LNG Shipping: a Descriptive Analysis By Susanna Dorigoni, Luigi Mazzei, Federico Pontoni, and Antonio Sileo IEFE – Centre for Research on Energy and Environmental Economics and Policy, Università Bocconi, Milan, Italy Phone: +39-02.5836.3817

e-mail: [email protected]

Abstract At present, throughout the world LNG investments are abounding. In particular, the US and the EU are trying to facilitate their security of natural gas supply. For this reason, LNG enables importers to extend their gas suppliers’ portfolio, considering that some producing countries can be reached only via sea transport. LNG will thus increase importers’ choice; at the same time it can widen the group of exporting countries and enhance the construction of a global gas market. In the LNG value chain one can identify three elements: liquefaction, shipping and regasification. While the first and the last element of the value chain have been deeply studied, too little attention has been paid to shipping. Nevertheless, the number of operative ships will be one of the key variables for the increase of an effective spot market and, consequently, for a greater market liquidity. The goal of this paper is thus clear: will LNG shipping be able to sustain a liquid market? The findings are not straightforward. In fact, while on the one hand shipping is not likely to represent a bottleneck in the foreseeable future, at least for what concerns transport capacity, on the other hand most of the ships are bound to long-term contracts and this can be detrimental to competition. Moreover, shipping is often part of vertical integrated projects. Thus, authors cannot hypothesize that in the short term spot sales will grow, as it happened with oil tankers.

Introduction: the increasing demand for LNG In these last few years one of the main concerns of both the EU and the US has been that of facilitating the safeguard of raw materials’ security of supply, especially that of natural gas. Notably, the European Council has identified a list of instruments to strengthen its security of supply: as for natural gas, the LNG chain has been recognized as the most secure means of transportation (Directive 67/04, concerning the “Measures to safeguard security of natural gas supply”). Liquefied natural gas (LNG) is natural gas that has been cooled to the point that it condenses to a liquid, which occurs at a temperature of approximately -256° F (-161° C) and at atmospheric pressure. Liquefaction reduces the volume by approximately 600 times, thus making it more economical to transport between continents in specially designed ocean vessels, whereas traditional pipeline transportation systems would be less economically attractive and could be technically or politically infeasible. Thus, LNG technology makes natural gas available throughout the world. To make LNG available, energy companies must invest in a number of different operations that are highly linked and dependent upon one another. The major stages of the LNG value chain consist of the following: • Liquefaction, to convert natural gas into a liquid state so that it can be transported in ships; • Shipping the LNG in special purpose vessels; • Regasification, to convert the LNG stored in specially made storage tanks, from the liquefied phase to the gaseous phase, ready to be moved to the final destination through the natural gas pipeline system. Indeed, import through LNG chain does not imply an indissoluble physical tie between producer and buyer, contrary to what happens with pipelines (Chernyavs’ka e Dorigoni, 2002). In other words, the investment in a pipeline is very specific: according to Williamson (1985), the greater is the switching cost for an alternative use of any asset; the greater is its specificity. As for pipelines, their degree of specificity is maximum (the “site specificity” kind). Given this specificity, the counterparts recognize as vital the continuity of their contractual relation. That is why any such agreement will guarantee strong safeguards to each party, which would be unnecessary for more common neoclassic (nonspecific) transactions. For the natural gas market, these safeguards take the form of longterm agreements with minimum off-take requirements (take or pay clauses), designed to preserve counterparts from ex-post contractual opportunism (hold-up problem), that is really likely in these circumstances. Such contracts definitely contribute to the “cartelization” of the market, hindering competition. As said before, the LNG chain presents a much lower degree of specificity: in fact, even though the construction of a regasification plant is generally tied to the stipulation of a long-term agreement (with take or pay clause), once the contract is expired and the investment is sunk, the importer may satisfy his gas supply needs on the basis of his relative gains. Moreover, LNG costs have significantly decreased over time, thanks to the technological innovation (Oil&Gas Journal 2006); at the same time, it is getting increasingly common that part of regasification plant capacity is made available for spot transactions (in some countries this is a regulatory requirement, CEER, 2006).

As far as LNG import contractual practices are concerned, significant changes have started to take place in the last few years, both in terms of agreements’ length – average duration has significantly decreased – and in terms of price indexation – in the most developed markets LNG price is tied to gas spot price (IEA, 2006). As for LNG costs, it is important to stress that some recent studies (such as IEFE, 2008) have demonstrated that from early 2007 there has been a clear reversal of trend. The main drivers for this turning point are the increase in raw materials costs and the oligopoly conditions that characterize the terminal construction industry, whose main players are already booked for the next three to four years. Nevertheless, this increase in costs concerns also pipeline costs, leaving the relative competitiveness unchanged. Another advantage given by LNG is that liquefied gas enables importers to widen their suppliers’ portfolio, considering that some producing countries (i.e. stranded gas) can be reached only via sea. A wider supply portfolio mixed with an increased integration of final markets, given by the possibility of redirecting cargoes depending on single countries’ supply-demand balance, would contribute decisively both to the security of supply and to the enhancement of competition (IEA, 2004). It is noteworthy to specify, though, that when referring to positive effects of LNG on competition, the authors mean competition between importers and only to a lesser extent between producers: in fact, upstream competition is not likely to occur, given the chronic deficit of liquefaction capacity (IEFE, 2008). Anyway, even if a bigger recourse to LNG would cause an increase in supply prices, this would be more than offset by sharp reductions in final prices, as shown in Graziano et al. (2008). When considering the possible advantages of LNG over pipeline transportation, it should be stressed that import via tanker is competitive for medium and long distances (the majority of experts indicate a critical threshold of 4,000-5,000 km, IEFE 2008). Anyway, given the current progressive depletion of both European and North American gas fields, it will be necessary to search for farther gas fields, so that pipeline and LNG costs will be increasingly closer. So far, the LNG stage that has received too little attention is shipping. Being the link between the producing/exporting country and the importing one, and having been subject to major changes in the last few years, it is particularly interesting to analyze it singularly, in order to understand how it is linked to the other two stages of the chain. In particular, it is important to comprehend the evolutionary trajectory this segment may take in the future: the number of operative ships will be one of the key variables for the growth of the spot market and, consequently, for greater market liquidity. The paper is organized as follows: in the first section LNG market will be introduced, with particular reference to the evolution of regasification and export capacity; section 2 and 3 will be dedicated to an analysis of shipping, both in current and in perspective terms. Then, in section 4 the proprietary and contractual situation will be addressed; section 5, finally, will present some concluding remarks.

3

1. International LNG trade In 2007, the world consumed more than 3.000 Bcm; 775 Bcm were exported, 550 Bcm via pipelines and 225 Bcm via LNG. At present LNG accounts for almost 30% of global natural gas exports. Japan and Spain alone imported about 115 Bcm (89 Bcm Japan and 26 Bcm Spain); thus, these two countries account for about 51% of world LNG imports. The main exporting countries are Qatar (39 Bcm) and Malaysia (30 Bcm). As we noted before, LNG market is witnessing a strong expansion; therefore, it should be analysed starting from producing countries’ export potential. With regard to this issue, we show below a table presenting IEFE hypothesis about liquefaction capacity for each producing basin for 2010, 2015 and 2020. Table 1: Current and expected liquefaction capacity. Data in Bcm. Area 2010 2015 2020 Atlantic Basin 119 208 226 Middle East 115 137 191 Pacific Basin 139 178 238 TOTAL 373 523 655 Source: IEFE 2008.

As we can see from table 1, between 2010 and 2020, world liquefaction capacity will increase by about 75%. Nigeria, Qatar – which, given its geographical position, will probably become the LNG world swing producer (Wagbara 2007) – and Australia alone will hold about 35% of total liquefaction capacity. Nevertheless, it must be stressed once more that this significant increase in liquefaction capacity is only potential: technical and economic constraints may affect the expected growth rate. As said before, to better assess the evolution of liquefaction capacity, authors have divided countries for their ocean basin of origin. This reflects the natural subdivision of LNG world market, which is due to the distances between producing countries and importing countries (the key driver for shipping costs): for instance, Europe and the US are supplied traditionally by Middle East and Atlantic countries. Nevertheless, this splitting may be partially overcome in the future, thanks to attractive arbitrage opportunities, fostered by the possible offsetting of differences in final prices and by the reduction in transport costs: in this case, LNG would be sold to markets with the highest willingness to pay and a proper global LNG market would then be established. The last stage of the LNG chain, regasification, is experiencing abounding investments, especially in Europe and in the United States. Despite the significant investments in liquefaction, the gap between regasification and liquefaction capacity is expected to grow remarkably, so that by 2020 the ratio between the two capacities could be 2:1 (see Figure 1). Yet, it should be stressed that many planned regasification plants lack of supply contracts, circumstance that put at risk their actual realization (while in the case of liquefaction terminals the problem does not hold: all the projects are covered by one or more long term contracts, at least for part of their expected production).

4

1400 Regasification

1200 1000

America Asia Europe

600

Middle East Liquefaction

800

400 200

Pacific Basin Atlantic Basin

0

2010

2015

2020

Figure 1: Expected liquefaction and regasificaction world capacity (Bcm). Source: IEFE 2008.

As a consequence, only little competition between importing countries will occur, thus leading to conservative estimates for the utilization ratio of world regasification capacity. 2. Shipping – preliminary remarks Likewise the other two stages of the LNG chain, shipping has undergone deep transformations in the last ten years. Moreover, ship construction costs are still falling (thanks to considerable economies of scale) in spite of the increase in raw materials prices. In other words, the biggest technological improvements are concentrated in shipping: thus, while both regasification and liquefaction are suffering an unexpected cost increase, shipping remains the key driver for LNG competitiveness (IEFE 2008). Construction cost reduction has brought about an increase in average ship size, that makes it possible to have a less numerous fleet to deliver LNG cargoes; meanwhile, it has become convenient to be supplied by farther liquefaction plants. To analyze the cost sensitivity of shipping to distance, we show below a scatter plot between distance and transport capacity. As one can see from the scatter plot below, the average distance is about 5,700 km. It will be shown later how the average distance changes depending on the routes considered. With regard to unitary transport cost, according to a IEFE survey (2008), it is about 0.6 eurocent for the first 1,000 km, and then it increases of about 0.2 eurocent every 1,000 km. This cost has been estimated considering tankers of 150,000 liquids cm with an overall construction cost of about 220 millions euro and an estimated life of 25 years.

5

Figure 2: LNG tankers capacity and distance covered. Source: Authors’ elaboration on Drewry data 2008.

At present, shipping does not represent a bottleneck to the market, but it should be borne in mind that also shipping may become an element of inflexibility in the LNG industry, given the amount of investments in liquefaction and regasification facilities and the entrance of new players in the industry. The reason for this is that an increase in the number of operators may complicate the coordination of investments all along the value chain (scheduling problems). The first problem is, thus, to evaluate if shipping capacity will catch up with liquefaction. 3. Current situation and future development Hereunder a table showing the current size of LNG tankers fleet and the number of ships under construction 1 (as of August 2008) is presented.

1

As a proxy, the number of European companies that have booked LNG tankers has been used.

6

Table 2: Data about cryogenic ships. Existing Of which serving the European market Under Construction Of which serving the European market Source: Drewry 2008.

Ships 278 57 89 17

Total Capacity (Mcm) 37.3 7.3 15.1 3.9

The previous table also shows the existing total transport capacity (given by the 278 LNG existent tankers) and the transport capacity forecasted for 2010, relative to further 89 ships (currently under construction). To verify that the transport capacity expected for 2010 is sufficient when compared to the liquefaction capacity forecasted for the same year, authors have made a proportion equating the ratio of liquefaction capacity to transport capacity at 2008 and 2010. The minimum transport capacity required relative to the liquefaction capacity is determined on the basis of the following formula:

Qmin =

L2010 × Q2008 , L2008

[1]

where Q stands for transport capacity and L for liquefaction capacity. Table 3: Forecasted capacity – minimum capacity. Data in Mton. 2008 Liquefaction capacity (existing and 203.9 foreseen) Transport capacity (existing and 16.4 foreseen) Minimum capacity required Source: Authors’ elaboration on Drewry data 2008.

2010 272.5 23.0 18.6

As shown in the table above, forecasted transport capacity is sufficient for handling the volume of natural gas (net of load limits) that will supposedly be liquefied. From this point of view, there will not be supply-side bottlenecks, at least at an aggregated level. Thus, we can say that there will not be problems in transporting LNG, which will be sold in most valuable markets. Hereafter, detailed analysis on the LNG tankers currently operating will be presented. From a technical point of view, 96 of the LNG existing carriers are provided with double membrane steel tank integral with the hull. The structure of this kind of tanks does not assure the uniformity of the inner pressure. During its transportation, the cargo is subject to continuous solicitation caused by the rolling of the ship. This phenomenon in known as “sloshing”: these solicitations are offloaded directly onto the hull compromising its service life (maximum 40 years) and amplifying the differences in tanks pressure, provoking cargo losses and the consequent boil-off. For these reasons and to ensure safe trips, this kind of

7

ships are subject to cargo limitation (usually, with a maximum capacity threshold of 80% of the total volume). Improvements in this technology have been represented by the introduction of prismatic membrane tanks. The latter are as usual made out of steel, but the prismatic shape tending to a spherical one ensure a more uniform pressure and less liquid gas boil-off-related losses; in this case, cargo limitations are smaller (maximum capacity threshold of 95% of the total volume). Spherical tanks are self-supported, not integral with the hull, but propped up by a cylindrical supporting structure. As a consequence, solicitations are not offloaded directly onto the hull. The employment of aluminium, a material characterized by a considerable elasticity, and tanks sphericity ensure a more uniform pressure inside the tank, limiting the sloshing. This technology ensures a greater stability and does not require cargo limitations. It must be recalled that each tank construction technique requires different costs and construction periods: generally, ships equipped with spherical tanks cost more and require longer construction periods, generally they cost 20% to 30% more than ships with membrane tanks. The main shipyards specialized in LNG carriers are located in Japan, Korea, Finland, France and Spain. LNG carriers delivery time depend on the technology adopted and on the market situation. For the construction of the most common size of LNG Carriers, that is, 145,000 cm, 24-26 months are required, if equipped with prismatic membrane tanks, and 26-28 months if equipped with spherical tanks. Prices range from 180-200 million dollars for a tanker equipped with prismatic membrane tanks to 200-220 million dollars for a ship equipped with spherical tanks. 3.1 Ship size and market segmentation After these previous general remarks, it is important to focus on ship size, on their usual routes and on the related ship-shore issues. In particular, it is extremely important to analyze the features of the existing and under construction tankers to evaluate the future market segmentation and, basically, the role of shipping in limiting or emphasizing the potential positive effects of LNG. In the figure below, world LNG carriers’ fleet is grouped by storage capacity.

8

Figure 3: Ship size frequency distribution. Data in Mcm. Source: Authors’ elaboration on Drewry data 2008.

As one can see in figure 3, most of the LNG carriers present a size in the range 120,000 cm150,000 cm. This happens for two reasons: the first is that most of them have been built in the last 15 years, when this size represented the technological standard; at the same time, just few regasification terminals can receive ships bigger than this size. As far as small LNG tankers, it should be stressed that not necessarily they are the oldest, since even recently a Japanese firm has booked a ship with a capacity of less than 20,000 cm. The relationship between year of construction and ship size will be addressed later. In the table below, in confirmation of what said before, receiving terminals capacity (both existing and under construction) is reported. At present, only few terminals (all in the UK, except one in the US) can receive ships larger than 200,000 cm. Table 4: Maximum receiving capacity of existing regasification terminals. Data in liquids cm. TERMINAL COUNTRY MAX CAPACITY Lake Charles US 160,000 Elba US 145,000 Everett US 126,000 Teeside UK 250,000 Barcellona Spain 140,000 Zeebrugge Belgium 135,000 Panigaglia Spain 40,000 Source: IEFE 2008.

9

As far as under construction terminals are concerned, only those being built in the US will be able to receive technological frontier ships. Therefore, it can be said that the average ship size is greatly affected by regasification plants capacity. Table 5: Maximum receiving capacity of under construction regasification terminals. Data in liquids cm. TERMINAL COUNTRY MAX CAPACITY Sabine Pass US 250.000 Golden Pass US 250.000 Free Port US 250.000 South Hook UK 250.000 Fos Cavou France 160.000 Rovigo Italy 150.000 El Ferrol Spain 140.000 Source: IEFE 2008.

To better evaluate the increase in ship size, authors have divided LNG tankers by their period of construction. Hereunder a table and a box plot regarding ship size evolution over time are shown. The characteristics of each group are the followings: group 1, which includes ships built before 1990; group 2, with ships built between 1990 and 2000 and group 3, with ships built in the last 8 years. One can appreciate from the table hereunder that ship size has constantly increased over time. Table 6: Ship size evolution. N° of ships built Average size Max size Min size

1 63 108.5 133.0 29.6

2 48 119.1 138.3 18.9

3 167 148.4 266.0 19.1

Source: Authors’ elaboration on Drewry data 2008.

The most interesting figures to highlight concern the number of ship for each period: in these last 8 years more ships entered the market than those built in the previous 35 years. It is worth also noting that ships of modest size have been ordered also in the last few years, as can be seen in the figure hereunder.

10

Figure 4: Ship size box plot (ships grouped by period of construction, outliers labelled by importing countries). Source: Authors’ elaboration on Drewry data 2008.

As it is evident in the figure above, the great majority of the ships built from 1990 onwards have capacities around 130,000 cm, while capacity of ships in group 1 has a bigger variance; the explanation for the latter pattern is that the first group includes ships built in a period lasting 20 years. The main finding derived from the above reproduced box plot concerns the outliers, labelled with importing country’s name. As for under-dimensioned ships, Japanese firms appear as the main buyers of this kind of ships; this finding can be explained considering that some of the many Japanese regasification plants (whose total number is 25) have tiny capacities and can admit only small-sized ships. These small-sized regasification plants are needed in the Asiatic country since they represent the only mean to fulfil peak gas demand, often represented by winter residential consumption and power generation demand. In fact, some small regasification plants are owned by electric companies. As for Spain, Italy and France, these are the first European countries to have imported LNG, which is one of the reasons why they have some small capacity regasification plants. As for Italy, there is only one terminal, about 30 years old, able to admit only ships with capacity of 40,000 liquid cm.

11

It is no wonder that most of the tankers with capacity greater than 200,000 cm, instead, are directed to Great Britain: in fact, British regasification plants have all been built after 2003 and have exploited at most economies of scale. Hereunder, the box plot shown earlier in the article is reported, this time with outliers labelled by exporting country.

Figure 5: Ship size box plot (ships grouped by period of construction, outliers labelled by exporting countries). Source: Authors’ elaboration on Drewry data 2008.

The box plot shown clearly highlights that small capacity ships directed to Japan come from Malaysia, Indonesia and Alaska; small capacity ships directed to Europe, instead, transport gas coming from Algeria. Finally, tankers with capacity over 200,000 cm that reach British and American regasification plants come from Qatar. These data suggest authors to analyze more deeply the relationship between routes, age and size of LNG tankers. First of all, it is possible to notice that 208 ships, out of 278 (thus, 75%), have a usual route from a single exporting country to a single importing country. The remaining 70 tankers have multiple destinations (51) or are currently uncontracted (19). As for routes distribution, the most frequent one is Malaysia to Japan (18 tankers); in the table hereunder, the main three routes for number of tankers and transport capacity are shown.

12

Table 7: Three main LNG routes. 1 2 For number of Malaysia - Japan (18) Indonesia - Japan (17) tankers For transport Australia - Japan Indonesia - Japan capacity (2.3 Mcm) (2 Mcm) Source: Authors’ elaboration on Drewry data 2008.

3 Australia - Japan (17) Malaysia - Japan (1.9 Mcm)

As can be seen in the figure above, the main three routes are all circumscribed to the Pacific basin; in particular, 58 ships (21% of the total number of tankers), with a total transport capacity of about 6 Mcm (about 17% of total transport capacity), usually move from Australia, Malaysia and Indonesia to Japan. A matrix of LNG routes divided by geographical area is presented below. Table 8: Number of ships dedicated to each trade route. TO Asia US FROM Pacific Basin 78 0 Middle East 48 10 Atlantic Basin 2 15 TOTAL 128 25 Source: Authors’ elaboration on Drewry data 2008.

Europe

TOTAL

0 15 40 55

78 64 63 208

The main destination for LNG tankers is by far the Pacific Basin: in particular, 92 ships are dedicated to Japanese imports, much more than those dedicated to the second biggest importer of the area, which is Korea (30 ships). As for Europe, 55 tankers are regularly employed for import, of which 24 are habitually destined to Spanish regasification plants. As the above table highlights, tankers are quite evenly distributed as for their trades’ area of origin. Instead, as far as single countries are concerned, Qatar is by far the main country of origin for LNG tankers trades, with 51 outgoing ships for habitual trade routes and other 4 available for spot trades. The second biggest country for outgoing number of tankers is Algeria (24), closely followed by Indonesia (23). As for capacities, the total amount of cubic metres transported by tankers leaving from Qatar is about 9 Mcm (24% of total yearly quantity transported with LNG tankers), followed by Nigeria (3.3 Mcm, 9%). As far as receiving countries are concerned, the biggest one is Japan with more than 11 Mcm imported (share of about 31%), followed by US (5 Mcm) and Korea (4.1 Mcm). In table 9 we recap few data on LNG tankers grouped by habitual trade routes.

13

Table 9: Features of LNG tankers, grouped by habitual route. AVERAGE STANDARD MAXIMUM CAPACITY DEVIATION Pacific Basin– Asia 118,519 34.50 155,000 Atlantic Basin – Europe 109,434 39.34 147,200 Atlantic Basin– US 137,564 6.63 147,208 Middle East – Asia 140,382 11.86 216,200 Middle East – Europe 172,740 39.24 217,000 Middle East – US 218,825 16.78 266,000 Source: Authors’ elaboration on Drewry data 2008.

MINIMUM 18,927 29,588 125,858 135,000 126,277 210,100

Two clear trends stem from the above reported data: trade routes Pacific Basin-Asia and Atlantic Basin-Europe involve ships with capacities pretty different from one another, since these were the first routes established and those with the most countries involved. As for Middle East, its entrance in LNG world market is quite recent, as the (pretty big) average size of tankers leaving from the region can testify. As said before, Middle East seems destined to play an increasingly crucial role in this market. Now let us have a look at the box plot for tankers grouped by trade routes.

Figure 6: Ship size box plot (ships grouped by trade route). Source: Authors’ elaboration on Drewry data 2008.

14

As shown on the above reported box plot, the variance for ships dedicated to European import is much greater than that for ships bringing LNG to Asia and to the US. More, this graph highlights how strikingly small is the minimum size for ships taking LNG from Atlantic basin to Europe when compared with the minimum size for ships taking LNG from Middle East to the Old Continent. This can be explained by the different average length of the two routes: the route Middle East-Europe is about 5,200 km, while Atlantic basinEurope is about 2,900 km. Moreover, in a few cases distances are less than 2,000 km, as for instance the case of gas liquefied in Egypt and Algeria and transported to Mediterranean Europe regasification plants. Instead, ships serving the Asiatic market present a comparable capacity (about 120,000150,000 liquid cm). Only Japan has smaller tankers, since 12 out of the 89 ships serving the Japanese archipelago have a capacity of less than 100,000 cm. More, four of them have a capacity of less than 50,000 cm, even if covering regularly a distance of about 4,500 km, since they upload gas from Brunei or Malaysia. This demonstrates that the employment of small-sized ships is only needed in case of peak demand. Instead, tankers exporting LNG from the Middle East are all of big-sized. Nonetheless, the picture above shows several outliers for observations tagged Middle East-Asia; the reason for this is that the 46 tankers employed in this route are almost identical in terms of capacity, with a strikingly low variance. As for tankers serving the US market, their capacity present a really low variance within the same route covered: tankers usually covering the Atlantic basin-US route all present a capacity in the range 125,000-150,000 cm, while those covering the Middle East-US route (notably, all but one) have a capacity of around 210,000-220,000 cm. This homogeneity is due to the recent development of US LNG imports: all but three tankers have been built after 2003, thus in a very few years. Now, let us examine the main features of the 89 tankers under construction. First of all, average ship size will be of about 169,000 cm, 35,000 cm more than the average size of tankers built in the last 8 years. 28 tankers under construction out of 89 have been ordered by Qatargas: this confirms, once again, how important the role of this country in LNG world market is. Qatar is also the country who has orderd the biggest ships: in 2009 three 267,000 cm tankers will be operating, all destined to serve the British market. As highlighted in the following box plot, under construction ships’ interquartile range is 150,000-210,000 cm, which means that there is going to be a significant increase in operating ships size. Yet, it must be stressed that these big-size tankers will not be receivable by all regasification plants. This could represent a serious limit for the development of a spot market. More, European terminals risk not to be able to receive ships otherwise directed to the US, even in case European countries could spend more than US for that gas. The only three outliers are represented by two tankers serving the Chinese market, ordered for speculative reasons, and an Exmar tanker, destined to supply US with LNG coming from Trinidad.

15

Figure 7: Ship size box plot (tankers under construction). Source: Authors’ elaboration on Drewry data 2008.

To confirm what it has just been said, the table hereunder highlights that 68% of the ships under construction have a capacity equal or greater than 150,000 cm, making it impossible, for most of the existing import terminal, to admit these tankers. Moreover, small-sized ships would make the supply of big import terminal uneconomical. Table 10: Size of under construction tankers. Tanker size Number of tankers Less than 150,000 cm 26 150,000-200.000 cm 31 More than 200.000 32 TOTAL 89 Source: IEFE 2008.

Share 29% 35% 36% 100%

In conclusion, one can state that, even if there are no supply-side constraints, capacityrelated features of under construction tankers seem likely to extend in the future natural LNG market segmentation (historically related to routes and transport costs). Even if this can be positive in terms of security of supply, it reduces significantly the chances to redirect

16

LNG cargoes towards markets with higher user value and to create a global market for LNG. 4. Tankers’ proprietary and contractual situation Hereunder, the LNG tankers’ proprietary structure is analyzed in order to see whether the contractual situation can have any influence on the enhancement of LNG competitiveness. In the traditional contractual practices relative to LNG, tankers were dedicated exclusively to specific trades. Moreover, it was common practice to require newly built tankers for new supply contracts, which contributed to a relatively inflexible tanker fleet: in fact, if a tanker were to be idled for any reason, it was very difficult to find another charter for it and it was likely to be laid up. Moreover, this lack of contractual flexibility, binding tankers to long period contracts, prevented cross shipping, practice that enables an LNG trader to minimize his transport costs 2 , and that could become increasingly important in the future, given the growing geographical dispersion of supply sources and markets. What’s more, at the beginning of the LNG industry, the view was that LNG tankers had a limited effective life, not outlasting the terms of the original contract, and new vessels would need to be ordered if the contract was renewed. Since last decade, instead, flexibility was established as the new trend in LNG contracting; not only in supply contracts (with a smaller share of long-term take-or-pay contracts in favour of short-term contracts and spot sales, thus a smaller average contract length), but also in shipping contracts, favoured by the growth in world tanker fleet. First of all, it has been recognized that these tankers may have a useful life of as much as forty years, and they do not need to be replaced with every new contract extension. As a consequence, “second-hand tankers” began to appear in the market. Another trend has been that of a gradual elimination or curtailing of destination restriction terms: this allows multiple destinations to be associated to each tanker, and LNG buyers to divert their own surplus on the short-term market. Moreover, comparing world tanker fleet capacity with world gas demand, a chronicle surplus of capacity results (traditionally present in this market, see J.T. Jensen, 2006), that may support extensive short-term trading (contributing to the globalization of the market). It is evident that the availability of shipping surplus capacity (and of liquefaction surplus not yet contracted) makes it possible to transport LNG economically for distances much greater than those possible with long-term contract. Moreover, it is interesting to notice that there is already empirical evidence of speculative investments in LNG shipping sector (mainly by LNG majors); that is, tankers ordered to be employed in the short-term market. Yet, some of these majors have integrated downstream through self-contracting with their marketing affiliates: thus, tankers that have been ordered to shuttle between various majors-controlled liquefaction and import facilities, may not fit either the traditional definition of “dedicated” trade, nor can be classified as “speculative”. As for the proprietary situation, authors analyze the data for the 57 tankers serving primarily the European market. According to the data, 2/3 of the capacity is detained by a 2 For instance, if an LNG trader wanted to deliver gas imported from Indonesia to a Californian receiving terminal and if it owned a liquefaction terminal in Alaska dedicated to export to Japan, it would be more convenient (distances to cover would be halved) to transport the LNG imported from Indonesia to Japan and that imported from Alaska to California.

17

small group of transport companies, the main being Bonny Gas Transport, fully controlled by Nigeria LNG, that holds a 25% share of the transport capacity.

Bonny Gas Transport 25% Others 33%

BW Gas 10% Peninsular Gas LNG 8%

Kristen Navigation 4% A.P.Møller 4% SNTM Hyproc 4%

Teekay LNG Partners 6%

Oman Gas/MOL 6%

Figure 8: LNG tanker capacity allocation. Source: IEFE 2007.

Most of these companies are not independent, being subsidiaries of a liquefaction or regasification company. In particular, more than 40% of the 57 tankers serving primarily the European market are owned by firms vertically integrated with gas producers, which created their transport company or acquired significant stakes in existing transport companies. Similarly, some transport companies are fully owned by importing companies: for instance, Messigaz, fully owned by Gaz de France, or LNG Shipping, fully owned by ENI. Authors hinted before at the globalization of the LNG market and at the chance for exporting countries to obtain economic gains from price differentials, which will lead, alongside a supply deficit and a receiving capacity surplus, to a certain degree of competition between importers. With regard to this issue, it is interesting to look at the contracts currently in effect, to evaluate the entity of the contracted quantities and the residual duration of the agreements. The latter binds tankers to cover predefined distances and therefore represent a hindrance to possible cargo diversions. Even if this is positive in terms of security of supply for companies holding these contracts, on the other hand it 18

results unfavourable for firms that will realize new infrastructures. That is because they will have to incur in investments on their own, instead of opting for tanker rentals, in case of no shipping capacity uncontracted. The graph below is based on the latest available data (2007). As it can be seen from the percentages reported, 27% of tankers’ current world capacity will be uncontracted by 2010, thus, in the next two years a massive reallocation of shipping capacity is likely. Yet, more than 60% of tankers’ current world capacity is contracted until 2020 or longer.

Uncontracted 9%

Expiring after 2025 18%

Expiring before 2010 18%

Expiring before 2025 21%

Expiring before 2015 13% Expiring before 2020 21%

Figure 9: Tankers contractual situation. Source: LNG Journal 2007.

As for the European market, it can be noticed that for newer tankers the average contract’s duration is shorter; anyway the data possessed do not specify if any renewal has occurred. In table 11, contracts’ duration for tankers serving the European market is analyzed: the number of ships featured is 64 instead of 57, since tankers often serving the European market, even if they have multiple destinations, are included in the dataset.

19

Table 11: Contract duration for LNG tankers serving the European market. Contract duration Number of tankers Share Period ≥ 40 years 11 17% ‘60s 30-40 years 14 22% ‘70s 20-29 years 26 41% From the ‘70s to the ‘90s 10-19 years 5 8% After 2000 Less than 10 years 3 5% After 2002 Not available 5 8% TOTAL 64 100% Source: Ernst&Young 2006.

This trend can be explained by several factors. In first place, in the last few years, investing in LNG tankers has become more and more attracting: in fact, thanks to technological innovations, significant economies of scale, that shorten the amortization period, can be achieved, as discussed before. Moreover, the expected increase in world gas demand, and the consequent development of arbitrage operations based on price differentials, makes investments in this sector increasingly interesting. 5. Conclusions From our analysis on the shipping sector, we can conclude that this stage of the LNG value chain will not represent a bottleneck in the next few years, at least in terms of transport capacity. More, thanks to economies of scale, new tankers are bigger than before and consequently investing in this sector is getting more attracting. Looking at habitual tankers’ routes, Pacific stands today as the biggest player, both in terms of fleet and capacity, in the LNG import market. We can also notice that Asia, the most mature LNG market, at the same time is served by big-sized ships, which keep a constant flow of natural gas, and by an important number of small-sized ships, to deal with peak demand. The second biggest market, Europe, does not have this flexibility; nonetheless, investments in LNG transport capacity have intensified in the last 7 years. Notably, in the United Kingdom big sums have been invested in LNG, which is perceived as the primary source of external supply. As for the American market, one can say that it represents a marginal share of the market. Yet, facing a rapid depletion of gas reserves, receiving terminal projects abound. American demand’s impact on the LNG world market may contribute decisively to the achievement of the necessary level of investments for the development of a liquid LNG market. This would be important, since many of the 255 tankers now operating are bound to habitual routes. As for contractual practices, in fact, even if the average duration of contracts is falling, most of the tankers are part of a project that stretches from the construction of a liquefaction plant to that of a regasification plant. This means that the company that incurs in the costs necessary for the construction of a regasification (or liquefaction terminal) probably will need to rent an existing tanker. More, it will have to consider that tanker shipyard is an oligopolistic industry, with mainly Japanese and Korean players that have waiting lists that are currently three-year long. 20

In conclusion, we can say that LNG shipping is intrinsically related to the LNG sector dynamics, especially since tankers’ owners are companies controlled by gas producers and, to a lesser extent, by gas importers. This implies that long-term contracts, similar to supply contracts, are often used; thus, authors cannot hypothesize that in the short term spot sales will grow, as it happened with oil tankers.

References Chernyavs’ka, L. and S. Dorigoni. 2002. Modalità di trasporto del gas naturale e liberalizzazione ei mercati in Europa: pipelines o GNL?, Economia delle Fonti di Energia e dell'ambiente, 2/03. Council of European Energy Regulators (CEER).2006. Annual Report, Brussels. European Commission. 2007. Energy Sector Inquiry. Bruxelles. Graziano, C., S. Dorigoni and F. Pontoni. 2008. Entry and LNG: a model, IEFE working paper series (forthcoming). Global Insight. 2007. The LNG Market – Globally now to 2025 and the Implications for the UK, Report for the Department for Business, Enterprise and Regulatory Reform, London. Hallouche, H. 2006. The Gas Exporting Countries Forum: Is it really a Gas OPEC in the Making?, Oxford Institute for Energy Studies, Oxford. Hubbard G., and R. Weiner .1992. Long-Term Contracting and Multiple-Price Systems, Journal of Business, 65: 177-198. IEA. 2004. Security of Gas Supply in Open Markets – LNG and Power at a Turning Point, OECD, Parigi. IEA. 2006. World Energy Outlook, OECD, Paris. IEFE. 2008. Il Potenziale Ruolo del GNL nell’Incremento della Concorrenzialità dei Mercati Europei e nella Sicurezza dell’Approvvigionamento Energetico, Work in progress. Jensen J.T. 2006. The Development of a Global LNG Market, Oxford Institute for Energy Studies. Oil&Gas Journal. 2006. various issues. New York.

21

Percebois, J. 2003. Ouverture à la Concurrence et Régulation des Industries de Réseaux: Le Cas du Gaz et de l’Electricité, Cahiers de recherche, CREDEN, 03.11.40 Wagbara, O.N. 2007. How would the gas exporting countries forum influence gas trade?, Energy Policy, 35:1224-1237. Williamson, O. E. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting, New York: The Free Press.

22