What a low rate can tell you
43 of the 145 covered economies have a latest reading below one intentional homicide per 100,000 people. A rate of one corresponds to ten deaths in a population of one million over a year. This puts the magnitude in context, but it is a population rate, not the probability that a particular visitor, worker or customer will be harmed.
Japan's reading is 0.23 per 100,000 in 2023. That supports a specific statement about recorded intentional homicide in that year. It does not, by itself, prove which policies caused the rate, describe every region or establish the current situation. A causal explanation would need additional evidence beyond this country ranking.
The World Bank's indicator metadata identifies criminal justice and public health systems as the sources collected and checked by UNODC. The measure is intended to capture unlawful deaths inflicted with the intent to kill or cause serious injury. Differences in recording, classification and coverage can affect comparisons; a decimal place is not a guarantee that the underlying systems are equally complete.
Why this table cannot establish overall safety
The distinction is especially consequential for West Bank and Gaza: its 2022 homicide reading of 0.62 per 100,000 appears in this dataset. That is a dated homicide statistic and does not include a count of deaths directly associated with armed conflict. Calling it evidence that the territory is safe would misstate both the period and the definition.
Homicide does not capture robbery without a death, assault, road injuries, fraud, cybercrime or interruptions to transport and essential services. It also hides differences between cities and neighbourhoods and between demographic groups. Several distinct risk measures are needed for a travel, relocation or business decision; this table supplies only one of them.
Small differences require careful interpretation
In a hypothetical economy with 100,000 residents, one additional homicide changes the annual rate by one per 100,000. In an economy of ten million, the same additional death changes it by 0.01. Small populations can therefore move sharply in a ranking after a small change in counts. Compare several years and, where available, the underlying counts before interpreting a single low or zero reading as a lasting condition.
The latest-available ranking compares readings from 2019 to 2023. Those do not describe a shared point in time. A same-year comparison is preferable when the question is which economy had the lower rate during a specific year. A newer article date cannot refresh an older observation.
Use the measure in a documented risk screen
For a company reviewing locations for staff or facilities, start with the relevant city and operating conditions, then use the country rate as background context. Record the homicide year, inspect the trend, and assess conflict, transport, health and crime exposures separately. Do not turn this table into a single safety score or use it to set an individual risk estimate.
The ranking uses the latest numeric reading held by SOTW for each economy, excludes the world aggregate, and sorts from lowest to highest without imputing missing values. The page does not expose a verified provider release vintage. For a client report, save the API response and retrieval date together with each row's reference year, and cite the original indicator. SOTW methodology explains the broader data workflow.