On 2022-06-23 Canada repealed the federal statute meant to set drinking water standards on First Nations reserves, and its regulations were never brought into force — as of 2026-08-15 no replacement law is in force, Bill C-61 died on the order paper at prorogation in January 2025, Bill C-37 is awaiting second reading, and 40 long-term drinking water advisories remain in effect as of 2026-08-13

What is happening?

There is no federal law setting drinking water standards on First Nations reserves in Canada, and there has not been one since 2022. The Safe Drinking Water for First Nations Act, passed in 2013, was repealed on 2022-06-23 through the Budget Implementation Act, 2022, No. 1. The regulations that would have given that Act its actual standards were never brought into force, so the statute spent its whole life as an empty frame. First Nations had asked for the repeal, on the grounds that the Act imported provincial standards without consultation and carried no funding guarantee. Four years on, the department page of Indigenous Services Canada describes the situation as a long-standing legislative gap.

The advisories the missing law was meant to prevent are still running, and the count went up in the last year. As of 2026-08-13 there were 40 long-term drinking water advisories in effect in 38 communities, with 156 lifted since November 2015. The Statistics Canada series on this indicator runs 103 in 2016-05, down to 37 in 2025-05, and back up to 40 in 2026-05. A long-term advisory is one that has been in place for twelve months or more. Indigenous Watchdog notes that the share of advisories resolved sat at about 80 percent in both April 2022 and June 2026, meaning four years passed without the proportion moving.

One community has been under an advisory for three decades. Neskantaga First Nation in Ontario, a community of roughly 350 people reachable only by air and by winter ice road, has held the longest-running advisory in Canada since February 1995. A 2024 news report put it at 29 years.

The deadline that used to exist was missed and never replaced. In the 2015 federal election the government promised to lift every long-term advisory on public systems on reserve by March 2021. That date passed, and no new end date has been set since.

Two attempts to write a replacement law have not produced one. Bill C-61, the First Nations Clean Water Act, followed consultation that began in 2018 and received unanimous support from the executive committee of the Assembly of First Nations in 2024, then died on the order paper when Parliament was prorogued in January 2025. A successor, Bill C-37, was introduced and passed first reading on 2026-06-16, announced alongside 4.6 billion Canadian dollars. As of 2026-08-15 it is awaiting second reading, with the House due to resume on September 21; committee, report stage and third reading have not been reached. The Assembly of First Nations has said the funding provision rests on a best efforts formulation rather than a guarantee, that the bill was developed without substantive consultation, and that a human right to water is not written into it.

And while one arm of the federal government proposes a rights-based law, another argues in court that no such duty exists. On 2025-12-05 the Federal Court held in Shamattawa First Nation v. Canada (Attorney General), 2025 FC 1927, that Canada owes a sui generis fiduciary duty, an ad hoc fiduciary duty and a common law duty of care with respect to safe drinking water on reserve, and that sections 15, 7 and 2(a) of the Charter are engaged. That was stage one of the case; breach and remedy are left to stage two. The class covers status First Nations members living on reserve or on modern treaty lands affected by an advisory since 2021-06-20. Canada appealed on 2026-01-05, and its written argument to the Federal Court of Appeal is that it owes no legal obligation to provide safe drinking water to individual First Nations members. Bill C-37 states as policy the progressive realization of a human right to safe drinking water.

Whose problem is this?

The party that would have to write the law is the same party arguing in court that it is under no obligation, and the communities affected are not in either room. Nothing in the arrangement produces a step that happens automatically when another year passes.

Where does this problem end?

The boundary matters because two different failures are easy to merge into one. Building a treatment plant is an engineering and funding question, and setting a binding national standard for the water that plant produces is a legislative question. Canada has now spent four years with neither a standard nor a date by which one is meant to exist.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: no source opened here gives a date by which a replacement statute is meant to be in force, a date by which the advisories are meant to end, or a numerical target for either. The March 2021 commitment was the last stated end date and it was not replaced after it lapsed. Bill C-37 describes a direction, being the progressive realization of a human right to safe drinking water, but a bill awaiting second reading sets no date and the sources are silent on what date is intended.

How big is it?

40 long-term advisories in 38 communities, as of 2026-08-13. That is the federal figure, and it counts advisories and communities rather than people. A second snapshot from the same year reports the same 40 advisories across 37 communities, and neither source states its counting basis, so the discrepancy is left visible here rather than reconciled.

No federal source opened here publishes the number of people living under a long-term drinking water advisory. The federal count is a count of advisories and a count of communities, and the population behind those counts is the figure that would actually say how large this problem is. The only bound available is an upper one, being the statement that a single advisory can affect up to 5,000 people with wide variation by community size, and the only individual community figure available is Neskantaga at roughly 350. Those two numbers bound nothing together, because the distribution between them is not published. The population file for this document therefore records the estimate as not derivable and names the missing term.

The federal count is also not the whole population at risk. It covers public water systems on reserve. The 2010 National Assessment recorded 13.5 percent of Indigenous community households on trucked water delivery, 13 percent on individual wells, and 1.5 percent with no water service at all. None of those households enter the count of 40. No refreshed national assessment of that kind was found in this round, so the most recent full picture of that separate population is sixteen years old.

For context on the denominator, Crown-Indigenous Relations and Northern Affairs Canada describes more than 630 First Nations communities in Canada. A ratio of affected communities to that total is not computed here, because the two figures come from different sources with different update cycles and no source opened here establishes that they share a reference date.

Under what conditions does it arise?

1. There is no rule to enforce, rather than a rule going unenforced. The 2013 Act was enabling legislation whose regulations never came into force, and then the Act itself was repealed. So for the entire period there has been no federal standard that a water operator, a department or a court could point to and measure against. Provincial drinking water standards do not extend onto reserve.

2. Nothing in the process forces a bill to become a law. Bill C-61 absorbed six years of consultation and the support of the Assembly of First Nations, and then a prorogation ended it. Legislation that dies on the order paper does not resume where it stopped; the successor bill starts again at first reading, which is where Bill C-37 was on 2026-06-16 and near where it remains.

3. The commitment was a date rather than a duty. A political target for March 2021 could be missed without any legal consequence, and after it was missed the response was to stop stating a date rather than to state a later one. An unmet promise with no successor promise leaves nothing to be late against.

4. Operating money is announced rather than owed. The concern the Assembly of First Nations raises about Bill C-37 is that its funding language turns on best efforts, which is the same shape of commitment that the repealed 2013 Act was criticised for lacking. A treatment plant that is built and then underfunded in operation produces an advisory just as reliably as one that was never built.

5. The federal count defines the problem narrowly enough to shrink it. Short-term advisories, individual wells, cisterns and trucked delivery are outside the tracked number, so households in those situations can neither appear in the count nor be lifted from it.

6. The legal question is being contested rather than settled. As long as the appeal is live, the existence of a duty is undetermined, and a duty whose existence is undetermined cannot be the basis of a schedule.

What has been tried?

Two directions ran at the same time and pointed opposite ways. One arm of the federal government tabled a bill declaring safe drinking water a human right to be progressively realised, and another arm asked an appellate court to find that no such obligation is owed to anyone.

What was found?

Why is it still unsolved?

Institutional gap — the instrument that would govern this does not exist, and no process is obliged to produce one.

An absent standard fails differently from an unenforced one. Where a rule exists and is ignored, there is a body that can be asked to act and an order that can be pointed at. Where the rule itself was repealed and never replaced, there is nothing to point at, and the question of whether anyone is even obliged to act is itself the matter now before the courts. That is why this document is filed as a gap rather than as a failure of enforcement.

The second part is that the repeal was correct and still left a hole. First Nations asked for it, because a statute that imported provincial standards without consent and without funding was worse than useless to them. The mistake was not the repeal; it was that repeal and replacement were treated as separable, and only the first half happened. Four years is long enough that the interim has become the arrangement.

The third part is that the process for producing the replacement has no memory. A bill that has absorbed six years of consultation carries exactly the same legal weight as a blank page the day after Parliament is prorogued. So the work survives only as a document that someone chooses to reintroduce, and the reintroduced version can differ from what was agreed, which is precisely the objection now being raised about the successor bill.

The fourth part is that the delay is invisible as an event. There is no day on which anyone announces that a law will not be written. The advisory count is published, capital projects are announced, settlements are paid, and each year closes with a number that looks a great deal like the number from the year before. In 2026 that number was higher than in 2025, and nothing in the machinery treated that as a trigger.

What observation would mean it is solved?

Candidates — (a) a replacement statute is in force with its regulations actually brought into force and enforceable standards named in them (b) the count of long-term advisories reaches zero and stays there across several consecutive years (c) the appellate courts settle whether Canada owes an enforceable duty, and the answer is yes.

(a) alone repeats the failure that produced this problem. The 2013 Act was in force for nine years and set no standard, because its regulations were never made. A statute counts as an answer only if the instrument beneath it exists and binds. That is the exact distinction this document turns on, and it is easy to lose in a headline about a law being passed.

(b) alone is fragile in both directions. A count that reaches zero in one year can rise again in the next, and the record of the last decade shows exactly that pattern, with the figure falling to 37 and then returning to 40. The count also excludes trucked water, cisterns, individual wells and short-term advisories, so it can reach zero while households outside its definition remain where they were. Zero on a metric that does not include you is not a solution to you.

(c) alone establishes an obligation without a schedule. A declared duty still needs a remedy stage, a compliance path and a body willing to supervise it, and the neighbouring record of binding orders on services to First Nations children shows how long a duty can stand while the thing it requires does not happen. The three have to be read together, and (a) has to be read down to the level of whether regulations exist.

What is it connected to?

Fills with researchon-reserve housing and infrastructure funding, wastewater as distinct from drinking water, the implementation of the United Nations Declaration on the Rights of Indigenous Peoples Act in domestic legislation, the parallel service gaps for Inuit and Métis communities, water security on tribal lands in other countries, and the relationship between this gap and the separate track of binding tribunal orders on services to First Nations children. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How many people live under a long-term advisory. Indigenous Services Canada publishes advisories and communities and stops there. No source opened here sums the affected population, and no per-community population is given for 37 of the 38 communities.
• How many households are affected outside the federal count. The last full picture of trucked delivery, individual wells and unserved households is the 2010 National Assessment. No refreshed national survey of that kind was found in this round, so what looks like an absence of the problem outside the count is an absence of measurement.
• What the 2025 follow-up audit actually found. The audit report of October 2025, its press release and its mirrors all returned automated-access refusals in this round, and the original press release URL now returns a not-found error. Figures circulating in search summaries, being 18 of 34 recommendations inadequately implemented and 9 of 35 advisories running longer than ten years, could not be confirmed against the primary document and are therefore not stated as findings anywhere in this dossier.
• The vintage of the health comparison figures. One analysis describes waterborne illness rates in Indigenous communities at many times the national average and a far higher likelihood of having no piped water at all. The reference year for those comparisons is not stated in the source, so they are recorded here as unusable rather than repeated as current.
• Whether the annual operating shortfall figure holds. A figure of 138 million Canadian dollars a year for operations and maintenance appears in a campaign source as a second-hand citation. The primary source behind it was not opened in this round.
• What has actually been spent from the settlement. The commitment of at least 6 billion Canadian dollars for infrastructure was confirmed as announced. No year-by-year record of what has been spent against it was found here, so this dossier can state the promise and not the delivery.
• Whether Bill C-37 will pass, and in what form. The bill is awaiting second reading. Whether the objections about consultation, the best efforts funding language and the missing human rights wording are addressed in committee has not happened yet and cannot be reported.
• When the appeal will be heard. One source describes a hearing expected during 2026. No scheduled date was found.
• How the affected share of communities has moved over time. The total of more than 630 First Nations communities and the annual advisory counts come from different sources on different update cycles, and no source opened here places them at a common reference date, so no time series of the affected proportion is computed.

See the evidence

No primary legal or audit document was read in full. The two departmental pages of Indigenous Services Canada and the parliamentary legislative record were opened directly and carry the repeal date, the statement that the regulations never came into force, the advisory counts and the status of Bill C-37. The Federal Court judgment itself was not opened; its holding is taken from two independent legal summaries that agree on the docket number, the date and the duties found, and the position of Canada on appeal rests on one of those two alone. The October 2025 follow-up report of the Auditor General is the largest hole in this evidence base: eight separate routes to it, including the report PDF, the press release, its mirror, a ministerial statement and news coverage, all failed on automated retrieval, and figures circulating in search summaries were left out of the body rather than repeated. Where opened sources overlap they agree on the advisory total of 40 and on the repeal date, and where they disagree the disagreement is left visible, being 38 communities in the federal page against 37 in the campaign page for the same 40 advisories. This is a Path A output, research-based definition, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below