Data checked: 7 August 2026. Safety is not one number. A homicide rate can compare lethal violence between countries reasonably well; it cannot tell you whether your phone will be stolen outside a particular supermarket on Friday night. Police files can show what entered an institution's workflow; they cannot reveal every unreported offence. And a department-level rate says little about the next block.
That distinction is the key to reading Paraguay honestly. The country is neither the uniformly dangerous place imagined by some outsiders nor a low-crime haven simply because daily life often feels calm. The best short answer is: Paraguay sits between its neighbors on lethal violence, with a very uneven internal map, while current official data are too coarse to rank Asunción neighborhoods or make clean international comparisons for theft and robbery.
The comparison that travels best: intentional homicide
For comparisons between countries, intentional homicide is usually the least-bad crime indicator. A death is harder to leave unrecorded than a stolen phone, and the international series applies a common statistical concept. The World Bank publishes the UN Office on Drugs and Crime series in deaths per 100,000 residents. In the latest common year available, 2023, the values were approximately:
| Country | Intentional homicides per 100,000, 2023 |
|---|---|
| Germany | 0.91 |
| Argentina | 4.49 |
| Paraguay | 6.78 |
| Brazil | 19.28 |
This supports three careful statements. Paraguay's rate was about seven and a half times Germany's. It was roughly half again as high as Argentina's. Brazil's was almost three times Paraguay's. Those ratios describe one severe form of violence in one year; they do not mean that every person, city or journey carries the same ratio of risk.
Do not mix this international series with a national table and assume the decimals must match. Paraguay's INE ODS platform reports 7.3 for 2023 and 7.5 for 2024 from domestic police records. Different publication vintages, classifications and validation processes can produce modest differences. For a country comparison, keep all four countries inside the same World Bank/UNODC series. For the newest domestic map, use the INE table.
Paraguay's 2024 national result—and the geography behind it
The national INE indicator records 7.5 intentional-homicide victims per 100,000 residents in 2024, up slightly from 7.3 in 2023 but below every national result from 2015 through 2022. The long view therefore matters more than a claim that crime simply 'rose' or 'fell.'
The 2024 departmental rates show how much a national average hides:
- Amambay: 41.2 per 100,000
- Canindeyú: 22.4
- Presidente Hayes: 15.3
- Concepción: 13.1
- San Pedro: 8.5
- Asunción: 8.1
- Alto Paraná: 7.8
- Central: 3.2
- Caaguazú: 3.9
- Cordillera: 4.0
This is not a complete ranking. The INE table prints Boquerón and Alto Paraguay per 10,000, because their populations are below 100,000, while the other jurisdictions are shown per 100,000. Copying the displayed numbers into one sorted list would compare different units. Small populations also make annual rates volatile: a few events can change the rate sharply.
The broad pattern is nevertheless clear. Lethal violence is concentrated far more heavily in parts of the northern and northeastern border geography than in the populous Central department. That does not make every place in Amambay dangerous at every hour, and it does not make every place in Central safe. It means that 'Paraguay's crime rate' is a poor substitute for location-specific due diligence.
Why there is no honest 'safest neighborhood in Asunción' table here
The current official sources used for this article provide an Asunción-wide homicide rate, not a recent neighborhood table with consistent boundaries, population denominators and offence definitions. The 2024 rate for the capital was 8.1 per 100,000, down from 10.4 in 2023. It cannot be broken into a reliable barrio ranking from these data.
Raw incident counts would not solve the problem. Downtown, nightlife and commercial areas receive commuters and visitors who are absent from the residential census denominator. A busy neighborhood can record more thefts because more people and transactions pass through it, not because each resident faces the highest risk. A quiet-looking area may have fewer reports but poorer lighting, longer response routes or more underreporting. Police-station jurisdictions also do not necessarily match barrio boundaries.
A useful housing check is therefore micro-local and recent: visit the exact block in daylight and after dark; walk the route to transport or parking; look at lighting and sightlines; ask several unrelated neighbors or building staff about recent incidents; check whether gates actually close; and ask which police station covers the address. Treat crowdsourced maps and estate-agent assurances as leads, not measurements.
Theft, burglary, vehicle theft and armed robbery are not interchangeable
Paraguayan criminal law distinguishes hurto from robo. Under Penal Code Article 161, hurto is taking another person's movable property with intent to appropriate it. Article 166 classifies the act as robo when the theft is carried out with force against a person or with a present threat to life or physical integrity. Article 167 makes robbery aggravated in circumstances including carrying a firearm or another weapon, putting a third person in present danger, or acting as part of a continuing robbery-and-theft gang.
What English speakers call 'burglary' is not a single matching headline category in the Police report cited here. Penal Code Article 162 treats forms of forced or irregular entry—breaking protection, climbing in, using a false key or hiding inside a building—as aggravated theft in the circumstances listed there. Translating every hurto agravado as burglary would therefore be too broad. Likewise, a reported 'armed robbery' should not be inferred from every robo agravado count without case-level detail.
The Police National annual report illustrates another data trap. Its Directorate for Investigation of Punishable Acts says it investigated 426 hurto cases, 1,120 aggravated-hurto cases, 111 robbery cases and 683 aggravated-robbery cases from January through December 2024. It also lists 22 vehicle-theft investigations. Elsewhere in the same report, the specialized vehicle-breaker-yard directorate records 1,018 vehicles reported stolen and 783 recovered. These are different operational tables from different units. They are not competing estimates of one national total, and they should not be added together.
The same warning applies to the report's 286 investigated intentional-homicide cases. The INE ODS rate is based on victims registered in a police complaint or intervention. A case workload, a count of victims and an internationally harmonized rate are related but not identical statistical objects.
Why theft comparisons with Germany, Brazil or Argentina fail faster
It is tempting to place national police counts side by side. That can be badly misleading. Countries define offences differently, split attempted and completed acts differently, record one event or several victims differently, and have different reporting behavior and police systems. Vehicle fleets, insurance requirements and online reporting also affect whether theft reaches an official file.
Paraguay's own 2019 Victimization Survey explains why household surveys and police statistics are complementary. Surveys can reveal offences that were never reported or registered and can measure perceptions; police data describe institutional activity and capture some serious offences that general-population surveys measure poorly. The survey covered selected police jurisdictions in Asunción and Central and referred to events from September 2018 through August 2019. It is valuable methodology, not a current nationwide neighborhood map.
Consequently, this article compares Paraguay with Germany, Argentina and Brazil only on intentional homicide from a common source. It does not claim that Paraguay has a specific multiple of Germany's burglary risk or Brazil's motorcycle-theft risk. The public data assembled here cannot support those statements.
A practical way to use the numbers
For relocation, travel or business decisions, work from broad to narrow:
- Country comparison: use a harmonized indicator and the same year. Homicide is useful for severe violence; it is not a universal safety score.
- Department: use the current official domestic series, check the unit and read several years. One small-population year's rate can swing.
- City and exact address: seek recent, local information and observe the route and time of day. Do not convert an Asunción-wide rate into a barrio claim.
- Type of exposure: a pedestrian with a phone, a rural landowner, a night-shift worker and a driver parking a motorcycle face different risks.
- Data provenance: ask whether a number counts reports, investigations, victims, convictions or survey responses. Never assume they are interchangeable.
Bottom line
On the latest comparable homicide series, Paraguay is clearly more violent than Germany and somewhat more violent than Argentina, but substantially less violent than Brazil. The 2024 domestic map is sharply uneven, with very high rates in Amambay and Canindeyú and much lower rates in Central. Asunción's citywide homicide rate does not justify a neighborhood ranking.
For property crime, the honest conclusion is less satisfying but more useful: current administrative figures describe parts of the police workload, not the full probability that a resident will be victimized. Read the category, unit, geography, year and institutional source before reading the number. Safety decisions improve when the national statistic is treated as a starting point—not a verdict on an address.
Three questions a national crime rate cannot answer
Do walls, cameras and guards make a home safe? They can reduce particular opportunities, but they are measures rather than guarantees. Their value depends on the exact block, lighting, access control, maintenance, response arrangements and the behaviour they are meant to address. A tall wall is not a statistical measure of neighbourhood danger.
Does the general rate describe women's safety? No. Homicide and robbery tables do not measure harassment, coercive control, domestic violence or the practical accessibility of emergency and support services. Women assessing a home, commute or workplace need recent local information and a separate plan for transport, communications and support; immediate danger belongs with emergency services, not an internet comparison.
Does organised crime at a border describe every visitor's risk? No. Trafficking networks, contraband routes and institutional corruption can concentrate serious violence without creating the same exposure for every resident or traveller. Border departments are internally diverse. Use route-specific official advice, current local reporting and the exact destination rather than converting a network map into a household-risk map.