Investment demand for gold will wane as PGMs gain in appeal

The following is the first of several instalments of The Commodity Refiner, published by London-based Barclays Capital Research. This week we examine precious metals markets; subsequent instalments will focus on base metals.

Commodities are at a crossroads. After a strong performance during the first quarter, prices are back at levels similar to where they opened the year. Speculative funds flowed into commodities early in the first quarter, contributing to price strength across the whole spectrum and sending market indices to multi-year highs. Extreme geopolitical uncertainty and the poor performance of other asset classes provided trading conditions that almost uniquely prompted markets as varied as energy and base and precious metals to move in unison.

Despite the fog of war beginning to lift, there is even more than the usual level of uncertainty hanging over near-term price direction for commodities. But one thing is certain: a period of more diverse trading patterns beckons for individual sectors. Investors will need to be much more discriminating in how they make their decisions in commodities markets over the next few months.

Gold: the perfect storm?

Even the most ardent gold market bull has begun to realize that demand has fallen sharply in the current gold rally. According to statistics released by Gold Fields Mineral Services (GFMS), total fabrication demand fell 11% in 2002 and is now down 15% since 2000. The bulk of this fall occurred in the crucial jewelry sector, which represented more than 75% of demand on average during the 1990s.

Part of the reason for this fall is simple: the price has risen. Between 1980 and 2000, there was a negative 0.66 correlation between the average annual gold price and jewelry demand. In 2001, demand fell broadly across a variety of markets. In 2002, however, there appears to be one major culprit — India, where demand fell a staggering 20%. Demand in gold’s largest market is actually more strongly correlated to the gold price (in U.S. dollars) than the world in general — since 1980 the negative correlation has been more than 0.8.

But Indian demand will recover, surely. Is this not what we saw in the fourth quarter of 2002 when demand rose 24%, versus the fourth quarter of 2002? Does the fact that the Q4 price was the highest of the year mean India is already adjusting to the new, higher price? Perhaps, but it is also worth noting that the gain was off a very low base — Q4 01 demand was down 27% on Q4 00. However, there is a more important reason for believing demand is unlikely to recover — history. Just before the 1993-1996 gold rally, Indian demand peaked at 143 tonnes in Q1 93. As gold rallied, Indian demand fell. Demand did not start to rise again until the gold price began to fall back from its peak of US$415 per oz. Demand only rose above the Q1 93 peak in Q1 97, when the price averaged US$351 — 15% from its peak. Therefore, history suggests that while demand may recover, Indian demand is unlikely to “snap back.” It is also worth recognizing that the recent rally has been greater — up 33% from before the rally by the first quarter of 2003 versus only 21% in the previous period.

Moreover, there are reasons for thinking that this dip in jewelry demand may be part of a broader trend. Put simply, the world appears to be spending less money on gold jewelry. After peaking at US$36 billion in 1996-1997, global jewelry spending has fallen to US$26.8 billion — this despite first the spur of a lower price (to demand in ounces) and then the recent sharp price rise.

If a broader macro shift is happening, what is driving it? This is even less clear, but there is one possible answer: gold is losing “market share” to platinum.

Platinum’s conquest of China, mirroring its success in Japan, has significantly eroded gold’s share of the Chinese consumer Renminbi. In 1997, Chinese consumers spent RMB 23.9 billion on gold jewelry and only RMB 1.5 billion on platinum (converted at average annual prices and exchange rates). By 2002, gold jewelry spending had fallen to RMB 15.7 billion while spending on platinum had risen to RMB 7 billion. While the overall jewelry pie had shrunk by 11% (worth bearing in mind for those who believe that economic growth equals more gold demand), platinum’s share had risen from 6% to more than 30%. As platinum statistics are difficult to come by, it is impossible to know if this is being repeated worldwide. However, platinum’s success in Japan (where it has surpassed gold in both U.S. dollars and ounce terms) provides at least some support for this argument.

Perhaps, say gold’s true believers, but any fall in physical demand will be more than offset by a sharp rise in investment. Faced with falling stock markets and derisory bond yields, investors will surely return to the stability of gold. Moreover, the sharp rise in global insecurity will rekindle desire for a “safe haven,” further drawing money into gold. Add on to this a dollar that is, in the eyes of some analysts, only beginning what will be a dramatic correction (and ignoring the fact that they have been forecasting this event for more than a decade) and the case for gold becomes overwhelming. Indeed, it is true that this is to some extent a “perfect storm” for gold. So how has the market responded? Is gold investment indeed surging?

Before answering those questions, it is important to define “investment.” If investment demand is to replace jewelry, demand it must be physical. Buying Comex or TOCOM futures does not replace physical demand unless the owner stands for delivery. As warehouse stocks have not taken a precipitous plunge, it does not appear that the average futures investor intends to take delivery. Obviously, futures-buying can have (and has had) a significant impact on the price, and it would be foolish to suggest that in the short term such buying will have no impact. It won’t, however, affect the supply-and-demand balance, which bodes ill for gold in the medium term.

Investing in gold equities will also not replace physical demand. It may, perhaps, lead to a reduction in supply if investors disapprove of hedging (and as equity issuance allows companies to raise capital without going to the banks, which usually require hedging in order to protect their lending). Of course, one might argue that providing mining companies with plentiful, cheap capital may encourage production growth (as was the effect of virtually free capital on the telecom sector). Regardless, an investment in the XAU or the HUI does not consume physical gold and so cannot offset a fall in jewelry demand. So what is investment? In its simplest form, coin buying, bar hoarding, and the like. In its more complex form, products such as those offered by the Perth Mint, which allow investors to purchase gold physically allocated at the Mint. Or even ideas such as Gold Money in which “eGold” is backed one-for-one by physical bars held in allocated form. So how has “real” investment fared in this gold rally? Poorly.

What “investment” gold has seen thus far has largely been paper — not physical. Comex open interest and LBMA clearing statistics tracked each other fairly closely from 1997 (when the LBMA began releasing statistics) to mid-2001 (when the current gold bull market began). Since then, the lines have diverged dramatically as Comex volumes (of paper gold) have soared while LBMA turnover (of physical gold) has continued to fall. Even small speculators seem to far prefer “paper” gold investments to physical ones.

In August 2001, the U.S. Mint sold 6,500 ounces of Gold Eagle coins and Comex small speculators (Commodity Futures Trading Commission category “non-reportable”) were long just over 23,000 contracts, or 2.3 million oz. Both “positions” hit their high at about the same time — January of this year for coin sales and the Feb. 4 reporting period for small speculator positions on Comex. Between August 2001 and January 2003, Eagle sales rose by 89,500 oz. Small speculator positions rose by almost 3.7 million oz. Speculators, therefore, remain firmly in favour of paper over bullion
. Perhaps this is because just as they are easy to enter, they are easy to exit. Paper positions like this are precisely the kind of “hot money” flows that some governments criticize for the destabilizing effect they could potentially have on their economies.

Indeed, although 2003 has started well for the Mint, 2002 ranked seventh out of the past 12 years. In fact, September 2001 was only the third best September since 1992. Despite the “perfect storm,” Eagle sales have never recovered to the levels hit in the run-up to Y2K — in 1999 Eagle sales were more than 2 million oz., almost 6.5 times 2002 levels.

Perhaps this is because, contrary to newspaper articles, gold is not really perceived as a “safe haven.” Contrary to popular opinion, gold’s response to “events” post 9/11 is no greater than it was before the attacks. Indeed, although rumours still have a powerful effect on short-term price fluctuations (as anyone trading Comex will testify), actual events have had a relatively muted impact on the price. Events that were expected to have a powerful impact on the gold price have not. Indeed, the only events that seem to have powerfully affected the gold price were negative ones, such as Hans Blix’s or Colin Powell’s testimonies to the United Nations.

But, the true believers will say, it is too early to judge. Gold’s 5,000-year history as a “safe haven” can hardly be invalidated by two years of derisory Eagle sales and a failure to rally after a random selection of events. Indeed, one need not go back that far to find evidence of gold playing a real role as a store of value.

In the 1980s, gold investment (coins, bar hoarding and estimates of physical investment) was almost 30% of gold demand and more than half the size of jewelry demand. This was equal to 0.2% of average annual gross domestic product and almost 0.6% of the market capitalization of the S&P 500. Translated into today’s world, this would equate to between 2,000 and 4,700 tonnes of annual gold demand. As annual supply is only 3,870 tonnes, to say that this would have a positive impact on the price would be something of an understatement.

Of course, were the price to explode, the impact on scrap supply would be dramatic — in 1980 scrap represented 30% of total supply, its highest proportion then or since. Interestingly, the second-highest level was in 1998, when it represented more than 25% of supply as the Asian crisis led to a sharp rise in local currency prices and massive dishoarding. One might, therefore expect a similar response to a surge in prices, such as happened in 2002 when scrap supply rose by 18%.

But we are unlikely ever to test such a thesis. Why? Because physical investment is not set to return to 1980s levels. As much as we hope that the examples above convincingly argue why, the best reason is simple. How many of our readers (or their friends or family) have significantly increased their holdings of (non-jewelry) physical gold recently? Indeed, how many of you actually hold any physical gold?

Silver: photo-finished or relief rally?

For reasons stretching back into the mists of time, this brief analysis of silver sits with precious metals and not base metals, where it belongs. Despite its monetary history, silver exhibits a tighter correlation with traditional industrial metals than with gold.

Indeed, at 0.66 since 1992, its correlation with the Metal Component of U.S. Durable Goods Orders is actually higher than that with the Component and the Economist Base Metals Index (0.53). Moreover, that relationship is multiples of that between silver and gold (0.12 correlation since 1992). It is, therefore, not surprising that silver prices have tended to rise during periods of base metals strength. Indeed, although market participants still tend to focus on the gold-silver ratio, the concept may have outlived its usefulness.

The gold-silver relationship has been breaking down over the past 10 years, largely due to gold and silver’s relative progress on the road to commodity status. Silver has become almost exclusively an industrial metal (as evidenced by the correlations above) while central banks only began to divest themselves of their legacy gold holdings in the middle of the 1990s. Moreover, the two metals have suffered from different microeconomic shocks, such as the “Buffet rally” in silver and Asian Crisis-related dishoarding in gold.

But silver is facing a problem akin to gold’s — it is losing its primary role. As gold lost its monetary status, the price fell (in both nominal and real terms), making jewelry cheaper and spurring consumption. The single largest source of silver demand is photography (film and paper), and one can argue that photography saved silver from irrelevance after it lost its monetary status between 1871 and 1900. However, it seems silver use in photography is in decline. Silver consumption in photography fell in 2001 by more than 4% — a fall that probably owes more to the collapse in travel after 9/11 than any switch to digital. Of greater concern, though, is the fact that silver consumption fell in 2000, when there was no 9/11 to dampen travel, and is now down to levels not seen since 1994.

A shift to digital is the likely culprit, for although sales of traditional cameras continue to dominate sales of digital ones, digital’s share of “snaps” is rising.

Digital cameras need not, strictly speaking, be negative for silver demand — after all, printing digital pictures requires just as much silver as printing traditional ones (though digital does not use silver halide film, this represents only about one-third of total silver photographic demand). However, only a fraction (estimated in Japan to be 10%) of digital pictures are actually printed. Perhaps the reality of the impact on silver demand is best summed up by the revenues of the world’s largest film company, Kodak, which have been falling since 2000.

So perhaps silver is a cheap bet on a U.S. economic recovery? Or even, perhaps, if such a bounce is unlikely, on a post-Iraq equity “relief rally.” But won’t silver be hit by a gold slide in such an environment? Perhaps, but by less than gold.

Consistent with this relationship, silver tends to outperform gold as equities outperform bonds. Over the past five years, the gold-silver ratio has demonstrated a minus 0.75 correlation with the ratio of the S&P 500 to the U.S. 10-year bond price. That is, as equities have gained relative to bonds (S&P 500 up versus U.S. 10-year price), the gold-silver ratio has declined (silver gains relative to gold). Moreover, the ratio has exhibited a negative 0.71 correlation with U.S. 10-year interest rate swaps. Again, as swap yields rise, the index falls (silver gains relative to gold).

Another reason for warming to silver is one of the things that make it unique — the fact that it is largely a byproduct of other metals.

The largest source of silver byproduct production comes from the base metal industry with arguably the worst economics — zinc. At current zinc cash prices, a substantial amount of global zinc mining is operating at a loss. Although there have been no mine closures as yet, the current price level has resulted in smelting cutbacks. As discussed elsewhere, with treatment charges at 10-year lows, a significant number of zinc smelters are operating at a loss. Although this has been true for some time, about 400,000 tonnes of smelting capacity have been cut this year, some of it permanently. If zinc production falls, there will be a significant effect on silver production, regardless of the level of the silver price. This has always, admittedly, been the argument of silver bulls (that and the supposed constant supply deficit) — an argument that has been wrong since well before Mr. Buffet’s foray into silver in 1997-1998. However, if zinc production actually starts to come off-line (and producers gain the discipline of, say, copper producers), then this will support the silver price.

But perhaps the best reason for expecting a short-term silver bounce is simple: at the beginning of the second quarter, speculators are net short. This is relatively rare –
– speculators have been net short on only 61 occasions (8.6% of observations) since 1986. On only 11 occasions have they been shorter than they are at present (3,680 lots, or 18.4 million oz.).

Of those 11 occasions, one-month later silver has been an average of 3.89% higher. On only one occasion has it been lower (0.54%), while the average gain on the 10 times it has been higher is 4.33% (peak gain: 8.31%). That works out to an 8-to-1 “bet.” Not bad odds.

For those concerned about the effect of the size of the speculative long in gold, although speculators have never been simultaneously as long gold and short silver as they are now, they have been long gold and short silver on nine occasions since 1986. Of those occasions, silver has been on average 2.14% higher in one week. With an average gain of 5.69% (seven gains) versus an average loss of 0.94% (two losses), at 6:1 the odds, while not as strong, are still good.

Ah, but what about the potential for post-Baghdad gold losses? The picture is not as negative as one might think. There have been 26 occasions when large speculators have been net short silver, and gold has fallen from report period to report period. Although the average gain in one month has been lower (0.91%) and the odds worse (1.5-to-1 based on a 4.02% average gain and 2.71% average loss), again history favours silver.

Indeed, putting it all together, there have been only four occasions when speculators have been net long gold, short silver, and suffered a period-over-period price fall. On average, the silver price was 3.94% higher in one month. The ratio of average gain to average losses of more than 4-to-1 is again a “decent” bet.

Whatever the reason, although it is hard to dispute the medium-term negative outlook, in the short term, silver is due for a bounce.

Palladium: darkness before dawn?

Palladium prices have fallen almost by 30% since the start of 2003 and are now down to levels not seen since May 1997. Although the fall is precipitous (down 84% from the all-time high of US$1,085 per oz.), it is worth bearing in mind that palladium has averaged only US$260 per oz. since 1991. Indeed, excluding the 1999-2001 period, the price has averaged only US$167. The key to what is happening now, and what may happen, is what took place during this period — when the palladium price averaged more than US$540 per oz.

Most important is not the absolute level of the palladium price but its spread to platinum — its primary competitor in autocatalysts. In order to understand the relative development of platinum and palladium, it is worth taking a brief look at the history of the use of platinum group metals (PGMs) in autocatalysts.

Emissions legislation beginning in the 1970s was largely targeted at reducing carbon dioxide emissions. For various reasons, platinum is more efficient than palladium at converting carbon dioxide emissions — for example, the fact that palladium has a lower melting point than platinum. This meant that palladium-based catalysts required two to three times more metal than platinum-based ones. New legislation in the U.S. and the European Union targeting hydrocarbon emissions, introduced in the early 1990s, favoured palladium because it is more efficient at converting hydrocarbons. This led to an increase in palladium demand. At the same time, large-scale Russian stockpile sales in the first half of the 1990s depressed palladium prices, encouraging a switch out of platinum and into palladium, even at the higher loadings then required.

As spending increased, automakers increasingly used palladium over platinum so that, by 1996, on the cusp of the introduction of the U.S. tier-1 standards, palladium’s share of autocatalyst PGM spending had risen from 4% in 1991 to more than 30%.

The announcement of European Stage 2 standards in 1996 led to another wave of PGM spending. It rose from US$890 million in 1996 to US$1.9 billion in 1998. Again, palladium was the major beneficiary, and its share of PGM spending rose to more than 70% in 1998. At first, palladium prices did not respond — they averaged US$130 from 1991 to 1997. In 1998, prices surged to US$283 in response to the rise in spending, as well as some well-publicized hedge-fund buying.

The impact of such a switch out of platinum and into palladium caused a dramatic swing in their relative prices. From 1999 to 2001, palladium averaged a US$61 premium to platinum, compared with an average discount of US$161 (1991 to date).

As palladium demand grew, not only did the large Russian sales of the early to mid-1990s cease; significant instability began to be injected into supply. Russian supplies suffered from bureaucratic-related shocks, which meant that export contracts (the approval of which followed a particularly tortuous route) were delayed in 1998 and 1999. Whether this was due to legitimate chaos (at various points export licences were said to be signed and then not signed — and in one case “lost”) or a conspiracy to drive prices higher is irrelevant. It injected a degree of instability into the supply of a metal that, by 1999, exceeded 80% of total PGM spending. It is important to note that there was never a problem with PGM production — only with the transportation of that supply. At that point, demand growth began to accelerate as consumers, concerned about the stability of Russian supplies, began to buy forward.

As palladium prices gained sharply, relative to platinum, in 1998 (when the platinum-to-palladium ratio fell below 2-to-1 for the first time), consumers began to switch back to platinum. Historically, the estimated time required for a switch from one metal to another is two years — hence the lag between the announcement of the new U.S. and European emissions standards in 1994-1996 and the bottoming of the palladium price in 1997. This logically prompted a switch back to platinum, and indeed 2000 saw demand for platinum autocatalysts rise for the first time since 1994.

Total PGM spending peaked at US$4.4 billion in 2000, spurred by palladium’s market share rising again and an average price of US$678 per oz. Since then, total PGM spending has collapsed — to an estimated US$2 billion in 2002 — and palladium’s share of that spending has fallen to less than 50%. While this is partially due to lower demand overall, a significant portion represents the running-down of stocks of palladium purchased during 1998-2000 in response to the Russian supply concerns.

Collapsing autocatalyst demand is not the only factor behind the falling palladium price. Scrap supply exploded in 2002 to 7% of the total — almost double its peak in 2000. Moreover, not only has palladium autocatalyst demand suffered, but electronic and dental demand have as well, with nickel and gold the prime beneficiaries. Another important culprit has undoubtedly been the need of Russian producers to re-establish their credibility with consumers.

After the supply shocks (hitting a market in which demand was already soaring), the risk premium required to use palladium over platinum undoubtedly increased. During the recent platinum price surge, Russian producers made it clear that the palladium price would not increase in line with platinum. Indeed, there were rumours in South Africa that Russian buying was behind the platinum price surge — trying to encourage consumers to switch back. While that seems unlikely, Norilsk’s pending purchase of Stillwater Mining in Montana included promises to keep Norilsk metal in the U.S. — a clear attempt to make supply more stable, thereby encouraging a switch back into palladium.

Spurred by strong jewelry demand, platinum prices have recently soared and are now at their highest premium to palladium since 1988. With platinum trading at more than 3.5 times the price of palladium (palladium is at a US$440 discount), it is difficult to imagine that autocatalyst manufacturers are not looking to switch back. Indeed, unofficial reports suggest precisely that. This is especially true given the technological advances autocatalyst manufacturers have made since the mid-to-late 1990s. Although most autocatalysts continue to req
uire more palladium than platinum, in some they are now interchangeable on a 1-to-1 basis. While a variety of factors will affect the loadings, the reality is that the relative ratio of palladium to platinum has declined, reducing the price ratio at which it becomes profitable to switch to palladium.

Although automobile production is likely to remain weak in the short term, emissions regulations are likely to continue to tighten, and in such an environment, PGM loadings will have to increase. In the current economic environment, it is unlikely that the costs of such an increase will be passed on to the consumer. This will necessitate further cost savings, thus encouraging a switch into palladium. The question, of course, is when? If the ratio remains broadly at current levels, the impact will begin in 2005. However, technology has moved on since the mid-1990s and many autocatalyst manufacturers claim to be able to switch in a shorter period of time (though carmakers still prefer to switch from one model year to the next). If the palladium price remains at such a discount to platinum, one can only assume that any switch will be accelerated.

Palladium prices have shown a 0.84 correlation with the metal’s share of total PGM spending since 1991. It is likely (and logical) that a switch back into palladium will boost its price. Adding to the potential for a spike higher is the fact that Russian palladium sales have fallen from 4.34 million oz. in 2001 to only 1.6 million in 2002.

Palladium is unlikely ever to trade over US$1,000 again. The example of rhodium suggests that PGM markets, although prone to spikes, are strongly reverting. Moreover, most metals markets that have suffered such an event never reclaim the panic highs (for example, copper in 1995 and gold in 1980). However, we favour a recovery in palladium in early 2005, spurred by a discount to platinum not seen since 1992.

Platinum — Demand shift

Platinum prices have recently retreated from their multi-decade highs of more than US$700 per oz. toward the US$600-per-oz. level. The severity of the move down, coming as it did during a wildcat strike at Impala Platinum, the world’s second-largest producer, suggests that speculative positions were significantly larger than had been estimated. Although the fall has been severe, it is worth noting that it has not yet hit the 38.2% retracement of the move from below US$420 per oz. to US$710.

While there are a variety of reasons behind platinum’s rally, continued strong growth in platinum jewelry demand in China stands out. Short-term factors, such as an increase in autocatalyst demand in response to palladium’s price explosion in 2000, “back loading” of South African production expansion plans, and an explosion at Lonmin’s smelter in South Africa, all exacerbated trends that are likely to continue to support the platinum price.

As we have already mentioned, platinum demand for autocatalyst use soared with the introduction of new emissions legislation in the 1970s — averaging 28% between 1975 and 1980. As supply was relatively static, this was largely at the expense of jewelry demand, which actually fell 10% on average. During the 1980s, however, jewelry demand began to recover as the initial boom in platinum catalyst demand slowed. By the 1990s, with automakers shifting away from platinum and into palladium, autocatalyst demand growth collapsed to average barely 1%. As production continued to increase (notably due to an increase in Russian exports), platinum was faced with a potential problem. What “saved” platinum was jewelry demand.Notably, unlike gold, where there is a strong (0.68) negative correlation between jewelry demand and the price, from 1980 to 2002 the correlation between platinum jewelry demand and the price was positive (0.12). That is, platinum prices and demand rose in tandem. Global platinum jewelry spending growth has exceeded that of gold, taking platinum’s share of global jewelry spending up to more than 5% by 2000, from an average of only 2.5% during the 1990s. As an aside, this suggests that a gradual demand shift may be occurring, away from gold jewelry and into platinum.While demand growth came originally from Japan (which was for many years the largest platinum market) and gradually became evenly spread, platinum, like so many other commodities, has now become dependant on China. Chinese platinum buying, despite 17% value-added-tax rates, exceeded Japanese platinum buying in 2000 and now represents more than half of total platinum jewelry demand.Chinese platinum jewelry has demonstrated a stronger positive correlation to the platinum price (0.32), which is likely to reflect the impact of Chinese demand on the price but possibly also its increasing status as a luxury consumer item where a higher price actually attracts more consumers. This is at least partially due to the efforts of the Platinum Guild. The Guild has been careful not to market platinum as a “mass market” competitor for gold — probably out of a desire to avoid the negative price correlation from which gold jewelry demand suffers. It has targeted the young, affluent urban consumer with specific products tailored to their income bracket — that is, “expensive” relative to national incomes, not those of Western Europe. It is intimately involved with micro-managing platinum’s image in specific markets (which, outside of Japan and China, include India and South Korea) to the point of choosing stores, educating staff, designing local language marketing material, and choosing separate spokesmen for different national markets — indeed, a visitor to the Guild is stunned by the seriousness with which they take their responsibilities.The result has been, thus far, a brand to which consumers seem attracted without apparent regard to price. This has implications for potential scrap supply when the price is rising. Unlike gold, where scrap makes up a significant portion of demand and is positively correlated to the price, the platinum “brand” appears, for now, to have protected the metal from the impact of a rising price on scrap supply — after all, would you sell your wedding ring if the price doubled? No. But you might sell gold chains bought for quasi-investment purposes (the reality of much gold jewelry demand in the developing world).While it is obvious that platinum jewelry spending will rely on global growth and, more specifically, on the ability of the Chinese government to deliver the 8% or more growth rates that have become common, barring a global economic collapse, demand growth is likely to remain strong. One significant concern is the impact of SARS on consumer spending in China and Hong Kong. Unless it explodes into a 1918-1919-style flu epidemic, the outbreak is likely only to delay jewelry spending in platinum’s core market for the short term. A strong bounce in demand should occur as the effects abate. We believe, therefore, that platinum prices are likely to be firmer for longer than most market participants expect.– The opinions presented are the author’s and do not necessarily represent those of the Barclays group. For access to all of Barclays’ economic, foreign-exchange and fixed-income research, go to the web site at barclayscapital.com. Queries may be submitted to the author at matt.schwab@barcap.com or to Kevin Norrish, head of commodities research/energy for Barclays, at kevin.norrish@barcap.com