Does a running mate matter?
Running mates rarely move the national vote, but history gives them one thing: their home state. Since 1948 a vice-presidential nominee's state has swung towards their ticket, by more in small states (6 to 8 points in Delaware or Alaska) than in big ones (nothing measurable in New York, Texas or California), and by more when the running mate had already won statewide there. So a running mate is worth most from a small state that is close.
This page shows where that would count in Kamala Harris v JD Vance. Tap a state or a name to see the map with that running mate on the Kamala Harris v JD Vance page, where you can pick running mates for any pairing.
Where a running mate would matter most electoral votes a home-state pick would add on average
For Kamala Harris
For JD Vance
The possible running mates what each would add through their home state, in electoral votes on average
With Kamala Harris
With JD Vance
How the home-state effect is measured
For every major-party running mate from 1948 to 2024 we compared their home state's lean (its margin minus the national margin) that year with its lean four years before and after. Across 30 cases the state moved 2.5 points of margin towards the running mate's ticket on average (standard error 1.1; median 3.1). Cases where a presidential nominee came from the same state, or both running mates did, are left out, as are the one-party Southern states before 1964. Studies of the same question find a similar small lift, larger in small states.
The average hides two things, so the lift we use depends on both. Size: the lift is largest in small states and fades in big ones (Dick Cheney and Sarah Palin moved Wyoming and Alaska 8 points or more; Kamala Harris and William Miller moved California and New York the wrong way). Record: the five running mates who had never won statewide in their state (William Miller, George H. W. Bush in 1980, Geraldine Ferraro, Jack Kemp, Paul Ryan) did worse than the rest. Fitted on the same 30 cases, the lift is 1.33 + 0.76 if the running mate had won statewide there − 2.98 × the natural log of the state's share of the national vote over 2%, and never below zero: about 1.8 in Arizona, 0.0 in Pennsylvania and nothing in Texas. The size term is well measured (standard error 1.02); the record term is not (0.76, standard error 2.95), because the five who had never won statewide mostly came from big states. A running mate picked as "anyone from" a state is taken to have won statewide there, as 25 of the 30 had.
The boost is added to that state in every simulation of the pairing on /us/2028, with its uncertainty drawn each run. Nothing is added anywhere else: we find no measurable national effect. Speculative; not part of our forecast. Pick running mates for any pairing on the 2028 page.
