Backwardation and Declining COMEX Inventory
1 June 2011 20 Comments
Silver inventory at the New York Commodity Exchange (COMEX) is declining. Since the beginning of 2011, in addition to the overall decline, a lot of inventory has been shifted from the registered to the eligible category.
For much of 2011, the COMEX silver futures market has been in backwardation. Similarly, the London OTC silver forward market is in backwardation as reported by the London Bullion Market Association (LBMA). Figure 1 shows the Silver Forward Offered Rate (SIFO) from November 22, 2010, to May 19, 2011.
Figure 1: The LBMA Silver Forward Offered Rate (SIFO) 2010-2011
We see from Figure 1 that the contango of the LBMA silver forward market collapsed on January 19, 2011, and silver has been in backwardation ever since – with the exception of about 4-5 weeks in March and April. For more background, please take a look at Backwardation in The Case of a Monetary Metal.
In this article, I briefly explain why it is perfectly natural if
- COMEX inventories are declining,
- Silver is being moved from the registered to the eligible category,
- Silver deliveries occur as late in the delivery period as possible,
- Silver deliveries suddenly appear out of nowhere and are checked into the registered category only at the last minute before delivery,
as long as silver is in backwardation. None of these indicates that there is a shortage of physical silver or that there is a short speculator about to default.
Managing Inventory
In order to understand COMEX inventory, it is helpful to take the point of view of a COMEX dealer or of a market maker. We have a fixed capital to operate with, we do not care about whether prices rise or fall, but we wish to book as many transactions as possible. What is expensive about our transactions is moving actual physical silver. We therefore prefer a large inventory which always stays in the same vault and on which we write warehouse receipts that change hands when our customers trade. The bigger the inventory the smaller is the amount of physical silver we have to move around and the more profitable we are.
Given our starting capital in US$, how do we initially get our silver to operate with? The answer is that we can borrow it from the market. We buy silver at spot and sell the futures contract maturing in, say, one year. This is a synthetic swap of US$ for silver: we lend our US$ for one year and accept silver as the collateral. This is the silver that goes into our vault as our initial inventory.
For the above argument, it does not matter whether this synthetic swap is the way we actually get our silver. It is sufficient that we could have done it this way and, by arbitrage, any other way of getting the silver would be equally expensive. Expensive? The swap is usually a good deal for us. As long as silver is in contango, i.e. the futures contract trades at a price higher than spot, we receive interest on our swap because we sell the futures contract and buy at spot. This interest is the market interest rate on a US$ loan that is collateralized with physical silver. Not only can we borrow physical silver from the market, we are even paid for it! We can use the interest we receive in order to fund our vault operations and to cover other operating expenses. With silver in contango, we can basically carry an inventory as large as we please.
The situation changes entirely when silver is in backwardation. With backwardation, we have to pay interest in order to carry our silver inventory. In the example of our synthetic swap in which we buy silver at spot and sell the future, we now pay a higher price for spot silver than the price we realize when we sell the future. With silver in backwardation, keeping inventory costs us money. We therefore have to compromise and reduce our inventory as far as possible without hampering our transactions.
We can summarize the situation and say that for a dealer or market maker, the opportunity cost of holding physical silver as opposed to US$ cash is the negative of the contango. If silver is in contango, it is therefore preferable to hold physical silver whereas if silver is in backwardation, it is preferable to hold a cash balance.
This consideration explains why COMEX inventory decreases when silver is in backwardation. It also explains why inventory is moved from the registered category (the dealers’ inventory on which warehouse receipts circulate) to the eligible category (inventory that is owned by somebody else and just left in the COMEX vault system for convenience).
When to Deliver on a Short Position
Let us now think about the delivery process. Every holder of a long contract who still has his position open on First Notice Day, must fund the contract in full and must be prepared to take delivery. Every holder of a short contract must either close it and buy it back or must deliver by Last Notice Day the latest. It is the holder of the short contract who decides about the precise date of the delivery.
Now imagine we are a speculator or a mining company, we own a large position of physical silver, and we are short the futures contract and prepared to deliver on our contract. When do we choose to deliver? We are free to select any day between First and Last Notice Day. Since we locked in the price when we sold the contract some time ago, at first it seems that this choice would not matter. But let us think about every single penny.
If silver is in contango, we would probably deliver as early as possible just because we can then save the storage fees if we get rid of the silver rather soon. But with silver in backwardation, this decision changes as well. We could postpone our delivery as long as possible and might still be able to lend the silver to someone else against a US$ collateral in the meantime. With silver in backwardation, we would receive additional interest on such a swap.
This consideration explains why in a contango situation, the holder of a short position has an incentive to deliver early whereas under backwardation, they have an incentive to deliver as late as possible.
Conclusion
Given that silver is in backwardation and assuming that every participant maximizes their profit, we have explained why inventories shrink and deliveries occur as late as possible. Indeed, both effects are observed. It would be wrong to interpret these observations as a shortage of physical silver or the impending default of a short speculator.
What these observations do not explain, however, is the reason why backwardation has occurred in the first place. For this, please take a look at Backwardation in the Case of a Monetary Metal.
Victor The Cleaner 
Victor,
Is there a marketplace where shorts can loan their silver for a period of less than 30 days?
If you have enough silver, yes, you can phone your favourite bullion bank and ask for an OTC swap. Also, if you are a large institution yourself, you can arrange your other silver positions in such a way that no physical silver sits idly in the vault before Last Notice Day. This sort of arbitrage will probably not be possible on the entire volume of deliveries, but it explains why a considerable amount of silver is delievered only late in the month.
Victor
The additional swap would likely be done with the BB that is displaying counterparty risk right? So there is potential for a chain reaction of bad events if the BB defaults on their obligation to return the silver when its called.
Good question. Depending on your business connections, you could phone a few banks and ask them what they would charge or pay you for such a swap. Even if only one of them is to ‘blame’ for the observed backwardation, I guess, all of them would give you more or less consistent rates. They know yesterday’s official SIFO, they know the current COMEX order book, and they know all of their own OTC positions. I don’t think any one of them would step out of line.
The data I would love to have is the SIFO quotes that each of the BBs reports to the LBMA before they compute the average. From these numbers, you should see who is under pressure and pushes SIFO into negative territory. Some people know this inside information and can trade on it.
Victor
I would also like to see that data, also for LIBOR. I doubt it will ever happen though. Could freedom of information act be used I wonder?
In the case of LIBOR, which is calculated by the British Bankers’ Association (BBA) from a survey of a number of leadings banks, there is now an investigation. Some of the banks are suspected to have rigged LIBOR. Apparently, they are now taken to court by investment funds that hold variable interest rate bonds whose payout is linked to LIBOR:
Libor: investigation by regulators (Financial Times, March 16, 2011)
Big banks investigated over Libor (Financial Times, March 16, 2011)
Funds accuse banks of manipulation (Financial Times, April 19, 2011)
Victor
It would have to be through some sort of governmental or quasi governmental agency that also has the data like the fed. But they would fight it tooth and nail. The lack of transparency is really terrible.
Victor,
Thank you for your response to my question. I am not very sophisticated in these types of financial transactions, but am trying to understand. Your assertion of financial incentive being greater than risk in the time frame of less than 30 days seems illogical to me.
On first notice day, aren’t parties on both sides of the transaction supposed to have their accounts fully funded?
If yes, then what advantage to a short who has US$ tied up until delivery of metal (I assume @ 0 interest earned) to delay delivery and assume additional risk by going into the OTC market for a short term swap? It would seem to me that the best risk/reward would be to close out the trade by delivering metal to settle the contract as soon as possible, if the short is in possession of the metal, and then redeploy the US$ held in trust by the COMEX into a new trade.
Thank you for this question!
First, I realize that there is a detail about the COMEX procedures that I am not sure about. I know that the holder of a long position needs to fully fund the position when he holds it into the delivery period. This makes sense because he may be assigned the metal at any day.
What I do not know is the precise requirement on the holder of the short contract. It would not make 100% sense to require him to deposit the full value of the contract, simply because he will never need that cash. What he needs is the silver.
If there is either no requirement or a cash deposit, then it is indeed the most profitable strategy to lend the silver for as long as possible and to deliver late. This is what I assumed in the article.
If the holder of the short position were required to prove that he owns unencumbered silver as soon as he goes into the delivery period (I am not aware of any such requierement), then it would of course be impossible to lend that silver, and that part of my argument would not apply.
Finally, all the optimizations described in my article are basically arbitrage that exploits the untypically high silver lease rate in the backwardation environment. Of course, if counterparty risk is your main concern, you should not perform that sort of arbitrage. Please also see my reply to the question by FJF below.
Victor
Your reasoning makes sense as constructed, but it ignores the bigger picture, which you clearly cover in the article “backwardation in the case of a monetary metal”. The reason for backwardation is non-negligible countypart risk. – i.e., the possibility that paper promises to provide silver in the future aren’t worth the paper they are printed on. When this occurs for brief periods, there isn’t necessarily a problem. When it becomes “permanent” for monetary metals, as explained by Antal Fekete, there is a very big problem. And things can get ugly in a hurry when market participants collectively recognize that backwardation is permanent and that counterparty risk is very real. So while there might appear to be natural incentives for the market makers at COMEX to do what they have been doing, these incentives derive only from the fact that they themselves created the very backwardation that threatens to annihilate them.
Yes, I agree. As long as you ignore the counterparty risk that is priced into the swaps, it seems there is a free lunch that you can pick up. When COMEX dealers and market makers adjust their inventory as I described in the article, they engage in arbitrage in order to pick up that free lunch. Nice.
But when the counterparty risk ever turns out to be justified, they are, of course, in a more vulnerable position because they have essentially lent their inventory to someone who was paying a premium to get hold of that inventory.
Victor
Thanks to everyone for the questions and the very constructive comments!
Victor
Thinking about your sifo comment earlier. If the BBs were to quote a potential swap using the previous days sifo as a barometer of sorts, then why would we expect the individual components of sifo to have any variation at all? They would simply use the previous days sifo to make the new sifo. Clearly, this is not how it works, so somebody must be participating in swaps at variable sifo rates in order for the sifo rate to vary day to day, right?
It might be worth it to call all the bullion banks for a swap rate quote just to see. Can this be done by john q?
What I wated to say above is this: Someone must be pushing down SIFO. If you see the spreadhseet of which rates the individual banks submitted each day, you might be able to extract the inforamtion about who it is that pushes SIFO down and who just follows. If bank X quotes -0.50 for 3 months because they desperately need more silver, and your bank (Y) is still alright, you would be quoting something close to -0.50, too, wouldn’t you? Say, -0.45. Everything else would be wasting money.
They will give you a (useful) SIFO quote only if you have a serious business relationship. Do you have a company that is an ISDA member?
Victor
Since you posit that silver is in backwardation because of counter-party risk, it would seem a bit daft for holders of short positions (metal sellers) to engage in additional transactions in the days between notice and delivery.
The situation you alluded to in one of your comments seems more likely: if part of the problem is low physical reserves in the BBs, then the more days they hold on to their bullion before delivery, the less days they have to borrow reserves from the market. So the delay in delivery could be an avoidance of a counter-party transaction in the meantime.
Thanks, yes, I fully agree.
Victor
Victor,
This is one of the links from Costata’s Open Silver Forum at FOFOA. It comes from the metalaugmentor site and adds to your discussion of COMEX procedures in the post above.
metalaugmentor
Thanks, that was well worth reading.
Victor
What is expensive about our transactions is moving actual physical silver.
Are you familiar with this ?
No, I am not familiar with this. I thought it would depend on whether you just have to transfer a warehouse receipt or whether you actually want to ship the metal to some other location.
Victor