By NIALL MCGEE
Saturday, July 13, 2019
Even though the price of gold bullion had tumbled by more than a third from its 2011 peak, and many of his competitors were struggling, his company was defying the odds.
Guyana Goldfields Inc. had managed to raise US$700-million from investors and put a high-grade gold mine into production in early 2016.
Mr. Caldwell, an avuncular mining engineer with a soothing tone, was happy to promote the company's Aurora mine, located in a remote Guyanese rain forest, as a cash machine.
Indeed, at the prevailing gold price of US$1,200 an ounce, Guyana looked like a surefire winner.
"A little less than US$800 an ounce [cost], US$400 an ounce margin," he said during a segment on Business News Network (BNN). "Pretty easy to figure out how we're going to do."
The company's share price soared as it ramped up production, and its market capitalization crested above $1.5-billion.
But last October, seemingly out of nowhere, the wheels came off. Guyana shed half its stock-market value in one trading session after the company raised doubts about the geology at Aurora.
A technical report, upon which the mine was built, had vastly overestimated the amount and grade of gold at Aurora. This past March, Guyana cut its reserves by more than 40 per cent, after releasing an updated study on the mine.
Guyana's chairman, René Marion, later admitted in an interview that some 1.5 million ounces of gold assumed by Guyana to be in the ground were "never there."
Ten months on, Guyana's share price is down 87 per cent from its peak. Its founder and almost its entire legacy management and board of directors have left. Mr. Caldwell will step down once a replacement is found. Nobody is sure whether the company can weather the crisis.
The meltdown at Guyana's isn't a one-off. Over the past few years, several other mining companies have shocked the market with nasty technical surprises.
Vancouver-based Pretium Resources Inc. has seen its share price whipsawed on multiple occasions by geological setbacks at its erratic Brucejack deposit in British Columbia; Toronto-based New Gold Inc. saw the economics of its Rainy River mine in northwest Ontario go up in smoke last year after it fell short on grade; and shareholders in Rubicon Minerals Inc. were almost completely wiped out after its deposit in Ontario's Red Lake camp turned out to be not mineable at all.
Virtually all of the incidents are occurring at technically demanding ore bodies that require exhaustive study.
While seniors, such as Goldcorp Inc. (now owned by Newmont Mining Corp.), haven't been immune to technical blunders, this is mostly a small company problem.
Many juniors have little or no experience in building mines and lack the technical talent that might head off calamities in advance.
Small mining companies rely heavily on external consulting firms that prepare resource models. The bigger companies have reams of inhouse talent - geologists, metallurgists and engineers - who vet the work of consultants. But juniors often don't have the same level of expertise to be able to push back if something seems off.
"[Smaller gold companies] don't have the human expertise to be able to steer away from those disasters. They don't have the technical bench strength.
They don't have people that can look at it and say, 'hey, this is wrong.' "said Andrew Kaip, mining analyst with BMO Nesbitt Burns Inc.
"They're reliant on external advice and that can be flawed. It can have wildly bad outcomes."
The industry's recent flops also raise the issue of accountability when things go wrong. It's very easy to blame the consultant when the mine plan falls apart, but the management and boards of troubled companies, often responsible for making questionable decisions, are no angels either.
"In order for these things to collapse, half a dozen constituents of people have to not do their jobs," said John Tumazos, chief executive of New Jerseybased Very Independent Research.
"And the reason they don't do their jobs is that no one wants to kill the golden goose, the gravy train. Even when the project sucks." Compared with almost any other mineral, gold is a geological nightmare - harder to find, harder to model and harder to mine. There is no MRI machine for finding gold. Prospectors still have to identify a promising property, drill test holes, send samples to a lab for analysis and cross their fingers.
Even if you find gold, invariably there will be hardly any of it in the ore. The term "high grade" is actually misleading. Eight grams of gold in a tonne of rock is considered high grade. That's eight parts per million. Low grade is one part per million - a grain of salt in a giant bag of Doritos.
The gold industry is perhaps unrivalled in its wastefulness. A producer has to dig up about 20 tonnes of ore for enough gold to make a wedding ring.
Sometimes gold plays nice, occurring as a fine powdery-like substance in rock, with consistent grades throughout the entire ore body - specks of salt uniformly spread across the Doritos. If drill samples confirm that consistency over and over, such deposits can be fairly straightforward to model.
But gold deposits can also be "nuggety" - low grade in most spots, but with the occasional highgrade cluster. And often there is no discernible pattern - like finding a random pretzel in the Doritos.
These ore bodies are among the toughest to model, because geologists can't be entirely sure whether the high grade is a statistical fluke, or a pattern across the entire deposit.
Since it's financially feasible to drill only a tiny proportion of any potential gold deposit, experts have to take sample data and try to figure out what the rest holds.
Correctly modelling a mine, based on a sample that is perhaps only 0.13 per cent of the total mineralized rock, requires immense skill. Such work is typically done by a select group of independent mining consultants. Combining geological field work, and a branch of mathematics called geostatistics, the job is a blend of art, science and luck.
In 2012, SRK Consulting (Canada) Inc. produced a model for Guyana's Aurora property. Like all gold deposits, Aurora had its charms and its challenges.
Early drilling revealed it was a little nuggety.
One way geologists deal with the presence of high-grade gold in what appears to be a mostly lower-grade deposit is to assume it's an anomaly. In constructing a geological model, consultants will routinely disregard high-grade drill samples above a certain level.
This practice, known as "capping," is supposed to prevent consultants from overestimating the overall average grade. But here's the rub. If a deposit is capped too low, that can kill the financial case for building the mine.
In 2012, SRK capped a section of Aurora, called Rory's Knoll, at 80 grams of gold per tonne. That meant Guyana could expect to find a certain amount of high-grade ore when it mined the area.
But last year, as it mined Rory's Knoll, the high grade simply wasn't there.
"We weren't seeing the grade that we thought we would, based on the original 2012 model," Guyana's CEO, Mr. Caldwell, told The Globe and Mail earlier this year.
Guyana's chairman, Mr. Marion, pointed the finger squarely at SRK. The consultant was "very aggressive" in capping the deposit, he said.
Late last year, Guyana asked another consultant, Roscoe Postle Associates (RPA Inc.), to redo the technical report on Aurora from scratch. In its report issued in March, RPA capped Rory's Knoll at just 35 grams per tonne. Guyana's current management team maintains that RPA's capping is much more appropriate.
But SRK isn't taking any of this on the chin. The consultancy points the finger back at Guyana. After an internal review earlier this year, SRK concluded that its 2012 report on Aurora was technically sound based on data available at the time.
Adam Nott, general counsel with SRK, disputes any notion that the consultancy was aggressive in its modelling. The report was produced when Aurora was at an early stage, and was never meant to be relied upon for the construction of the mine, which came some four years later.
SRK would have had discussions with Guyana about the need to update the model and get lots more data before building Aurora. That would have required more drilling and the outlay of significant amounts of additional capital from Guyana. "For whatever reasons, internal to Guyana Gold, that update wasn't done until 2018, when new management came in," Mr. Nott said.
If SRK had access to the same data RPA did in 2018, including three years of actual mining, the consultancy "probably would have come to different results," he added. Of course, any allegation that a consultant was too aggressive in its interpretation of the geology of a deposit hits a nerve in the Canadian mining industry.
Consultants are supposed to provide an unbiased and impartial view of an orebody. But the reality is more nuanced.
"Some [consultants] look at deposits and imagine all kinds of good things happening, and others, and we're among them, try to be more realistic," said Graham Farquharson, veteran mining consultant with Strathcona Mineral Services Ltd. in Toronto.
(In the late 1990s, when doubts arose about Bre-X Minerals Ltd.'s 70-million-ounce gold find, the industry turned to Strathcona to investigate. Mr. Farquharson himself later made what he calls the "sixbillion-dollar phone call," to Bre-X's board, definitively declaring Busang a hoax.)
There is also an inherent conflict of interest. Because consultants are paid by the mining companies, they face financial pressure to be positive. Having a negative stand on a project, even if it's spot on, can result in the consultant getting canned.
"It's a very hard battle telling your client that we think they need to go back to the drawing board," SRK's Mr. Nott says. "Especially when the clients know there are other consultants who are willing to use those [data points] and say that's within a reasonable range."
Mr. Nott added that SRK has lost work to rival consultants who were willing to provide a more bullish outlook on a deposit.
The technical reports themselves are also heavily influenced by clients. Consultants and management go back and forth on many issues, such as appropriate capping levels, the distance between drill holes and what long-term gold prices to assume in projecting returns.
Sometimes technical reports aren't as thorough as they could be, either, and that is often because of money. A client may not want to spend more on drilling and will choose to live with the added risk that entails. "SRK, in a lot of ways, is driven by what the client is willing to pay for, and what the client feels its risk-reward balance is," Mr. Nott said.
Most of the time, these kinds of behind-closeddoors discussions between consultants and mining companies are kept secret. But once in a while they become public. High up in the mountains of northwest British Columbia, Pretium Resources' Brucejack property was an enigma from the get-go. Early work in 2012 pointed to an extremely high-grade gold deposit. Some drill holes came back with as much as 41,000 grams of gold per tonne.
Despite extensive drilling, Brucejack was incredibly difficult to pin down. "You could come back with one sample that would have spectacular results and then 10 samples all around it that had nothing," said Mr. Farquharson, whose consultancy did a bulk sample on the deposit.
In 2013, Pretium shares shed half their value within two weeks after it revealed that Strathcona's analysis didn't square with a far more optimistic study by an Australian firm, Snowden Mining Industry Consultants. Strathcona insisted that Pretium disclose the discrepancy to its investors, then resigned in the aftermath.
Pretium, in turn, stuck with Snowden and trashed Strathcona's work as subpar.
Snowden felt Brucejack had similarities with deposits in the South Pacific with similarly eccentric geology. The consultant used a mathematical model called multiple indicator kriging (MIK) to predict the grade and location of the high-grade gold.
MIK is well suited to "mosaic" deposits such as those at Brucejack, where extremely high-grade gold occurs next to low grade, or even no grade, said international geologist Ashley Brown, who's now based in Kazakhstan. But MIK is extremely challenging. "The implementation of MIK is very difficult," he said. "It's easy to screw up."
What struck Mr. Brown as odd about Brucejack is that Snowden decided against capping the grade. By forgoing capping, SRK allowed the pockets of high-grade gold samples to strongly influence the average grade for the entire deposit. Brucejack's reserve grade was pegged at 14.4 grams per tonne, which made it among the highest-grade gold mines in North America.
Snowden's approach didn't sit well with Haywood Securities Inc.
analyst Kerry Smith, either. A former mining engineer, he's seen his fair share of geological goofs in his almost 40 years in the business.
About four years ago, Mr. Smith attended an information session with Snowden about Brucejack.
"Snowden spent the whole day trying to rationalize why they should model it the way they did, which was basically to model those high-grade numbers and use them to influence the ore around it," Mr.
Smith said. "I came away thinking 'I wouldn't do that. That makes no sense,' because these numbers are not going to have any continuity."
Mr. Smith was right to be wary. In January of last year, Pretium said Brucejack's grade was only corresponding 75 per cent to Snowden's model. The stock lost more than a quarter of its value.
"The high-grade mineralization was in narrower corridors than originally thought," Pretium CEO Joseph Ovsenek said in an interview.
Earlier this year, after undertaking a review of Brucejack, Pretium cut the mine's grade to 12.6 grams per tonne, increased its cost projections by 12 per cent and reduced its expected mine life by four years.
Snowden declined an interview request from The Globe. Ivor Jones, who had responsibility for the technical report on Brucejack, also declined to comment beyond saying, "It is easy to criticize other people's work. Especially something as challenging as Brucejack."
Pretium's CEO meantime refuses to play the blame game. Mr. Ovsenek instead points to the baffling geology, calling Brucejack a "beast."
"I can tell you from talking to a lot of people in the industry and others, there is no orebody like ours out there," he said. "I challenge anyone to say that they could have done better."
While Pretium has been wounded, even with a materially lower grade, Brucejack is still plenty profitable. Over the past 18 months, amid a recovery in bullion prices, the company's share price has regained most of its losses since early 2018. The trouble for many other juniors is that they don't have deposits with grades that come anywhere close to Pretium's Brucejack, or the financial cushion to recover from geological setbacks.
It is possible for a gold producer to make lots of money from a low-grade mine if costs are kept in check and the geology is sound. But it's crucial that there be a margin for error built in, in case things go wrong. Otherwise, a small slip can spell big trouble.
New Gold Inc.'s Rainy River mine is exhibit A.
Midway through 2018, less than a year into production, New Gold said it was seeing a roughly 11per-cent shortfall in the grade at Rainy River. With that, the mine's profit margin vanished.
New Gold also made a basic engineering error in designing the tailings dam at Rainy River and had to build a drastically strengthened structure. The episode blew its capital budget to smithereens.
New Gold now loses hundreds of dollars on every ounce of gold it produces at Rainy River, its debt load is US$780-million and it isn't expected to produce any free cash flow until 2021.
"Some of these things should just never ever get built. That mine was one of them," said Rob Cohen, manager of the Dynamic Precious Metals Fund.
If Rainy River's economics were so dicey, why did it get built? A close reading of the mine's technical report would have shown how thin the margins were. The projected average grade was just 1.12 grams per tonne and the return on mine was forecast at 11 per cent. But technical reports for the most part are impenetrable, and few investors are skilled enough to understand them. Reports can be penned by as many as a dozen authors, run 700 pages or more and are laced with terms such as "kriging" and "variogram."
Here's a passage from New Gold's 713-page report in 2014, describing Rainy River: "The volcanic rocks have been intruded by a wide variety of plutonic rocks including synvolcanic tonalite-diorite-granodiorite batholiths, younger granodiorite batholiths, sanukitoid monzodiorite intrusions and monzogranite batholiths and plutons."
The seeds of some mining disasters are buried in technical reports, there for the world to find them before a cent is spent on a mine. But these reports are written by geeks for geeks. The common investor doesn't stand a chance. New Gold declined an interview request for this story. In addition to technical challenges, however, an old chestnut plays a role in some, if not all, of these cautionary tales. The gold industry is renowned for its culture of exaggeration, hype and promotion, and even the smartest among us can fall victim.
Gold mines are almost always built off a feasibility study (FS), which entails extensive drilling to confirm the existence of gold.
But Rubicon Minerals built its Phoenix underground mine in Northern Ontario off a preliminary economic assessment - a much more rudimentary early stage study.
Despite the obviously materially higher risk profile, Rubicon raised more than half a billion dollars from investors. It even attracted one of Canada's most sophisticated institutional money managers: The Canada Pension Plan Investment Board put $50-million into the miner.
In late 2015, mere months after starting production at Phoenix, Rubicon suddenly halted production, citing complications with the geology. Over time, it emerged that Rubicon hadn't done nearly enough drilling to confirm the gold was actually in the ground. The company, which at one point was worth $1.2-billion, never recovered.
Shareholders lost almost everything. In this case, they should have known better.
While most of these catastrophes involve small mining companies, there are a few outliers in the junior and intermediate sectors that have demonstrated both geological prowess and sound judgment.
In 2011, junior gold company Osisko Mining Inc. put what is now Canada's biggest gold mine into production. While the Canadian Malartic mine in Quebec is low grade, it is very profitable.
The technical team behind Osisko did their homework, including drilling the deposit like crazy. Two of the company's top three executives were geologists and the other was a mining engineer. (Osisko was acquired by Agnico Eagle Mines Ltd. and Yamana Gold Inc. for $3.9-billion in 2014.)
Vancouver-based B2Gold Corp. is another example. Founded in 2007, the company acquired, developed and built Fekola in Mali, now one of the world's most profitable gold mines. Instead of outsourcing mine construction to external engineering firms, as is industry practice, B2 builds its own mines with a tight-knit staff CEO Clive Johnson has worked with for decades.
But of all of Canada's gold miners, Toronto-based senior Agnico Eagle Mines probably has the strongest reputation for technical excellence over the long term. Over more than 60 years, the company has never experienced a serious geology mistake, despite dealing with many technically demanding orebodies.
To access ore at its LaRonde mine in Quebec, the company mines three kilometres underground. Agnico built two mines in Nunavut, despite having no access to power, or roads, and operating in a brutally harsh climate. In Finland, the company deals with complex metallurgy.
Agnico is known for its conservative approach. It's stacked with technical staff, and renowned for its airtight chain of command that starts at the top, with CEO Sean Boyd, and extends through the entire organization.
"Sean Boyd knows how to delegate responsibility. He understands the importance of his technical guys, understands about getting the mine engineers talking to the metallurgist, talking to the electricians.
Everyone," Dynamic's Mr. Cohen said.
"That's what brings success to these projects. Having a sharp pencil and being no nonsense." A decade ago, Pretium, Guyana Goldfields and New Gold might well have been bought by a bigger miner, well before major problems occurred. Within a technically stronger and better capitalized senior, basic geology mistakes could have been averted or minimized.
But in 2012, the mergers and acquisitions (M&A) market in mining went into a deep freeze.
A vicious gold bear market in the first half of this decade, and terribly timed acquisitions during the most recent bull market, forced the majors onto the sidelines.
Smaller companies have been forced to hang around as stand alones longer than before. That has forced many of them into the uncomfortable terrain of building mines by themselves - often for the first time.
The risk of something going wrong was always going to be higher.
While M&A has taken off again in a limited way among the seniors, for the most part it's crickets further down the ladder. If that dynamic doesn't change, more mines will invariably be built by the tenderfoot, and investors will be left to wonder where the next geological shock lies.
Before being acquired by Agnico Eagle Mines and Yamana Gold, Osisko Mining put the country's biggest gold mine, the Canadian Malartic, into production in Quebec.
Despite being faced with challenges from the start, Pretium Resources' Brucejack mining operation in British Columbia is still profitable.