Skip to content
Welcome, thoughtbot's Giant Robots listeners!Start here
US Midterms 2026
15SEP

Generic ballot hits D+6.9, above 2018 wave mark

3 min read
14:21UTC

Silver Bulletin recorded a D+6.9 congressional generic ballot average on Thursday, the highest Democratic reading of the cycle and the first to exceed the D+6.5 margin that delivered Democrats a 40-seat House gain in 2018.

PoliticsDeveloping
Key takeaway

D+6.9 clears the 2018 wave benchmark, but Democrats still face a structural Republican map advantage built to absorb that margin.

Silver Bulletin, the forecasting publication founded by Nate Silver, recorded a D+6.9 congressional generic ballot average on Thursday, up from D+5.9 in the prior reading and the furthest Democratic advantage of the cycle⁠1. The generic ballot measures which party voters prefer for Congress rather than testing any individual candidate; the aggregate tracks national political mood across dozens of polls, weighted by historical accuracy. Republican support fell to 41.6% in May, the lowest of the cycle, marking the third consecutive monthly decline. The prior D+5.8 reading from late April has moved a full point in four weeks.

The D+6.5 threshold is the comparable reading from the 2018 wave cycle that preceded a 40-seat Democratic House gain⁠2. The current D+6.9 reading sits above it. The counter-argument holds structural weight: the 2018 wave hit maps drawn before the redistricting harvest that has now banked a double-digit Republican structural floor without any Democratic offset. A uniform D+6.9 national swing hitting a map engineered to absorb that environment may produce gains in the low twenties rather than the 40-plus of 2018. The efficiency gap, how many Democratic votes pile up in already-blue districts versus reach the marginal seats the maps now insulate, is the decisive variable the topline does not capture.

Three consecutive months of Republican decline is not a single-week spike reversible through a single news event. The Q1 GDP contraction of 0.3% anchored the tariff-driven approval collapse that Cook first registered in April; the generic ballot's direction tracks that macro sequence rather than any single political event. Whether D+6.9 holds, widens, or softens over the next 158 days is the question neither party controls.

Deep Analysis

In plain English

The 'congressional generic ballot' is a standard polling question asking voters which party they would support in a generic House election, without naming any specific candidates. Political forecasters use it as a summary measure of the national political environment. A reading of D+6.9 means roughly 6.9 percentage points more voters say they would support a Democrat than a Republican in a generic congressional race. In 2018, a D+6.5 reading preceded a 40-seat Democratic gain and a House majority. However, the 2018 comparison has a catch. Republicans redrew congressional maps across multiple states in 2026 after a Supreme Court ruling removed a requirement to draw minority-majority districts. That redistricting added a structural advantage worth 12-15 seats in Republicans' favour. So Democrats would need to win more votes than they did in 2018 to gain the same number of seats.

Deep Analysis
Root Causes

Three structural drivers produced the third consecutive monthly Republican decline to 41.6% support.

The Q1 2026 GDP contraction of 0.3% anchored a tariff-driven approval collapse that began in January 2025 when the generic ballot stood at R+3.3 and has produced a 10.2-point cumulative swing. Consumer sentiment surveys from the University of Michigan show the largest four-month confidence drop since the 2008 financial crisis.

Special election results in Georgia's 14th District (25-point Democratic swing) and New Jersey's 11th District (20-point Democratic swing) have demonstrated that the generic ballot is not a lagging indicator in 2026: both results ran ahead of contemporaneous ballot readings, suggesting the aggregate understates Democratic intensity.

The third driver is compounding: each monthly Republican decline makes the 2026 environment resemble 2008 and 2010 more closely, pulling late-deciding campaign donors toward the party showing upward momentum.

First Reported In

Update #7 · Paxton wins; maps lock

Silver Bulletin· 29 May 2026
Read original →
Different Perspectives
Department of Justice (DOJ)
Department of Justice (DOJ)
DOJ appealed to the 9th Circuit after Judge David O. Carter did not rule by a noon deadline on 22 October that DOJ itself had set, seeking to have California set aside mail ballots from more than 20,000 listed voters. Its case is that ineligible votes cancel lawful ones; the panel dismissed the appeal 2-1 on 26 October.
Lower Costs Now
Lower Costs Now
Lower Costs Now reported $7.37m in Kansas on 24 and 25 October, part of $65.65m in Senate races since its first independent expenditure on 1 October, including $17.21m for Josh Turek in Iowa. Its FEC filings show $50.18m raised by 14 October.
MAGA Inc
MAGA Inc
MAGA Inc, the super PAC aligned with Donald Trump, filed $7.20m against James Talarico on 22 October, taking its Texas spending to $37.65m. Of that, $30.12m is against Talarico and $7.53m is for Ken Paxton, the Republican nominee.
Senate Majority PAC (SMP)
Senate Majority PAC (SMP)
SMP's pre-general report shows it spent $108.70m from 1 to 14 October, including $65.0m to WinSenate and $10.0m to Texas Forever, and ended with $32.72m. Texas Forever has reported $19.40m against Ken Paxton since 1 October.
Senate Leadership Fund (SLF)
Senate Leadership Fund (SLF)
SLF's pre-general report shows $88.42m in cash on 14 October and $18.98m raised in the fortnight, none of it from Elon Musk. It has since reported $12.07m against Sherrod Brown in Ohio, and its $2.7m to Plains PAC preceded that committee's first Kansas buy on 25 October.
National Republican Senatorial Committee (NRSC)
National Republican Senatorial Committee (NRSC)
The NRSC's pre-general report shows nothing spent with Ashley Hinson in Iowa, or with its nominees in six other states, from 1 to 14 October, after $5.54m with Hinson in September. It took no new loan and ended the fortnight with $25.20m in cash and $20m of debt.

Written with AI from Lowdown's published reporting. Not individually reviewed by an editor. Editorial standards.