By Andy Home
LONDON (Reuters) - In the beginning there was Detroit.
Looking back now on the genesis of the London Metal Exchange's (LME) warehousing woes, the original aluminium load-out queue in Motown wasn't rocket science.
Global financial crisis. Manufacturing collapse. A deluge of unsold metal into the "market of last resort." Super-contango in the LME forward aluminium curve. Cheap money. Stocks financing.
Metro, the dominant LME warehouser in Detroit, found itself with most of the free-float aluminium in the LME delivery system. The ensuing surge to get hold of that metal, led by stocks financiers, got jammed in the exit door because of the exchange's then minimum load-out rate of 1,500 tonnes per day.
Revenues from the queue allowed Metro to out-bid everyone else for fresh metal supply, creating a circular money-making machine.
Cue Goldman Sachs, which bought Metro and fine-tuned the strategy. There are now 983,925 tonnes of aluminium in the queue at Detroit, with a notional rental value of over $220 million.
Modern-day alchemy, turning base metal into gold!
But it seemed to be a localized problem. One metal. One location. One company.
Five years on, though, the queues have spread to other LME good-delivery locations and to other metals such as copper, where surplus is still a highly marginal phenomenon.
The LME has tweaked its load-out rules, to no obvious effect. Now the entire global user-base is mulling over the exchange's latest proposed fix.
But the simple fact is that this metastasizing problem, to coin a phrase used by my colleague John Kemp, is as much about the changing landscape of the LME warehousing business as anything else.
This is a still unfolding story of the rise of the merchant-warehouser, a new breed of operator that has evolved the Motown queue metric to a new level.
THE DECLINE OF THE INDEPENDENT WAREHOUSER
When Goldman bought Metro back in 2010, it triggered a wholesale consolidation of the LME warehousing business.
Prior to then only one warehousing company, Henry Bath, had been owned by a bank, JP Morgan. It was a historical anomaly dating back to the purchase of Henry Bath, which issued the first ever LME warrant in 1883, by LME broker Metallgesellschaft in the early 1990s.
No one really cared that much since there were plenty of other LME warehousers to use.
Once Goldman got into the business, though, others rushed into the same space. Glencore, as it was then, bought Pacorini, Trafigura bought NEMS, Noble Group bought what is now called Worldwide Warehouse Solutions (WWS) and Louis Dreyfus bought GKE.
CWT, a logistics company, reverse-engineered the process by buying up trading houses MRI and, more recently, LN Metals.
Barclays Capital, together with steel trader Metalloyd, set up its own LME warehousing company, Erus Metals.
The number of independent LME warehousing companies has accordingly steadily declined to around 35 percent of the total.
And most of them are small and location-specific such as BLG (Bremen), Zuidnatie (Antwerp), Keystore (Hull) and Vollers (Antwerp and Rotterdam).
The only truly global independent LME warehouser is Steinweg, the grand-daddy of the LME storage business.
Maybe that should read "independent", given its linked ownership with trade house Raffemet. Or as Steinweg successfully persuaded the Minor Metals Trade Association, which prohibits trade ownership of its warehousing facilities, "services provider" Raffemet.
The consolidation process has now pretty much run its course, largely because there are so few independent warehousers left to buy.
The only recent development has been the purchase of Scale Distribution in Liverpool by Macquarie Bank and fund manager Red Kite, a deal everyone concerned describes as a bespoke business.
Similarly, the build-out of LME storage capacity has, for now at least, also run its course.
The number of LME-registered units increased from 584 in March 2010 to 689 in August 2012.
Over the last year, however, the total has marginally contracted to 678 units, based on the LME's warehouse list dated July 11, 2013.
That levelling-out of total storage capacity masks some very sharp shifts in market share, though.
The graphic below shows the net changes in warehouse unit ownership by major operators since August last year.
There is always going to be a certain amount of noise in this data-series. Warehouse operators, most of whom rent storage space, are always fine-tuning their capacity as leases expire and metal flow dynamics change.
But there are still three very clear trends at work.
The first is the shrinkage of the smaller, mostly independent warehousing sector, a net loss of 17 units.
The second is the shrinkage of the bank-owned warehousing companies.
Henry Bath's foot-print, for example, has declined by a net 26 units in what looks like a major portfolio tidy-up.
So too has that of Goldman's Metro to the tune of a net 11 units, including a complete withdrawal from Johor in Malaysia, where it has closed all six registered units.
While the banks retreat, the merchant-warehousers advance.
Glencore Xstrata's Pacorini has added a net 24 units over the last year. Trafigura's NEMS has added a net 9 units.
The rise of these merchant-warehousers becomes even more evident in a comparison with March 2010, before the carve-up of the LME warehousing business had really got going.
The two stand-outs are again Pacorini, up a net 92 units, and NEMS, up a net 20 units, as shown in the next graphic.
LOCATION, LOCATION, LOCATION
Even more telling is where these merchant-warehousers have expanded their storage capacity.
Trafigura's NEMS has consolidated its operating focus on Antwerp, where it has opened 16 new units over the last year, while trimming the number of those elsewhere.
It now operates 26 of a total 57 LME-registered sheds in the Belgian port. The next largest operators are Steinweg and Metal Terminals, each with seven units.
Glencore Xstrata's Pacorini has expanded into Vlissingen in the Netherlands, New Orleans and Johor.
Pacorini effectively "owns" LME storage in Vlissingen, operating 53 out of a total 55 units and turning what was once a port famous for its herring industry into a city of aluminium.
In New Orleans it operates 32 out of a total 55 units. The next largest operator is Metro with 18 but the battle for control is swinging in favour of Pacorini. It's opened five new units over the last year, while Metro has closed two.
Pacorini is equally dominant in Johor with 11 out of 20 units. Following the exit of Metro, the next largest operator is Henry Bath with four units.
Now, it's surely no coincidence that the five locations with load-out queues in excess of 100 calendar days, the LME's proposed threshold for faster load-out, are Detroit, Antwerp, Johor, New Orleans and Vlissingen.
Dominating a good-delivery location is a pre-requisite for controlling both inflow and, crucially for those queues, outflow, as Metro showed all those years ago.
Metro, incidentally, is still the largest warehouse operator in Detroit with 27 units, although both Noble's WWS and Pacorini are muscling in with five and two units respectively.
THE RESISTIBLE RISE?
The LME, as it has repeatedly pointed out, has no legal jurisdiction over who owns the warehouse companies that service the world's foremost metals exchange.
It does control the approval of new storage units. The technical pre-requisites are multiple and onerous, as any warehouser will attest. Indeed, warehousers ascribe the high cost of LME storage to their own costs of complying with exchange rules.
However, beyond the technical tick-boxes, the LME seems to have never had a broader set of criteria about the build-out of storage capacity at any one location.
Or if it has, it has been a case of worrying about too little rather than too much storage. Back in 2008-2009, for example, Detroit was something of a back-water in the LME delivery system and Metro itself was gradually shrinking its presence. It had to scramble to open units ahead of the flood of aluminium deliveries from North American producers.
That, though, was a unique moment in time.
There is no comparison with what Pacorini, for example, has done in Vlissingen. In early 2010 it had just nine units in the port. It now has 53.
This may be good business for it and its owner but does the aluminium market, either the LME or the physical market, actually need such huge capacity in a location surrounded by other good-delivery points like Antwerp and Rotterdam?
The LME has always espoused free-market principles, which in the case of warehousing has translated into a refusal to contemplate capping storage capacity by location and/or operator.
Yet a relatively simple solution to its warehouse woes was always to link capacity to load-out requirement, emulating the largely successful way it handles dominant futures positions. Put simply, if a warehouse operator wants to dominate a good-delivery location, it comes with a quid-pro-quo in terms of how much it must deliver out.
It's a shame that such an option didn't make it into the current consultation document.
Because this LME warehousing landscape is going to keep changing and, based on current trends, it's going to keep shifting in favour of the merchant-warehousers.
Both JP Morgan and Goldman have to sell their warehousing businesses by 2020 under U.S. banking rules on commodity investments.
In reality neither is going to wait that long. Indeed the for-sale signs are already up.
Who, though, is going to buy?
Other banks are unlikely to, given the accumulating regulatory pressure on proprietary trading and physical commodity assets.
There will be few logistics companies with the financial muscle to afford what are likely to be high-priced assets.
The most obvious contenders are another generation of physical merchants, who often now come disguised in the garb of hedge funds.
The risk is that they evolve further that original Motown queue metric.
(Andy Home is a Reuters columnist. The opinions expressed are his own)
(Editing by James Jukwey)