Who would vote for Elon Musk's America Party?
Crunching the numbers on the potential coalition for an anti-spending, anti-Trump, anti-Democrats, anti-Republicans political party... wait what?
Will Elon Musk be the next Ross Perot, or Cornel West?
For those getting caught up: After being frustrated by the increase in the federal budget deficit that Republicans have caused with their "One Big Beautiful Bill,” the richest man in the world last week declared independence from the two-party system. He ran an insta-poll on Twitter to assess support for a new political party, which received 80% of votes, and subsequently announced he was forming one. He calls it the "America Party."
According to Musk's social media posts, the movement will draw support from people who want to shrink government spending, are disillusioned with the two-party system, and think the Democrats are too ”woke.” Musk has taken increasingly strong stances against Donald Trump in recent weeks, making opposition to the president — or even soft support, relative to Musk’s — another cornerstone of the party agenda.
Musk appears to be serious about making the America Party work. In recent days, he has sought the advice of Andrew Yang, founder of the pro-electoral reform Forward Party (as of 2025, it has won one state legislative race) and the right-wing thinker Curtis Yarvin (here's a Guardian interview from before he got a rebrand for the NYT). Yarvin is famous for saying U.S. democracy needs to be overthrown via violent coup and replaced with a so-called "corporate monarchy." Strange bedfellows. Yang and Yarvin.
Alright, but is this really going anywhere? What is the constituency for (1) cutting spending, (2) opposing Trump, (3) liking Musk, and (4) otherwise being centrist?
In this week's Chart of the Week, I load up some polling data to make a quick approximation of the size of Musk’s potential constituency. Spoiler alert: it's even smaller than you think.
This won’t be a long post. Maximally, I think we want to figure out the percentages of Americans that have each of these traits:
Strongly identify as Republican or Democrat
OR have a positive view of Trump
OR oppose cuts to government spending
OR have a negative view of Musk
The sum of these four percentages gives us the share of voters that we can reasonably assume Musk (or some other candidate running in his party) is not winning in a future election. Because we don't have polling data about the America Party yet, we can use this method to define support for Musk in the negative.
The only poll I could find that comes close to asking all of these questions is the 2024 American National Election Study, which unfortunately doesn't ask anything about Musk. But the ANES does give us individual-level data on the other three questions (plus a whole lot more), which we can use to sort Americans into different multi-party buckets.
Here’s the method;
First, I take the ANES sample and remove people who aren't registered to vote. Removing people who aren’t registered gets us closer to the demographics and attitudes of the voting population, which is ultimately what we care about with an analysis of electoral coalitions.
Now, you could start with a naive prior that every registered voter in the country is a possible Musk voter. But I think you first need to remove people who say they identify as strong or solid Democrats or Republicans, since they are unlikely to bail on their party. Unfortunately for Musk, this is the vast majority of America. About 67.5% of voters in the ANES identify as a Democrat or Republican (ie not counting the independents who say they lean toward one party). The remaining pool of voters, 32.5%, are potential America Party voters (at this, the first step, in the analysis).
Support data-driven journalism!
Become a paying member of Strength In Numbers today and get exclusive posts, early access to data, and more.
Your support helps fund polling trackers, original polls, deep analysis, and data-driven journalism that cuts through the noise of today’s media landscape. Support our site and tell the world that when it comes to the news and democracy, there is strength in numbers.
Next, I remove Americans who aren't solid partisans, but have a positive impression of Trump. Researchers should do this because, in a fight between Musk and Trump, the reasonably strong prior belief is that these people are likely to stick with Trump. This accounts for another 14% of voters, bringing the total non-Musk pool to 81.5%. Musk can now draw on 18.5% of the electorate.
Now, in a third culling, I remove people who are pro-spending. A big part of Musk's agenda is cutting (non-existent) waste and massively rolling back social programs. The ANES has a variable loosely proxying this, too — it asks Americans to rate their preference for government spending on a scale from 1, being anti-spending, to 7, being pro-spending. I'm going to take out the people who aren't super partisan, don't like Trump, and are pro-spending (at a 4 out of 7 or higher) from our remaining America Party pool. This group is another 16% of Americans.
That leaves Musk with 2.5% of the public that might potentially vote for his America Party.
However, one final thing we might want to do is account for the Musk factor. The ANES doesn't ask about Musk, but I can add in the fact that just 30% of adults in this week's YouGov/Economist poll say they have a favorable opinion of him. So now we take away 70% of that 2.5%.
That leaves Musk with a whopping 0.75% of the American public, which corresponds to just 36 out of 4,855 registered voters who took the ANES in 2024.
That's the coalition for a party that is (a) generally anti-spending, (b) not the Democratic or Republican Party, (c) anti-Trump, and (d) pro-Musk.
Nobody is building a successful third party with less than 1% of voters. Musk will need to make serious inroads with Americans to make it work. He is starting from the absolute bottom.
Okay, but maybe, you say, this is unfair to the America Party; you think I’ve chosen an approach that is too rough on Musk or something. After all, it’s called the America Party! People love America! So how about another measurement strategy?
Using the ANES data, I can also sort Americans into different groups based on a combination of attitudinal questions. Instead of party ID, I'll use feeling toward the Republican Party, Democratic Party, preference for spending, and feeling toward Trump — one of the variables from the last analysis. But then instead of using a variable specific to Musk, since his America Party is also campaigning against the Democrats for being too woke, I'll include a variable that measures how ideologically extreme people rate the Democrats.
I use this data to sort people into one of five buckets:
Hardcore Republican: Someone who rates Republicans positively, the Democratic Party poorly, and opposes government spending (at or less than a 4 out of 7 on the spending scale)
Hardcore Democrat: Someone who rates the Democratic Party well, Republicans poorly, and supports government spending (at 5 or higher)
America Party: Someone who rates both the Republicans and Democrats poorly, wants to severely cut spending (3 or lower on the spending scale), feels either indifferent or negatively to Trump, and thinks the Democrats are extreme liberals (a 1 or 2 out of 7 in ideology, with 1 representing extremely liberal and 7 extreme conservative)
Everyone else is sorted into the Soft Republican or Soft Democrat camp based on how they rate the parties — specifically if they give either party a rating higher than (but not including) 50 on the feeling thermometer scale (0-100). Finally, if any remain, they are sorted based on how they feel about spending.
I get the following breakdown:
Hardcore Republican: 24%
Hardcore Democrat: 32%
Soft Republican: 22%
Soft Democrat: 20%
America Party: 2%
Two percent (it’s actually 1.7%, but I’m rounding up to be nice) is certainly higher than 0.75, but it’s not going to be large enough for Musk to accomplish his goal of winning swing House and Senate seats. I would also consider this a generous upper bound, since it assumes that Musk's unpopularity does not matter (we can be reasonably confident it does, though).
So, that’s my prior for the pool of votes the America Party would draw on: somewhere between 0.75 and 2% of registered voters. And since these voters are distributed roughly equally across the country, the party shouldn’t hope to build a coalition large enough to win any seats (they’d need to be concentrated in one geography to win any given seat).
One lesson from all of this: If you’re going to call yourself the America Party, it’s a good idea to offer something America actually wants.
I'm open to alternative definitions of who qualifies as a potential America Party voter. You can look at my R code for this analysis and download the ANES data at their website.
Thanks as always for reading. Have a nice weekend,
Elliott
There's a statistical problem with the repeated assertion that a majority of Americans want a third voting option.
That group doesn't all want the SAME new party option. Some want more government involvement. Others want less. Still others want none. And for many that position depends on what the government would be doing about what.
Maybe in some future chart you could analyze voters based on issues and assess the various potential coalitions?
This research confirms my intuition.
The "No Labels" organization raised $70 million for a 3rd party effort, advocating no more than, "Let's just be nice." But with no policy platform -- they failed to attract political talent to lead.
The GOP and Democratic Parties are out of gas. There is a hunger for thoughtful, modest, centrist, stable, reliable, reasonable political leadership -- that will govern toward the middle, which has most always been the case in America, but not practiced since Bill Clinton. A well-funded effort with modest goals would get traction.