If everyone voted, would the Democrats win? Not any more
Until 2016 the people who stayed home leaned Democratic, so higher turnout helped the party. In 2020 and 2024 they leaned Republican. On our figures, every citizen voting in 2024 would have moved the result about 2 points towards Trump.
Written in house from our poll data and our forecast.
It is one of the oldest ideas in American politics: the Democrats do better when more people vote. The reasoning sounds solid. White Americans lean Republican, Black Americans vote heavily Democratic, and white Americans turn out more often. Bring everyone to the polls and the electorate gets younger, less white and more Democratic.
For most of this century our figures agree. Since 2020 they do not.
What we measured
For every presidential election since 2000 we hold, for each state, how many citizens each group has (white, Black, Hispanic, Asian American and everyone else), how many of them voted, and how they voted. The Cooperative Election Study also asks people who did not vote who they would have chosen, and checks against the voter files whether they really voted. That second question is the one that matters: the people who stay home do not have to think like the people of their own group who turn up.
So we ask what happens if every eligible citizen votes, with the newcomers splitting the way their group's non-voters said they would.
| Election | Turnout | If everyone had voted | |---|---|---| | 2008 | 64% | 4.0 points more Democratic | | 2012 | 60% | 3.2 points more Democratic | | 2016 | 61% | 2.9 points more Democratic | | 2020 | 68% | 3.0 points more Republican | | 2024 | 66% | 1.9 points more Republican |
Why it flipped
The groups did not change places. Black non-voters still lean Democratic, and white non-voters still lean Republican. What changed is how far each group's non-voters sit from its voters.
In 2024:
- White voters went Republican by 15 points; white non-voters by 21.
- Black voters went Democratic by 67 points; Black non-voters by 43.
- Hispanic voters went Democratic by 16 points; Hispanic non-voters were split almost evenly, at 2.
- Asian American voters went Democratic by 34 points; non-voters by 16.
In 2008 Hispanic non-voters leaned further Democratic than Hispanic voters, and white non-voters were evenly split. The people Donald Trump brought into politics were mostly occasional voters, and the occasional voters who still stay home now look more like them than like the regular Democratic electorate.
There is also simple arithmetic. Black turnout is lower than white turnout, but white citizens are about two thirds of the adult population. Most of the people who do not vote are white, and in recent elections most of them lean Republican.
What it means for 2028
Turnout still matters, but not in one direction. On our 2028 map you can set turnout from 30% of eligible citizens to everyone. Lower turnout than 2024 helps the Democrats a little (at 2012's 60%, about 0.8 points nationally); higher turnout helps the Republicans a little. The effect is uneven: each state moves by its own mix of voters, and some move against the national trend.
These are survey answers from people who did not vote, so they carry more doubt than a count. They are also the best evidence there is, and they point the same way in both of the last two elections.
Try it on the 2028 map.
Polling Forecast reads every poll in ten countries, corrects for how pollsters have missed before, and turns them into seats. Our forecasts, the polls and our methods are free to explore.
Elsewhere on Polling Forecast
Victoria votes in 23 days, and the Coalition is close to a majority
Victoria's state election is on 28 November. Our projection gives the Coalition 48 of the 88 seats, three more than a majority; One Nation's vote decides how many it really wins.
Is the NDP on 8 per cent or 14? It depends which pollster you ask
Liaison and Nanos have the NDP on 13.6 to 14 per cent; Abacus and Léger on 8 to 9. At 8 the party would hold a handful of seats; at 14 it starts to win back some it lost in 2025.
Only a repeat of 2023's polling miss would cost the right its majority
PiS and the two Konfederacjas have 240 seats in our projection. If the polls miss like 2023, they fall to 226 and the governing coalition, with the PSL back in the Sejm, reaches 234.