Unsafe Drinking Water by Country: What Global Data Shows

Access to safe drinking water is a basic requirement for health and development, yet large drinking-water service gaps remain. According to the WHO/UNICEF Joint Monitoring Programme (JMP), 74% of the global population used safely managed drinking-water services in 2024, up from 68% in 2015. That still left 2.1 billion people without a safely managed service. Those figures describe levels of service, however; they do not mean that every person counted among the 2.1 billion was drinking contaminated water.

Water safety is not a simple national yes-or-no issue. Even within the same country, water quality can vary widely between urban and rural areas, higher- and lower-income communities, centralized and informal systems, and individual supplies. A country can also have severe service deficits in some regions while other communities receive reliably treated water. For that reason, global data are more useful for identifying populations and service gaps than for producing a blacklist of countries with universally unsafe water.

What global data can—and cannot label as unsafe

Why “unsafe country” is the wrong unit of analysis

The phrase “countries with unsafe drinking water” is convenient search language, but it can obscure what the available evidence actually measures. International datasets generally estimate the share of a population using different drinking-water service levels. They do not test every tap, well, tanker, private supply or household storage container in a country.

This distinction matters because unsafe drinking water can arise at many points between source and consumption. Untreated surface water or a shallow well may be contaminated at the source; treatment can fail; broken or intermittently pressurized pipes can introduce contamination; or initially safe water can become contaminated during transport and storage. Conversely, a community in a country with substantial national service gaps may have a well-managed supply.

For broader comparisons of what country-level water-quality information can show, see the water quality by country guide. The complementary guide to countries and safe drinking-water data applies the same caution to claims about safety.

What 2.1 billion people lacking safely managed services actually means

The JMP defines a safely managed drinking-water service as one using an improved source that is accessible on premises, available when needed and free from contamination. All three dimensions matter. Infrastructure alone therefore does not establish that drinking water meets the highest international service category.

In 2024, the 2.1 billion people without safely managed drinking water were not one homogeneous “unsafe-water” population. According to the JMP 2025 household report, about 1.4 billion had basic services, 287 million had limited services, 302 million used unimproved sources and 106 million used surface water. A household can fall below safely managed status because of accessibility or availability as well as contamination. In particular, “basic service” should not be read as another term for contaminated water.

Safely managed coverage also cannot support a complete worldwide ranking. For 2024, JMP safely managed estimates were available for 160 countries and six of eight SDG regions, representing 72% of the global population. Absence of a safely managed estimate is a data limitation, not evidence that a country’s water is safe or unsafe.

Unsafe water, service gaps and water scarcity are different questions

It is important to distinguish water scarcity from water quality. A place may have limited freshwater resources but deliver reliably treated drinking water. Conversely, abundant rivers or groundwater do not guarantee a safe supply. Scarcity concerns the relationship between available water and demand; drinking-water safety concerns hazards and exposure, while service indicators also encompass access and reliability.

These issues can interact—for example, unreliable supplies can complicate water treatment and household storage—but they should not be collapsed into a single ranking. See the separate guide to water scarcity worldwide for the availability side of the problem.

How unsafe drinking water is measured

The JMP drinking-water service ladder

The JMP service ladder is the most widely used framework for comparing household drinking-water services internationally. Its categories indicate progressively different service conditions rather than providing a universal laboratory verdict on whether every glass of water is safe.

Risk/service indicator	Meaning	Dataset	Caveat
Safely managed	Improved drinking-water source that is on premises, available when needed and free from contamination	WHO/UNICEF JMP; latest household estimates through 2024	Coverage estimates were not available for every country; falling below this category does not by itself prove contamination
Basic	Improved source meeting the JMP collection-time threshold	WHO/UNICEF JMP	Basic does not mean contaminated; the complete safely managed criteria are not established or met
Limited	Improved source requiring more than 30 minutes for a round trip to collect water	WHO/UNICEF JMP	Primarily identifies an accessibility and collection burden, not proof of contamination
Unimproved	JMP service-ladder category below improved-source services	WHO/UNICEF JMP	The category should not be converted into a claim that every sample from every such source is contaminated
Surface water	Drinking water obtained directly from surface-water sources	WHO/UNICEF JMP	This is the lowest JMP household drinking-water service category; 106 million people were in it in 2024

JMP methodology treats an improved source requiring more than 30 minutes for a round trip as a limited service. Improved water collected within the threshold qualifies at least as basic under the service ladder, while safely managed service requires the additional accessibility-on-premises, availability and water-quality criteria.

Service level is not the same as a water-quality test

Water can look clear and still contain microorganisms or chemical contaminants that cannot reliably be identified by sight, taste or smell. Direct water-quality evidence requires appropriate sampling and analysis. Microbial testing may examine indicators of fecal contamination, while chemical testing is selected according to relevant hazards such as arsenic, fluoride, nitrate or lead. Operational measurements can also help determine whether treatment and distribution controls are working.

Improved infrastructure does not automatically guarantee safety. A protected source or piped supply may still experience contamination or unreliable service. Likewise, a service classification should not be treated as a substitute for results from a particular supply, utility or household.

From source-to-consumer monitoring to water safety planning

Testing should not focus only on a final drinking-water sample. Effective management can include assessing the source for hazards, checking treatment processes, monitoring distribution, investigating sanitary conditions and, where relevant, considering household transport and storage.

The World Health Organization’s 2026 Guidelines for Drinking-water Quality emphasize preventive, risk-based management rather than relying only on end-product testing. WHO’s framework combines health-based targets, proactive management through water safety planning and independent surveillance. The current guidance also strengthens attention to small supplies, sanitary inspections and support for supplier risk management.

Countries and populations facing the largest service gaps

Where basic-service coverage gaps are especially high

For comparisons across many countries, basic drinking-water coverage is more widely available than the more demanding safely managed indicator. In 2024, 696 million people worldwide lacked even a basic drinking-water service. This population included 287 million using limited services, 302 million using unimproved sources and 106 million using surface water.

JMP reported that 28 countries had more than one quarter of their population lacking basic drinking-water services in 2024. They were concentrated in sub-Saharan Africa; the exceptions identified in the report were Haiti, Papua New Guinea, Solomon Islands and Sudan. This is evidence of large national service gaps. It is not evidence that every resident or every drinking-water supply in those countries is unsafe.

Where large populations face service gaps

Absolute numbers tell a different story from percentages because populous countries can contain large underserved populations even when their national percentage deficit is smaller. For limited drinking-water service alone in 2024, JMP reported about 37 million people in Ethiopia, 34 million in India, 26 million in the Democratic Republic of the Congo, 18 million in Indonesia and 15 million in Sudan.

These numbers should retain their indicator label: they refer specifically to limited drinking-water service, meaning an improved source with a collection trip exceeding 30 minutes. They should not be restated as numbers of people known to be drinking contaminated water.

Why percentages and population totals answer different questions

A coverage percentage helps describe how widespread a service deficit is within a population. An absolute count helps describe the scale of the population affected. Both can matter for public health and infrastructure planning, but neither alone establishes the quality of a specific supply.

Drinking water also sits within a wider set of sanitation, hygiene, infrastructure and development challenges. For that broader context, see the global water crisis overview.

Why infrastructure, contamination and reliability differ

The reasons communities face unsafe or inadequate drinking-water services are often interconnected. Natural conditions, human activity, infrastructure limits, sanitation, institutional capacity and social inequality can all play a role. The relevant failure may involve obtaining water, keeping it available, protecting it from contamination, treating it adequately or maintaining its quality through distribution and storage.

Why risk differs	Infrastructure	Monitoring	Source	Inequality
Accessibility and infrastructure	Location of improved sources and the burden of collecting water	Household surveys, administrative data and JMP accessibility estimates	WHO/UNICEF JMP	Large rural–urban and subnational differences can occur
Reliability	Whether water is available when needed	JMP availability element	WHO/UNICEF JMP	Service levels can differ even where households use improved sources
Water quality	Source protection, treatment and distribution influence water reaching consumers	JMP free-from-contamination element and water-supply monitoring	WHO/UNICEF JMP; WHO 2026 drinking-water guidelines	National averages cannot establish the quality of every local supply
System management and surveillance	Supplier controls and sanitary conditions affect risk from source to consumer	Water safety plans, sanitary inspections and independent surveillance	WHO 2026 drinking-water guidelines	WHO highlights particular management needs for small supplies
Geographic and socioeconomic inequality	Service conditions differ between communities	Disaggregated and subnational monitoring where data are available	WHO/UNICEF JMP	Differences can occur by residence, subnational region and other dimensions of inequality

Infrastructure and access

A safe source that is difficult to reach still imposes a substantial household burden. Infrastructure can also influence quality: damaged distribution systems, inadequate treatment and poor source protection can increase risks. Even if water leaves a treatment plant in good condition, conditions during distribution, transport or storage can change its quality before consumption.

Reliability and continuity

Availability when needed is one of the three additional criteria required for safely managed service. This is why two households using nominally similar improved-source technologies can receive different service classifications. Reliability is an important service dimension in its own right and should not be mistaken for a direct laboratory measurement of contamination.

Contamination and source-to-consumer risk

Contaminants can be microbial, chemical, physical or radiological. Metals such as lead and arsenic are chemical contaminants rather than a wholly separate scientific category. Some contamination originates naturally in local geology; other hazards can arise from sewage, agriculture, industrial activity, plumbing or failures in treatment and distribution.

WHO identifies microbial contamination as the most significant drinking-water safety risk. Fecally contaminated drinking water can transmit diarrhoeal disease, cholera, dysentery, typhoid and polio, among other infections. Chemical risks differ markedly by substance, concentration and exposure duration, so the presence of a chemical should not be translated into a health claim without considering those factors.

Management, surveillance and small supplies

Durable protection depends not simply on building infrastructure but on operating, maintaining and monitoring it. WHO’s current guidelines emphasize water safety plans and sanitary inspections as tools for identifying and controlling hazards proactively. This approach recognizes that occasional end-point sampling alone cannot characterize every change occurring throughout a water system.

Rural–urban and within-country disparities

The global rural–urban gap

National averages can conceal major inequalities. In 2024, global safely managed drinking-water coverage was 60% in rural areas compared with 83% in urban areas, according to JMP. The rural–urban gap exceeded 25 percentage points in sub-Saharan Africa, at 35 points; Latin America and the Caribbean, at 30 points; and Eastern and South-Eastern Asia, at 27 points.

These are service-coverage differences rather than evidence that all rural water is unsafe or all urban water is safe. Urban systems can experience contamination and reliability failures, while well-managed rural supplies can provide safe water.

National averages can hide extreme subnational differences

The Democratic Republic of the Congo illustrates the limitations of a single national figure. JMP reports recent subnational estimates ranging from 97% basic drinking-water coverage in Kinshasa to 2% in Mongala, a 95-percentage-point difference. A country-level label would erase precisely the geographic variation that matters for understanding access.

Similar principles apply at smaller scales. Conditions can differ by utility service area, neighborhood, source type and household. The appropriate unit of analysis therefore depends on the question being asked.

Fragile contexts and uneven service coverage

JMP also reports substantially lower coverage in fragile contexts: safely managed drinking-water coverage was 38 percentage points lower than in other countries. Conflict, displacement and institutional disruption can affect infrastructure and service delivery, but the consequences still need to be assessed at the relevant population and system level rather than assumed to apply identically to an entire country.

Health and household implications

Microbial contamination and infectious disease

Unsafe drinking water remains an important public-health concern because contamination can cause both acute and chronic harm. The exact effect depends on the contaminant, concentration, duration and route of exposure, as well as characteristics of the person exposed.

For microbial hazards, fecal contamination is particularly important because contaminated drinking water can help transmit infectious diseases. Young children and other people vulnerable to severe dehydration can face especially serious consequences from diarrhoeal illness. Clear appearance is not evidence that water is microbiologically safe.

Household access burdens are not the same as contamination

Some drinking-water deficits impose burdens even when contamination has not been demonstrated. Under the JMP ladder, for example, a limited service involves more than 30 minutes for a round trip to collect water. That is a significant accessibility deficit, but it should not be described as laboratory evidence of unsafe water.

This distinction is one reason the safely managed indicator is multidimensional. Accessibility, availability and freedom from contamination are related aspects of adequate service, but they measure different problems and can require different interventions.

Interpreting WHO’s drinking-water health-burden estimate

WHO estimated that unsafe drinking water was attributable to 505,000 diarrhoeal deaths in 2019. The reference year is important: this is not a 2024 or 2026 mortality estimate. It is also distinct from WHO’s broader burden estimates for unsafe water, sanitation and hygiene, which cover additional exposures and health outcomes. The two figures should not be added or treated as interchangeable.

Travel interpretation without simplistic rankings

Why global service data cannot tell a traveler whether a specific tap is safe

Household service statistics are useful for understanding population-level conditions, but they are not a traveler-specific tap-water rating. A national safely managed percentage cannot determine the current quality of water at a particular hotel, neighborhood, utility, private well or temporary accommodation.

The reverse is equally important: high national service coverage is not a guarantee about every individual supply. Distribution failures, localized contamination events and building plumbing can matter after water enters a service area. Piped water is therefore not automatically safe simply because it is piped, nor should piped water be presumed unsafe because other communities in the same country have lower service levels.

What to check for a specific destination

For travel decisions, use current destination-specific public-health guidance and advisories from relevant local or national authorities. Where there is a current boil-water, do-not-drink or other water-quality advisory, follow its specific instructions because appropriate precautions depend on the hazard. Global country comparisons are useful background, not a substitute for local, time-sensitive information.

The same caution applies to bottled water and household treatment. Boiling is effective against many microbial hazards but does not remove dissolved metals, salts or many chemical contaminants. No single household treatment method removes every possible hazard, so treatment needs to match the contaminant and conditions involved.

How the latest authoritative datasets should be cited

Cite the observation year, not just the publication year

The principal current global household service source is the WHO/UNICEF Joint Monitoring Programme report Progress on household drinking water, sanitation and hygiene 2000–2024: special focus on inequalities, published in 2025. Its latest household estimates refer to 2024. A statement such as “2.1 billion people lacked safely managed drinking water in 2024” is therefore more precise than calling the figure “2025 data” or “2026 data.”

Use JMP for service coverage and WHO guidance for risk management

The datasets serve different purposes. JMP is the appropriate source for internationally comparable household service estimates and service-ladder definitions. The WHO Guidelines for Drinking-water Quality: fourth edition incorporating the first, second and third addenda, published in June 2026, provide current normative guidance on drinking-water risk management, including health-based targets, water safety planning and surveillance.

WHO guideline values and risk-management recommendations should also be distinguished from enforceable law. Drinking-water standards, monitoring duties and enforcement structures are established by the relevant national or subnational authorities and vary by jurisdiction.

State dataset coverage and metric limitations

Responsible country comparisons should name the indicator, observation year and unit being compared. They should also say whether a value is a percentage or an absolute population count and acknowledge significant missing data. In particular, the 2024 JMP dataset did not provide safely managed estimates for every country.

The issue of countries with unsafe drinking water is therefore not a simple matter of identifying which nations are safe and which are not. Water safety exists on a spectrum shaped by infrastructure, sanitation, natural geology, pollution, reliability and system management. Within any country, some communities may have well-managed, high-quality water while others face serious service or contamination problems. Global data are most informative when they preserve those distinctions rather than turning them into binary national labels.