What Prompted the LME Spreads Rule Change?

Outright prices on the LME are impacted by a multitude of factors. Global balance projections, supply shocks, changes in demand, macro/micro factors, geopolitics, and of course speculation. Because of the huge variety of impacts on the outright price of a metal, it's extremely hard for any one position holder to influence the trend of the market. It's not impossible as we saw as recently as 2021 during the nickel crisis, but it's certainly harder.

LME spreads on the other hand are a much more delicate situation. They highly correlated with physical trading. While there is certainly speculation spreads, they are generally driven by physical demand and supply, LME warrant stocks, and traders' positioning on a given prompt date.

A bit of background on spreads first. As discussed previously, contangos are capped at what is known as 'full finance'. Owning an LME warrant in a warehouse has an implied cost that is comprised of LME warehouse rent, and the finance cost of owning the warrant.

Finance cost: ~7% = $15/mt/month + Warehouse rent: ~$0.55/mt/day = $16.50/mt/month So, the full carry cost is ~$31.50/mt/month.

If a month-over-month contango on aluminium today breached $31.50 and say went to $35/mt, traders would be able to make a risk free profit. They would simply borrow at $35, let their long position expire and take delivery of warrants. They would hold the warrants until the short position expired the following month and deliver them against the short. They would have made a risk free $3.50/mt on whatever volume they were able to borrow at $35/mt.

What happens in reality: as contangos widen and start approaching full finance, more and more borrowers will enter the market as they deem these levels good enough to borrow at. Many traders will also be holding material in non-lme warehouses or at non-lme rates and therefore can already make a profit by simply holding metal at lower contangos than full finance - this trade is known as a 'cash-and-carry' trade.

In general, contangos tend to be widest when there is ample supply of material in the market, and often ample supply of LME warrants. There is no fear that liquidity will dry up and so nobody is willing to pay more for material today when there is plenty available, than in the future when supply is less certain.

Month-over-month backwardations however are different. Up until a couple of weeks ago, there was no limit on how large a month-over-month backwardation could become. Backwardations are typically driven by a deemed lack of supply in the market. Traders are willing to pay more for material today if metal is scarce, than they are in the future when more metal may become available.

Traders are in general long physical product before they sell it. Those that hedge are inherently short futures and then square those positions when they hedge their physical sales. But often they will have to carry stock for a varying amount of time after purchasing it before they can physically sell it. This may be due to long transit times between purchase and sale locations, they may be stocking up on metal for an expected increase in demand in the future, or they may simply not be able to find a sale. Whatever the reason, this often means that traders need to borrow short futures positions month-over month.

In a healthy contango market, this is actually a profitable trade, buying nearby at a lower price and simultaneously selling further out at a higher price. However, in a backwardation this borrow creates a loss where they are buying at a higher nearby price and selling at a lower price further out.

The LME monitors, and publishes position reports when traders amass positions over a certain size of the open interest a given prompt date. While it doesn't name traders, it alerts market participants when trader(s) are building positions. This is known as a dominant position. Traders were under no obligation to lend those long positions, particularly if they were using the argument that they may wish to let those long positions expire in order to pick up warrants to use for physical sales.

Counterparties holding short positions on that same prompt date may begin to face liquidity issues in trying to borrow those positions if there were fewer and fewer lenders in the market. If the short holders could not come up with, or did not want to deliver metal in the form of warrants against their short positions, they would have no choice but to keep increasing the number they would be willing to pay in order to borrow the position. At the same time, the holder of the length could also keep increasing number they would be willing to lend at. Essentially the dominant position holder has taken advantage (within the rules) of a perceived lack of nearby metal, and the need for short holders to borrow their positions.

When backwardations spike in these specific scenarios, they can very quickly snowball as panic borrowing sets in among low liquidity. Some short holders wait in the hopes that things will ease but then can find themselves chasing an ever increasing backwardation. This is similar to a short squeeze in outright price trading.

So why would a trader do this? Well firstly, because up until 2 weeks ago it was perfectly within the rules. The LME has always been known as a lender of last resorts. This means that should you not be able to access metal for your physical sales, the LME would be there to provide warrants that you could deliver to your customer. The metal may not be in the perfect location, or the perfect brand, but there would be metal there.

A trader borrows in a contango, amasses a large enough position, and then sits on the position waiting to see if a backwardation would develop. If it does, the trader can then lend their position and capture the profit between their original borrow and their lend. If a backwardation does not occur, they can still let the position expire and pick up warrants to use for physical sales. The LME struggled with a good argument to prevent this. Some traders have large enough physical sales books where they could genuinely say they could use 100KMT+ warrants in a clip. While this is not a surefire way for a trader to create profits, if timed right the risk/reward can be very good.

This pattern of amassing a large long position and taking it close to expiry in the hopes of a backwardation occurring has become a common enough issue that the LME made the decision to change the rules. The LME after all is the merchants market. It was not simply other large traders on the other side holding these short positions. The shorts could be comprised of many different types of companies from small to large with varying sized short positions, that would still be feeling the impact of the trader's squeeze. While not against the rules, it was deemed against the spirit of the function of the LME, and often crossed over to the spreads becoming 'disorderly'.

The rule was changed so that if a dominant position became larger than the total available warrant stock for each metal, the holder of the length would be obliged to lend their position at level until it was back under the available stock level. The argument being that a trader would not have use for any length over and above the available stock other than to push the market into a large backwardation.

While this change does not prevent backwardations occurring (as we are currently seeing on copper), it does disincentivize some of the huge long positions we've seen in the past from being built. If traders can only profit by lending up to a certain volume, it should at least prevent some of the larger, more drastic spikes in backwardations.

The caveat to this lending/position reducing requirement is if available LME warrant levels drop below certain thresholds. Realistically the LME cannot expect long positions to remain near zero if stock levels start approaching zero as well. At that point they'd essentially be saying that backwardations were no longer allowed. Backwardations are often actually a good way for LME warrant stocks to be replenished as holders of metal take advantage and deliver warrants into the exchange system.

It is a very tricky situation for the LME to navigate, trying to keep all stakeholders happy while at the same time keeping the market functioning as it should - with high liquidity, high volumes, and healthy levels of warrants, that allow for the physically settled function of the LME to continue to work and thrive the way it has for the last 150 years.