5050 Fifty Fifty Politics ← Go to Homepage
Media & Information

What Democrats and Republicans Actually Agree On

Fifty Fifty Politics · Background & Data
It is a comforting idea that the two sides of American politics agree on more than they admit. We were in a position to check. This site carries 900 arguments across 90 debates, written to a fixed standard that requires the strongest case from each side. We measured every one of them for places where left and right accept the same underlying fact. The result was less comforting and more interesting than the premise.

What we measured, and what we did not

A note on method first, because it determines what the numbers mean. We did not measure whether Democrats and Republicans agree. We measured whether the strongest left-leaning and right-leaning arguments on a given debate rest on a shared fact.

Those are not the same thing. Party positions and ideological positions diverge often, and plenty of arguments on this site are not any party's platform. What follows is a finding about how political arguments are constructed, not a survey of voters.

Explicit agreement is rare

Of 900 arguments, 54 use language that explicitly concedes a point to the other side: words like already, still, even, or a direct acknowledgement that something the other side claims is true. That is six percent.

Only eight of the ninety debates have both sides conceding something. If you expected a hidden reservoir of common ground, the corpus does not show one.

How often each side concedes a point Explicit agreement is rare, but it follows a pattern Arguments that concede a point to the other side 54 of 900 (6%) Topics where both sides concede something 8 of 90 Measured across every argument on the site. Concession was counted only where an argument explicitly grants a fact rather than merely discussing the same subject.

But where it appears, the pattern is consistent

In every case we found, the shape is the same. Both sides accept an identical fact, and disagree entirely about what follows from it.

Take hate speech. The left-leaning argument notes that existing exceptions for true threats and harassment show the First Amendment has never been absolute. The right-leaning argument notes that narrow exceptions for incitement, true threats and harassment already reach the conduct that causes concrete harm. Both accept the same legal fact. One reads it as proof the line is movable; the other reads it as proof the line is already correctly drawn.

One shared fact, two conclusions Hate speech: one shared fact, two conclusions BOTH SIDES ACCEPT Narrow exceptions for true threats and harassment already exist THE LEFT CONCLUDES The First Amendment has never been absolute, so the question is where the line sits, not whether there is one. THE RIGHT CONCLUDES Those exceptions already reach the conduct that causes concrete harm, so no further category is needed. Both arguments are quoted from the Hate Speech & the First Amendment page on this site.

The same shape, on other debates

On tuition-free college, both sides accept that most low-income community college students already pay no tuition. The left concludes that the word free is doing useful work anyway, because aid goes unclaimed. The right concludes that waiving tuition would mainly subsidise students who can pay.

On non-compete agreements, both sides accept that trade secret law and confidentiality agreements already exist. The left concludes that non-competes are therefore redundant. The right concludes that states narrowing them by statute shows the system is self-correcting without a federal ban.

On bail reform, both sides engage the same question of whether defendants appear for court dates without cash bail. Here they genuinely dispute the evidence, which is a different and rarer kind of disagreement.

Why this matters more than a longer list would

The popular version of this idea holds that if people just talked, they would discover how much they share. Our corpus does not support that. Ninety-four percent of arguments make no concession at all, and the eight debates where both sides concede are not the eight you would guess.

What the corpus does support is something more useful. When agreement exists, it is agreement about facts, and the disagreement that remains is about values and inference. That is not a failure of communication. It is what a real disagreement looks like once the factual questions have been settled.

Two people can accept every number in a policy debate and still reach opposite conclusions, because they weigh liberty against safety differently, or the present against the future. No amount of additional evidence resolves that, and treating it as an information problem misdiagnoses it.

Where the two sides never meet

The other half of the finding is worth stating. In eighty-two of the ninety debates, neither side concedes anything explicitly, and the arguments run past each other rather than into each other.

On most of those, the two columns are not answering the same question. One side argues about outcomes and the other about process; one about the median case and the other about the edge case. Immigration enforcement is typical: the left-leaning argument concerns how local police resources are best used, the right-leaning argument concerns whether federal law should preempt local policy. Both are reasonable. Neither is a reply to the other.

That is a more accurate picture of political disagreement than either the optimistic version, in which we secretly agree, or the cynical one, in which one side is arguing in bad faith. Most of the time both sides are arguing carefully about slightly different questions.

What to do with this

The practical value is in the question it suggests. Faced with a political argument, it is worth asking whether the disagreement is factual or evaluative. If factual, there is usually a primary source that settles it. If evaluative, more facts will not help, and the honest move is to identify which value is being traded against which.

Most political argument conducted in public never makes that distinction, which is why so much of it goes nowhere. The six percent of arguments in our corpus that concede a point are the ones doing it properly.

Sources
Browse All Blogs