Note No. 182 April 1999 Trends and Markets in Liquefied Natural Gas Rob Shepherd "It was the best of times, it was the worst of the United States receives a trickle (soon to be aug- times, it was the age of wisdom, it was the age mented by the startup of the Trinidad project). of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of LNG commands a significantly higher price in Light, it was the season of Darkness, it was the Japan, the Republic of Korea, and Taiwan spring of hope, it was the winter of despair, we (China) than it does in Europe or the United had everything before us, we had nothing States. So more supply has been economic to before us . . ." Quoting the opening of Charles develop, and since the Pacific Rim has both Dickens's Tale of Two Cities in connection with ample gas reserves and limited local markets, the current state of the liquefied natural gas that region dominates LNG trade, with more than (LNG) industry may, if anything, be overly opti- three-quarters of total supply. mistic, beset as the industry is with low prices and stuttering demand in its Asian stronghold. In LNG, history matters But it is hard to resist calling on these contrasts to characterize the LNG industry, for despite its To find the roots of the current situation, it is problems, there are glimmerings of change that necessary to go back to the 1980s. The 1970s could profoundly improve its lot. had been years of expansion for LNG, and by the end of the decade Japan was receiving LNG LNG is essentially a niche fuel. Liquefying and from Alaska, Brunei, Abu Dhabi, and two shipping gas is expensive, so the LNG route is Indonesian plants at Arun and Bontang, all attractive for developers only where there is no under long-term take-or-pay contracts closely local market or where capacity in the local mar- tied to crude oil prices. The first Malaysian plant ket is insufficient to take all the available local was under construction and would start up in supplies. LNG requires large investments by the 1983. But the second oil shock of 1979 and the buyers in terminal and regasification facilities, so restructuring it engendered set back demand in it generally flourishes only where there is a short- Japan. The buyers--particularly the power age of indigenous gas supplies and where com- companies--found out just how rigid those petition from pipeline gas is limited. In bulk, long-term take-or-pay contracts could be. They LNG is suitable for transport only by sea, so its took the full volumes, but were not happy. New use in landlocked areas is confined to small peak LNG became difficult to sell. The Australian pro- shaving plants or isolated locations such as cen- ject did not come on stream until 1989, after sig- tral Australia. nificant delays. By that time oil prices had dropped, Japan had recovered, and power Not surprisingly, there are only a handful of LNG demand in the country was growing so rapidly projects, and most supply East Asia, which lacks that it could be met only by building gas-fired indigenous resources (table 1). But the earliest power plant. Suddenly LNG was in demand LNG supplies went from Algeria to Europe and the again. Korea, in 1986, and Taiwan (China), in United States. Europe still takes significant quanti- 1990, had begun to take LNG, having bought ties (just under a quarter of world demand), and incremental capacity from the Indonesian The World Bank Group Finance, Private Sector, and Infrastructure Network 2 Trends and Markets in Liquefied Natural Gas TABLE 1 CURRENT AND POTENTIAL LNG SUPPLIES Current Potential Country Project Country Project Algeria Arzew Pacific Skikda Australia Darwin Australia North West Shelf Gorgon Brunei Brunei NWS Expansion Indonesia Arun Indonesia Bontang I Bontang Natuna Libya Marsa el Brega Tangguh Malaysia Malaysia I and II Malaysia MLNG Tiga Nigeria Nigeria LNGa Oman Oman Expansion Oman Oman LNGa Qatar RasGas Expansion Qatar Qatargas Russian Sakhalin 1 Ras Laffan LNGa Federation Sakhalin 2 Trinidad and United States Alaska North Slope Tobago Atlantic LNGa Yemen, Rep. Yemen United States Kenai (Alaska) Atlantic Nigeria Third Train Trinidad and Tobago Atlantic Expansion a. Expected to start production in 1999. plants. The Korean market began growing at a ity, and Australian and Malaysian plants routinely phenomenal rate. produce 25 percent more). Greenfield plants also seemed uneconomic when comparted with The resurgent demand was met largely by expansion, particularly after the fall in oil and LNG expansion of existing plants. A second plant was prices in 1986. While an expansion might need constructed in Malaysia alongside the first, and only marginal additional investment, a greenfield Bontang continued to be expanded. All the other LNG plant involves not only a central gas pro- existing plants managed to squeeze out more cessing unit, but also site preparation, harbor, LNG. Why were no new plants built? For three marine, tankage, accommodation, utilities, and the main reasons. general infrastructure to establish and support the operation in a remote location. First, the cost of constructing LNG plants had risen sharply. Because few plants are built, there are few Second, an expansion does not have the same contractors and process licensors with a proven scale problems as a greenfield project. A typical track record, and thus little competition. High LNG liquefaction train by the late 1980s was about 2.5 prices before 1986 and the emphasis on reliability million metric tons (3.5 billion cubic meters) a of supply reinforced this tendency. Buyers and year--a volume that the market could easily project sponsors insisted on proven technology digest. But the minimum scale for economic via- and experienced contractors. Designs were lav- bility on a new site had come to be seen as 6 mil- ishly gold-plated (an LNG plant can often produce lion metric tons a year--a much harder prospect at least 15 percent more than its nameplate capac- to place even in quickly growing markets. The World Bank Group 3 Third, speed was important, and it is quicker to and the rest of this Note focuses on where they expand existing plants than to build new ones. might lead us. LNG projects are extremely complex, and it nor- mally takes at least two or three years to set up Potential supply--and its implications the venture structure of a new one. More than 100 million metric tons a year of By the early 1990s all the expansion possibili- potential LNG is seeking a market. Given open ties had been soaked up. By this time, encour- markets and enough finance, the industry could aged by the buoyant demand if not by the more than double in size in half a dozen years. prevailing prices, several new projects were emerging, mainly in the Middle East. Project The traditional markets in Asia will be unable to sponsors were heard to say that the buyers absorb the potential supply before 2015--or even needed the LNG and that prices would there- 2020. The Pacific projects, all advertising startup fore have to rise to make new projects eco- dates between 2001 and 2005 (though with vary- nomic. Qatargas, based on the enormous ing degrees of unreality), face an unpalatable reserves of Qatar's North Field, got in ahead of prospect. The performance of an exploration com- any real competition and sold 4 million metric pany depends not only on its ability to find oil and tons a year to Chubu Electric in Japan, quickly gas but also on its ability to commercialize dis- followed by 2 million more to seven other coveries as rapidly as possible. In a highly capital- Japanese buyers. It had started up in 1997, eight intensive industry the discount rate relentlessly years after the last greenfield project in ticks away value. Something must be done to res- Australia, and was supported by a guaranteed cue the projects. Three conclusions are emerging. minimum price (or so it appeared). First, the projects must be made more competi- Not surprisingly, buyers were resistant to higher tive, not just with one another but now also prices, and Japanese power companies shifted against low oil prices. This conclusion is being their preferences toward coal. Some of the pro- accepted only reluctantly. LNG projects have ject sponsors started to consider whether costs scarcely had to compete with one another in the could be reduced to make greenfield plants eco- past and have generally had little problem com- nomic without increasing prices. peting with oil at US$18 or more a barrel. For most of the life of the LNG industry the available gas Then demand growth started to ease, at least in barely sufficed to meet the needs of importers. the Japanese market, coming to a crashing halt in Competition takes three main forms: competing 1998. Yet gas continued to be found, and prospec- on cost, offering more market-friendly terms, and tive projects to increase. By 1995 it had become calling on established buyer relationships. apparent that there was more LNG than the tradi- tional markets could absorb. Projects would have Second, new markets must be opened up. Oman to become more competitive and find new mar- LNG tried to open a new market in Thailand but kets. Nevertheless, on the strength of soaring ultimately failed against competition from pip- Korean demand, two new projects, Oman LNG eline gas imports and the contracting economy. and Ras Laffan, will start up this year. The emphasis is now on India and China. Meanwhile, there was at last some activity in the Third, producers can get out while the going is Atlantic basin. After some thirty years of trying, good. BP is the only producer to have done this, the Nigerian project finally began to supply LNG selling its gas resources in Papua New Guinea. buyers in Europe. And a rejuvenated Trinidad Two other projects, Pac Rim in Canada and project will supply the U.S. market as well as Cristóbal Colón in Venezuela, have stopped Spain. All four projects point in new directions, trying to market LNG, having failed to put 4 Trends and Markets in Liquefied Natural Gas economic schemes together. In all these cases tage can be eroded. Moreover, the number of there are possible alternative uses for the gas. used ships available has declined, while the demand has increased to the point where the Finding ways to compete on cost benefits of used ships have virtually disappeared. LNG is forced to be more competitive in the Plant cost Atlantic trade than in the Pacific. There is com- petition from pipeline gas in the target markets Although BP was probably the first to call in Europe and the United States, and prices are attention to the need for improving the cost and lower than in East Asia. Not surprisingly, the economic performance of new LNG schemes, Nigerian and Trinidad projects lead the way in Trinidad made the real breakthrough. Although the pursuit of low cost. new to LNG exports, the Trinidad partners were determined not to build a high-cost plant. They Shipping applied the cost savings lessons that low oil prices had forced on offshore developments in Nigeria's main innovation was to use idle ships. such high-cost areas as the North Sea. At the same For many years there has been a pool of unem- time Phillips, with Bechtel, was attempting to ployed ships, built speculatively on the assump- market an updated version of the cascade lique- tion that a spot trade in LNG would develop, or faction technology developed for the early freed up by the failed Algeria-U.S. project or the Alaskan plant and not used since. The Trinidad failed Indonesia-California project. Because buy- team not only produced a design suited for the ers in Japan insisted on new ships for new trades, purpose, it sought bids for two front-end engi- the ships languished except for occasional short- neering design contracts, one for the Phillips term charters. technology and the other for the APCI process that has been used for all other recent plants. This Shell acquired some of these ships cheaply for strategy enabled it to obtain truly competitive Nigeria LNG at the end of the 1980s--well before bids for the main contract for plant construction. the project needed them, as it turned out. It was a brave move that paid off handsomely in the end. The results were startling. All the bids came in at Nigeria LNG earned enough from short-term char- less than US$250 per metric ton a year of installed ters to cover the cost of the ships before they capacity--30 to 40 percent less than the costs in reached Nigeria. And high demand for LNG in the late 1980s. The Phillips cascade technology Japan and Korea that could be supplied from spare probably had little to do with the low bids. The capacity in existing LNG plants created a need for real breakthroughs were in design philosophy ships that Shell was only too pleased to fill. and, perhaps most important, in engendering real competition among the contractors. In most LNG The Trinidad project has also benefited from the projects the construction contract goes to the con- use of secondhand ships, two retired from the tractor that carries out the front-end engineering Phillips Marathon Alaska-Tokyo route, one from design because of its significant information the Abu Dhabi project, and one of the U.S. Marad advantage. And since there are few contractors, ships, now owned by Trinidad partner Cabot. But the advantages of bidding have been limited. under normal conditions, cheap secondhand ships are not necessarily the bargain they appear The producers in the Pacific basin do not seem to be. Usually they have to be acquired well to have fully absorbed the lessons of Trinidad, before they are needed, and they need upgrad- although both RasGas and Oman LNG benefited ing to ensure that they last for the life of the new from relatively low bid prices, possibly from project, or at least for much of it. If they then must contractors trying to avoid losing out again. be laid up for a year or two, the initial cost advan- Shell, probably the leading LNG supply com- The World Bank Group 5 pany, is pursuing its own route to cost reduc- Attention shifted to India and China, but both tion, largely through scale economies. It is talk- present formidable obstacles to establishing a ing of single liquefaction trains approaching market for LNG. In traditional markets LNG can capacity of 4 million metric tons a year. Not only rely on powerful, creditworthy buyers that can is this an unwieldy scale for a project, Shell also underwrite a twenty-five-year take-or-pay con- appears to be struggling to get costs down to tract. No such buyers exist in the new markets. US$250 per metric ton a year. Neither country has a fully convertible currency. There is virtually no gas infrastructure, particu- Financing larly in the target areas for LNG. But there is huge potential demand, particularly in power genera- Financing has seen some innovation, although tion, a sector in crisis in both countries. With the not all the developments have been positive. traditional route to developing LNG trades Even with highly creditworthy buyers, most of closed, a new way of conducting business had the early LNG projects were equity (or at least to be found. shareholder) financed, and it was generally large oil companies that developed LNG schemes. Initially, supply to independent power producers More recently project financing has increased, (IPPs) in India was expected to be an easy mar- presumably because companies with smaller bal- ket to develop. But finance proved to be an obsta- ance sheets are becoming involved. Project fin- cle. IPPs are generally project financed and rely ancing is not a cost savings route and is also time on long-term electricity sales contracts. In India consuming. Nigeria LNG gave up its attempts to most state electricity companies supply electricity raise project finance and reverted to equity below cost to the rural sector and are loss mak- financing. RasGas moved to bond financing, rais- ing. The federal government is unwilling to issue ing US$1.2 billion on the U.S. bond market at sovereign guarantees. And the complexities of remarkably good rates. But Korea's economic dual project financing--with an IPP at one end of problems and the decline in its debt rating have the chain and a new LNG development at the led to a downgrading of the bonds' rating (al- other, and with different borrowers--are prob- though not below investment grade), with a cor- ably insurmountable. responding impact on their price. The bond route has probably closed for LNG finance, at least tem- A second obstacle was the lack of gas infra- porarily. Oman LNG had intended to go that structure in all but a limited area. There was no route but changed course after the East Asian obvious strong utility company to buy LNG and financial crisis. develop the nonpower market. LNG sellers would probably have to get into local marketing, Opening new markets but while they were prepared to invest in receiv- ing terminals, few were willing to go much fur- With demand low in the main East Asian markets, ther. Yet LNG imports were unlikely to be limited the industry has tried to open up new markets in to power demand: even with a serious shortage Asia. India and China have always been seen as of electricity, demand at any one location would the main prizes, though the first progress was in not grow fast enough to fully load an LNG ter- Thailand, where Oman LNG and RasGas tried to minal (or a large LNG plant) quickly enough. sell LNG. Price was a sticking point: pipeline gas set a marker, and Thailand wanted indexation Complicating the situation in India, the host gov- linked to coal for power generation. Oman LNG ernments have tended to put the cart before the proved more flexible on this point and a deal was horse, calling for tenders for LNG supply before concluded in principle, only to be overturned as tackling the market absorption and finance ques- more pipeline imports appeared and Thai de- tions. Because of the complexities of LNG devel- mand collapsed. opment, no tender can be unconditional on either 6 Trends and Markets in Liquefied Natural Gas side; usually there are major reservations by both Gulf than from any of the Pacific Rim projects parties over financing, timing, and commitment (except Arun to Ennore). Competition among by the other side. So the process has been of dubi- Gulf producers should keep their f.o.b. (free on ous value, at best only an invitation to negotiate. board) prices close to the equivalent netback from traditional buyers, making it unattractive for Enron's scheme to supply Dabhol Power Com- Pacific-based projects to compete in the Indian pany, in India, will probably be the first LNG market. But the Gulf-based projects suffer a freight supply in either India or China. It appears that disadvantage in supplying Japan and Korea. this scheme will be able to use the tested method for opening new Asian LNG markets--taking Less progress has been made in China, despite spare capacity from existing projects, in this case intensive study of the market by several potential Oman and Abu Dhabi. The suppliers will prob- suppliers, including Shell and Mobil. Most attrac- ably have to take more risk and provide more tive is the fast-growing coastal strip between contract flexibility than in traditional contracts. Guangdong and Shanghai. With Shanghai now The saving graces: the demand is apparent, and appearing to be within economic reach of the state is prepared to give some support to the pipeline gas from Siberia, the focus has shifted to state electricity board. the Guandong area. The government has said that it favors LNG imports and has called for a major Enron will have a major stake in the receiving feasibility study, but nothing will happen until this terminal and power plant, but not in LNG sup- study has been completed. The stronger central ply. The company also plans to market gas to control in China lends a different flavor than in other industries in the region. Financial closure, India, but many of the same issues will have to be the key step, is reported to be imminent. faced and there seems to be no prospect of cen- tral government guarantees to support imports. Elsewhere in India, a different approach is being tried by the Petronet group, which includes most Weathering liberalization in of the largest oil and gas companies in India, pre- established markets sumably in an attempt to assemble stronger cred- itworthiness. This group called for tenders to In the Japanese market the effects of the eco- supply 7.5 million metric tons a year and to be nomic downturn on energy demand and LNG involved in the receiving terminals at several loca- prices may be short term and coped with fairly tions. RasGas won the bid and is reported to be easily, but the effects of liberalization in the gas moving toward a sales contract. But the scheme and, particularly, the power sectors are essen- raises all sorts of questions and there is a long way tially unpredictable. Power buyers in particular to go. A major expansion for RasGas, it will have cannot be sure of their future market share, to be financed, and the two sides will have to which makes it distinctly risky for them to make work out an acceptable way of distributing the long-term take-or-pay commitments and favors risks. Even so, with Qatar boasting in December fuels that can be purchased as and when needed. of 4 million metric tons a year of spare capacity, Small wonder that Japanese buyers have appar- even the Petronet project will not rely entirely on ently decided to take on minimal new long-term new LNG. RasGas also won a bid for the Tamil LNG commitments. But this is largely a problem Nadu state project planned for Ennore (this time of transition. In the long run a liquid spot mar- as the fuel source for a group interested in invest- ket should remove the volume risk even for gas, ing in the terminal and associated power plant). as it has in the United States. Even so, it takes years for such a market to develop, and LNG sell- That all the supplies for India originate in the ers could face a decade of uncertainty. The Middle East is not insignificant. The shipping dis- future is further clouded by the emissions reduc- tance to India is considerably shorter from the tion obligations Japan accepted at Kyoto, which The World Bank Group 7 tend to put fuel choices in conflict with those that European gas markets are also under pressure to follow from liberalization. liberalize--pressure that is being strongly resisted in some quarters. Liberalization does not sit eas- The Korean LNG market--highly seasonal and ily with the traditional way of trading LNG, which bedeviled by conflict between the two users of relies heavily on long-term contracts and take-or- LNG, KEPCO (Korea Power) and KOGAS (Korea pay. In the longer run, with liquid trading systems Gas)--has suffered a decline that has been exac- removing volume risk, long-term contracts and erbated by the conflict. This decline led to rephas- take-or-pay can be combined with a floating gas ing of some contracted purchases and to concern market price, but the uncertainties of the transi- about Korea's capacity to absorb contract volumes tion are unsettling. Moreover, liberalization usu- from Rasgas and Oman LNG that start this year. ally pushes prices down--an uneasy prospect for But Korea, which has done more to put its eco- a high-cost source of supply. nomic house in order than most countries in the region, should be able to meet its contractual Price wars? obligations. LNG pricing is an area where novelty and tradi- In Korea too liberalization is in the air, but the tion are likely to come into conflict--with unpre- timing and extent are uncertain. POSCO, a major dictable results. steel maker, will be allowed to build a terminal and import LNG for use in electricity generation, In Europe LNG needs to compete with pipeline mainly for its own use. Whether POSCO will gas at the point of entry, and current prices are cooperate with KOGAS to avoid worsening the low enough to frighten all but the very brave or problems of temporary oversupply and seasonal foolhardy. The big questions for the future are storage remains to be seen. KOGAS has been how long gas prices will be coupled to oil prices planning a third terminal of its own, and there in continental Europe and what will happen to seems no need for both this and a POSCO ter- gas prices when there is a decoupling. In North minal. Nor does there appear to be any immedi- America and Britain decoupling has tended to ate need for newly contracted supply to meet reduce prices. POSCO's requirements. The problems of sea- sonal supply and demand could be addressed in In Asia the pricing signals also point to innovation, several ways, including introducing interruptible and perhaps confusion. After nearly three decades industrial tariffs that would reduce summer val- in which prices in Japan, Korea, and Taiwan leys and thus increase total supply. And there is (China) moved in parallel (under the general con- inherent unsatisfied demand that new initiatives trol of Japanese buyers), there are now seeds of could uncover. real competition among suppliers. The "floor price" that was essentially agreed for Qatargas sup- Taiwan (China) has suffered little from the eco- plies has already been dropped (RasGas dropped nomic disturbances in the region. Here too there a parallel provision for Korea in order to enlarge are thoughts of energy liberalization. There are the supply contract), and the recent results of price also new IPPs, and severe strains in the rela- renegotiation with existing suppliers suggest that tionship between CPC, the government-owned the apparently inexorable upward creep of prices monopoly operator for both oil and gas, and has been halted and probably reversed. Taipower, the government-owned power utility. Political positions will take time to unravel, and The traditional pricing formulas ensure that LNG as the future growth of LNG supply depends becomes less competitive with oil at low oil largely on the timing, cost, and ownership of the prices, causing gas companies to suffer and dis- proposed second LNG terminal, the watchword couraging power companies from using any is "wait and see." more LNG than their contracts call for. If low oil 8 Trends and Markets in Liquefied Natural Gas prices persist, there may be pressures for price tial for freight saving deals (although the benefit is changes that are difficult to resist. And if East not easy to capture). Clearly, a more flexible trad- Asian buyers overcome their reluctance to buy ing pattern would benefit both buyers and sellers, new LNG, competition could result in a new, but extreme caution, particularly among the more buyer-friendly deal, giving established Japanese buyers, has inhibited its development. buyers a new yardstick and a reason to renego- tiate across the board. The opening of new markets such as India could also lead to a more flexible trade. The players in East Asian pricing structures may also come the chain have to accept more risk that the mar- under strain as a result of deals in India, where ket will not perform as expected. To deal more novel pricing structures and levels are being pro- flexibly with the Indian market, they may look posed. For example, the winning bid for Ennore at alternative ways of disposing of surplus LNG is reported to offer LNG at a fixed price, a rather or acquiring LNG on short notice. But for this to eccentric choice. Indexation, which is more be a real option requires a market of last resort. closely tied to real competition in the end use In winter there is generally a market in Europe market, must be a real possibility. The ultimate that could absorb some surplus LNG at a rea- end is a price linked to gas prices in a liberalized sonable price, as long as there is shipping capa- and liquid gas market (as in the United States). city to get it there. The only truly liquid gas But this is a long way off, and how the industry market, however, is in the United States. It is gets there will be interesting to watch. probably too far and its price too low to support Viewpoint is an open forum intended to an Indian or Pacific traded market, but it could encourage Will spot trading develop? provide the Atlantic trade with yet another dissemination of and opportunity to innovate. There are signs that this debate on ideas, innovations, and best So far there has been no real spot trading in LNG, opportunity is being exploited. Cabot, one of the practices for expanding although there have been many short-term deals Trinidad partners and also the U.S. buyer of the private sector. The between established buyers and sellers based on Trinidad gas, has on-sold part of its supply to an views published are those of the authors and spare plant and shipping capacity. Nevertheless, IPP in Puerto Rico promoted by Enron. Thus a should not be attributed several forces could lead to more extensive trad- new market has been opened using LNG whose to the World Bank or any ing that might just result in a spot market. development was underwritten by a sale into the of its affiliated organiza- tions. Nor do any of the United States. This might eventually evolve into conclusions represent The first is the Korean seasonality problem. Gas a much more flexible trade, although there are official policy of the demand in Korea has a strong winter peak, but still many obstacles to overcome, not least the World Bank or of its Executive Directors or LNG contracts require even deliveries through the availability of adequate shipping capacity. the countries they year. This can be handled in part (though expen- represent. sively) by storage. But the growth of the Korean Conclusion To order additional market has threatened to exceed the storage copies please call capacity and the country has run perilously close These are obviously difficult times for LNG. But 202 458 1111 or contact to stock-outs in winter. Much of the Korean sup- they are also exciting times. Difficulties lead to Suzanne Smith, editor, Room F11K-208, ply has been based on short-term supplies, and new ideas and to attempts to rewrite the rules. The World Bank, these can be biased toward winter, although not Not all these efforts will succeed, of course, but 1818 H Street, NW, without diverting some cargoes originally there is certainly plenty to maintain the interest Washington, D.C. 20433, or Internet address intended for Japan. Japanese buyers have been in the LNG business today. ssmith7@worldbank.org. reluctant to participate in swaps, but in the current The series is also market stress Osaka Gas has provided Korea with available on-line This Note was prepared with the help of James Ball and David (www.worldbank.org/ a winter cargo this year. There is an obvious syn- Spottiswoode, Gas Strategies, London. html/fpd/notes/). ergy with Taiwan (China), where the load peaks in summer, but the buyers have not organized to Rob Shepherd (R.Shepherd@gas-strategies.com), Printed on recycled paper. take advantage of it yet. There is also much poten- Gas Strategies, London