Republicans lost the popular vote but won the presidency in 2000 and 2016, and almost did it again in 2020. How is this possible? The answer is that Republicans have a built-in advantage in the rules of the electoral college.
In fact, Republicans also have a built-in advantage in the House and the Senate, which are even harder for Democrats to win. This means that Republicans can lose the popular vote but still win the presidency, House, and Senate, while Democrats need a landslide popular vote victory to win all three.
The graphics below show the outcome under different popular vote scenarios, imagining that states and districts vote according to whether they lean Democrat or Republican after including the popular vote scenario. For example, in a "D +3" scenario, Democrats would win a state like Pennsylvania which leans one point more Republican than the nation. For a detailed explanation, see the About the Data section.
Use the slider to see how different scenarios lead to different outcomes, and read on to understand why Republicans have such a strong built-in advantage.
The Presidency is biased in favor of Republicans because of the electoral college, where states are winner-take-all. In each state, the losing party's votes are "wasted" - whether they lost in a landslide or a slim margin. Currently, there are more wasted Democratic votes than Republican votes. For example, in states like Florida and Texas, Democrats usually do well but not well enough to win.
An additional reason for pro-Republican bias is that each state gets a number of electoral votes equal to the number of congressional districts plus two, favoring small states.
The electoral college hasn't always favored Republicans, and could shift to favor Democrats in the future if demographic changes result in more wasted Republican votes in states like Florida and Texas.
One solution to this problem is state-level initiatives like Maine and Nebraska have adopted - in these states, instead of a winner-take-all system, the winner of each congressional district gets one electoral vote, and the statewide winner gets two votes.
Republicans have an even larger advantage in the House - Democrats need to win the popular vote by five points just to win a majority. This is because each district is winner-take-all, and as with the presidency, there are more "wasted" Democratic votes than Republican votes.
One reason for this is gerrymandering, where a party draws the district boundaries to favor itself. By packing the other party's voters into one district, a party can lose that district but win more districts overall. The other reason for this is the natural pattern of where Democrats and Republicans live - urban areas tend to be heavily Democratic.
The solution to this would be for Democrats to win state legislatures and redraw the district boundaries, but this only happens every decade, and Democrats failed to win the needed state houses in 2020.
Last but definitely not least, the Senate has an extreme structural bias towards Republicans, far more than the presidency or the House. In the Senate, each state has two senators regardless of population. This means a small population state like Wyoming has the same power as a large state like California. There are more Republican-leaning states than Democratic-leaning states, so the Republicans have an advantage.
The advantage is enormous - Republicans can lose the popular vote by six points and still maintain control of the Senate. Having a filibuster-proof majority of 60 senators, as the Democrats did in 2008 when they passed Obamacare, is inconceivable in the near term.
The only possible way for Democrats to even the playing field in the Senate is to create multiple new states that lean Democratic, for example, giving statehood to D.C. and Puerto Rico. But this would require - wait for it - control of the Senate.
The scenarios presented here imagine that each state and district votes according to whether it leans Republican or Democrat after factoring in the popular vote scenario. Each state and district has a partisan lean, which is the difference between how it votes and how the country votes overall (in percentage points).
For example, Pennsylvania’s partisan lean is +1 R, meaning it votes one percentage point more Republican than the national average. In an “Even” popular vote scenario, the national popular vote is tied (50% to 50%), and Republicans win the state of Pennsylvania by 1 percentage point (50.5% to 49.5%). In a "D +3" popular vote scenario, Democrats win the popular vote by 3 points (51.5% to 48.5%) and win Pennsylvania by 2 points (51% to 49%).
In the event of a tie in a state or district, Democrats are assumed to win, in order to provide a more conservative estimate of the level of bias.
In reality of course, which party wins a state or district also depends on the strength of the individual candidates, not just party identification. Additionally, elections for each office aren’t all held at the same time, and the presence of third party candidates also affects the outcome.
The partisan lean dataset comes from FiveThirtyEight, and uses election results from recent elections to calculate each state and district's partisan lean. The dataset weighs more recent elections the most, but does not yet include data for the 2020 election.