Shipowners structure and fleet distribution in the liquefied natural gas shipping market. Siyuan Wang* and Theo Notteboom

Transcription

1 488 Int. J. Shipping and Transport Logistics, Vol. 6, No. 5, 2014 Shipowners structure and fleet distribution in the liquefied natural gas shipping market Siyuan Wang* and Theo Notteboom ITMMA, University of Antwerp, Kipdorp 59, 2000 Antwerp, Belgium *Corresponding author Abstract: The liquefied natural gas (LNG) shipping market has witnessed a rapid development over the last decade in line with the rising natural gas demand worldwide and the ongoing transformations in the globalising gas market. The purpose of this paper is to explore unique features of the LNG shipping market, more in particular to investigate shipowners structure and fleet distribution. We specifically focus on the following research questions: Who are the market participants and what tanker ownership scenario is to be expected? How does the shipping market structure look like, in terms of number and size distribution of shipowners? Several concentration measures are applied to examine the latter question. Moreover, the characteristics of the LNG shipping market are further examined by looking at the fleet distribution. By answering the above questions, this paper aims to provide a deeper understanding of the current LNG shipping market. Keywords: LNG shipping market; market participant; market structure; concentration; fleet distribution. Reference to this paper should be made as follows: Wang, S. and Notteboom, T. (2014) Shipowners structure and fleet distribution in the liquefied natural gas shipping market, Int. J. Shipping and Transport Logistics, Vol. 6, No. 5, pp Biographical notes: Siyuan Wang joined ITMMA of the University of Antwerp as a PhD student under the CONNEC scholarship programme of the European Commission in She holds an MSc in Transport and Maritime Management from ITMMA and a Master in Maritime Law from Shanghai Maritime University. Her research activities at ITMMA particularly focus on the LNG market. Theo Notteboom is a Professor and the President of ITMMA, Institute of the University of Antwerp. He is also a Part-time Professor in Maritime Transport at the Antwerp Maritime Academy and a Visiting Professor at various universities in Europe and Asia. He is currently the President of International Association of Maritime Economists (IAME) and Chairman of the Board of Directors of Belgian Institute of Transport Organizers (BITO), an institute of the Belgian Federal Government. He has published widely on (trans)port and maritime economics. This paper is a revised and expanded version of a paper entitled Shipowners structure and fleet distribution in the LNG shipping market: a comparative study with container and dry bulk markets presented at ECONSHIP 2011, June Chios, Greece. Copyright 2014 Inderscience Enterprises Ltd.

2 Shipowners structure and fleet distribution in the LNG shipping market Introduction Natural gas has become a key fossil fuel that accounts for 24.2% of the world energy consumption. It is either internationally transported via pipelines to consuming regions or in the form of liquefied natural gas (LNG) carried by specially designed tankers to distant markets. In 2010, global natural gas trade increased by a robust 10.1% (BP, 2010). With the ongoing process of liberalisation in the world gas market, LNG plays an increasingly important role in the security of countries gas supply, because it offers a flexible solution to diversify the importing sources, and moreover conveys price signals across isolated regions to compete with the regional piped gas and promote the formation of a competitive global gas market (Wang and Notteboom, 2011). In the past ten years, the LNG industry transformed from an infant towards a maturing industry. The LNG business was traditionally characterised as a highly-structured and capital intensive one, tied in long-term contracts for 20 years or more. In order to commit such high investments, the LNG projects formed the basis for the business with long-term planning and close cooperation between gas sellers and buyers. Such business model was regarded as the guarantee for a secure gas supply and a stable delivery of long-term revenues. However, with falling transportation costs and a fast evolving short-term trade, the restructuring of former rigid systems seems necessary and inevitable. Literature pointed out that a liberalised, integrated global natural gas market is in the making. Mazighi (2003) considered that the globalisation of LNG markets requires four conditions to be met; natural, economic, technical and institutional conditions. Natural conditions relate to the existence of surplus and deficits in different regions of the globe. Economic conditions refer to the increasing willingness of producers to sell more LNG on a short-term basis in line with huge cost reductions along the supply chain. Technical conditions concern the free third-party access to regasification terminals. Institutional conditions mean the emergence of organised markets for both the trading of gas molecules and their shipping segment. Mazighi underlined that the globalisation of the LNG trade basically required uncommitted shipping and regasification capacity, and he also indicated that before the LNG market becomes global, arbitrage trades would become the main way to create a linkage between different regions in the world. According to Jensen (2004), there are some potential similarities between the oil market and the natural gas market. However, the emergence of the global LNG market is found to be different. The high costs of LNG transportation are one of the main reasons preventing the LNG market to be as flexible as the oil market. He also pointed out that, although long-term contracts in LNG have been the vehicle for sharing the large up-front investment risks that characterise LNG projects, short-term trading has seen fast growth in recent years, partly as a result of declining costs along the supply chain, a growing diversity of supply sources and a loosening of the traditional rigid industry structure. A flexible LNG market can create a system which can transmit price signals freely between isolated regional gas systems. The sellers can take advantage of price differentials to process the arbitrage trade. As the author noted, in spite of the fast growth of the short-term market, LNG is by its nature highly capital-intensive with substantial financial risks. As a consequence, no developer has been willing to build a new facility on speculative basis without a contracted outlet. Thus, long-term contracts still remain a mainstay of international LNG trade.

3 490 S. Wang and T. Notteboom Looking at the dynamics of the LNG business model in the last decade, Von Hirschhausen and Neumann (2008) argue that asset-specific investments and uncertainty along the LNG supply chain lead to high transaction costs. Long-term contracts can however help to minimise transaction costs for two parties engaging in a commitment involving significant specific assets. In the course of the transformation from a monopoly industry to a competitive market structure, long-term contracts lost some market share for the uncommitted volume and the average contract length shortened significantly (from an average of years in the 1990s to approximately 8 15 years today). Nevertheless, long-term contracts will continue to play a more important role in deals where large-scale, asset-specific investment decisions are at stake. Iniss (2004) claimed that the coexistence of long-term and short-term trade is increasingly accompanied by vertical integration in the LNG industry. According to an empirical analysis of Rüster and Neumann (2006), high transaction costs along the LNG value chain induce a higher degree of vertical integration. In response to market deregulation, vertical integration was driven by upstream producers moving downstream with ownership of transportation capacities to exploit arbitraging possibilities and distribution, while power companies moved upstream to ensure margins and security of supply. Market restructuring and the emergence of spot trade have significantly reduced barriers to entry and increased the number of participants active in the markets. In this paper, we only focus on the LNG transport segment which is known as a relatively unrestrained component part of the supply chain. In contrast with the sectors of liquefaction, regasification and storage which are mainly controlled by national gas companies or local power and utility monopolists, the LNG shipping market appears to some extent be free and open to the new entrants. However, the high price of LNG tankers, compared with other merchant ships, makes the LNG shipping industry only accessible to players who have access to the necessary capital. In this sense, the LNG shipping market seems to have some unique features not found in other maritime sectors. On the basis of the literature review, we conclude that many papers investigated dynamics of LNG market and its pricing mechanism, e.g., Aune et al. (2008), Jensen (2004), and Wang and Notteboom (2011), but little notice was given to the features of the market structure in LNG shipping. To fill this gap, the present research seeks to explore the unique features of the LNG shipping market with specific attention to shipowners structure and fleet distribution. The following three research questions are put forward: 1 Who are the market participants and what tanker ownership scenario is to be expected? 2 How does the shipping market structure look like, in terms of number and size distribution of shipowners? 3 What features does the LNG fleet distribution reveal? To address the above questions can help us to get deep insights on the current LNG shipping market and identify its characteristics. With such a specific view of the market features, this study is expected to facilitate incumbent market participants and potential new entrants to find the right strategies and to make the right decisions for capturing optimal throughput based on the given competitive setting. This paper is organised as follows: Section 2 provides a brief overview of LNG trade dynamics over the last decade, especially underlining the market changes driven by

4 Shipowners structure and fleet distribution in the LNG shipping market 491 short-term trade; Section 3 identifies the market participants in the LNG shipping market and explores tanker ownership scenarios; Section 4 examines the LNG shipping market structure in terms of number and size distribution of shipowners by conducting a market concentration analysis using three concentration measures. Meanwhile, in order to well understand the LNG market structure and stand out its uniqueness, we include a comparative study with container and dry bulk markets; Section 5 analyses the features of the LNG fleet distribution; Section 6, finally, provides some concluding remarks and avenues for future research. 2 Dynamics in world LNG trade an overview LNG is natural gas stored and transported in liquid form at atmospheric pressure at a temperature of 161 C ( 256 F). Liquefied gas occupies a volume corresponding to 1/600 of the product in the gaseous state, which brings a remarkable economic value for transportation and storage. LNG not only can be used as a solution, instead of piped gas, in geographical constrained areas like mountain ranges or deep sea but also can be a more economic approach for medium and long distances. For instance, the LNG project break-even distance is estimated at 2,500 miles (4,000 km) for a 42-inch onshore pipeline and 1,240 miles (2,000 km) for an offshore pipeline (Jensen, 2004). Figure 1 Development of LNG trade pattern and shipping capacities (see online version for colours) Source: Own compilation based on Poten & Partners data ( The LNG industry originally developed as a niche business where a relatively small number of sellers supplied specific regional markets, using a traditional method with tankers dedicated to bilateral trades for long-term contracts (Gkonis and Psaraftis, 2009). There was little scope for short-term trade or cargo diversions from the originally intended destination. In the early 1990s, LNG world trade was firmly divided into two basins: Atlantic and Asia-Pacific. The two regions were isolated with few trade connections and price interactions. However, from the late 1990s, as result of project delays in regasification plants and a capacity surplus during ramp-up period in producing

5 492 S. Wang and T. Notteboom countries, more redundant volume was available for diverting to the markets with high profit. The increasing amount of cargo traded for arbitrage purposes formed the basis for a short-term/spot market. As a result, the old pattern of intra-regional trades started to break down and inter-regional trades developed (Figure 1). Natural gas is becoming a global commodity competing with oil and coal. As illustrated by Figure 2, Short-term trade witnessed fast growth in 2010 which occupied 22.5% of total LNG trade, compared to only 5% ten years ago. Figure 2 Evolution of LNG short-term trade (see online version for colours) Source: Poten & Partners data ( With the start-up of the Qatar project in 1997, the LNG trade got divided into three main trading bases: Asia/Pacific, the Atlantic (including the Mediterranean) and the Middle East (especially Qatar). The Middle East is positioned between the two other trading blocs and thus benefits from its geographical advantage of being able to trade with both markets as a swing producer. The Asia/Pacific market is the largest and longest established of the three markets. The region is dominated by Japan, which imported 31.9% of all LNG trade in 2010, and 53.1% of Asia s imports. In second place comes South Korea, with Taiwan and India forming a third tier of importers. China will soon become a major importer in view of many new regasification plants being constructed. Exports in this market are dominated by Indonesia and Malaysia, but Australia is not far behind having increased its exports significantly in 2009, by 19.4% over 2008 (Clarkson Research Services, 2010). The market share of short-term trade in this region is relative low, as most of the big importers like Japan and Korea use LNG as the only viable solution to gas import. They are thus willing to take long-term contracts in order to secure supply. Also, the producing countries in this region mainly serve their regional consumers on a long-term commitment basis with less cargo diverted to the Atlantic basin. The destructive earthquake and tsunami in Japan in March 2011 resulted in a very extensive short-term

6 Shipowners structure and fleet distribution in the LNG shipping market 493 LNG demand to cope with the energy shortfall. The extreme high price attracted producers in the Atlantic region and the Middle East to divert their cargo to this profitable market. The main consuming nations in the Atlantic market are in Western Europe and North America. Spain, the largest importer in Europe, accounted for 9.4% of all imports in The UK, the fastest growing market, imported 6.4%, and France represented 4.7%. In the past two years, the USA decreased its LNG imports following an increase in its domestic production of shale gas. On the export side, Algeria is the largest producer in this market, with Trinidad and Tobago and Nigeria forming a second tier leaving Egypt a step behind. In the Middle East, Qatar is to date the largest exporter in the world trading cargoes to both Asia/Pacific and Atlantic markets. It constituted 25.5% of world LNG trade volume in Oman and Abu Dhabi are also fast growing sellers in this region. In addition, the Atlantic and the Middle East are more active in the short-term trade with a share of 83.3% of world spot volume, mostly flowing to the Asian market (GIIGNL, 2010). 3 The LNG shipping market participants and tanker ownership scenarios Traditionally, because of the large capital requirements of LNG projects together with the high risks involved, the market participants were normally limited to the so-called super majors. Gas buyers were either government monopolies or franchised utility companies, and gas sellers were typically either national oil companies of producing countries or international oil majors. In order to structure a risk sharing system among the participants, the long-term contract between buyer and seller was central to the project. The risk sharing logic was embodied in the phrase buyers take the volume risk and sellers take the price risk (Jensen, 2004). Most contracts, therefore, featured take-or-pay provisions to assure the buyer offtake at some minimum level and a price escalation clause to transfer responsibility for energy price fluctuations to the seller. As regards the point of delivery, it was either F.O.B or D.E.S, depending on which party assumed the transportation responsibility. Tankers, owned by the buyers or the sellers or independent shipowners, usually were dedicated to a specific import or export project for the life of the contract. However, in the late 1990s, competition, mirrored by spot trade and flexible contracts, started to gradually move the LNG industry away from monopolistic structures. The market deregulation favoured vertical integration along the LNG value chain to amortise the capital intensive investments, reduce transaction costs and secure gas supply in times of a growing world gas demand. Thus, gas buyers began to go upstream and participate in activities such as shipping and liquefaction. The sellers, on the other hand, tended to move downstream and become minority owners in shipping and regasification plants. Consequently, the LNG shipping was under control of different parties. Nowadays, along with the global trend towards privatisation of the energy markets, the transformations taking place in the LNG market are heavily influenced by liberalisation in both the upstream and downstream gas markets. Figure 3 indicates that the ongoing changes have recently resulted in more strategic partnerships and the setting up of consortia between market players. Moreover, it is worth noting that some

7 494 S. Wang and T. Notteboom independent shipowners start middle stream activities to integrate their LNG value chain by either cooperating with upstream gas sellers in liquefaction projects (e.g., in 2004, Golar LNG had a cooperation with the BG group) or investing with downstream buyers in import terminals or even taking over the role of refrigeration terminals by converting ships to be gas floating production or regasification units (Engelen and Dullaert, 2010). Figure 3 Evolution of LNG business model (see online version for colours) Source: Own representation based on Engelen and Dullaert (2010) In addition, one of the most noticeable changes in the market structure (which has both contributed to and benefited from the new short-term market place) has been the emergence of companies that invest throughout the LNG supply chain. These portfolio players have a portfolio of liquefaction interests which they use to supply the product to a portfolio of import terminals using vessels they control specifically for this purpose. This group mainly includes a number of international oil majors (e.g., Shell, BP, ExxonMobil, and ConocoPhillips), utility companies (e.g., Suez de GDF, ENI and Osaka Gas), private companies (e.g., Mitsui) and quasi-government companies (e.g., Sonatrach, Qatar Gas and Statoil) (Clarkson Research Services, 2010). According to Nissen (2004), the evolving commercial LNG model implies that players control capacities in all stages along the value chain, participating in different export projects, owning undedicated transport capacity and controlling multiple import facility access. Thus, they are able to use these capacities flexible and benefit from arbitrage. Given the dramatic changes in the LNG business model in the past twenty years, the shipping segment has become the most promising area for upstream and downstream activities. Shipowners can be classified into the following four categories: 1 LNG integrated projects with gas suppliers including national oil companies and their partners who own the tankers and transport their own LNG for sales based on delivered ex ship (DES) terms to their customers. Most of this fleet is committed to long-term contracts for bilateral trades with low flexibility.

8 Shipowners structure and fleet distribution in the LNG shipping market Gas buyers, like Japan and Korean who are dependent on LNG imports. They own and operate some shipping capacities for diversifying gas suppliers to secure market demand. This kind of operation also involves long-term contracts and low flexibility. 3 International oil majors, portfolio players, use own vessels to sail between the liquefaction and regasification plants where they have portfolio interests. The flexibility of the ship operations facilitates cargo swaps for cost optimisation and encourage arbitrage trade for speculative profits within their own portfolios. 4 Independent shipowners who previously secured long-term contracts from LNG integrated suppliers or buyers. The growing spot gas price coupled with the rising charter rates in recent years has attracted some independent owners to order new capacities with uncommitted contracts (see Table 1). Table 1 Type of tanker ownership and its operation characteristics (see online version for colours) Type of ownership Operation characteristics Flexibility/operation level Integrated project (Gas suppliers, NOC) (QatarGas, Petronas, etc) Gas buyers (Japan, Korea, etc) International oil companies (BP, shell, etc) Independent shipping companies (Teekay, Golar LNG, etc) Specific project on a long-time basis Long-term commitment to a specific project, but use of surplus capacity for spot transaction, etc. Involvement in the entire LNG supply chain, and operation adapted to the situation as regards demand, etc. Long-term commitment for buyers and sellers or short-term transactions Source: Own representation based on Ndao (2004) Low flexibility, low risk or low operation Low-medium flexibility, low risk of low operation High flexibility, medium risk of low operation High flexibility, high risk or low operation for short-term transactions As shown in Figure 4, independent shipping companies are taking up an ever larger share in terms of LNG tanker ownership. The share increased from 38% in 2004 to 49% in This group represents 64% of the capacity on order (figures May 2011) or 25 ships of 44 new buildings. Moreover, there are 15 speculative ships with no clear long-term committed contracts ordered by independent owners. These figures clearly demonstrate the ambitions of the independent shipowners to extend the spot market. Mazighi (2003) argues that one of the prerequisites for the globalisation of the LNG market is to have enough uncommitted shipping capacities for spot trade. However, the current strategies undoubtedly make independent owners to be the most important players for establishing an international gas market in the future.

9 496 S. Wang and T. Notteboom Figure 4 LNG tanker ownership changing scenario (see online version for colours) Source: Own compilation based on Clarksons data ( 4 LNG shipping market structure As discussed above, the traditional long-term contracts between gas sellers and buyers did not require vertical integration in the sense of capacity ownership, while with ongoing liberalisation in the worldwide gas markets, the producing companies started to integrate downstream to benefit from marginal rents, and in contrast, the downstream players integrated upstream to ensure gas supplies in times of a developing competition between importing regions. As for such integration model, how does the LNG shipping market structure look like? Rüster and Neumann (2006) pointed out that Vertical integration along the value chain limits competition at the horizontal level thus counteracts liberalisation efforts in downstream markets. This section is therefore based on our hypothesis that the higher the degree of vertical integration along the LNG value chain, the higher the market concentration in the shipping segment. We provide an empirical analysis of the LNG shipping market structure by examining market concentration in terms of the number and size distribution of shipowners. Since market concentration is a key element of market structure which reflects to some extent the nature of competition in the market concerned, we aim to apply three market concentration measures to explore the unique features of the LNG shipping market structure. In order to have a better understanding of the unique features in the market structure of the LNG shipping segment, we use two traditional shipping markets, i.e., container and dry bulk, as references. The comparison among these three shipping segments not only highlights the unique nature of the LNG shipping market, but might also point some future development paths of this emerging maritime sector. Both the container and dry bulk shipping sectors are two relatively mature markets with their own inherent business models, but, in contrast, the LNG shipping market is still in its infancy stage.

10 Shipowners structure and fleet distribution in the LNG shipping market Market data The analysis is based on an extensive database of Clarkson containing the situation for March As shown in Table 2, there are 56 owners in the LNG market holding 363 tankers with a total capacity of 51.9 million m 3. Compared to the container and dry bulk markets, the LNG tanker market is quite smaller in terms of total deadweight tonnage. It only constitutes 15% of the container market and 5% of the dry bulk market. Table 2 Market data in three shipping sectors (March 2011) LNG tanker market Containership market Dry bulker market No. of owners ,608 No. of vessels 363 4,985 8,326 Total DWT 28,339, ,885, ,564,433 Total cargo capacity 51,910,768 (m 3 ) 14,257,291 (TEU) 544,567,433 (dwt) Source: Clarksons data ( Most of top 10 LNG tanker owners are oil and gas companies (Table 3). QatarGas, MISC and Nigeria LNG are national oil companies. Shell and BG are international oil majors. The remaining five companies are independent shipowners. Three Japanese shipping majors (NKY, MOK and K-line) built a consortium and cooperate with QatarGas in liquefaction projects, employing dedicated tankers for importing gas to Japan with long-term contracts. Teekay and BW are both large shipping companies operating variable fleets. They entered into the LNG shipping segment just a few years ago, but managed to gain market share by engaging in joint ventures or strategic partnerships with upstream and downstream entities. Apart from long-term commitments and high capital intensity, another unique feature of LNG shipping is that it involves less speculative activity, which implies most of the tanker owners are also operators and managers. This is however not the case in the container and bulk markets where the vessel owner is not necessarily the vessel operator and the supplier of vessel services to the customer base. Charter contracts (voyage, time and bareboat) are therefore very common in the dry bulk market. Considering the direct link between vessel ownership and vessel operations in the LNG market, the ownership of tanker capacity can be seen as a good indicator to examine concentration in the LNG shipping market and thus the level of market competition. 4.2 Methodology-concentration measures Concentration measures are the most widely used indicators of market structure. In this section, we use the most common measures (N-firm ratio, HH index and Gini coefficient) to examine the market concentration of the LNG shipping market. This empirical study focuses on one aspect of market structure, i.e., the concentration level of transport capacity owned by shipowners. We thus look at the vessel supply side of the market, not the distribution of actual ship operators or the competition between these shipping lines. As mentioned above, in other shipping sectors like container and dry bulk, shipowners structure does not necessarily reflect the level of competition between shipping lines since various forms of chartering contracts make that shipowners are not necessarily the ship operators.

11 498 S. Wang and T. Notteboom Table 3 Top 10 shipowners in the LNG, container and dry bulk markets (March 2011) (see online version for colours) LNG market Container market Dry bulk market Owner Capacity ( m 3 ) % Owner Capacity (TEU) % Owner Capacity (dwt) % 1 QatarGas* 6,063, % 1 Maersk# 1,108, % 1 COSCO 24,856, % 2 MISC (Petronas)* 3,675, % 2 MSC# 862, % 2 NYK 17,270, % 3 NYK 3,252, % 3 CMA CGM # 496, % 3 MOL 14,983, % 4 MOL 2,843, % 4 Reederei C.P 484, % 4 K-Line 13,260, % 5 Teekay 2,362, % 5 COSCO# 375, % 5 Zodiac 7,312, % 6 Shell* 2,094, % 6 Evergreen# 342, % 6 China shipping 7,029, % 7 K-Line 2,027, % 7 NSB N. 341, % 7 Angelicoussis 5,584, % 8 BW Ltd* 1,918, % 8 NVA N. 334, % 8 Enterprises 5,339, % 9 BG Group* 1,834, % 9 Peter Dohle S. 329, % 9 Hanjin shpg. 4,897, % 10 Nigeria LNG* 1,743, % 10 Hapag-Lloyd# 325, % 10 HOSCO 4,756, % Total 27,814, % 5,001, % 105,290, % Market total 51,910, % 14,257, % 544,567, % Note: *indicated as an oil and gas major; # indicated as a shipping line. Source: Own compilation based on Clarksons data (

12 Shipowners structure and fleet distribution in the LNG shipping market 499 Three concentration measures for examining shipowners structure in the LNG shipping market are applied. The first one is: the N-firm concentration ratio, which measures the shares of the N largest shipowners in total LNG transport capacity: CR n = n i= 1 S i S i is the share of the i th largest shipowner in total transport capacity in the LNG industry and n is the number of large shipowners to be included in the calculation of CR n. The most common concentration ratios are n = 4 or 8, which refer the cumulative market share of the four or the eight largest shipowners. Concentration ratios are used to show the extent of market control of the largest shipowners in the industry and to illustrate the degree to which an industry is oligopolistic. Bailey and Boyle (1971) find that an attractive property of the n-firm concentration ratio is that it requires size data on the top n firm only, together with the corresponding aggregate size measure for the entire industry. Hence, the data requirements are less demanding than for the other concentration measures. However, such a ratio cannot take into account the number and size distribution of firms that are outside and inside the top n firms, therefore it provides limited detail about the overall competition in the industry (Lipczynski et al., 2009). Another method to assess LNG tanker owner concentration consists in applying the much used Herfindahl-Hirschman (HH) index. Compared with N-firm ratios the HH index provides a more complete picture of industry concentration (Lipczynski et al., 2009). The HH index is calculated as follows: HH = N 2 S i= 1 i S i is the market share of shipowner i, and N is the total number of shipowners involved in the LNG shipping segment. The maximum value of HH (i.e., a value of 1) occurs when there is only one single shipowner in the market. The minimum value of HH is 1/N and represents a situation in which the industry consists of n equal-sized shipowners. In this case, each owner has a market share of S i = 1/N. A practical difficulty with the HH index is its requirement for individual size data on all of the industry s member firms. In addition, there are several drawbacks to the use of this index. Kwoka (1977) argues that this index embodies both size inequality and firm numbers with weights which are assumed a priori instead of being derived. Most of the other critiques are empirically based: for instance, Borenstein et al. (1999) have shown that at least in the case of electricity markets the HH index is a poor measure of competition. According to Matsumoto et al. (2011), the index may not provide a correct picture on competition in the industry in case of N-firm oligopolies with partial cooperation among firms. A normalised HH index (HH * ) is usually used to allow comparison between different markets which each have a different number of firms. The index then ranges from 0 to 1, instead of 1/N to 1. The normalised HH index is computed as follows: HH * 1 HH = N 1 1 N

13 500 S. Wang and T. Notteboom The third applied method for examining shipowners concentration is the Gini coefficient in combination with the Lorenz curve. The Gini coefficient is a widely used index that measures per cent departure from a perfectly uniform distribution, and it is shown graphically using a Lorenz curve. A Lorenz curve displays the variation in the cumulative market share in terms of transport capacity of the n largest shipowners in a shipping market, as n varies from 1 to N (where N is the total number of owners). Figure 5 illustrates a typical Lorenz curve. The shipowners are ranked from the largest to the smallest (from left to right) along the horizontal axis. The vertical axis shows the cumulative market share (the sum of the market shares of all shipowners from owner 1 to owner n, as a function of n). If all of the shipowners, holding the same market share in transport capacity, are equal of size, the Lorenz curve is the 45 degree diagonal line (OCA). If the owner size distribution is skewed, the Lorenz curve is the concave OBA curve, and in case one dominant shipowner accounts for the total volume of the shipping capacity (full concentration), then the Lorenz curve approaches the line ODA (Lipczynski et al., 2009). The Lorenz curve can be used to define a concentration measure, known as the Gini coefficient. With reference to Figure 5, the Gini coefficient is defined as follows: area of crescent between OBA and OCA G = area of the triangle ODA According to the above definition, the value of the Gini coefficient has a range from 0 to 1. The maximum value of G = 1 corresponds to the case of one dominant shipowner with a market share approaching one, and the minimum value of G = 0 corresponds to the case of N equal-sized shipowners in the market. It is thus concluded that the greater the deviation of the Lorenz curve from the diagonal line, the greater is the inequality (Notteboom, 2006). There are different ways to calculate the Gini coefficient, but the formula applied in this study, is shown below: N n Xi n= 1 i= 1 N + i i= 1 G = 0.5(N 1) X ( Xi id the soze of firm i) where G is the Gini coefficient and X i is the owned transport capacity of tanker owner i, expressed in cubic meter. The Gini coefficient s main advantage is that it is a measure of inequality by means of a ratio analysis. This makes it easily interpretable to use for comparison of capacity distributions across different shipping sectors. But the Gini coefficient has a number of drawbacks, such as sensitivity to ordering (Cowell, 1988), sensitivity to data errors, sensitivity to inequality around the mean of the size distribution (Champernowne, 1974) and small sample bias (Deltas, 2003). According to Scherer (1980), the Gini coefficient can produce misleading results when examined in an industry with a small number of evenly matched firms (this is not case in LNG shipping sector). Moreover, a high inequality generates a strongly skewed shape of the Lorenz curve with heavy tails in the size distribution. A long tail effect is visible when the dataset contains a large number of small shipowners and only a handful of larger players. As demonstrated below, the long tail effect does not appear in the LNG shipping market, but is quite visible in the container and dry bulk markets.

14 Shipowners structure and fleet distribution in the LNG shipping market 501 Figure 5 The Lorenz curve Source: Own compilation based on Lipczynski et al. (2009) 4.3 Empirical results Table 4 presents the empirical results obtained by using Clarkson market data in Section 4.1 and applying the three concentration measures introduced by Section 4.2. Table 4 Empirical concentration results for the three shipping segments (March 2011) Market segments LNG tanker (Cu.M) Containership (TEU) Bulker (DWT) n-firm concentration ratio CR CR Herfindahl-Hirschman (HH) Index HH HH * Gini coefficient Source: Own compilation based on Clarksons data (

15 502 S. Wang and T. Notteboom Figure 6 The Lorenz curve for LNG tanker markets (based on date for March 2011) (see online version for colours) Source: Own compilation based on Clarksons data ( According to the n-firm concentration ratio in the table, it explicitly shows that both values of CR4 and CR8 in the LNG market are higher than those of the other two markets. This suggests that the LNG shipping market is more concentrated than the container and dry bulk segments. However, as discussed in Section 4.2, the concentration ratio may give a limited picture of market structure. Therefore, the normalised HH index is used to explore more complete detail on market concentration across these three sectors. Such index in Table 4 indicates a roughly same but more precise result that the LNG market is slightly more concentrated than the container market but far more concentrated than the bulk market. The reason why the LNG shipping market is so concentrated in terms of shipowners structure is linked to the hypothesis introduced earlier in this paper. The trend towards vertical integration along the LNG value chain has led to a higher market concentration in the shipping segment. However, with more new entrants entering the market, and more

16 Shipowners structure and fleet distribution in the LNG shipping market 503 partnerships and cooperation among the incumbent owners, the picture of market structure is expected to change accordingly. Figure 7 Owner size distribution in the three markets (based on date for March 2011) (see online version for colours) Source: Own compilation based on Clarksons data ( In regard of the Gini coefficients, the three shipping markets all possess relatively high values that imply a certain level of inequality in the size distribution of shipowners. However, the LNG market shows the lowest values among the three (i.e., ), which implies that the owners size appears a bit less unequal than those of the bulker (i.e., ) and containership owners (i.e., ). As discussed in Section 4.2, such degree of inequality can be elaborated by the Lorenz curves where. Thus, it is clearly observed in Figure 6 that the container ship market has the highest degree of inequality followed by dry bulk market and the LNG tanker market. Nevertheless, the LNG segment has a higher concentration (a higher CRn and HH index) but a lower inequality of size distribution (a lower Gini coefficient), while these two criteria of market concentration normally show the same tendency, as it is true in the container and dry bulker markets. In order to explain the reason for this, we make scatter plots (Figure 7) to show the size distribution of each owner in the three markets by average size and number of vessels. It

17 504 S. Wang and T. Notteboom is notable that the average ship capacity of most LNG tanker owners is intensive within the range of 130, ,000 m 3, but the market average size is 140,000 m 3 which indicated by the dotted line in the graph. So the relative average difference between shipowners are small which lead to the owner size distribution are more equal. However, why the average owner size in LNG market is so intensive within the range of 130, ,000 m 3? The feature of LNG fleet distribution in the next section may give the explanation. To sum up, through analysing the owner concentration in the three markets, some notes can be thereby concluded as follows to characterise the LNG shipping market: The LNG shipping market is a rather small market compared to the container and dry bulk markets. This implies that LNG shipping still is an emerging maritime sector. However, with the rapid growth of global LNG demand the market is expected to transform from an infant towards a more mature market in the future. The current LNG shipping market is more concentrated than the container and bulk shipping sectors in terms of distribution of transport capacity among the owners. It is noted that LNG shipping embodies the characteristics of an oligopoly where few oil majors are dominating the market. And the ongoing vertical integration along the LNG industry chain also contributes to the high market concentration. However, the LNG shipping market is less unequally distributed than the container and bulk markets as shown by the Gini coefficient and the Lorenz curve. The reason for this seeming contradiction between concentration levels and inequality levels is that the average ship capacity of most LNG tanker owners is intensive within the range of 130, ,000 m 3. Hence it can be concluded that in the LNG market, the fleet size of the owners is mostly determined by the number of ships not the ship size (most common average ship size is 130, ,000 m 3 ), except for Qatar Gas who is the only LNG tanker owner possessing the largest tankers (more than 200,000 m 3 ) at this moment. In contrast, the container market is filled with some large shipowners (many of which also act as shipping lines, see Table 1 earlier) who hold a huge vessel capacity both in number and size of the vessels. At the other end of the market there are a large number of small shipowners who typically own smaller vessels. A similar situation can be observed in the bulk market. 5 Features of the LNG fleet 5.1 Development of the LNG fleet The development of the LNG fleet is in accordance with the growth of LNG trade. Due to high energy prices and a collapse in demand for natural gas, there was a bleak period for the LNG business in 1980s, while, up to the 1990s, the revival in gas demand brought a lot of new projects, with accompanying ship orders promoting a steady fleet growth until the early years of the 21st century. The glut of new projects that were approved at that time sent the order book spiralling. In the period , around 20 vessels a year were ordered at a cost of in excess of $12 billion, and subsequently a record of 70 orders was reported in 2004 with a contract value of $23 billion (Clarkson Research Services, 2010).

18 Shipowners structure and fleet distribution in the LNG shipping market 505 The LNG fleet appears to resume a period of unspectacular growth in the wake of the economic crisis in According to economic research by Platou Economic Research (2009), the supply of LNG shipping capacity has outweighed demand during the last three years. To date, the fleet has increased to 363 vessels totalling about 51.9 million m 3, with the economics recovering starting from the end of 2010, more new buildings are ordered with 44 vessels on the order book as of August, Furthermore, the increasing number of speculative orders, most of which are placed by the independent shipowners, will facilitate the development of spot trade. Figure 8 World LNG fleet development and forecast (see online version for colours) Source: Clarkson Research Services (2010) Figure 9 LNG fleet age profile (see online version for colours) Source: Poten & Partners data (

19 506 S. Wang and T. Notteboom Figure 10 LNG tanker new building price for a 147,000 m 3 LNG carrier (see online version for colours) Source: Clarkson Research Service (2010) The traditional LNG vessel, as built between the early 1960s and the early 2000s, could be grouped into three size ranges. The first category contains the small carriers of between 25 50,000 m 3 which were used for short-distance trades, especially in the Mediterranean. Vessels between ,000 m 3 falls into the second range regarded as the larger sized class. Moreover, with the first delivery of a 140,000 m 3 vessel in 2002 and a 150,000 m 3 tanker in 2006, the third type of ship size has been consistently extended even to 165,000 m 3 in However, the changes to the fleet have not just been a simple up-scaling. There are now a significant number of so-called Q-Flex ( ,000 m 3 ) and Q-Max ( ,000 m 3 ) vessels operating in the fleet. These vessels were initially designed to service the new projects in Qatar (owned by Qatar Gas). The first of the Q-Flex vessels was delivered in late 2007, and the first Q-Max vessel in October So far, there are 13 Q-Max vessels and 31 Q-Flex vessels in the LNG fleet. The year 2010 has seen the first vessels in the ,000 m 3 range delivered, narrowing the gap between the larger of the more traditional vessels and the new Qatari designs. With delivery of larger tankers, the average size of the LNG fleet has been growing from 112,649 m 3 in 1998 to 136,367 m 3 in 2008 and 142,092 m 3 in 2010 (Figure 8). The fast growth of LNG fleet in recent years implies that such a fleet is one of the youngest fleets with average age of 9 years and 50% of the fleet is under the age of four years. All large ships (Q-Flex and Q-Max) are within this class (Figure 9).

20 Shipowners structure and fleet distribution in the LNG shipping market 507 An LNG carrier might be the most technical sophisticated and expensive cargo vessel in the world. The rising cost of steel, changes to the size of the vessels being built and technological advancements all added to the escalating new building prices for LNG carriers from 2003 until the economic recession hit in late But the difficult economic situation caused a sharp downward movement in Clarkson s assessed price for a 147,000 m 3 LNG carrier from $220 million in October 2008 to $187 million a year later and $185 million at the end of June 2010, a fall of 15.9% (Figure 10). It was because that individual vessel prices varied according to the individual specification of the contract. There was a complete dearth of contracting of LNG vessels in 2009, one of the few vessel sectors to report zero new vessel contracts for the year, while after twenty one months, the drought ended in February 2010 with the contracting of a 147,100 m 3 vessel by China LNG Shipping at Hudong Zhonghua for delivery in 2012 (Clarkson Research Services, 2010). 5.2 LNG fleet distribution It is clearly illustrated by Figure 11 that the ship size distribution in the LNG fleet is characterised by a high concentration within one range of size from 130,000 to 150,000 m 3. Figure 12 shows the corresponding concentrated range of size from 70,000 to 89,999 dwt which accounts for 69.1% of the whole LNG fleet. In contrast, the container and bulk fleets show a much more diversified ship size distribution catering for different routes and trade flows. For instance, in the liner shipping sector, large container vessels are deployed between hub ports and smaller ships are deployed on secondary trades and feeder services. Also, different parcel sizes in dry bulk determine the various ship sizes in bulk shipping. Thus, it is believe that such concentrated size distribution in the LNG fleet can well explain why the average owner size in this market is so intensive within the range of 130, ,000 m 3, and why the level of inequality of shipowner size distribution is lower in the LNG market than that in the other two markets. The distribution of LNG ship size by sailing distance is presented in Figure 13. There are still some small ships operational on regional trades, e.g., Algeria-France (515 miles) and Malaysia Japan (2,208 miles). However, the most common vessel size (130, ,000 m 3 ) is deployed on a wide distance range from 1,500 miles to 6,000 miles, some serving regional and others cross-regional trades (e.g., Qatar to Japan 12 ships and Korea 8 ships). The newest classes of large ships, Q-Flex and Q-Max, are deployed on the routes between Qatar and Europe/US. Qatar has become the biggest LNG exporter in the world, despite its location far away from the markets: the average distance to Europe amounts to 5,737 miles, to Japan is 6,508 miles away and to the US is 9,680 miles. Qatar Gas uses large ships to reduce unit transport costs for the sake of keeping its competitive position in the world LNG industry.

21 508 S. Wang and T. Notteboom Figure 11 Ship size distribution in the three market segments (see online version for colours) Source: Own compilation based on Clarksons data ( However, it is worth noting that the concentrated ship size in the LNG fleet is expected to change as the world LNG trade increases. As an environmental friendliness energy source with adequate reserves for the next 100 years, natural gas will certainly change the world energy mix in the future. The uniqueness of its early development stage will change with the fast rising world LNG market in the future. The maritime transportation of LNG between large terminals by standard large tankers, will not meet the growing demand for natural gas (e.g., LNG as a fuel for ships and trucks). Thus, the expected wider usage of natural gas needs to be supported by the development of a feeder system based on existing large hub terminals and small LNG distribution networks fed by dedicated LNG feeder tankers and even smaller LNG barges. We expect that there will be a more diversified LNG fleet in the future and that the LNG shipping market is more likely to rely on the hub-feeder model. The market structure of the LNG shipping segment (as discussed in Section 4.3) is expected to change with some owners operating large tankers between hubs and small owners dedicating their activities to local feeder services.

22 Shipowners structure and fleet distribution in the LNG shipping market 509 Figure 12 Comparison of ship size distribution in the three markets (in dwt)(see online version for colours) Source: Own compilation based on Clarksons data ( Figure 13 LNG tanker size and distance covered (see online version for colours) Source: Own compilation based on Clarksons data ( 6 Conclusions and future research The LNG shipping sector has witnessed a dramatic change coupling with the fast growing demand for natural gas. This shipping segment was mainly characterised by dedicated long-term contracts and regionally-based trade flows. Tankers were mostly purpose-built for special projects and trades dedicated to the life of the contract. However, the emergence of the short-term/spot trade and the liberalisation of the gas market made LNG shipping a more promising area in the shipping industry. Such significant changes are reflected on the tanker ownership structure. More market participants are being involved in the LNG shipping segment, and in particular, independent shipowners fast gained market share in terms of ownership of transport capacity and increased the uncommitted orders. Additionally, they start middle stream to integrate their LNG value chain by either cooperating with upstream gas sellers in

23 510 S. Wang and T. Notteboom liquefaction projects or investing with downstream buyers in import terminals. It thus suggests that independent shipowners will play an important role in establishing an international gas market in the future. According to the study of shipowner concentration, the current LNG shipping market is an oligopolistic market where there are a small number of owners with relative small transport capacity. The market is more concentrated in terms of vessel ownership than the container and bulk shipping sectors (examined by concentration ratio, HH index), which may caused by the vertical integration along the LNG industry. But according to the Gini coefficients and Lorenz curves, the shipowner size is less unequally distributed in the LNG market. The reason for this rather interesting combination of higher concentration and lower inequality is that the average tanker size is centralised in the range of 130, ,000 m 3, which used to be represented as a LNG project standard. The result of this is that large players cannot heavily weight their market share. On the contrary, the container and bulk markets are filled with some large companies holding huge capacities both in number and vessel size and many smaller companies with smaller vessels, and the ship sizes of these two markets are diversified by trading routes, parcel size and trade flow. Figure 14 Ecosystem of Independent shipping company (see online version for colours) Source: Own elaboration based on Gkonis and Psaraftis (2008) However, the LNG shipping market is still at an early development stage. The current market features are expected to change as the world LNG market grows. The LNG shipping market structure is likely to evolve towards the segmentation between owners who operate large tankers between hubs and small owners dedicated to local feeder services.