I also surprisingly get some of the few hits to this blog for people searching on topic like "copper cancelled warrants", "Johor" and "warehousing".
This post aims to provide some background and what it all means. The key take outs are:
- Warehousing for aluminium has not been functioning as "a market of last resort" for a number of years. This has crept into other base metals
- Traders are earning risk free returns through the arbitrage between low cost of carry and the higher future delivered metal prices ("carry trades").
- This requires combination of surplus markets, contango pricing in futures curves and a low cost of funding.
- Without one of these factors, warehousing would return to being a market of last resort.
On high premiums
- Low load-out rates from warehouses are exacerbating physical shortages and supporting premiums for delivery. This is not necessary for "carry trades" to work, which is central to the current physical tightness.
- Concentrated metal stockpiling can create market squeezes and backlogs for delivery. But this is risky and is unlikely to be persistent. Current copper market conditions fall into this category.
- Changing load-out rules at backlogged warehouses, (or Goldman's offer to source metal for those in queues) should temper premiums, although metal could just be delivered into other warehouses for carry trades.
- The much bigger impact will come when short-term financing is not so cheap, which is perhaps still a couple of years away.
So what went wrong? Aluminium provides the template
*At the end of this post is some of the basics of warehousing and the LME for reference
What has changed in the last few years is that a combination of market forces has ensured that storing metal has now become a source of risk-free returns and that storing metal is no longer a "market of last resort".
Simultaneously, the cost of short term funding collapsed as central banks scrambled to mitigate the impact of collapsing demand.
So if you had access to cheap financing and somewhere to store the metal relatively cheaply, you could lock in a risk-free return by simultaneously buying physical metal at low spot prices and hedging the future delivery price at a higher level.
This differential is shown on the chart on the left. The profit is the difference between the spot and futures price minus the cost money and the cost of storage (or the cost of carry). This trade doesn't work if you can't get preferential rents, although their is some profit and some parts of the curve. It also requires borrowing at ~libor rates. But if you have these two factors in place then it is very profitable to undertake these carry trades beyond a 3 month horizon.
This differential is shown on the chart on the left. The profit is the difference between the spot and futures price minus the cost money and the cost of storage (or the cost of carry). This trade doesn't work if you can't get preferential rents, although their is some profit and some parts of the curve. It also requires borrowing at ~libor rates. But if you have these two factors in place then it is very profitable to undertake these carry trades beyond a 3 month horizon.
Note that this "carry" trade does not require any speculation on the price in the future. Rather, the desired return is locked in when the trade is put on. It also didn't mean that prices were higher as a result, rather than trade worked precisely because the market was so weak rather than strong.
It is also argued that owners of the warehouses are also those participating in these carry trades, so they can make the warehouse costs low enough to make the venture more profitable (although warehouse rates are not disclosed).
As a consequence, queues for delivery of metals can become extremely long once warrants are cancelled. As a basic example, consider inventory spread over 10 warehouses that would take 3 months to clear at minimum(maximum) load out rates. Now if all that inventory is located in one warehouse, its going to take 30 months, or 10 times as long. This is effectively what has happened in aluminium, with inventory accumulated in a couple of key locations (namely Detroit and Vlissingen, which are owned by Goldman Sachs and Glencore respectively).
Premiums for delivery become skewed
In the last 12 months or so we have now gotten into a situation where because traders want to hoard metal for carry trades and existing cancelled warrants in warehouses are sitting behind huge queues, consumers are finding it more difficult to get their hands on physical metal.
As a result, premiums for delivery (which are negotiated between suppliers and buyers for physical delivery from warehouses) have risen rapidly to record high levels.
Previously, rising premiums were seen as bullish for flat prices. But in this case, its a more manufactured outcome than suggesting markets are tightening in a traditional sense.
High premiums are not a prerequisite for hoarding of metal to happen, rather its a consequence at a different point in the cycle. To be sure, hoarding of metal began well before there was any sense that aluminium premiums would end up jumping so much.
It has been suggested that increasing the minimum load out rules would change the game in hoarding metal and this is currently under consideration by the LME (and seems likely to come into play). This will probably reduce premiums and queues for metals, but perhaps doesn't solve the problem of the warehouse system not operating as a market of last resort.
Carry trades catch on in other metals, some hoarding just about premiums
Aluminium has been the poster metal for this activity, but it is increasingly seen in other metals too. Zinc was next cab off the rank and more recently copper appears to becoming more affected by hoarding of metal.
The arbitrage in Chinese markets also seems to be more important to copper. SHFE prices are currently higher than LME prices, so there is a big incentive to move refined metal into China.
To be sure, the accumulation of metal is probably more about taking advantage of Chinese demand for refined copper rather than the carry trades we have seen in aluminium and zinc. To be sure, the epicentre of this trade has been Glencore's Johor warehouse in Malaysia. Absorbing metal here and forcing longer queues has had the impact of pushing premiums for delivery into China much higher.
But the difference here seems to be that a higher proportion of the warrants has been cancelled at Johor than aluminium has been in Detroit or Vlissingen, suggesting this is more a shorter term trade by Glencore to make money out of stronger Chinese demand than it is about earning a profit on the carry. Johor copper stocks are starting to come down, suggesting delivery for consumption rather than rolling metal in carry trades is the aim of this game.
Copper hoarding for cheap finance in China?
A further variation on this theme is using copper stocks for cheap foreign currency financing in China. This done by a Chinese entity financing the purchase of offshore copper though a low interest rate Letter of Credit from a Chinese bank. The copper is then immediately sold for RMB. The letter of credit is then paid back in 3-6 months time.
This transaction does require someone to buy the copper in China, so unless the demand is there on the other side this transaction may not work. So this kind of activity is unlikely to be so pervasive that it will affect copper prices for a nearly as long as carry trades have done to aluminium.
Solutions? Changing load-out rates helps but doesn't affect core carry trade issue
To address these problems the LME has floating lifting load out rates for those warehouses with extreme backlogs i.e. queues of several years of metal. This should help tip the balance and shift premiums as it should reduce concentration of where metal is stockpiled.
That said, it doesn't affect the incentives of financial players to undertake carry trades. So while some metal may find its way to customers, this maybe offset in the overall balance by those with cheap warehousing bidding for metal. So while premiums may go down, they may not fall to pre carry trade levels.
Finally the Goldman Sach's offer to source metal for those waiting in queues at its Detroit warehouse may not be as altruistic as its made out to be. There is enough aluminium in warehouses that don't have queues, which Goldman could easily swap warrants with. But effectively it gives them metal in warehouses they own, which they can obviously acquire at low rent. So perhaps its just swapping more metal into profitable carry trades for the bank.
Basics of LME warehousing
The warehousing network for base metals traditionally operates as a market of last resort. In times of unanticipated weakness in manufacturing cycles, excess supply can be kept in these warehouses and then drawn down once demand recovers faster than supply.
The largest network of warehouses for base metals is maintained by the London Metals Exchange (LME). Warehouses are required as contracts can be physically. Most of the time, transactions are not physically settled, with the exchange mostly used for hedging or speculating on price movements. But in the event of a surplus or deficit in the market for physical metal, the LME warehousing system usually acts as the balance.
Warehousing comes at a cost, so buyers and sellers would prefer to avoid them altogether. But in the event you do need metal from the warehouse, you know that it must be LME approved quality and that the warehouse must meet the standards required for the safe storage of metal.
When metal is delivered into an LME warehouse, the LME provides a warrant for that metal to the owner. When the metal is sold, it is the warrant that is sold to the buyer. When the buyer wants to take delivery of the metal, the warrant is specified as cancelled until the metal has been removed from the warehouse, which can take time. The LME provides data on cancelled warrants on their website.
The warehouses in the LME network are privately owned and fees for warehousing are contractually negotiated between the warrant holders and the warehouses owners. While warehouses are free to negotiate these terms, there are rules on how metal must leave warehouses each day if warrants have been cancelled. This is through the specification of minimum load out rates, which become de facto maximums given warehouses have every incentive to keep metal in warehouses for as long as possible.
Not all key warehousing points are in the LME network. In particular, warehouses in China are either part of the Shanghai Futures Exchange, or are in private bonded warehouses. There are also some warehouses outside of the LME ex-China holding aluminium in particular.
