisconsin have been 124,566 (2016) and 159,293 (2020) ballots apart in Trump’s two races, for an average of 141,930.

Trump should win Iowa by a minimum (shown above) of 188,901, which gives him Wisconsin on the low end by 47,990 (1.4%), which is right in line with my pessimistic forecast of the Badger State, and if he hits the high end estimate, which is probably overly optimistic, he would win Wisconsin by about 143,219 votes (4.1%). Trump will need to be held to a victory of less than 9.5% in Iowa for Wisconsin to have a shot at being stolen for Harris.

Ann Selzer certainly knows all of this by now, and we will see how she responds in about 6 weeks.

---

Seth Keshel, MBA, is a former Army Captain of Military Intelligence and Afghanistan veteran. His analytical method of election forecasting and analytics is known worldwide, and he has been commended by President Donald J. Trump for his work in the field.