Does Sharing a Border Help or Hurt a Relationship? What the Data Actually Shows
2026-09-19 · by Jordi Fuster
There's an intuitive assumption that proximity breeds familiarity, and familiarity breeds goodwill. WorldRelat's own data says otherwise — and the gap is large enough to be worth explaining rather than just noting in passing.
The headline number
Across every documented relationship on the site, pairs of countries that actually share a land border average a sentiment score of 5.55 out of 10. Pairs that don't share a border average 6.11. That's a real, consistent gap: neighbors, on average, get along measurably worse than countries with no shared border at all. It's not a subtle effect — it holds up across 273 distinct neighboring pairs compared against thousands of non-neighboring ones.
Why that's not actually surprising
Distant countries mostly interact through trade, diplomacy and the occasional formal summit — low-friction contact that rarely produces genuinely bad blood. Neighbors are the ones who actually have something to fight over: contested borders, shared rivers, competing claims on the same resources, refugee flows, and centuries of accumulated history that distant countries never had the opportunity to accumulate. Our look at WorldRelat's territorial dispute data makes the same point from a different angle — virtually every active dispute on the site is, by definition, between neighbors.
The worst-scoring borders
Some of the coldest relationships anywhere on WorldRelat are between direct neighbors: Poland and Belarus score a flat 0.0. Rwanda and the Democratic Republic of the Congo, Pakistan and Afghanistan, Thailand and Cambodia, Lebanon and Israel, and Ukraine and Russia all score 0.07 — as close to the bottom of the scale as this site gets. Chile and Bolivia (0.10) and India and Pakistan (0.14, covered in more depth in our dedicated piece on that relationship) round out some of the lowest-scoring shared borders anywhere in the data.
The best-scoring borders
Plenty of neighbors buck the trend entirely, usually through deep economic and political integration: Slovenia and Croatia (9.94), the United Kingdom and Ireland (9.93), Ukraine and Poland (9.93, covered in our look at Poland's shifting alliances), Pakistan and China (9.93), Greece and Bulgaria (9.89) and Luxembourg and Belgium (9.87) all sit near the very top of the entire site's rankings, not just among neighbors.
One result the model gets wrong
In the interest of being straightforward about the model's limits: Serbia and Kosovo currently score a perfect 10.0 — the single highest score on the entire site, and, on the actual merits, an obviously wrong read of one of Europe's most genuinely unresolved and tense relationships. It's the clearest example we have of what happens when sentiment analysis runs into thin, narrowly diplomatic source text: a small amount of formally neutral language can outscore a much larger volume of text that would, in aggregate, read as far more contentious. It's a useful data point for interpreting every other number on this page with the right amount of caution — the score describes documented tone, not verified reality, and every so often those two things diverge sharply.