Growing LNG trade accelerating integration of global gas markets
A cliché of the energy business has been that "oil markets are global, while gas markets are continental."
In a world where only a handful of pipeline gas crosses from one continent to another1 and the LNG market is significantly smaller than the pipeline market, there has been good reason to be skeptical about the global integration of gas markets.
However, price and trade data suggest that a global market for gas is not a scenario for the distant future. By responding to customer needs, integration already has begun. For example, in 2002, gas prices fell in Japan, South Korea, southern Europe, northern Europe, the US Gulf Coast, and Western Canada, compared with 2001 prices. In all of these markets, gas prices rose in 2003. Furthermore, developments in gas trade around the world point to a growing role of gas in ways that increase the global nature of the business.
In short, a global gas market is already here, although the degree of integration is modest. To make the case for global integration of gas markets, two elements are needed:
- A mechanism to transmit market forces physically from one market to another.
- Price shifts that reflect active arbitrage in the market.
Logistics and LNG
Logistical issues matter when integrating markets. At first look, the LNG market appears unlikely to provide the needed trade mechanism. Oceans create long distances between gas markets in different continents. There are only roughly 150 LNG tankers in the world as of early 2004, a small fleet compared with the fleet of oil carriers. Further, the history of LNG contracting points to rigidity, rather than flexibility.
Long-term contracts for LNG, often with take-or-pay terms, limit the impact of one gas market on another. With volumes determined by contract and prices set by formula, there is little opportunity for transmission of market conditions from one market to another. Arbitrage opportunities are virtually dead if take-or-pay contract rights are enforced—unless contracts show some flexibility.
Such flexibility is now showing up in LNG contracts in several ways:
- Contracts increasingly provide for FOB delivery, rather than CIF delivery, allowing the customer to divert a cargo to an alternative market. Volume holders from Middle East plants, roughly equidistant from Mediterranean Europe and the Far East, are best poised to profit from global market shifts.
- Customers for long-term gas increasingly own (or charter) their own LNG tanker fleet, allowing greater control over delivery and destination.
- Terms for volumes in LNG contracts are increasingly flexible, allowing for a wider range of deliveries within a long-term contract. Some contracts spell out explicit tranches of base and swing deliveries.
- The European Union is insisting that terms for supplying LNG to European importers allow for trading of cargoes and do not lock in one destination for contract gas. Negotiations between the EU and Sonatrach have been aimed at increasing the flexibility of older contracts signed in earlier, less flexible times.
- By several counts, spot cargoes have risen to 8-10% of all LNG deliveries in the past year,2 or nearly 20 billion cu m, or bcm (2 bcfd).
- LNG liquefaction plants are now being built without firm contracts for the full output. In one case (Nigeria's Train 4, under construction), a liquefaction plant is being built with contracts covering only 60% of the output.
This increase in contractual and operational flexibility allows players in the LNG business to be more responsive to market shifts than in the past. This increase in flexibility is taking hold of short-term supply and trade decisions and is having a measurable affect on gas balances around the world.
Pop quiz
What do the following events have in common?
- Cold weather in North America in early 2003.
- The closure of nuclear power plants in Japan over the past year.
- A surge of gas-fired electric generation in northern Europe in late 2003.
LNG traders and brokers know the answer: The volume of LNG imported into each of these markets increased in the wake of the events listed here, each of which increased the demand for gas. These episodes show directly the oil-like response of LNG to local market conditions. While trade of LNG continues to be dominated by long-term contracts—where the size, stability, and creditworthiness of the buyer are key concerns—trade at the margin is being guided by Adam Smith's invisible hand.
Examples of market-influenced trade abound, particularly in the Atlantic Basin. Trinidad and Tobago, as a supplier to the less-regulated markets in Europe and North America, has been the most active supporter of arbitrage-led trade, heavily influenced by its short distance to a market where endusers are not necessary and where volume can be hedged. Elsewhere, cargoes from Algeria traveled to the Far East when South Korea needed additional supply, while cargoes from the Pacific Rim traveled to Lake Charles, La., on the US Gulf of Mexico coast when the US market was tight.
Although the volumes are admittedly limited (10% of the LNG market is roughly 18 bcm/year, or 1.8 bcfd), these episodes demonstrate that LNG is providing a degree of market integration to the global gas market today. When one market is tight and another market is in surplus, ships are diverted to the strong market. These market-based reactions, while still limited in volume, testify to the responsiveness of LNG as a global commodity.
Although long-term contracts from gas supplier to enduser are still the backbone of the LNG market—and are likely to continue in that fashion—swing volumes of LNG to end-markets already provide a degree of integration to the world's gas markets. The degree of that integration can be measured by a look at the gas spot price data.
The growth of the LNG market over the coming decade will serve only to aid such market-responsive trade. Contracts already signed will raise the volume of LNG by 25% by the end of the decade, while projects under development could easily double that quantity. And with these projects showing greater flexibility than past LNG projects, the volume of price-seeking natural gas is on a strong upswing.
Price correlation
A central theory of economics—perhaps the most important building block of economics following supply and demand—is the "law of one price."3 In its simplest form, it argues that a traded good will have the same price in every country that allows its trade. In broader form, allowance is made for taxes, transportation costs, import duties, and other transaction costs, so that the price of a commodity will be higher in importing countries than in exporting countries, and distant markets will be priced higher than markets closer to supply. Further, when the price rises in one market, it should rise in all markets by an equivalent amount.
For natural gas, the law of one price means that the short-term ups and downs of gas prices should be seen in all parts of the world that trade gas. The simplest test of this is whether the prices are correlated over a short-term timeframe. For this analysis, we have calculated the correlation of gas prices, using monthly prices at various locations around the world for 1998-2003.
When we look at this data set, we find that the price paid for natural gas at markets around the world already exhibits a degree of integration (see table). Gas prices show a degree of correlation, generally 0.6-0.85. All of the correlations are statistically significant at the 95% level (and are thus unlikely to be due to random luck). This leads to the inevitable conclusion that spot prices in major gas-consuming nations already rise and fall together.
To put this relationship in perspective, we have calculated the price correlation for a number of energy commodities: North American gas, propane across the Atlantic, and crude oil around the world.4 The price link between global gas markets is not as tight as the price links between the other energy commodities: gas markets in North America (that are united by the pipeline grid), propane markets across the Atlantic, or global crude oil markets (see figure). The higher degree of correlation of pipeline gas markets (0.85-0.99), LPG markets (0.93) and crude oil markets (0.98-0.99) around the world indicate a greater degree of integration than in the global gas market.
The comparison example for pro- pane is of particular note. Like natural gas, it is primarily a local or regional product—only limited volumes travel long (oceanic) distances (except to the Far East). Like LNG, specialty tankers must transport propane with high costs per mile (relative to oil carriers). The high correlation of propane prices between the US Gulf Coast and Northwest Europe is likely due more to competition with oil than due to the strength of arbitrage between the two markets.
The nature of energy markets around the world also plays a role in the measured correlation shown in the figure. Natural gas competes with oil in all the import markets. The technology to use gas is largely the same around the world. Further, a significant portion of contracts for pipeline and LNG gas supplies in Europe and the Far East are indexed to oil prices. As a result, a significant share of the correlation is due to competition between gas and oil, whose prices rise and fall in tandem all around the world.
Think global
Is there further to go for the global integration of gas? Certainly. As swing volume increases and flexibility in contracts grows, arbitrage opportunities will expand and the price link between markets almost certainly will grow stronger.
The volume of gas moving to arbitrage-favored markets is likely to double and redouble in the coming decade, even though much of the swing volume will originate with contract customers. The countries with the least-regulated natural gas markets, the US and UK, are starting to build new import terminals in anticipation of dramatically increased LNG imports. Japan is likely to play a rising role, primarily through more flexible contracting, which will allow them to divert cargoes out of the Middle East to Europe when market conditions dictate. This will tend to push the gas price correlations around the world closer to the levels seen in pipeline markets.
The logistics inherent to LNG, however, suggest that there are limits to global integration of this commodity:
- Shipping cost. The cost per mile of shipping LNG is greater than moving oil.
- Quality. The North American market uses leaner gas (less ethane and propane) than Europe or the Far East, adding a cost to meet local standards.
- Boil-off. LNG shipments are subject to boil-off, the loss of roughly 1% of the cargo per week on board. As a result, LNG tankers rarely change destination after beginning a trip. A corollary is that the response to a market shift must wait until unloaded tankers make their way back to the liquefaction plant.
- Spare tankers. There is a much smaller pool of spare LNG tankers in the world than the pool of crude carriers. There is little ability to accommodate a shift to longer LNG routes in the short run, even if the market demand (and prices) were high. By way of contrast, the crude tanker fleet has greater short-term flexibility in the face of market shifts.
In addition, the time horizon of current contract gas supplies in some markets (e.g., Japan) may limit the degree of integration. With virtually 100% of gas supply imported via term contracts, the Japanese market is unlikely to closely follow global gas prices until a meaningful share of imports arrive on a spot basis (a difficult number to estimate, but likely over 10% of imports). This development will require time (for the expiration of current contracts) and future flexibility (less rigid terms of new contracts). However, Japanese importers already have begun to show an interest in such developments, which would tend to lower their average cost of LNG imports.
There certainly are new risks in an integrated global gas market. Events anywhere in the world have an impact on today's oil market, and that impact is felt globally. Demand spikes, supply shortfalls, or tanker constraints could squeeze LNG deliveries around the world, raising prices simultaneously. A cartel along the lines of the Organization of Petroleum Exporting Countries could be formed. However, there are mitigating factors: LNG exporters, with higher capital costs, rely on outside finance and good customer relations; LNG exporters, as a group, have lower political risk, compared with oil exporters; and LNG will need to continue its current growth pace for a generation before rivaling the magnitude of the oil trade.
In summary, the integration of global gas markets is already significant, but falls well short of the integration of the global oil market. LNG already provides a link between global gas markets, allowing a modest volume of arbitrage trade, enough to affect gas supply at the margin. Growth of LNG in volume and flexibility is likely to raise this integration over time, making a global gas market increasingly important. There are limitations to market integration, however, including the inherent logistics of gas trade and the contractual basis for such trade.
References
1. Two pipeline routes from Algeria cross to Europe. Siberian pipelines to Europe could be added to this number if Asia and Europe are counted as separate continents.
2. The transactions counted as "spot" in the LNG market might look alien to a Henry Hub gas trader or a Rotterdam oil trader. They include deals with formula prices, generally tied to oil prices as well as term deliveries up to 1 year long.
3. Owen Lamont and Richard Thaler stated, "The first law of economics is clearly the law of supply and demand, and a fine law it is. We would nominate as the second law 'the law of one price,' hereafter simply the Law. The Law states that identical goods must have identical pricesUbut in practice, details about market institutions are important in determining whether violations of the Law can occur." Journal of Economic Perspectives, Vol. 17, No. 4, Fall 2003, pp. 191-202.
4. North American gas prices included Henry Hub, Appalachia, New England, and Sumas. Propane prices were compared at Mont Belvieu and Northwest Europe. Crude prices included West Texas Intermediate, Alaska North Slope delivered to California, Brent first month, Dubai, Oman, Tapis, and Duri. Locations were chosen to illustrate the range of correlation (or the degree of integration).
The author
Thomas A.Z. Howard, PhD, is managing director of PIRA Energy Group (www.PIRA.com) and heads PIRA's Global LNG service, which includes the Global LNG Monthly, a comprehensive analysis of the LNG market; forecasts of gas markets around the world; an online database of the LNG industry; and the interactive PIRA LNG Netback Calculator, with over 1,000 existing and potential trade routes.