The mine that never closes: Joburg’s acid water is a R17bn business we keep paying to throw away

The Witwatersrand’s mines can’t close. The water rises regardless. Since 2002 the state has pumped and neutralised it on a budget line labelled “interim”: R562m this year, R728m by 2029/30, with nothing earned back. In the 34th of his Gold Papers, Dr Duarte da Silva asks the obvious question: why not run it as a business? Germany pays for the Ruhr’s “eternity costs” out of a €15.9bn endowment. Britain sells heat from its flooded mines. Da Silva prices the basin as a water utility, a power station, a materials plant and a land company. His figure is R17bn, paid for from mine rehabilitation money that is already set aside. Only the law stands in the way..By Dr Duarte da Silva.Closure is the wrong word, and the goldfields have paid for it for a generation. A Witwatersrand mine cannot close. The water rises on its own timetable, the dumps stay where they were poured, and the certificate meant to turn a permanent flow into a finished event went to twenty-seven large mines out of 803 issued in six years. This paper puts the alternative on a balance sheet: the basin run as a permanent enterprise (a water utility, a power station, a materials plant and a land company), funded by an endowment the provisions already hold, staffed by the second shift, and offered as a model to every goldfield that will one day need one..Every abandoned goldfield has taught the same lesson at the same price: the mine leaves, the water stays, and the public pays for both. The Ruhr answered with a foundation and Britain with an authority. Montana learned it from three thousand dead geese, California from water more acidic than battery electrolyte, and Yellowknife from 237,000 tonnes of arsenic dust it now keeps frozen. Each has stopped pretending its mines closed and runs them as permanent water-management enterprises with a public balance sheet. The Witwatersrand has three treatment plants, a decant in the west since 2002 and another in the Central Basin at Boksburg Lake in each of the last two winters, a long-term solution costed at ten to twelve billion rand and deferred for a decade, and a closure regime that has never reached the end it describes. Priced as a business, the same obligations look different: the pump that must run forever is a utility, the void is a store of energy, the dump is a quarry and a building plot, and the workforce the closure model let go is a second shift. Costed that way, the mine that never closes is the basin's largest remaining business rather than its burden, and one the rest of the world will want to buy..I. Closure, as practised: a finish line nobody has crossed.The problem starts with the word. The statute treats a mine as a project with an ending: extract, rehabilitate, certify, release. The land goes back to something like its prior state, the company is excused and the file is closed. Every rehabilitation provision in the country is written to fund that sequence, and on the Witwatersrand none of it applies. There is no prior state to go back to. A hundred and forty years of deposition and undermining have made the central Rand a manufactured landscape, and it has to be managed as one, the way the Dutch manage a polder. Rehabilitation here is continuous: the void refills on its own schedule, and the water has to be lifted or treated on the same schedule, with no end date. The closure certificate, the instrument meant to mark the end, went to twenty-seven large mines out of the 803 certificates the Department signed between 2011 and 2016; the rest went to mining permits, prospecting rights and borrow pits. The state's own register shows the result: some six thousand one hundred derelict and ownerless sites, each a mine that closed in fact without closing in law, its owner gone long before any certificate.The word sets the wrong objective. A management team told to close a mine will minimise the cost of closing it, deferring every expense it can to a date after the certificate. The certificate never comes, so the deferral becomes permanent. A regulator told to certify closure will not certify what cannot be shown to be finished, and nothing on this basin can be, so it withholds. Each side is acting rationally inside a framework that describes something that does not exist. The mine sits in the limbo the law calls care and maintenance, the province of shells and liquidators, while the water rises through it. Change the word and the incentives change with it. A management team running a permanent water and land enterprise will invest in the pump, because the pump is the business. A regulator licensing a utility will license it, because that is its job.The world's older goldfields and coalfields have already made the change, pushed by the water rather than by foresight, and they have built the institutions, endowments and accounts to run mines that never close. South Africa has the same water, a larger endowment already lodged, and a workforce the closure model has been shedding since 1987. It lacks the word..A mine told to close defers every cost past a certificate that never comes. A mine run as a water and land enterprise invests in the pump, because the pump is the business..II. Seven places that stopped pretending.The Germans have had a word for it since the coal ran out: Ewigkeitskosten, eternity costs. When the last Ruhr colliery closed in December 2018, the water kept rising through workings beneath a conurbation of five million people, in a region that had subsided by as much as twenty metres in places and could not let the water find its own level. The 2007 law that wound down hard-coal mining therefore created a foundation, the RAG-Stiftung, with one purpose: to pay, indefinitely, for lifting the mine water, pumping the polder areas that subsidence left below the rivers, and cleaning the groundwater. A 2006 actuarial estimate put the capitalised eternity cost at twelve and a half to thirteen billion euros. The foundation had to hold at least eight billion by the end of 2018 and took over the payments from 2019. Built from the Evonik shares it received at inception and the industrial holdings it has bought since, its assets stood at seventeen billion at the end of 2024 and 15.9 billion at book value at the end of 2025, with some ten billion of that set aside as provisions against the perpetual tasks. It spent 309 million euros on the perpetual tasks in 2025, and more than two billion since the payments began. Six central stations lift some seventy million cubic metres of Ruhr mine water a year, and RAG's twelve stations across the Ruhr, the Saar and Ibbenbüren about 110 million; the Saar alone is permitted up to twenty-six million. Nobody in Germany calls this closure. The foundation describes a permanent obligation, and its annual report reads like a utility's.Britain reached the same place by another route, and in November 2024 changed its name to say so. The Coal Authority, created in 1994 to hold the liabilities of a nationalised industry being privatised, became the Mining Remediation Authority. It runs eighty-two mine-water treatment schemes and cleans more than 122 billion litres a year (122 million cubic metres, well above the Ruhr's central stations) to keep iron and acidity out of the rivers of Durham, Yorkshire, South Wales and the Midlands. A quarter of Britain's homes sit on the coalfields it maps, and it has started to sell what it used to pump away: the warm water in flooded collieries is now a heat source, and the first district-heating scheme drawing on a flooded mine opened at Gateshead in 2023 and now serves more than 645 homes and two dozen large buildings. The authority did not set out to manage eternity. The alternative was orange rivers, and once it faced that, it built the balance sheet.The American cases were learned from disasters rather than planned. At Butte, Montana, pumping at the Berkeley Pit stopped in 1982 and the largest open pit in the state began filling with water that leached copper, cadmium and arsenic from its walls. The record of decision set a protective level of 5,410 feet above sea level, beyond which the water would begin to flow into the aquifer beneath the town, and by the late 2010s the water was within about a hundred and fifty feet of it. In November 2016 some sixty thousand snow geese, forced down by a storm, landed on the pit, and three to four thousand died after drinking the water. A treatment system handling seven to ten million gallons a day (the two published figures) began discharging in the autumn of 2019 and will run for as long as the pit exists. At Iron Mountain in California, an old copper and gold working discharges water measured in 1990 and 1991 at a pH of minus 3.6, the most acidic water ever recorded in the environment. The federal agency has run a treatment plant there since 1994 against a discharge it expects to last some three thousand years; a settlement in 2000 with the mine's corporate successor, valued at about a billion dollars, funds treatment for decades rather than millennia, and the public carries the gap. At Yellowknife the Giant gold mine, whose roaster produced arsenic trioxide for half a century, left 237,000 tonnes of the dust in fifteen sealed chambers seventy-five metres underground. When the owner went bankrupt in 1999 the federal government took the liability. In 2014 it approved a plan, at around a billion Canadian dollars, to freeze the chambers permanently with the refrigeration technology of an ice rink, and in 2022 it revised the estimate to 4.38 billion. By June 2026 1.58 billion had been spent; active remediation runs into the late 2030s, and long-term care is put at about ten million dollars a year. The project calls the period after the freezing "perpetual care".Two more cases show the revenue side. By the 1970s Sudbury, in Ontario, was the most visibly damaged mining landscape in the industrialised world, its hills stripped to black rock by a century of nickel smelting. Since 1978 the city and the companies have limed thousands of hectares and planted more than ten million trees, and the regreening has become a civic programme with its own budget line and its own visitors. At Mount Morgan in Queensland, the gold and copper mine whose dividends paid for William Knox D'Arcy's Persian oil concession, the beginning of BP, stopped mining in 1981 and stopped retreating its tailings in 1990, and the open cut has filled with acid water the state has managed at public cost for three decades. A private tailings retreatment, financed in September 2025 with a federal loan of up to 66 million Australian dollars and, on 15 October 2026, with a 40 million dollar prepayment from the trader Mercuria against its first 210,000 ounces, is planned to restart in the last quarter of 2026, though no start had been announced by 18 October; it already runs the site's water treatment plant and is to pay for it from the gold it recovers, while the state still funds the pit. In all seven places the mine did not close, the public accepted that it would not, and an institution was built to run it. Where that institution was allowed to sell what it managed (heat in Britain, gold at Mount Morgan, trees and visitors at Sudbury), the eternity cost began to earn..The Ruhr is not called closed. Some sixteen billion euros of endowment, seventy million cubic metres a year through the central stations, a 2025 bill of 309 million, and an annual report that reads like a utility's..III. What South Africa already pays, and calls temporary.The Witwatersrand has carried an eternity cost since 2002 and has never budgeted for one. That year the water in the flooded Western Basin reached the surface at the old Black Reef incline near Krugersdorp and began to decant, acid and iron-red and carrying uranium and sulphate, into the Tweelopiespruit and through a nature reserve towards the Cradle of Humankind. The state built an emergency neutralisation plant because nobody else would. Beneath Johannesburg and Germiston, the Central Basin was rising towards an environmental critical level set to protect the city's foundations and old shafts; a plant at Germiston, averaging seventy-two megalitres a day with a peak of eighty-four, began lifting and neutralising it in 2014. Even so, the Central Basin surfaced at Boksburg Lake, near the capped New Blue Sky incline, in May 2025 and again in May 2026, with the Germiston plant running above ninety per cent of capacity and shaft levels no longer falling in the dry season. The Eastern Basin, under Springs and Nigel, followed with a plant averaging about 106. The three are rated together at some 229 megalitres a day, more than eighty million cubic metres a year and more than the Ruhr's central stations lift, although the Western Basin plant treats only part of its decant and about twenty megalitres a day still reach the Tweelopiespruit untreated. The plants neutralise with lime, precipitate the metals as sludge and discharge the water, still salty, into rivers that feed the Vaal system the province drinks from.For fifteen years the spending has been classed as temporary. The Trans-Caledon Tunnel Authority, which builds the state's water infrastructure, was directed in 2011 to deliver a "short-term intervention", the three neutralisation plants, pending a "long-term solution" that would desalinate the water to a saleable standard, costed in 2016 at ten to twelve billion rand. The long-term solution has not been built. It has been re-scoped as a packaged reverse-osmosis proposal and re-costed, and with two-thirds of its funding unavailable it has no timeline in the department's 2026/27 performance plan. The interim spending, meanwhile, has become permanent. The operating budget for the three plants is 562 million rand in 2026/27, up from 436 million two years earlier, and the medium-term estimates take it to 728 million by 2029/30, a compound rise of about eleven per cent a year, with no revenue against it: treated water discharged into a river earns nothing, and the sludge is dumped. The state is carrying an obligation of the Ruhr foundation's kind on a departmental budget line labelled interim, with no endowment, no revenue and no name for what it is doing.The mines' own provisions run alongside the state's spending and never touch it. Just under a hundred billion rand is lodged against the rehabilitation of operating mines: about half in trust funds that regulation reserves for post-closure impacts, nearly half in bank guarantees that can be called only on a default the Minister has never declared, and less than half a per cent in cash. The Auditor-General's 2022 performance audit of derelict and ownerless mines found 6,100 sites, more than fourteen billion rand of need, a budget of 124 million and a rehabilitation rate of about two sites a year; the derelict budget for 2026/27 is 140.87 million. The three treatment plants are the most expensive line in the country's mining legacy, and none of these instruments covers them, because the water comes from mines that closed in fact before the provisioning regime existed and were never certified closed in law. Itemised, the basin's eternity is a permanent pumping obligation on the Ruhr's scale, funded year by year from the fiscus, while the endowment that could carry it sits in instruments written for a closure that will not happen..Three plants rated at more than eighty million cubic metres a year, treating water from mines that closed before the provisioning regime existed. None of the hundred billion rand of provisions reaches them..IV. What the closure model did to the people.The gold mines employed 553,000 people at the 1987 peak. The Minerals Council's latest count for gold alone, for 2022, is 93,841. More than four hundred and fifty thousand jobs left the goldfields in thirty-five years, and the towns built for them (Welkom, Virginia, Carletonville, Stilfontein, Orkney, the mine villages of the East Rand) kept the housing, hospitals and water infrastructure of a population that no longer had an employer. That is the human side of closure as the law imagines it: the company rehabilitates, certifies and leaves, and the people, who were never in the closure plan, stay. The 1987 workforce was told that the ore was uneconomic at the price, which was true, and that there was nothing further for them, which was not. Nobody told them that the alternative to a mine that closes is a mine that is run permanently. The price that made the ore uneconomic touched 252.80 dollars an ounce in July 1999; on 16 October 2026 it closed at 4,165, about 2.21 million rand a kilogram.What has happened in those towns since is the strongest argument for the change, and an uncomfortable one. The abandoned shafts of the Free State and the Klerksdorp field were reopened, informally and lethally, by men who understood better than the statute that a mine with gold in it does not close, and by the syndicates that finance, arm and buy from them. In January 2025 the Stilfontein shafts, on a working that had been under care and maintenance for years, gave up seventy-eight bodies and 246 survivors, with ninety-three dead in all on the Police Minister's count. The police operation that sealed the entrances to starve the men out is the nearest thing that mine has to a closure certificate, issued by a tactical response unit rather than a regional manager. Every derelict shaft on the register is a site the state must guard, at a cost it does not count, against the people the closure model let go, because an unemployed miner's knowledge of where the reef lies does not expire with his contract. The result is a labour market with no legal employer, working an orebody with no legal operator, in a basin the law says has closed.The second shift answers the first. Treatment plants employ people: the three basin plants carry a few hundred operators, technicians and laboratory staff between them, and a desalination stage would double that. Surface retreatment of the kind the listed companies run employs several thousand across plants and pipelines, using the skills the underground mines trained into the workforce and then discarded: pumping, milling, chemistry, tailings management. A storage plant in a shaft needs the winding, ventilation and rock-engineering trades the shaft already had. A land company needs the surveyors, civil crews and clerks of works who built the towns in the first place. None of this restores half a million jobs; the arithmetic below puts direct employment in the permanent enterprise in the low tens of thousands across the basin. What it restores is an employer whose business does not depend on the gold price, since pumping water, storing energy and selling land are not gold businesses, and a reason for the towns to keep their hospitals. The first shift was told the mine was closing. The second can be told it works for a utility, and utilities hire..The first shift was told the ore was uneconomic, which was true, and that there was nothing further, which was not. The second shift works for a utility..V. Water utility, power station, materials plant, land company.The mine that never closes has to perform four functions whatever happens, and each earns something if it is run as a business. The order matters: the first pays for the second, and the second protects the first. Together they are environmental arbitrage at the scale of a basin. The same pump, void, dump and footprint are carried as costs on the state's ledger and as assets on the enterprise's, and the difference between the two ledgers is available to whoever writes the second one.The water utility comes first because it already exists, as three plants and an interim budget. Gauteng is short of water. Rand Water abstracts up to its licence, Johannesburg loses 44.8 per cent of the water it distributes to leaks and unbilled use, and the second phase of the Lesotho Highlands scheme has slipped again, to December 2029. Against that shortage, the state lifts more than eighty million cubic metres of mine water a year, neutralises it and pours it into a river. Desalinated to potable or industrial standard, which is the long-term solution the department costed at ten to twelve billion rand and deferred, it would be about a twentieth of the 4,520 megalitres a day Rand Water supplies, sold into a system whose bulk tariff is R16.73 a kilolitre from July 2026, up ten per cent on the year, and whose industrial users in Germiston and Springs pay more for water trucked in when the pressure drops. Eighty-odd million cubic metres at half the bulk tariff is some seven hundred million rand a year, earned against a treatment cost the fiscus already pays, and that turns a permanent budget line into a permanent utility. The sludge the plants precipitate (gypsum, iron, manganese and the rare earths that come with them) is feedstock for the materials plant, treated below. Once the utility has a tariff, its balance sheet is bankable: water utilities borrow against thirty-year concessions at rates no mine has seen since the 1970s.The void is the power station. A Witwatersrand shaft offers a vertical head of a thousand to nearly four thousand metres. The water that has to be lifted for the utility can be pumped up at night on cheap power and released by day at the peak tariff. That is pumped storage, the technology Eskom runs at Drakensberg and Ingula and Cape Town at Steenbras, built in a hole that has already been dug. No underground mine anywhere yet runs a pumped-storage plant at scale. The most cited plan, at Pyhäsalmi in Finland, has no investment decision; the mine's pumps were removed in the autumn of 2025, and a development partner was named only in August 2026. The physics is settled and the heads on this basin are the deepest on earth, but no turbine yet built works a head beyond about 1,900 metres in a single stage, so a deep plant has to be built in stages. The demand is there too. Eskom's evening peak is the most expensive electricity in the country, the corporate power-purchase market has opened since the licensing threshold was lifted, and Harmony has set out a pathway to about eight hundred megawatts of its own and contracted renewable power for its mills. Put solar on the dumps (flat, fenced and close to the substations the mines built) and storage in the shafts beneath them, and the goldfield becomes a dispatchable power station on land with no other use, supplying the mines by day and the grid at the evening peak. The first plant is a pilot on one shaft; the business is the shafts behind it.The materials plant works the sludge and the sand. The treatment plants produce gypsum by the hundred thousand tonnes and dump it, while the country imports gypsum for plasterboard and cement. The iron and manganese in the same sludge can be sold at a grade a magnetic separator can reach. The rare earths, cobalt and nickel that the acid leaches from the Witwatersrand's pyrite are in the water at concentrations a resin column can recover, and pilot work on the basin's own water has demonstrated the capture. The dumps themselves, once the gold has been taken, are construction aggregate and brick feedstock: Gauteng has 374 mine-residue areas, within thirty kilometres of the largest building market on the continent. None of these lines is a gold business. The same plants, water and people run all of them, and they are the revenue a closing mine leaves on the floor.The land company is what is left when the other three have done their work: the footprint. A dump that is retreated and removed frees the ground beneath it; a dump capped and covered in solar frees the ground beside it; a treatment plant run as a utility cleans the river frontage the mines poisoned. The central Rand's mining belt is a strip the width of a city between Johannesburg's northern suburbs and its southern townships. Vacant land in Houghton, on the northern side, lists at some twenty-five times the price at Olifantsvlei in the south, and the belt is much of the reason. Every hectare moved from a mining right into a township register becomes the land company's stock, sold once into a market that has waited a century for it. In any one year land is the smallest of the four revenues; across the century it is the largest. It exists only if the other three run first, because nobody buys a plot beside a dump that is still blowing..Eighty-odd million cubic metres a year at half the bulk tariff is seven hundred million rand, earned against a cost the fiscus already pays. A budget line cannot borrow. A utility can..VI. An annuity, an endowment and the arithmetic that joins them.A mine that never closes is financed as a perpetuity, and perpetuities are the simplest instruments to value. A cost that runs forever is capitalised by dividing it by the real rate at which the money set aside for it can be invested; a revenue that runs forever is capitalised the same way; the enterprise is worth the difference. The Germans did this sum in 2006, arrived at twelve and a half to thirteen billion euros for the Ruhr's eternity, and funded it with an endowment now worth some sixteen billion at book value, from which the foundation pays the bill out of investment income without touching the principal. Done for the Witwatersrand, the sum has a result the Ruhr never had, because the principal already exists.The state's treatment cost is 562 million rand in 2026/27 and is budgeted to reach 728 million by 2029/30. Take the higher figure as the steady-state cost of neutralisation alone and capitalise it at a real rate of five per cent, a conservative long-run return for a fund invested, like the German foundation, in industrial equity and bonds, and the eternity cost of the three plants is 14.6 billion rand. Add the long-term desalination solution at the department's own ten to twelve billion, and the whole of the basin's water eternity comes to twenty-five to twenty-seven billion rand. That is a quarter of the 99.6 billion rand lodged in rehabilitation provisions against the country's operating mines, instruments that fund neither the derelict sites nor the plants, because regulation reserves them for a post-closure phase the certificate regime does not reach. Move a quarter of the provisions into an endowment charged with the basin's water in perpetuity, on a statutory basis and against release of the holders' matching liability, and the interim budget line leaves the fiscus for good. The Germans had to build their endowment from a share portfolio. South Africa has already collected its own and is holding it in the wrong place.The revenues go on the other side. The water utility at seven hundred million rand a year, capitalised at the same rate, is worth fourteen billion. The materials plant, at a modest hundred and fifty million a year from gypsum, metals and aggregate, is worth three billion. The power station on a single shaft is a pilot earning tens of millions; across the basin, at a few hundred megawatts of solar and storage selling into Eskom's peak, it could earn a billion a year and be worth twenty billion, but it is carried here at pilot value until the first shaft has run. The land company sells its stock once rather than forever. At township prices the central Rand belt is worth more than the rest of the enterprise combined, but it is realised over decades and is the last revenue to arrive. Let the endowment carry the perpetual costs, count the land at nothing until the dumps are gone, and the enterprise is worth about seventeen billion rand from water and materials alone, rising into the tens of billions once the power line is proven. It runs on an endowment the country already holds and replaces a budget line the country already pays. The closure model has kept that business out of sight for fifteen years.The revenue side depends on tariffs for water, power and gypsum, set by regulators and markets that have never been asked to price this enterprise. The figures apply prevailing bulk and peak prices to volumes the plants already handle; they are not a forecast of what a concession would negotiate. The endowment transfer is a legal act, and the statute that would permit it does not yet exist; the sum assumes a clause that has been drafted but not enacted. Neither caveat changes the sign of the result. The eternity cost is real and already paid, the eternity revenue is real and currently thrown away, and the endowment is real and currently frozen. The arithmetic connects them. The law keeps them apart..Capitalised at five per cent, the basin's water eternity is twenty-five to twenty-seven billion rand, about a quarter of the provisions the country already holds. The Germans built their endowment from a share portfolio. South Africa has already collected its own..VII. A Basin Authority, and who works for it.Neither a department nor a mining company can run the mine that never closes, and the examples abroad show why. A department budgets one year at a time and cannot hold an endowment. A mining company answers to a share price driven by the gold price and will sell or shed its perpetual obligations when the price turns, as this basin's mining houses did in the 1990s. Every jurisdiction that has accepted eternity has built a third kind of institution: a foundation with an endowment and a statutory purpose in the Ruhr, a public authority with a grant and a licence in Britain, a trust under a federal record of decision in Montana. The Witwatersrand needs its own. The proposal is a Basin Authority combining the two European forms: a statutory body holding the water and land licences for the three basins, funded by an endowment carved from the provisions on the statutory basis described above, licensed to sell water, power, materials and land, reporting to Parliament each year as a utility does, and taking over the mines' remaining liabilities pro tanto against certification of what has actually been remediated. The German foundation is the model for the money and the British authority the model for the operations. The closure certificate, which neither has ever needed, gives way to the transfer.The Authority's workforce is the second shift, and the numbers are simple. The three plants and a desalination stage employ in the high hundreds, and a materials plant on the sludge and aggregate a few hundred more. Solar on the dumps employs thousands in construction and hundreds in operation; the first shaft storage plant needs a few hundred, and a basin programme several thousand. The land company employs surveyors, civil crews and clerks in the low thousands over its decades. Add the surface retreatment industry, which already employs several thousand and whose plants are the Authority's natural contractors, and direct employment in the permanent enterprise is fifteen to twenty-five thousand across the basin: about a twentieth of the 1987 workforce and a quarter of the latest count, in jobs that do not end when gold falls, in towns that already have the housing. It will not restore the goldfields. It gives them an employer with no closure date, and for Carletonville and Springs that decides whether they remain towns.The model can also be exported. The Witwatersrand will not be the only goldfield that needs it; it is the first to have every element on one property. Ghana's Obuasi, Western Australia's Kalgoorlie, Nevada's Carlin trend, the Abitibi in Quebec and Ontario, and South Africa's own Free State field around Welkom all combine old workings, surface tailings by the hundred million tonnes, treatment obligations with no end date and workforces being told the ore is running out. Each will need an authority, an endowment, a set of tariffs and a second shift, and each will buy the method from whoever runs it first. The Dutch sell polder engineering to the world because they had to build it at home, and other coalfields send their engineers to study Britain's mine-water heat. A South African Basin Authority that had run three basins' water as a utility, put storage in a shaft, sold gypsum from a treatment plant and returned a mining belt to a city would hold the operating manual for the largest unwritten business in mining: what to do with a goldfield after the gold. That is the gold standard this country can set for the world, an operating standard rather than a monetary one..Every jurisdiction that has accepted eternity built a third kind of institution: an authority with an endowment and a licence to sell. The Witwatersrand needs one, and could be the first to write the manual for a goldfield after gold..VIII. What the sceptic rightly takes.The bear's first blow is the endowment, and it lands: the provisions are not the state's to move. Companies lodged them against their own obligations, and a statute pooling a quarter of them into a Basin Authority would be challenged by every holder as expropriation without compensation. Conceded as a matter of law. The structure answers it: the transfer is made against release of the holder's matching liability, which the holder cannot otherwise escape, and a holder released from a perpetual obligation for a quarter of its provision has been well compensated. The mining houses will still litigate; the litigation is carried here as two years of delay with no change in the outcome. The second blow is the tariffs. Rand Water does not want eighty million cubic metres of mine water in its system, Eskom does not want another dispatchable supplier at its peak, and the regulator that would set a mine-water tariff has never set one. Conceded as the institutional position, and answered by the shortage: a province trucking water to its industrial users and paying the country's highest tariff at its evening peak is in no position to refuse either, and a corporate market for water will open the day a concession exists, as one opened for power. Then the pilots. No shaft storage plant runs anywhere, the desalination solution has been deferred for a decade, and the power and water revenues rest on plant that has not been built. Conceded. The power line is carried at pilot value in the table and the water line waits on the desalination stage; the neutralisation cost, the endowment and the gypsum are not pilots, and on those alone the sum is positive.Then the state's record. The Germans had coal money and a state that pays its bills; a South African Authority would be a state-owned enterprise in a country whose record with those includes Eskom, Transnet and the Post Office. Conceded as the heaviest blow of the set. The answer lies in the design rather than in trust: the foundation model works because the endowment is ring-fenced from the fiscus and the operations are contracted to private operators (the surface retreatment companies, water concessionaires, solar developers) against tariffs, with the Authority as licensor and trustee rather than operator. The Ruhr foundation keeps a small staff; a company does the pumping. Britain's authority has just let a £203 million contract for a water company to run its treatment schemes. Last, the sceptic says the whole case argues for never letting a mine finish, which suits operators who would rather run forever than pay to close. Partly conceded: the operators do prefer it. But they have been running forever already, on care and maintenance, with the public paying. The case does not create the eternity. It names it, funds it and puts revenue against it..The bear takes the litigation, the tariffs, the pilots and the state's record with its own enterprises. He leaves the water, which is rising whether or not anyone agrees with him..VIII. What the sceptic rightly takes.The bear's first blow is the endowment, and it lands: the provisions are not the state's to move. Companies lodged them against their own obligations, and a statute pooling a quarter of them into a Basin Authority would be challenged by every holder as expropriation without compensation. Conceded as a matter of law. The structure answers it: the transfer is made against release of the holder's matching liability, which the holder cannot otherwise escape, and a holder released from a perpetual obligation for a quarter of its provision has been well compensated. The mining houses will still litigate; the litigation is carried here as two years of delay with no change in the outcome. The second blow is the tariffs. Rand Water does not want eighty million cubic metres of mine water in its system, Eskom does not want another dispatchable supplier at its peak, and the regulator that would set a mine-water tariff has never set one. Conceded as the institutional position, and answered by the shortage: a province trucking water to its industrial users and paying the country's highest tariff at its evening peak is in no position to refuse either, and a corporate market for water will open the day a concession exists, as one opened for power. Then the pilots. No shaft storage plant runs anywhere, the desalination solution has been deferred for a decade, and the power and water revenues rest on plant that has not been built. Conceded. The power line is carried at pilot value in the table and the water line waits on the desalination stage; the neutralisation cost, the endowment and the gypsum are not pilots, and on those alone the sum is positive.Then the state's record. The Germans had coal money and a state that pays its bills; a South African Authority would be a state-owned enterprise in a country whose record with those includes Eskom, Transnet and the Post Office. Conceded as the heaviest blow of the set. The answer lies in the design rather than in trust: the foundation model works because the endowment is ring-fenced from the fiscus and the operations are contracted to private operators (the surface retreatment companies, water concessionaires, solar developers) against tariffs, with the Authority as licensor and trustee rather than operator. The Ruhr foundation keeps a small staff; a company does the pumping. Britain's authority has just let a £203 million contract for a water company to run its treatment schemes. Last, the sceptic says the whole case argues for never letting a mine finish, which suits operators who would rather run forever than pay to close. Partly conceded: the operators do prefer it. But they have been running forever already, on care and maintenance, with the public paying. The case does not create the eternity. It names it, funds it and puts revenue against it..The bear takes the litigation, the tariffs, the pilots and the state's record with its own enterprises. He leaves the water, which is rising whether or not anyone agrees with him..IX. Sequencing the enterprise.The programme follows the order in which the four businesses fund one another. The tariff comes first. The department applies to the water regulator for a bulk tariff for treated mine water and directs Rand Water to take it, so that the three existing plants earn from the day the determination is published. This needs no statute and no capital, and it turns the interim budget line into a utility's revenue within a year. Then the endowment: the provisions statute is amended to allow a holder's provision to be transferred into a Basin Authority against release of liability, and the Authority is constituted with the three basins' licences and a first tranche of endowment; two years, most of it litigation. Alongside it, the materials plant: the gypsum and metals are put out to a private operator on concession and aggregate from the retreated dumps is licensed for sale, so that the sludge stops being dumped; a year. Then the pilots: one desalination module on the Central Basin plant, one solar field on one dump and one storage plant in one shaft, each contracted on a tariff and each reported publicly; three to five years. And then the land: the first footprint cleared by retreatment is transferred through the Authority into a township register and sold, with the proceeds going to the endowment, so that the land company makes its first sale within the decade. After the first step the sequence pays for itself, and the first step is a tariff determination. The Germans took eleven years from statute to first payment. The basin has already had fifteen years of interim..X. The word that costs the most.Closure meetings do not close mines. They produce a form of words that lets everyone in the room leave (the company to its shareholders, the regulator to its register, the liquidator to his fee) while the water rises through the workings beneath them. The form of words is "closure", the most expensive word in South African mining: it describes an event that has not happened on this basin since the first shaft was sunk, and it lets everyone act as though it will. A permanent water enterprise cannot be walked away from. It has to be run, and whatever has to be run gets budgets, tariffs, an endowment and staff.The world has already conceded the point in the form that counts, money set aside. Germany has some sixteen billion euros in a foundation. Britain has an authority that treats more water than the Rand pumps. Canada has budgeted 4.38 billion dollars for a mine it will keep frozen indefinitely. Each reached eternity through a disaster and built the institution afterwards. The Witwatersrand has had its disasters, among them the 2002 decant into a nature reserve and the bodies in the Stilfontein shafts, and has built three plants and an interim budget line. It holds a hundred billion rand of provisions in the wrong place, eighty million cubic metres of water a year that it pours into rivers, 374 mine-residue areas beside the largest building market in Africa, the deepest heads on earth, a workforce that knows where the reef is, and a closure regime that has certified twenty-seven large mines. Everything the mine that never closes needs is already on the property except the word, and the word is the cheapest thing on the basin to change.So change it. The goldfields are not closing and never were. They are a permanent enterprise that has been run for a generation as a funeral, and what separates the two is a tariff, an endowment, an authority and a sentence in the statute that says so. The first goldfield to make that change will be the one the others buy the manual from. I would rather it were this one..THE MINE DOES NOT CLOSE. THE WATER RISES ON ITS OWN TIMETABLE, THE DUMPS SIT WHERE THEY WERE POURED, AND THE WORLD'S OLDER FIELDS HAVE STOPPED PRETENDING OTHERWISE AND BUILT FOUNDATIONS.A water utility, a power station, a materials plant and a land company, run on an endowment the country already holds, by the second shift. The only thing still missing is the word..Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox every morning on weekdays. Register here.Support South Africa's bastion of independent journalism, offering balanced insights on investments, business, and the political economy, by joining BizNews Premium. Register here.If you prefer WhatsApp for updates, sign up to the BizNews channel here.

The mine that never closes: Joburg’s acid water is a R17bn business we keep paying to throw away

Related Stories

R100 in 1940, R10 million by 2025: the investment risk no spreadsheet captures

Duarte da Silva: The city on the dumps

Duarte da Silva: Joburg's mine dumps hold $215bn of gold, two-thirds is on nobody's books

Dr Duarte da Silva on environmental arbitrage: The asset in the flood