This is something that I’ve only touched upon briefly in the recent past, but it is now clear that industrial users are currently engaged in some very substantial reductions of raw material inventories. Perhaps no single commodity illustrates this better than rhodium, a platinum group metal that has fallen from a high of $10,000 this July to $1,600 last Friday, a drop of 84% in the course of 3 months.
The interesting thing about rhodium is that there is no exchange or “market” for it like the COMEX, TOCOM, LME, LBMA, etc. It is almost exclusively sold by refiners to industrial users through production offtake agreements. So we cannot blame paper manipulation or speculators for the decline. Instead, the crash in the price of rhodium, and platinum and palladium as well, is due primarily to supply management of these metals in the industry in which they are predominantly used: automobile manufacturers, and in particular, manufacturers of catalytic converters.
The following Resource Investor article by Jack Lifton dated August 21, 2008 provides some interesting insights into the rhodium situation:
Can One Man’s Actions Take $6 Billion In Value Out Of A Minor Metal Market In A Month?
The global OEM automotive industry, by far, is the largest user of rhodium. It’s been estimated that as much as 85% of the annual rhodium supply—both from new mine production and from recycling of spent auto catalyst—goes into the production of automotive exhaust-control catalytic converters. In 2007 this would have amounted to some 850,000 ounces of the 1,000,000 ounces of rhodium mined and recovered from recycling for that year.
. . .
As I understand it there was a deficit ‘scenario’ projected for the rest of 2008 as there was supposed to have already been for 2007. This would have meant that, on June 21, 2008, the price of rhodium could have been expected to continue to rise. This was mainly due to OEM automotive industry consumption exceeding production and recycling. Moreover production would be impacted negatively by South African mining ‘woes.’ There was really only a single woe at the time: the supposed reduction of electric power being supplied to the PGM mining industry in the Republic of South Africa (RSA). This was due to the power shortage there diverting electricity supplies to consumer rather than industrial uses. This, of course, turned to be inaccurate. It turned out, as I have written here before, that the RSA’s government allowed almost full power to those industries that garnered the most export-sourced foreign exchange earnings per kilowatt. This meant platinum, rhodium, and palladium producers kept their electric-arc furnaces going while the ferroalloy furnaces went cold.
By August 20, 2008, when prices for rhodium (and platinum, but that is another story) were in free fall, there was little production shortfall caused by South African woes. Nonetheless the understatement was issued by Platt’s (above) that, “trade sources [are] saying there have been more sellers than buyers lately.” Ya think?
Why? It certainly wasn’t due to the reduction of annualized production of new cars in North America as many, many experts are claiming. Even if the worst nightmares of GM, Ford, and Chrysler come true and North American car production declines from a projected 16.5 million to 12.5 million for 2008, this will take only 4 million cars out of the global market. This will amount to 5% of global production. Wouldn’t it be logical then for rhodium prices to stabilize-if the metal’s supply were, prior to this slow down, in ‘slight deficit’ as Platt’s and Johnson-Matthey implied? In the worst case perhaps the price might decline by 5% on an actual demand basis. So what happened?
Let me tell you what people in the end-user sector in my home town, Detroit, are saying. It is being said here that the global crash of the price of rhodium is due to the foolish and unskilled actions of a single employee of the purchasing or finance staff of a once-great American OEM automotive end-user of PGMs. That once great company, it is said—until fairly recently—had the most skilled strategists in the global OEM automotive industry in its purchasing group. Their expertise on the insurance of the risk of price and availability of the PGMs was applied by car makers on their exhaust emission control systems.
The article goes on to essentially state that the purchasing manager of this automotive OEM (Original Equipment Manufacturer) has been selling the company’s existing rhodium and platinum inventory into a falling market and that is what caused the prices of these metals to accelerate downward. Please allow me to point out, however, that the allegedly reckless selling by this single OEM appears to have been completed a while back, and that was several hundred dollars ago for platinum and several thousand dollars ago for rhodium. So what happened since then? It seems that this particular OEM company is not the only one that has been getting rid of rhodium and platinum at fire-sale prices.
Rhodium Redux
In fact, this isn’t just a story about the bad purchasing and inventory management decisions of one company and one or two commodities. This very same thing is happening at many companies involving many commodities. Indeed, numerous companies with raw material inventories of every kind have been unloading the excess in the past several months. Part of the reason for this is to correct an overbuild of inventories that occurred as raw material prices rose during the past several years. It turns out that many companies deal with rising raw material costs not by making forward purchases at fixed prices but rather by purchasing raw materials outright ahead of the expected price rise. Now that prices are on their way down, these companies believe they are just as “smart” to get rid of the excess inventory as they were to build it up when prices were rising.
Another reason companies are unloading inventory is, of course, the ongoing credit crunch. As the commercial paper markets dried up and traditional sources of credit and financing disappeared, many companies have resorted to selling whatever isn’t bolted down. Out of desperation, they’ve been selling whatever can be liquidated just so they can make payroll and keep the lights on.
What I’m describing above, of course, is a different form of speculation that market observers rarely talk about. It is speculation by industrial users. But the net result is essentially the same as the other type of speculation: a buildup of supply stockpiles during rising prices and a reduction of these supply stockpiles during falling prices. Had these companies maintained lean inventory levels under just-in-time best practices, a significant portion of the rise in commodity prices would never have occurred.
I’m not saying just-in-time inventory would have been the absolute right thing to do because maintaining barely adequate inventory levels would have exposed these companies to some serious potential consequences such as: (1) losses they may incur due to their inability – as a result of global competition and increased manufacturing capacity – to pass through the higher cost of raw materials; and (2) the very real prospects of not being able to secure just-in-time inventory as required, which could result in a shutdown of the entire production line. The possibility of a business disruption would seem to be a particularly valid concern given that only a few months ago LME registered warehouses held just enough metal to satisfy a few days of global industrial demand. While these stockpiles have improved over the past several months, even today they are not sufficient to absorb global demand should there be a substantial disruption in mine supply lasting more than a few days. On the other hand, most companies now recognize that there have been very few if any such mine disruptions in the past several years and that there exist significant unregistered stockpiles of most commodities in addition to LME registered stocks.
The Missing Piece of the Puzzle
Surprisingly, there has been very little discussion so far about the implications of the industrial inventory cycle even though it is turning out to be quite a critical consideration. It seems most analysts prefer to continue to focus almost exclusively on mine supply, recycling rates and implied demand. The perfect case in point is copper, a market that until just a couple of weeks ago was still supposedly in tight supply because of intermittent labor strikes and production difficulties at copper mines in Chile and Peru. Never mind that annualized copper production, problems and all, is now most likely running more than a couple million tons ahead of just-in-time demand. Never mind that reductions in raw material inventory levels by industrial firms is likely putting hundreds of thousands of tons of copper back on the market.
No wonder it seems like most analysts don’t know what they are talking about! Should we really be surprised that they are surprised by how far commodities have fallen? These industry experts aren’t even looking at the right things or in the right places! Worse yet, some of the best-and-brightest commodity pundits are even claiming that industrial users have started to stock up (or will do so shortly) on raw materials to take advantage of lower prices despite evidence to the contrary. To wit, industrial users tend to only stock up on raw materials when prices are rising.
Must We Talk Contango Again?
So, if industrial users are actually reducing raw material inventories, where is this supply of commodities going? Well, that’s the problem. It doesn’t seem like anybody wants to buy these newly-available supplies coming on the market. This has undoubtedly been a major reason why prices have fallen by amounts that would have seemed impossible just three months ago. Sure, hedge fund speculators have been responsible for a portion of the carnage as they have liquidated positions to raise cash. But crude oil would not have fallen by 50% in three months and many other commodities would not have fallen by even larger percentages if we were simply talking about fund-based speculators fleeing for the hills. Indeed, the commodity market regulators along with many economists have studied the influence of speculative demand in the futures and forward markets, concluding that speculators were responsible for only a moderate portion of the runup in commodity prices. What their “studies” seem to have failed to point out is that speculation by industrial users (in the form of building up excess inventories) may have been responsible for a much larger chunk of the runup and consequently the crash.
Let’s now look at the question of the day: when will the commodity price decline be over and what will stop it? I believe the answer is actually quite simple. Absent a confirmed economic recovery (confirmation by definition occurs after the fact), commodity prices will need to go into above-normal contango in historical terms before we can expect a final bottom. In other words, spot prices will have to go substantially lower than the futures or forward prices. Why? Simply put, high contango will allow dealers and warehouse operators to purchase and hold inventories off market. Specifically, the positive spread between the forward and spot price will give them profit motive to hedge long physical positions by going short in the paper markets.
Unfortunately, the most reliable way to get a rise in contango would be to have commodity speculators return in significant numbers. Another way would be to have industrial users start to buy raw material supplies on a forward basis to replace the excess inventory they’ve been selling on a spot basis. A third way would be the gradual reduction in new production as commodity prices decline below their marginal cost of production. Neither of the first two prospects appears particularly bright at the moment. The third — shutdown of mines, wells and farms — is already occurring in some commodities but it can take years for there to be an appreciable reduction in output. For example, let consider OPEC, whose very purpose is to exert monopoly control over the exporting of oil by member countries. Even this cartel has had difficulties reducing oil output in the past (recall oil actually dropped to $10/barrel in the late 1990s) and will have difficulties in the future.
The best chance, in my personal opinion, for a near-term rise in contango to a sufficient level such that the commodity sector puts in a final bottom, is industrial users of these commodities regaining a modicum of confidence in the economy. Assuming their confidence level is quite low today, it is possible this could happen even as we continue to slide further and further into a recession, but only as long as the logjam in the credit markets is unblocked to a sufficient extent.
Crunching Data
Here is the bottom line. If my hypothesis is correct, we could perhaps get a profitable jump on the rest of the crowd by looking for an unusual rise in contango within the commodity sector while everybody else is looking for an improvement in general business conditions. Why? Because rising business confidence will probably not register in the official data until long after it has already registered in the form of soaring contango in commodity prices.
If you would like to help look for this unusual rise in contango (admittedly I haven’t had the time), please let me know your ideas. The method I’ve come up with so far leaves something to be desired in terms of its complexity and the amount of effort that would be required, but I’ll provide it just in case some of you are charting mavens or gluttons for punishment and would like to take a shot at it.
- Construct long-term charts that track the normalized spread (see step 2 below) between the front contract month and the first outer contract month, the second outer contract month, the third outer contract month, etc. for each commodity that may have been subject to industrial stockpiling in the past few years and de-stockpiling in the past few months. Both the front and outer contract months should be active settlement months.
- The normalized spread for purposes of these “spread charts” is the spread divided by the contract price and then annualized. In other words, it is necessary to create comparative data and that requires the spread to be stated as an annual percentage of the contract price. For example, let’s assume that December 2008 Crude Oil is $80 and March 2009 Crude Oil is $81 on September 1, 2008. The absolute spread is $1 and the normalized spread is 5%. How did I compute this? It’s actually quite easy. First, divide $1 into $80 and you come up with 1.25%. Then annualize this by dividing the 1.25% into a daily rate and multiplying by the number of days in a year. For the sake of simplicity, it would be okay to assume the front contract month expires on the first day of that month. That means we can use December 1, 2008 as the expiration date of the December 2008 Crude Oil contract. Thus, we have 90 days or so after September 1, resulting in a normalized spread of 1.25% divided by 90 times 360, or 5%. Note that normalizing and annualizing a spread is no different from normalizing and annualizing an interest payment in determining the annual interest rate on a loan.
- Alternatively, it may be possible to analyze historic spreads for many commodities by comparing specific contracts. This would require calculating a set of absolute spreads for each year and charting these spreads against each other. It should be relatively simpler to construct the charts using this approach because we can ignore normalized spreads completely. For example, we could look at the spread between the active December contract and the subsequent March contract for each year in Crude Oil. This would involve pairing the December 1998 and March 1999 contracts, the December 1999 and March 2000 contracts, the December 2000 and March 2001 contracts, etc. The spread for each contract combination would be displayed on the same chart allowing direct historical comparison. The advantage of this approach is that we can ignore normalizations completely assuming that we can line up the dates for each data spread (this is complicated a bit by leap years and the fact that holidays don’t fall on the same date each year).
- Because the use of daily closing prices can result in the introduction of significant timing errors other discrepancies, separate charts should be constructed using the open, high and low prices as well. Alternatively, one can try to average together the open-high-low-close prices for each date and then construct the charts using these averages.
- It might be useful to construct the charts on a daily, weekly and monthly basis to better identify trends.
- My experience is that moving averages can provide significant clarity to charts where the spreads seem to fluctuate with inscrutable randomness. Different moving averages work better or worse for different commodities so it may take some experimentation to come up with the best ones (for example, in silver I have found that 20 day moving averages work particularly well).
- Charts should ideally be constructed for each of the following components of the GSCI and CRB indeces (excluding those components without a significant U.S.-based futures market such as Gasoil and Nickel). The commodities in bold are the ones that I would specifically expect to have a “contango effect” at some point:
- Aluminum
- Brent Crude Oil
- Cocoa
- Coffee
- Copper
- Corn
- Cotton
- Crude Oil
- Gold*
- Heating Oil
- Lean Hogs
- Live Cattle
- Natural Gas
- Orange Juice
- Silver*
- Soybeans
- Sugar
- Unleaded Gas
- Wheat
*Constructing these charts for gold and silver would of course be a good idea for its own sake.
You might perhaps recognize that the above is nothing more or less than an exercise in constructing the basis. In this case, the basis is masquerading as a spread between two sets of futures prices so it is not the precise definition of the basis we would use in gold and silver trading, but the key concepts are interchangeable. Some of you have asked me in the past how to calculate the basis and you have even shown me charts you’ve created that didn’t quite get it right. Well, here is your chance to try again.
For those of you attending the GSUL seminar in Canberra next month, you might consider the above as a hands-on demonstration of many of the basis concepts that we’ll be discussing (in simpler terms) and building up to during the live lectures. If you were to try constructing just one of the above charts, it would be a tremendously useful experience. Remember, your GSUL registration fee and Founding Membership in The Metal Augmentor includes direct assistance from me, so please take advantage of it.
For Founding Members who will not be attending the GSUL seminar either in person or remotely, please consider the above as the first installment of basis instruction under your Metal Augmentor subscription (though I suspect many of you will curse my name for getting so wildly technical before covering many of the basics of the basis).
I suspect casual SILVERAXIS readers will let the above go over their heads without protest so that I hope no apologies will be necessary.
Supply Destruction
There is one other point I’d like to make before actually getting to what this all has to do with silver, and that point is the mistaken claim by some people that the mere reduction of raw material inventories at the industrial level is by itself a form of “supply destruction” equivalent to the closing of mines. In reality, the sale of excess inventory by industrial users is nothing more than a temporary change in ownership of existing supplies. And while some of these supplies are already headed back to registered warehouses and stockpiles, most will continue to be held in undisclosed locations close to the industrial source. Thus, there will be little actual physical movement of these raw material inventories from the warehouses where they are already stored by the industrial user. The main reasons for the lack of movement are the cost and logistics of transportation and the likelihood that the very same industrial user who is now selling these raw materials will end up buying them back when they are needed for future production.
So when will we get supply destruction of these industrial inventories? Basically, it will happen if and when supplies are sold forward (back) to industrial users, but only as long as the industrial users have gone back to just-in-time inventory management. In other words, the forward repurchases of these inventories by industrial users will correspond to actual need for raw materials and not just rebuilding of excess inventories. And even though these forward industrial repurchases will not have much effect on visible stockpiles (just as the physical industrial sales are not having much effect on visible stockpiles), we should know with a fair amount of certainty when such forward purchasing is occurring because contango will be rising.
The Silver Connection
Getting to silver at long last, there is no doubt whatsoever that some of the price weakness in silver, especially in relation to gold, has been the result of some industrial offloading on the physical market in the form of raw material inventories of silver. This source of supply has tended to offset the strong retail investment demand (clearing the shelves of bullion dealers worldwide) as well as the solid wholesale investment demand (via the ETFs and other public investment vehicles acquiring 1,000 oz. bars).
I think it is relevant to note here that some raw silver inventory is not in the form of 1,000 oz. wholesale bars but rather shot, cast ingots, sheets, etc. Industrial offloading of such raw silver would not make its way immediately to the physical markets but it could create additional physical supply for these markets in the future. Indeed, it is quite possible that much of this raw silver is now being fabricated into retail bullion products. The rest is probably being held by dealers who have likely hedged these physical silver positions by going or staying short in the paper markets.
Whatever the quantity involved, the impact of reductions in silver inventories by industrial users has clearly not been as major an influence on the price of silver as it has been for the platinum group metals, base metals and other materials. Moreover, once the raw material liquidation cycle has run its course, the monetary and investment qualities of silver will likely start to take command of the price and we should expect the gold-silver ratio to decline as a result. Therefore, Mr. Sinclair’s recent point about silver being an industrial metal and not a precious metal is more backward looking than forward thinking.
Silver Basis May Be Deceiving
The above discussion also brings up an interesting angle in terms of the silver basis. Recall that I mentioned earlier that the bottom in commodities would probably be within sight once contango has risen substantially above historic norms. Well, this might mean an interesting twist in the case of silver, and only silver. The reason is that silver, unlike commodities, can be tugged in separate directions by its dual nature. On one side is industrial demand and on the other side is investment/monetary demand. The industrial demand, as it recovers, should result in a higher contango for silver just like the commodities in general. At the same time, continuing growth in monetary demand for physical silver should result in a lower contango (movement toward backwardation).
What I’m trying to say is that two diametric factors may be exerting equal but opposite force on the silver basis, thereby canceling each other out. As a result, we may need to adjust how we look at the basis in silver while the industrial inventory liquidation and forward repurchase cycle plays out. Specifically, we should expect that the contango in silver may be artificially high while the tug of war plays out. And just how long should we expect this to continue? Once again, I suspect that the main clue that the impasse between opposing forces might be coming to an end will be a rise in contango substantially above historic norms across the commodity sector.
What about gold? Well, industrial stockpiling of gold as a raw material does not occur to the same extent as silver and so the impact on the gold basis of inventory liquidations and subsequent forward repurchases is unlikely to be (as) significant. Under these circumstances and for the time being, the gold basis may very well be the superior indicator of rising monetary demand and/or forthcoming economic calamity.