Everyone assumes big cities are more dangerous. We tested it across 460 cities in our index — and it's barely true. Population predicts safety only weakly (Pearson r = -0.16): megacities average just 9.7 points below small cities, and Seoul, home to 10 million, is safer than most small towns. Which city you pick matters; how big it is, hardly at all.
Across all 460 cities in the Kakapo index with more than 200,000 residents, city size barely predicts safety. The correlation between population and overall safety score is weak and negative — Pearson r = -0.16 — meaning population explains only a sliver of the difference between a safe city and a dangerous one.
"Big cities are dangerous" is one of the most durable ideas in travel. The mental image is fixed: the sprawling megacity, the crowded metro at night, the pickpockets and the no-go districts. So we ran the actual test. We pulled every city in the Kakapo index with more than 200,000 residents and a live safety score — 460 cities — and measured how tightly population tracks safety. If the assumption were true, the correlation would be strongly negative. It isn't.
The Pearson correlation between population and overall safety score is -0.16 — weak, and only slightly negative. Sorted into population bands, the pattern is a gentle downhill slope, not a cliff: each step up in size shaves a couple of points off the average, and the whole range spans under 11 points.
| Population band | Cities | Mean safety | Mean night |
|---|---|---|---|
| 200k – 500k | 133 | 82.0 | 70.1 |
| 500k – 1M | 106 | 77.8 | 64.3 |
| 1M – 5M | 172 | 76.7 | 63.2 |
| 5M – 10M | 30 | 75.1 | 64.2 |
| 10M+ | 19 | 72.3 | 59.9 |
If "big" predicts danger anywhere, it should be after dark — that's where the fear lives. But population tracks Kakapo's dedicated night sub-score even more weakly, at r = -0.10. Mean night safety does dip in the very largest cities (59.9/100 for the 10M+ band versus 70.1 for small cities), but the middle bands are essentially flat. A city of eight million is not, as a rule, scarier at night than a city of eight hundred thousand.
Averages hide the point. Look inside the extremes and the "size" theory collapses:
| City | Population | Safety | What it shows |
|---|---|---|---|
| Seoul, South Korea | 10.3M | 89 | A 10-million-plus megacity that outscores the average small town |
| Nara, Japan | 360k | 97 | A small city near the top of the entire index |
| Lagos, Nigeria | 15.4M | 54 | A megacity that fits the stereotype — but is the exception, not the rule |
| Port Moresby, Papua New Guinea | 284k | 50 | A small city more dangerous than nearly every megacity |
Seoul and Port Moresby are separated by nearly 40 points despite Seoul being roughly 36 times larger. Population sits on the wrong side of that comparison. What actually drives the score is governance, policing, road safety, healthcare and local crime — none of which scale neatly with headcount.
Stop screening destinations by size. A metropolis is not a red flag and a small town is not a green one — the correlation is too weak to plan around. Seoul, Tokyo and Singapore are among the safest places on earth and among the largest; plenty of modest-sized cities score far worse. The useful move is to check the specific city's score and its sub-scores — personal safety, night safety, transport, healthcare — rather than lean on a rule of thumb that the data says is barely there. "Is this city safe?" is a real question with a real answer. "Is it big?" is close to irrelevant.
We queried the live Kakapo index for every city with a population above 200,000 and a non-null safety score, deduplicated by city (keeping the largest-population entity and excluding New York's boroughs so the city isn't counted five times) — 460 cities in all. We computed the Pearson correlation coefficient between population and the overall safety score, and separately between population and the night sub-score, then took the mean safety and mean night score within five population bands. Scores draw on national travel advisories from seven governments plus crime, transport and healthcare data; full weighting is at https://kakapo.travel/about/methodology. Scores are comparative model estimates, read live, so the numbers match the rest of the site.
Free to cite in journalism, research or content — all we ask is a link back. Copy the attribution below:
Kakapo Editorial Team (2026). The Safety Sweet Spot: Does City Size Actually Predict Danger? Kakapo. https://kakapo.travel/blog/safety-sweet-spot-2026
Writing about urban safety, megacities or travel risk? We give journalists custom data pulls at no cost — the full band table, the city-by-city scatter, or a specific country or region. Email [email protected] and we usually reply the same day.
Kakapo publishes real safety scores for every destination, drawing on national travel advisories, local crime data, healthcare infrastructure, and night-safety measures. Kakapo Studies is our research imprint — independent editorial, no advertiser influence on scores. Contact: [email protected].