Two years ago on this date, we were still in the waning days of Brat Summer. Kamala Harris reached her peak in our presidential forecast on Aug. 14, though she’d never really achieved escape velocity, and by Aug. 29, Trump was the favorite again.1
That’s not the only summer memory that can give Democrats anxiety. In August 2018, Democrat Phil Bredesen led polling averages against Republican Marsha Blackburn in Tennessee, but the advantage flipped to Blackburn in September, and she eventually won by 11 points. In the 2020 South Carolina U.S. Senate race, Democrat Jaime Harrison was tied in some polls with Lindsey Graham (RIP) well into September, raising a truly insane amount of money … before losing by 10 points.
But is this part of a broader, predictable pattern of polling tending to overrate Democrats in the summer? And could that be a problem for our model?
Patrick Ruffini, a pollster at Echelon Insights, published a newsletter yesterday at his Substack detailing how summer polls over the past four cycles (2018 through 2024) have fairly consistently overrated Democrats. That doesn’t necessarily mean that the pattern will repeat itself in 2026 — and FLIPR already has various defense mechanisms against this. But these questions are tailor-made for a Model Talk column.
But first, a little bonus content: 800 words on the results of the primaries last night in Florida and Alaska.