Who Pays When the Mine Closes?

Start with the arithmetic, because it explains behaviour better than ethics does. Closure and reclamation costs are a deferred negative cash flow, and in an investment model that flow gets discounted. The further out the closure date sits, the less it weighs today. That in itself is not an abuse — bringing future obligations to present value is standard practice, and governments use it too. The Canadian audit says outright that environmental liabilities are measured at the present value of future obligations.

From this, three recurring risks show up across every jurisdiction examined below.

First — underestimation at the outset. A closure-cost estimate made early in a project's life is structurally prone to being too low and then growing. Not necessarily through bad faith: Giant Mine's 2010 estimate of CAD 1 billion was essentially a construction estimate. The revised figure of CAD 4.38 billion includes inflation, contingency, project management, water treatment, pit backfill, long-term monitoring, and rights-holder requirements. The Canadian audit also documents the scale of this as a systemic pattern: the cost of remediating the North's eight largest and highest-risk abandoned mines rose 95% since they were grouped under one federal program in 2020–21.

Second — assurance lagging the obligation. The Wolverine mine in the Yukon is the textbook case. The initial 2006 security was CAD 7.7 million; by 2013 it had risen to CAD 10.7 million, which the company posted with chronic delays. After the closure plan was updated, the territorial government reassessed the liability at CAD 35.5 million — and the company never posted the shortfall. The gap didn't open at the moment of bankruptcy. It opened years earlier, the moment the obligation was recalculated and the money didn't follow.

Third — the cheapness of a legal entity. Setting one up costs a few thousand dollars. Winding one down costs about the same. The pit, the waste-rock piles, the tailings facility, the water-pumping system cost orders of magnitude more. An actor choosing between these two liquidations will choose the cheaper one. This is exactly what the author of the Telegram post that prompted this piece was describing.

The common claim that in Russia "the duty exists but the money doesn't" is, by now, imprecise. Financial assurance for closure activities does exist in Russian law — it's just built around the perimeter of hazardous industrial facilities, not around the perimeter of subsoil use.

The core of the construction is Chapter VII.1 of the Federal Law "On Environmental Protection," introduced by Federal Law No. 446-FZ of December 30, 2021 — known as the "Usolye law." Article 56.1 requires owners of designated industrial facilities to submit to the oversight authority four things: a pollution-prevention-and-remediation action plan, confirmation of a state environmental review approving it, confirmation of an opinion validating the estimated cost, and documents confirming financial assurance for carrying out those measures. The deadline is no later than five years out, but the starting point differs by facility type: for most, it's the expiry of the operating life stated in the design documentation for the buildings and structures; for coal mines the law carves out a separate clause and ties the deadline to whichever comes earlier — the subsoil-use term under the licence, or the mine-life term under the technical development project. If a facility is decommissioned earlier than those dates, the plan is still mandatory.

Article 56.3 spells out exactly what that assurance can consist of. There are three instruments: an independent guarantee, a suretyship, and a reserve fund whose money sits in an escrow account. Requirements for the organizations entitled to issue such guarantees and act as escrow agents were set by Government Decree No. 828 of May 27, 2023.

The details of this construction reward a close read, because they answer exactly the questions Canada spent years litigating.

The aggregate size of the assurance must match the estimated cost of the measures — and that cost estimate itself requires a separate opinion validating how it was determined. In other words, the figure the assurance is calculated from cannot simply be self-assigned: an independent cost-review mechanism is already built into the law. Money in the escrow account is blocked by the bank and transferred into the federal budget if the depositor goes bankrupt, after which it is spent on those same measures by separate government decisions. When rights to the facility transfer — including through universal succession or through enforcement action against the debtor's property — all rights and obligations under the escrow agreement pass to the new owner. The deposit term is tied to the facility's operating life, and for coal mines specifically to the subsoil-use term or the mine-life term under the technical project, and it may exceed that term by no more than six months.

Put simply: for a narrow class of facilities, Russian law has already done what it took Canada's Supreme Court a case called Redwater to do. The assurance is placed beyond the reach of bankruptcy creditors, and a change of ownership through enforcement action doesn't zero out the duty.

What's more, Article 56.1 also answers the other half of the problem — the empty-shell scheme. On a transfer or reorganization less than five years before the deadline, the acquirer is itself obliged to submit the full package: plan, review, cost estimate, and assurance — or documents proving payment of a compensatory fee instead. On a transfer more than five years out, the acquirer submits documents on its own financial standing, the authorized body issues a finding on them, and forwards it to the tax authority and to the property registry. This is, in substance, the same solvency test Queensland's scheme applies to an acquirer, just built into registration procedures.

And the most interesting part is clause 11. On a transfer within a corporate group — between a parent and subsidiary, or between "sister" companies — the parent bears joint-and-several liability for the subsidiary's execution of the pollution-prevention-and-remediation measures. The rule has exactly one carve-out, stated plainly in the text: the case where the parent company itself goes bankrupt. So the corporate veil is pierced — but exactly up to the level where it usually tears.

The question, then, is not whether Russian law knows how to build such constructions. It does, and rather inventively. The question is what circle of facilities they apply to.

The precise formulation of the problem, then, is this: Russian law already recognizes financial assurance for closure measures for specific categories of facilities, including coal mines, but no general regime of assurance for closure and reclamation obligations exists in subsoil use. This is not "nothing exists." It is a far more interesting and harder task — completing a fragment into a system.

There is also substantive regulation that shrinks the obligation itself. Article 23.5 of the Subsoil Law and the procedure for using subsoil-use waste, overburden, and host rock, approved by Order No. 247/04 of the Ministry of Natural Resources and Rosnedra dated April 25, 2023, did something fines never do: they changed the economics of design. Dumping overburden in external waste piles has become less advantageous than putting it back into mined-out space. Engineers confirm this is working, and it is perhaps the single biggest practical success of Russian regulation in recent years — because it shrinks the future obligation during extraction itself, rather than after it.

What the Monitoring Shows

In 2025, Rosprirodnadzor (the federal environmental oversight agency) ran a targeted review of closure-obligation compliance across 1,289 subsoil licence areas whose licences expired that year. The results, published in March 2026, say more than any estimate of aggregate damage could.

Twenty-two sites out of 1,289 is 1.7%. The value of this figure isn't its drama — it's its methodological transparency: the denominator, the period, and the criterion are all known. Earlier discussion of the bill had cited a different figure — 352 abandoned subsoil sites identified over 2023–2025, concentrated most heavily in the Magadan, Amur, and Irkutsk regions, Buryatia, and Kuzbass. Both figures point to the same fact: there is no complete historical registry of abandoned sites in the country, and its absence does not mean the problem is small.

Draft Law No. 1050064-8

In October 2025, a group of deputies led by the chair of the relevant committee introduced a bill in the State Duma making financial assurance for land reclamation mandatory. Licences for gold exploration and extraction — lode, placer, and man-made deposits alike — as well as for widely occurring minerals, would be issued only after assurance was posted: an irrevocable bank guarantee from a systemically important bank, or a reclamation fund in a special account or deposit. The amount would be calculated from average regional reclamation-cost rates approved by the government every three years; regional estimates put those rates at 600,000 to 1.5 million roubles per hectare. The bill's proposed effective date was September 1, 2026.

The status of its passage deserves a careful word. On February 12, 2026, the responsible committee held what is called a "zero reading" of the bill, involving agencies, mining companies, experts, and regional authorities; according to the committee chair, 96% of the feedback received from regions was supportive, and the meeting also discussed adjustments to specific provisions ahead of a second reading. In early June 2026, trade press reported that deputies hoped to pass the bill in its first reading during the spring session. We were unable to confirm through the State Duma's official bill tracker, at the time this piece was prepared, that a first reading had actually taken place — so we assert neither that it did nor that it didn't, and describe the September 1, 2026 date as the bill's proposed date rather than an approaching one.

Industry objections come down to two. First, the freezing of working capital: for small artel-style operators, tens of millions of roubles in a special account before work even begins can mean simply being unable to operate. Second, selectivity: the requirement catches gold and widely occurring minerals but not non-ferrous ore projects or most of the rest. The Union of Prospectors and the Union of Gold Producers call the bill discriminatory.

It's important here not to conflate two different mechanisms. Coal genuinely isn't covered by the proposed reclamation-fund regime — but it doesn't remain entirely without a financial mechanism: coal mines are explicitly named in Article 56.1, and the assurance requirement under Article 56.3 applies to them. These are two distinct obligations: land reclamation under the new bill, versus pollution prevention and remediation upon decommissioning under Chapter VII.1. The discrimination argument retains its force with respect to ore projects and to facilities outside the perimeter of Class I–II hazardous industrial facilities.

In parallel, the state is paying for what has already happened. The federal "General Clean-Up" (Генеральная уборка) project provides for more than 140 billion roubles through 2030 for inventorying and eliminating sites of accumulated environmental harm; in February 2026, the first three sites selected for federal funding were named. The budget mechanism for cleaning up is built and operating. The mechanism for preventing new sites from appearing is still being designed.

Canada is the country where the cost of getting assurance wrong has been calculated most precisely — because it had to be paid out of the budget, and because an independent auditor published the number.

In April 2024, the Commissioner of the Environment and Sustainable Development released a report titled "Contaminated Sites in the North." Its overall financial finding: since the Federal Contaminated Sites Action Plan launched in 2005, the aggregate liability for known sites had grown from CAD 2.9 billion to CAD 10.1 billion, of which more than CAD 6 billion sits in the North. In the Commissioner's assessment, this is an enormous burden on the taxpayer and a failure to implement the "polluter pays" principle — because so many private-sector sites had to be taken over by the federal government.

Faro. A lead-zinc mine in central Yukon that operated from 1969, at its peak accounting for more than 30% of the territory's economic activity. Its last owner, Anvil Range Mining Corporation, entered creditor protection and was placed into receivership in 1998. The assurance in place: CAD 14 million. In 2002, the Commissioner estimated remediation cost at CAD 200 million. By fiscal year 2022–23, the discounted liability estimate stood at CAD 5 billion, with CAD 791 million already spent by that point. The site still holds 70 million tonnes of tailings and 320 million tonnes of waste rock; it requires ongoing care for the foreseeable future to keep contaminated water from reaching surrounding waterways.

Giant Mine. A gold operation on the outskirts of Yellowknife that roasted arsenopyrite ore and left 237,000 tonnes of highly toxic arsenic trioxide dust in underground chambers. Assurance in 2002: CAD 7.4 million, against a then-estimated remediation cost of up to CAD 400 million. By 2022–23, the discounted liability stood at CAD 3.9 billion; the project budget approved by the Treasury Board is CAD 4.38 billion, including CAD 710 million in historical costs since 2005. Of the eight sites in the federal program, the active remediation phase had begun, as of the audit, only here. The chosen strategy — artificially freezing the arsenic in place — implies perpetual site care. Not "for a long time." Perpetually.

Wolverine. An underground zinc mine in southeast Yukon, on Kaska territory. It operated from 2011 to 2015. In 2018, the territorial government took over care of the flooded site, whose water is contaminated with cadmium, selenium, copper, and lead; in 2019, the company was declared bankrupt. Receivership alone cost more than CAD 21.5 million over four years, nearly CAD 20.5 million of which the Yukon government advanced. A judge described the operation as an irresponsible mining venture. PwC's audit review, produced in the wake of the case, recommended that any assurance determination account for the company's and its parent's financial condition, the full life-of-mine plan, and the project's sensitivity to metal prices. That is, in effect, a ready-made design brief for any new assurance system.

Redwater: The One Genuinely Strong Answer

On January 31, 2019, the Supreme Court of Canada ruled in Orphan Well Association v. Grant Thornton Ltd. The bankruptcy trustee for oil-and-gas company Redwater Energy had disclaimed 107 licensed wells carrying onerous environmental obligations, preserving liquid assets for secured creditors. The Court, five to two, held that abandonment and reclamation obligations are not claims provable in bankruptcy, and therefore are not subject to the priority scheme of the federal Bankruptcy and Insolvency Act. The trustee bears no personal liability, but the bankrupt estate is obliged to perform them — ahead of settling with secured creditors.

Legally, this is the most radical of the existing mechanisms: it makes site closure not a debt, but a condition of the right to extract at all. Economically, it turned out to be costly — creditors reassessed the value of collateral, financing for small operators got more expensive, and disputes over the destabilization of credit markets haven't died down to this day. That is an honest price: if the closure obligation is truly senior, someone has to reserve capital against it in advance — either the borrower or the lender. And this is exactly the question the Russian bill is raising for the first time, by touching Articles 131 and 133 of the bankruptcy law.

Western Australia. The Mining Rehabilitation Fund Act 2012 introduced an annual, non-refundable levy for all mining tenement holders, except those operating under state agreements. The levy is calculated from disturbance area and the outstanding rehabilitation liability as of the reporting date; it became mandatory from July 1, 2014, and replaced unconditional performance bonds, which were returned to companies. The money goes into a pooled state fund and can be used if an operator is unable to meet its obligations; interest on the fund finances a program of work at historically abandoned mines.

There are two known weak points, both confirmed by the fund's mandatory statutory review. First: the levy runs at roughly 1% of the estimated liability per year — meaning, all else equal, a mine would have to operate for a hundred years for the fund to accumulate the cost of its own closure. The pool functions as insurance against isolated defaults, not as assurance for every individual site. Second: the abandoned-mines work program is tied to interest income, and during periods of low rates it was underfunded — the review stated outright that a significant share of planned work went undone for exactly that reason. The minister retains the power to additionally demand an unconditional guarantee on sites carrying elevated risk of the liability shifting to the state, which is itself an admission that the pure pool isn't enough.

Queensland. The financial-assurance scheme in force since April 1, 2019 sorts facilities by risk. Permits with an estimated rehabilitation cost below a threshold require full assurance for the entire sum — a bank guarantee, an insurance bond, or cash. Facilities above the threshold undergo a risk assessment that accounts for the holder's financial soundness, including its parent company, and the project's characteristics; the outcome determines whether the holder pays an annual contribution into the fund or provides individual assurance. From October 1, 2025, the threshold was raised from AUD 100,000 to AUD 10 million. A separate residual-risk fund has been set up for what remains after formal rehabilitation is complete.

Queensland's logic is sharper than Western Australia's: the pool covers the reliable, individual assurance covers the risky, and the parent structure's solvency is folded into the assessment. This is a direct answer to the empty-SPV scheme. There's counter-pressure too — the latest review of the scheme is explicitly framed as a search for unnecessary barriers to investment.

Season and logistics. Construction work is possible only a few months a year. Materials come in over winter roads or by ship in a short shipping window. Every scheduling error costs a whole year.

Permafrost. A large share of northern closure designs — dry tailings covers, freezing, waste isolation in frozen ground — is engineered for a temperature regime that is ceasing to exist. This isn't a hypothesis: the 2024 Canadian audit explicitly called weak the way the department reported on Giant Mine's climate risks, and noted that the closure plan was insufficiently developed on climate-change adaptation. A strategy built on artificial freezing is a perpetual obligation to keep something cold in a warming region.

Water. Acid drainage and metal leaching don't stop when a mine closes. Water treatment isn't a project — it's an ongoing operation with ongoing costs. Hence the audit's formula: Faro requires ongoing care for the foreseeable future.

Low population density. Rarely said out loud, but political attention to an abandoned site is roughly proportional to how many people can see it. Giant Mine sits on the edge of the territorial capital — and its budget quadrupled after a public impact assessment. A site three hundred kilometres from the nearest settlement gets no such assessment.

Greenland's Maarmorilik case shows a different facet of the same problem. A lead-zinc mine operated from 1973 to 1990, discharging tailings into a fjord; at closure, a clean-up was carried out that was reasonably thorough by the standards of the era, and the town was dismantled. And yet monitoring by Danish research institutes continued for decades: lead content in transplanted mussels declined by an average of 5.5% a year, and effects from the earlier mining were still detectable in 2009. That's the rate of change in one specific indicator, not the rate at which the ecosystem heals itself. Closing a mine is not an event — it's a process that spans generations, and someone has to fund it the whole time.

The same logic operates in the Russian Arctic in the absence of assurance for most facilities. Vorkuta is the rare case where coal mines are actually covered by Article 56.1, but the trajectory itself is telling. Today Vorkutaugol operates four underground mines and one open pit; around the city sits a ring of settlements built for mines that no longer exist. In April 2026, the company announced a production-cut program amid a surplus of coking coal and low prices, naming as an added risk factor the expiry, on April 30, 2027, of its agreement with its main buyer. This is exactly the moment the closure obligation stops being distant: the closure horizon shortens faster than the assurance accumulates.

Flooding worked-out mine workings — the cheapest and most common way of "closing" a site — pushes the problem into hydrogeology, where it becomes diffuse and legally ownerless. It's the perfect form of accumulated harm: it has no owner, no cost estimate, and no deadline.

The list from the first figure reads differently with one extra column added. Faro and Wolverine sit on the traditional territory of the Kaska people. Giant Mine sits on Yellowknives Dene land, whose hunting and berry grounds it destroyed. Between 1948 and 1951, the mine ran without any emissions control, with arsenic releases estimated at 7.5 tonnes a day. In April 1951, on Latham Island — today Ndilǫ — a two-year-old boy died of acute arsenic poisoning after drinking melted snow; his family was paid $750. The community says there were several deaths that period; government records confirm one. Dust-control systems were installed at the mine only after this. Blaiken sits within Sami reindeer-herding territory. Maarmorilik sits in a fjord system where the nearest settlement is 25 kilometres away.

This overlap can be counted. Researchers at the University of New Mexico find that more than 600,000 Native Americans live within roughly ten kilometres of an abandoned mine. And the same 2024 Canadian audit noted that the federal action plan lacked key targets for Indigenous engagement and for the socio-economic benefits due to them — meaning that even where clean-up money exists, the involvement of the people living nearby remained optional.

For the framework of free, prior, and informed consent, there's an uncomfortable consequence here. In practice, consent to a project is often locked in at an early stage — the point at which jobs, compensation, and the initial scope of impact are being discussed. A closure plan at that stage either doesn't exist, or exists as an estimate that's undersized for the reasons set out earlier. What ends up happening is that extraction becomes the actual object of consent, while its consequences are committed to a perpetual period about which almost nothing was said at the time of agreement.

Canada has begun to acknowledge this: the Giant Mine project includes an independent oversight board with Yellowknives Dene and North Slave Métis Alliance participation, a benefits agreement worth up to $20 million, a ministerial special representative examining how the historical mining affected treaty rights. All of it arrived more than twenty years after the operator's bankruptcy. It would have been simpler and cheaper to bring the community into closure planning before extraction ever began.

While the reclamation-fund bill sits at "zero reading," the substantive part of the debate has moved into industry channels — a rare case where practitioners' discussion is ahead of the statute's text.

The argument isn't over whether assurance is needed, but over which document it should be tied to. One position: the technical mine-development project and the feasibility study behind it. The position of a practicing engineer writing in the "Mining Business" channel: drop the feasibility study from the construction entirely — for a great many deposits of widely occurring minerals, no feasibility study is ever produced, and for previously explored deposits the feasibility study may have been done decades ago and contains none of the calculations the new mechanism would need. His proposal is a two-stage scheme: the technical project sets the total volume and estimated cost of future closure obligations, and the annual mine-operations plan determines what share of that sum must be assured at a given stage, based on the actual progress of mining.

The argument for anchoring to the annual plan is strong and comes from practice: no quarry develops over five years the way it was drawn on the technical project's calendar schedule. Throughput, the direction mining advances in, equipment, geological conditions, overburden volumes, the position of the working face — all of it shifts. Tying the annual reserve amount to the technical project would mean that every change in the calendar schedule requires amending the project and getting it re-approved — with a corresponding load on the review bodies.

The comparative material in this piece confirms the core intuition of that debate. The most persistent failure that recurs across different systems is the timing failure: the obligation changes faster than the assurance does. Wolverine didn't collapse because the original figure was wrong — it collapsed because the obligation was recalculated and the assurance never caught up. Western Australia recalculates its levy annually against actual disturbance. Queensland reassesses both the cost estimate and the holder's risk category. The Russian industry idea — "the technical project sets the volume, the annual plan sets the pace" — structurally matches these solutions.

But this is also where the central fork in the road sits. The scheme works only if the annual plan answers the question "how much money must be assured today," not the question "how many hectares will we disturb this year." The scale of planned technogenic disturbance is not the same thing as a financial obligation. A hectare of external waste pile, a hectare of tailings facility, a hectare of a workings due to be flooded, a hectare of a site requiring infrastructure demolition and ongoing water treatment — these cost fundamentally different amounts. This is exactly where the Russian construction of average regional per-hectare rates — 600,000 to 1.5 million roubles — risks repeating the mistake Western Australia made with its flat 1% levy: collecting a sum that is politically acceptable and technically inadequate.

The objection to using the feasibility study is worth agreeing with in the narrow administrative sense and disagreeing with in the economic sense. A reserves feasibility study genuinely isn't the document on which a legally binding assurance amount can be calculated. But thinking about the closure obligation at the feasibility-study stage isn't premature at all. If a project is economically viable only on the assumption that closure costs get counted later, or not fully, then the feasibility study isn't showing the project's real economics. Two things need separating here: the estimate of the obligation can start very early, while the legally binding assurance amount can be refined later, once the technical project and the actual course of mining exist.

And a third thing the industry debate doesn't touch at all yet: protection against a default that occurs before the fund has finished accumulating. Building up assurance gradually is the most humane construction for the industry and the most vulnerable one for the state, because the gap is largest in a project's early years. There are exactly three answers to this, and all three have already appeared in this piece: a pool to cover early defaults, a group solvency assessment feeding into the schedule, and protection of the assurance from the bankruptcy estate. The Russian bill is already attempting the last of these, by touching Articles 131 and 133 of the bankruptcy law.

1. What's the estimated cost of closure, and who calculated it?If the operator itself did the math and it never went through independent review, that's not a cost estimate — it's a model parameter. Ask what it includes: water treatment, backfilling workings, monitoring, demolishing infrastructure, removing equipment, long-term operation after work ends. For facilities covered by Chapter VII.1, a separate opinion validating the cost estimate is required — ask for that document specifically.
2. Does assurance exist, and in what form?A statement of intent is not assurance. Assurance is an independent guarantee from a named organization, money in a named escrow account, or a contribution to a named fund, with the amount and recipient specified. For facilities covered by Article 56.1, these documents must exist five years before the end of operation.
3. Who reassesses the amount, and how often?The most persistent failure recurring across different systems is timing: the obligation changes faster than the assurance does. A rule requiring reassessment at least every two or three years, and after every project change, is worth more than a higher starting figure.
4. Who stands behind the operator?Find the ultimate beneficial owner and the parent structure's jurisdiction. A project sitting in an asset-free SPV is a design decision made about you.
5. What does the plan say about the period after closure?Distinguish active reclamation from perpetual care. The latter is an obligation with no end date, and it needs its own dedicated funding source.
6. Who among the people living nearby was involved in preparing the closure plan?If the answer is "no one," consent to the project was obtained on incomplete information. This applies equally to territories of traditional land use, where the closure plan is almost never opened up for discussion with communities.

The author of the Telegram post that started this piece closes it with a thought that's hard to argue with: order always ends up cheaper than disorder. The record backs this with numbers — the rise in Faro's discounted liability, from CAD 200 million in 2002 to CAD 5 billion in 2022–23, is the price of what wasn't done earlier.

But the formula has a hidden parameter — time and subject. Order is cheaper in aggregate and more expensive today, for the specific operator. Disorder is more expensive in aggregate and free today, for that same operator. The economics of closure isn't a dispute between the environment and business — it's a dispute over who holds the risk in the gap between those two moments. As long as the operator holds the risk, regulation works. The moment the operator ceases to exist, the risk shifts to the budget, and then to people who never signed anything. Canada's auditor called this a failure of the "polluter pays" principle, and that's the precise formulation: the principle isn't repealed, someone else is simply appointed to pay it.

Russia sits not at the point of "nothing exists," but at the point of completing an unfinished structure. Financial assurance for closure measures already exists — for designated industrial facilities and coal mines, with a set of instruments comparable to Queensland's. The question is whether that perimeter will be extended to subsoil use as a whole, and by what design. The reclamation-fund regime under discussion covers gold and widely occurring minerals; left outside it are non-ferrous ore projects and everything that doesn't fall within the narrow transitional definition of a "designated industrial facility" in force until March 1, 2035. That is the substantive objection — not "too strict," but "too fragmented."

Experience across four jurisdictions converges on three points. Assurance must be reassessed mandatorily and regularly, not at the holder's discretion — which is precisely what Russian engineers are arguing about right now, debating the technical project versus the annual mine plan. Progressive reclamation during extraction is the one mechanism that actually shrinks the obligation rather than deferring it; in Russian law it has already appeared in the rules on placing overburden into worked-out space. And finally, assurance must be protected from the bankruptcy estate, and liquidating a legal entity must not be a way of discharging the obligation. As long as it remains one, everything else in the regulation is advisory.