Lloyd C. Blankfein, chairman and CEO of Goldman Sachs Group, center, walks with Gary D. Cohn, Goldman's president and chief operating officer.

A book hitting the stores Tuesday by CNBC's Kate Kelly takes a look inside the powerful group that runs the global commodities trade. The following is an excerpt.

During the summer of 2011, officials at the London Metal Exchange got an unexpected complaint from The Coca-Cola Company. The amount of physical aluminum in storage was piling up, said a representative of the soda maker, and, along with it, so was the expense of buying the metal for beverage containers.

The culprit, as Coke saw it, wasn't simple supply and demand—in fact, there was plenty of aluminum sitting in warehouses. It was the shrewd tactics of Goldman Sachs, the bank that owned a network of metal-storage facilities in the Detroit vicinity, where waiting times for extracting aluminum were longer than ever. Every day those metal bars sat idle, Goldman's warehouse company effectively drove up the premium amount that aluminum producers could charge for delivering supplies to beverage-packaging factories, a cost that amplified the expense of the actual metal and, thus, the prices Coke and others paid for soda cans.

"The situation has been organized artificially to drive premiums up," said Dave Smith, Coke's head of strategic procurement, at an industry conference that June. "It takes two weeks to put aluminum in, and six months to get it out."

Coke wasn't alone in its frustration. More than a year earlier, eight players had complained privately to the LME, the obscure London metals bourse that set the benchmark price for aluminum, zinc, copper, and other important nonprecious, or base, metals that were key in manufacturing. The U.S. warehousing system that the exchange oversaw was so inefficient that it was hurting corporate profits, they had argued.

The Midwest premium, the regional U.S. rate for getting metal from a seller to a buyer, was a cost imposed on Coke and other manufacturers in addition to the cash, or spot price of aluminum, which was by then bouncing back from its lowest levels in some time. Aluminum prices had increased 13 percent since the beginning of 2010, when Goldman had paid half a billion dollars to acquire Metro International Trade Services, the metal storage business it now owned in Romulus, Michigan. Bought relatively cheap at a time when commodity prices were low, Goldman took on Metro as a way to broaden its suite of physical commodity holdings, which had become an important complement to its derivatives trading in London and the U.S.

In addition to the revenue Metro offered, it presented Goldman with the possibility of a free look at what was happening on the physical side of commodities through ground-level operations. Tweaking existing contract trades in a commodity based on feedback from colleagues who worked in the physical markets was by then commonplace in banking. "We had pipeline capacity all over the place and we would call up and say, 'How's gas flowing this morning?'" remembers a trading manager who worked for years at one of Goldman's competitors. If flows were weak, he added, "We'd go, 'Oh, we're not going to get the pop we thought, so let's reposition.' " Still, Goldman spokespeople consistently denied that employees familiar with Metro's business operations shared information about it with internal traders.

Goldman wasn't the only one dipping into the physical commodity world, a cornerstone of companies like Glencore International, the Swiss trading firm originally known as Marc Rich & Co., that considered owning hard assets to be a critical aspect of trading commodities. JPMorgan had recently bought a metal-storage firm called Henry Bath & Son, among other physical assets. Broadly speaking, the bet was that metal traders would stock up on physical product while prices were relatively low and hold their stocks until the market improved, generating steady rental incomes for the warehouse owners in the process.

The new metal-warehouse owners quickly proved right—and had huge stacks of metal bars to prove it. The nine hundred thousand tons of aluminum Metro stored in its Detroit facilities when Goldman bought it in 2010 were growing daily. Meanwhile, incentive fees Metro paid to hedge funds, physical commodity traders and other tenants—usually one or two hundred dollars per ton each year—encouraged them to leave the metal there for longer periods.

Cohn's winning bet

For Goldman Sachs's president, Gary Cohn, the aluminum warehousing trade was a bit of déjà vu. Two decades before, in the early 1990s, he had moved to London to help expand Goldman's commodities business, a onetime New Orleans coffee trader that was known as J. Aron & Company. While in London, Cohn stumbled upon a lucrative aluminum trade that involved storing huge amounts of the metal until prices in the market allowed him to sell at a profit.

Aluminum, which was a popular futures contract to trade, was at the time in deep contango, a situation in which its future price was far higher than its cash price. This meant that investors in the Goldman Sachs Commodity Index — the commodities version of the Standard & Poor's 500-stock index, which had exposure to aluminum and other major futures contracts—were constantly paying extra money simply to maintain their aluminum exposure as they replaced their existing near-term contracts with new ones each month.

But every futures contract had an expiration date, and as that date neared, its holder had a choice: either take physical delivery of the aluminum, or roll the contract forward and replace it with a similar futures contract dated farther out.

Cohn realized it would be cheaper to take delivery of the aluminum than roll a futures contract forward. But doing so would mean breaking a long-held tradition of leaving the logistics of transporting bars of metal and barrels of oil to the professionals who actually used them. Sourcing and shipping a commodity like crude oil was elaborate enough that many futures traders would never bother with it. And storing metal until somebody else wanted to turn it into cans or cars was also a highly nuanced endeavor with its own rulebook.

Cohn dove into the world of metal storage and financing, visiting warehouses and researching metal insurance. "I spent a month sort of due-diligencing a trade that took me thirty seconds to figure out," he says.

Convinced of the potential profits, he took the idea to one of J. Aron's senior managers, Lloyd Blankfein. Blankfein, a former gold salesman, was dubious.

Cohn was insistent. He had thought through every issue, he told Blankfein, and this was a "riskless" opportunity. Cohn would hedge every potential problem he could find—interest rate movements, currency movements, anything that could hurt the trade's performance—and would definitely make money. Blankfein told him to run an experiment using no more than $300 million, a substantial portion of the firm's $4 billion in capital during that period.

The following Monday, Cohn bought $100 million worth of physical aluminum—about a hundred thousand tons. Even with all the homework he had done, and despite his boasts to Blankfein, he honestly expected the potential profits to vanish as a result of a sudden physical price drop, a surprise hike in storage fees, or some other unforeseen complication. But the following day, nothing had changed much, so he bought another $100 million of the metal. The next day, he bought another $50 million. By Friday, when the spread, or difference between what he would have spent to roll the contract forward and what he was actually spending to store the metal, was still intact, he knew he was on to something.

Cohn continued taking physical delivery of aluminum, stockpiling it in storage facilities in the Dutch city of Rotterdam. He also began calling the firm's aluminum-producer clients, seeing if they wanted to sell down their physical inventories; many were happy to oblige. Goldman Sachs was relieving them of excess metal and helping them to free up cash without tipping off the rest of the market. Meanwhile, its physical aluminum holdings offset the futures it sold in the markets, giving Goldman what amounted to a neutral position.

For more than two years, Goldman built enormous caches of aluminum. Its hoard helped prompt the biggest aluminum price rally the market had seen since 1988. During the years when Cohn was purchasing the commodity, physical market prices nearly doubled—amplifying the value of J. Aron's stockpiles and creating over half a billion dollars in profit to Goldman over several years.

Late in 1994, a few months after his thirty-fourth birthday, Cohn made partner, joining the ruling 2 percent echelon of a 9,600-employee firm. Around the same time, he also received a strange complaint: the metal masses in Rotterdam had grown so extensive that their reflection of the sun was creating confusion for local air-traffic controllers. Airport officials asked Goldman if it could throw a tarp over its aluminum stash to make navigating the local skies a little easier.

Complaints spur investigations

By 2010 Metro, the metal-storage business Goldman had bought, was storing one to two million tons of aluminum, zinc, and lead in its Detroit corridor, and had leases with about two dozen clients. But, thanks to the restrictive LME rules and the market's growing penchant for storing metal stocks, Metro's rental fees—about 40 cents per metric ton per day—were generating more than $100 million in revenue each year.

When Coke, a longtime investment-banking client that had recently hired Goldman to advise it on a $12 billion acquisition, initially complained about the escalating storage costs of aluminum and slow delivery times, the firm tried to assuage its concerns. The trouble was the LME system, Goldman argued, which was inefficient and needed updating. Goldman itself was simply following the rules.

Truthfully, shipping out even the required amount of aluminum per day was a huge undertaking. One day in September 2012, a few months after the LME had raised minimum removal mandates, the eighteen Metro workers in the company's Chesterfield, Mich. plant were very busy. Train boxcars sat with their doors open on the tracks that ran inside the building as workers loaded slabs of metal into them. Because the aluminum was so heavy, the stacks in each car were surprisingly low, filling up only a fraction of the space. Beeping forklifts breezed by as drivers picked up the bars specified by individual warrants, or certificates of ownership, for metal holders who wanted their wares transferred.

Rental fees had risen by that point to about 45 cents per ton per day, generating some $200 million per year in revenue on Detroit aluminum stores alone. Nonetheless, the premiums at work in the Midwest market made the ownership lucrative, explained a Metro executive, especially when aluminum markets were in contango.

"The ownership of metal is a control game," he said, adding that the markets were easy to squeeze because there were so few units in circulation. Metro's typical customer was a hedge fund, the executive said, that bought physical aluminum from a producer like Alcoa, shorted it through the futures markets as it was being shipped to the warehouse, and then took ownership of the warrants on the aluminum once they arrived in Detroit. Then the aluminum-owning hedge fund could sell its warrant to another party, who might wait for aluminum spot prices to rise, locking in a profit.

The LME's increased removal rate had done little to assuage Coke and other aluminum users. Premiums had gone from about 6.5 cents in 2010 to 11 cents by then, and would rise to a record of nearly 12 cents by the summer of 2013.

As the premium prices rose, Goldman and the LME were sued for violating U.S. antitrust laws in the Detroit warehouses. A parallel Justice Department inquiry was soon under way.

Excerpted from "The Secret Club That Runs the World: Inside the Fraternity of Commodity Traders," to be published in June 2014 by Portfolio, a member of Penguin Group (USA) LLC. Copyright © 2014 by Kate Kelly.