Sticky substance with sticky power: Oil in global production and financial networks

Introducing the Global Production and Financial Network of oil

Oil is the world’s most traded commodity and fundamental for the economic development of virtually all economies, whether they produce or consume it. This makes understanding oil markets an essential task for economic geography. In this paper, we combine Global Production Network (GPN) and Global Financial Network (GFN) approaches to analyse the interrelations between physical and financial oil trading. We apply the resulting framework of the Global Production and Financial Network (GPFN) to analyse the long-term evolution of the global oil industry, as well as mid-term and short-term oil market fluctuations and anomalies. Attention to the role of finance helps us see transnational companies and their production networks in broader networks with world government and financial and business services as leading actors, and financial centres and offshore jurisdictions as key geographies. We show the oil GPFN as multi-scalar, from global to local, in which small towns, for example, at the crossroads of major pipelines, matter in addition to global financial centres.

The second generation of GPN theory already includes finance in its framework as financial discipline, defined as financial actors, markets and institutions which organise and condition the geographies of production throughout the global economy (Yeung and Coe, 2015). Yeung (2016) also refers to financial discipline as challenges and opportunities presented to firms by financialisation. As such, financial discipline is proposed as one of the five key drivers of GPNs (Coe and Yeung, 2019). This, however, risks underplaying the many manifestations of finance in relation to oil. Consider, for example, the role of oil revenues in state (rather than just corporate) finance or the formation of oil price via futures markets as a financial process. Value flows through GPNs as money, and where it comes from into a GPN and where it goes to when it leaves a GPN is crucial to the very existence, operation and geographical footprint of GPNs. Likewise, the materiality of production, distribution and consumption shapes and constricts the operations of finance. For these reasons, we seek to complement GPNs with GFNs (Coe et al., 2014).

Figure 1 presents a stylised conceptual framework of the GPFN in the context of the firm-territory nexus established in economic geography. The GPN portion (represented along the vertical axis) focuses on Transnational Companies and their relationships with governments representing territorial interests.

OPEN IN VIEWER

Figure 1.

GPFN conceptual framework.

Building on the work on Haberly and Wójcik (2022), the GFN portion focuses on two groups of actors from GPNs (Financial and Business Services and World Governments) that operate across different key geographies (Financial Centres and Offshore Jurisdictions) shown on the horizontal axis. In the case of the oil GFN, Financial and Business Services are companies managing financial instruments and contracts related to oil including not only financial services firms, but also oil producing companies that engage heavily in financial activities such as currency hedging, or oil using companies, such as airlines, seeking to limit price volatility (Bridge and Le Billon, 2017). These activities are organised geographically including Financial Centres housing markets for oil-based financial instruments used for reliable valuation and exchange, creating liquidity and concentrating information about the oil industry. Offshore Jurisdictions are also key locations and used to bypass policy constraints elsewhere (related to World Governments’ ability to tax, regulate and mandate transparency of financial instruments and contracts) and achieve better legal protection.

A particularly key overlap between GPN and GFN (at the centre of the figure) are the actions by World Governments to intervene in these production and financial networks. World Governments are governments and inter-governmental organisations that exercise extra-territorial authority, and contrast with Other Governments that cannot act at the global scale. This authority is sometimes through violent means (such as wars, colonial expansion and sanctions) but also via their own capital (e.g. through national oil companies or sovereign wealth funds). These World Governments conduct fiscal and monetary policies affecting oil markets; and intervene when private markets malfunction, thus offering sovereign protection over Financial and Business Services, Financial Centres and Offshore Jurisdictions, and providing confidence to the whole oil GPFN.

The GPFN is not a GFN superimposed on GPN or vice versa. Instead, the GPN and GFN overlap and co-exist in a close relationship. It is about finance with production, not one over the other. Money used by Transnational Companies comes from the financial sector, which creates credit money backed by governments. Money leaving Transnational Companies is managed by the financial sector and some of it reaches governments in the form of tax. Business services, including law, accounting and consulting help create financial assets (law), manage stocks and flows (accounting) and establish standards and conventions of the industry (consulting and other) (e.g. Pistor, 2019). All these services, financial and business, are rendered for companies and governments alike, and essential to interactions among them. Likewise, both companies and governments exercise their power through Financial Centres and Offshore Jurisdictions (Haberly and Wójcik, 2022).

Without stages of development or a time axis, the GPFN framework can look static. However, while the categories of key actors and geographies depicted in Figure 1 go back centuries, to the roots of capitalism itself (Haberly and Wójcik, 2022), specific actors and geographies change with time. Change in GPFN is slow. This sometimes related to the momentum of materiality found in fixed infrastructure (although this varies considerably by industry) but GPFN is particularly affected by the inherent inertia and stickiness of power in the GFN. While new geological finds and technological change can shift the geography of production in a matter of years, World Governments and leading Financial Centres take from decades to centuries to develop and decline.

We apply the GPFN approach to the global oil industry to offer several contributions to literature. First, this approach brings together economic geographical studies that have typically focused on either production or finance but rarely on both aspects of oil at the same time. In a seminal paper, Bridge (2008) outlines a global production network (GPN) approach to oil, but not the role of finance. In an empirical application of his framework, Breul et al.’s (2019) analyse Singapore’s gateway role in the Indonesian and Vietnamese oil and gas GPN, but hardly mention Singapore’s role as an international financial centre and the hub of commodity trading (both financial and physical). Similarly, Stephenson and Agnew (2016) describe differentiated networks in the Russian oil and gas sector with little consideration of financial networks beyond a short reference to Chinese financing for Arctic oil and gas exploration.

Conversely, research that focuses on the oil-finance nexus can suffer from a physical reality-financial fiction dichotomy. As Hanieh puts it: ‘much of the literature on the financialisation of commodities tends to adopt a dualistic approach to financial markets and physical producers, where financial and non-financial activities are assumed to be externally related and counterposed to one another’ (Hanieh, 2021: 70). In a seminal study addressing this problem, Labban characterises the 2008 boom and bust in oil price as driven by the financial logic (with commodity traders as the lead actors) rather than resulting from supply-demand conditions. In his words: ‘finance has emancipated the circulation of oil in the world market from its circulation in physical space, fragmenting the oil market into a physical and a financial component, but reintegrating both under the dominance of financial logic without transcending their duality and their differences’ (Labban, 2010: 541). Crucially ‘This is not a dualism between a “real” space-time of material circulation and a “fictitious” space-time of financial representations. Both are real enough and have their own materiality’ (Labban, 2010: 542). The goal of our paper is to use the GPFN approach to be explicit about these physical and financial materialities by identifying key actors and geographies in oil networks and their interactions.

By proposing GPFN we engage with the close relationship between GPN and GFN approaches, which has been pointed out since the inception of the GFN (Coe et al., 2014). We do not see GPFN as a new theory, but rather a framework embracing both GPN and GFN and building on decade-long theoretical development of GPN and GFN. Our aim is to highlight a specific articulation of GPNs and GFNs, and how this has been key to the development of the oil industry in space and time.

Integrating production and finance the oil GPFN offers opportunities to follow and understand the flows of money in oil-related networks. While such investigation can start with the financial statements of oil companies, it needs to consider the cost of government investment in energy infrastructure and government revenues from oil. It must also include financial flows related to a wide array of Financial and Business Services active in oil trading and advisory. Through their activities and those of Transnational Companies, value in oil networks flows from exporting regions to Financial Centres and Offshore Jurisdictions. In this spirit, Loginova et al. (2020), for example, show how Australian oil and gas corporate networks connect Australian cities with global cities and tax havens overseas.

Understanding the actors and geographies involved in the oil GPFN and their interrelationships also allows a better understanding of price formation in oil markets. To be clear, the GPFN is not a substitute for econometric studies on the determinants of oil price (e.g. Dahl and Yucel, 1991) or a model that helps forecast it. Instead, it is a conceptual framework emphasising the institutional geography of the oil market. In this sense our approach shares commonalities with economic studies that go beyond a narrow focus on supply and demand, highlighting the role of broader market structures (Adelman, 1972; Verleger, 2015). It also builds on research that charts the political economy of oil (e.g. Davis, 2006), and the role of various groups of actors in oil markets and their influence on oil price (Fattouh, 2011; Turner, 1983).

To analyse the oil GPFN we use a wide range of data on oil production, long-term and short-term price fluctuations, oil infrastructures such a pipelines, ownership of oil companies and location of oil trading companies. We combine such data with historical narratives on the evolution of the oil industry, as well as industry, media and policy reports relevant to explaining recent developments. The next section presents the history and geography of the oil GPFN, focusing on the evolution of the six key actors and geographies from Figure 1 over five historical stages of oil market development. This long-term longitudinal perspective is complemented with an application of the GPFN to explain two major recent oil price anomalies in the following section. The first is the mid-term anomaly in the early 2010s, when the two major oil price benchmarks diverged from each other for several years defying theory and expectations of arbitrage. The other is the negative oil price in April 2020. While the long-term GPFN perspective revolves around the macro-economic geographies of oil production, trade and consumption, the mid- and short-term analysis requires attention to micro-economic geographies of oil pricing. The oil GPFN as a whole evolves slowly and with a lot of inertia, but short-term price anomalies represent space-time compression, when local material conditions and momentary financial behaviour of key actors can shake the market. The GPFN framework can be insightfully applied to both perspectives.

A short history and geography of the oil GPFN

In this section we present the evolution of the oil GPFN by focusing on its six key groups of actors and geographies over five historical stages. This chronology is inspired by existing work on the history of oil, including Guo et al. (2021), Parra (2004) and Turner (1983), and informed by the seminal geography of oil by Bridge and Le Billon (2017). The start and end of each stage are marked by specific events, but as in any periodisation, they are only indicative. Figure 2 does not aim at covering all institutions involved in the oil GPFN, and the inclusion and significance of those covered cannot be easily quantified. Where possible and relevant we use numbers, but mostly we apply the GPFN as a new lens to review existing work on oil and finance. Our goal is not to write a definitive financial history or geography of oil, but to highlight the evolving, inseparable and ever close relationship between production and financial sides of oil networks, and their role in the distribution of power and price formation.

OPEN IN VIEWER

Figure 2.

The evolution of the Global Production and Financial Network of oil.

From origins to monopoly – 1859–1911

The modern oil industry started when oil was struck in a well in Titusville, Pennsylvania in 1859. Wooden pipelines were used to move oil to railways for further transport to refineries. ‘Informal oil exchanges, where buyers and sellers could meet and agree on prices, developed in a hotel in Titusville and at a curbside exchange, near the railway tracks, in Oil City’ (Yergin, 2012: 17). In the 1870s more formal oil exchanges were established in Titusville, Oil City and New York, where oil was traded on a spot basis, with immediate delivery, as well as futures (Yergin, 2008). In the same decade, John D. Rockefeller, a refinery owner from Cleveland, Ohio, succeeded at developing the world’s first long-distance pipeline network, and purchased stakes in most existing US refineries. To consolidate his power, Rockefeller registered Standard Oil as a trust in New Jersey in 1882. Anti-trust laws forbade companies to own stocks in other companies, but New Jersey trusts offered a way to bypass these constraints, whereby stocks were just held in trust and owned by the shareholders of Standard Oil, not by the company itself. In 1885, Rockefeller moved the headquarters of Standard Oil from Cleveland to Lower Manhattan in New York (Yergin, 2008).

Relying upon New York’s status as the US leading Financial Centre, and the legal protection offered by New Jersey as an Offshore Jurisdiction within the US, Standard Oil monopolised the downstream and the midstream part of the oil production network in the US. This attracted growing criticism, with the company described as ‘really a bank of a most gigantic character – a bank within an industry, financing this industry against all competitors’ (Yergin, 2008: 84). Eventually, in 1911, the US Supreme Court ordered a dissolution of the company. It was divided into several entities, with Standard Oil of New Jersey (aka Esso) inheriting almost half of total net value (Yergin, 2008: 94).

Outside of the USA, the oil production network developed through the activities of imperial powers via presence and ties to leading Financial Centres. Key examples include the formation of Royal Dutch in 1890 in Amsterdam (to extract oil in Sumatra, part of the Dutch Empire) and the incorporation of Shell Transport and Trading company in London in 1897 by Marcus Samuel, later the Lord Mayor of the City of London. From its perch in London, Shell secured financial backing (including from the Rothschilds) and succeeded in transporting oil by tankers, mainly from the rich Baku oil fields in the Russian Empire (Blas and Farchy, 2021). Motivated by a common interest in shaping the oil GPN, namely transporting oil via the Suez Canal to compete with Standard Oil, the companies merged in 1907 to form Royal Dutch Shell. The only other significant producer of the era was Romania.

From its inception, oil production networks were inseparable from its financial networks. Financial Centres joined centres of production virtually as soon as the first wells became operational, and then became hubs of industrial consolidation. International mergers and acquisitions indicate a globalising scope of these production and financial networks. Oil companies working with Financial and Business Services also linked the production and Financial Centres with Offshore Jurisdictions. World Governments of the day played a key part, as they sought new sources of energy for their economies and militaries and found new ways of exploiting their colonies (Nowell, 1994).

The oligopoly of Seven Sisters – 1911–1951

Surging demand for oil from booming automobility and world wars was met by new discoveries (with Mexico, Venezuela and Middle East in the lead) and a GPN dominated by seven increasingly international companies. The Seven Sisters, as they were known, included: three descendants of Standard Oil (Standard Oil Company of New Jersey (Esso), Standard Oil Company of New York (Socony), Standard Oil Company of California (Socal)); two companies near Houston in Texas (Texaco and Gulf Oil); Royal Dutch Shell; and the British controlled Anglo-Iranian Oil Company. Production in Mexico was controlled by Royal Dutch Shell and US firms, while in Venezuela US firms were alone (Yergin, 2008). As Figure 3 shows the US remained the world’s dominant producer. The oligopoly was accompanied by vertical integration, with Seven Sisters growing upstream, midstream and downstream activities (Goldthau and Hughes, 2021). As such these firms moved oil from wells to refineries and customers, leaving little need to trade oil (Blas and Farchy, 2021). This meant they could control the price of oil and bred relative price stability (see Figure 4 below), though this was punctuated by the world wars and the great depression.

OPEN IN VIEWER

Figure 3.

Oil Production by Selected World Governments, 1918–2020.

OPEN IN VIEWER

Figure 4.

Oil – Crude prices 1861–2020 in 2020 US $.

The Rockefellers remained the largest shareholders in the oil industry, with their wealth expanded via the flotation of Standard Oil companies in the 1920s (Labban, 2010) on stock exchanges in Financial Centres, with New York in the lead, and Esso becoming one of the first US companies to exceed $1bn valuation in 1925. The heterogeneity of oil as a physical commodity, with different local costs, chemical features and local prices also spawned the industry of collecting, assessing, and reporting price information (Fattouh, 2011). In 1923 Warren C. Platt launched Platts Oilgram, a daily newspaper with oil prices. In 1928, Standard Oil companies, Royal Dutch Shell and Anglo-Persian Oil started using US Gulf Coast prices plus freight published by Platts as a benchmark in their transactions (S&P Global, 2023). Thus, Platts company established itself as key Financial and Business Services firm supporting the globalisation of the oil networks.

Growing oil wealth motivated governments of oil producing countries, such as Mexico, to seek better control over their resources. The first major milestone took place in Iran, where in 1951 the Parliament voted to nationalise the Anglo-Iranian Oil Company.

The rise of national oil companies – 1951–1980

Although the Iranian nationalisation of oil was prevented by US intervention, the larger context of decolonisation made nationalisation inevitable in the long run. This included the gradual development of national oil companies (NOCs) such as the relocation of the headquarters of Aramco’s, controlled by Socal, from New York to Dhahran in 1952 (Aramco, 2022), and nationalisation of Royal Dutch Shell assets in Indonesia (Yergin, 2008). To further increase their influence in oil markets, Iraq, Iran, Kuwait, Saudi Arabia and Venezuela established the Organisation of the Petroleum Exporting Countries (OPEC) in 1960, joined later by Indonesia and major producers in Africa, including Libya, Algeria and Nigeria. As crude oil production grew quickly in the Middle East and the Soviet Union, the USA lost its title as the world’s largest producer, surpassed by the Soviet Union in 1974 and Saudi Arabia in 1976 (see Figure 3).

The rise of NOCs accelerated in the 1970s, when the price of oil tripled during the 1973 oil crisis, triggered by the Yom Kippur War, and doubled during the 1979 crisis, triggered by the Iranian Revolution (see Figure 3). Beyond the immediate political events, the oil price rises were driven by OPEC decisions and expectations of impending oil scarcity, though the contribution of the three groups of factors has been debated (see e.g. Gately, 1984). In response to the growing power of OPEC and the oil crisis, the largest oil-importing countries founded the International Energy Agency, based in Paris, in 1974 (Bridge and Le Billon, 2017). The oil price surges contributed to high inflation internationally, influencing monetary policies of the era, ending with an unprecedented hike in interest rates led by the US in 1979.

The gradual erosion of the Seven Sister oligopoly and their impact on prices accelerated in the 1970s and pushes them and other international oil companies (IOCs) to seek new opportunities. As NOCs focused on upstream, IOCs specialised in downstream activities, resulting in a de-verticalisation of the oil GPN (Goldthau and Hughes, 2021). IOCs now needed to buy much of their raw material, which increased the scale of oil trading, run by specialised companies and IOCs themselves (Goldthau and Hughes, 2021). As the GFN of oil expanded, petrodollars earned by oil exporters were deposited and managed outside of the USA (to ensure that the US was unable to seize these assets, a particular concern of the USSR). These deposits were mainly in the Financial Centres of London, Luxembourg and Singapore, boosting their functions as Offshore Jurisdictions, and fuelling the development of Eurocurrency markets. Meanwhile, Argus, founded in London in 1970 developed into a leading price reporting agency alongside Platts, which moved from Cleveland to New York as it was acquired by McGraw Hill media company (Argus, 2023; S&P Global, 2023).

Globalisation – 1980–2000

During the 1980s and 1990s demand for oil hardly increased defying expectations and leading to falling prices (Huntington, 1994). The collapse of the Soviet Union in 1991, added Russian companies and other companies from post-Soviet states to the list of the NOCs, and pushed Saudi Arabia to the top of oil producing countries. In the same period, production in other OPEC countries shrunk, while Canada, China, Norway and the UK joined as major producers, undermining the influence of OPEC on prices (Fattouh, 2011; Mabro, 1989). Russian NOCs and those in other post-Soviet countries like Kazakhstan, needed access to foreign capital to keep their oil flowing to global markets (Blas and Farchy, 2021). These factors accelerated globalisation on both production and financial sides of the oil GPFN.

The continued decline in the IOC’s share of crude oil production led to a wave of mergers and acquisitions. The largest five emerging by 2000 were US-based ExxonMobil and Chevron (both acquiring descendants of Standard Oil companies), London-headquartered BP (descendant of Anglo Iranian) and Royal Dutch Shell headquartered in London, and Paris-based Total, putting Paris among major nodes in oil financial networks (Davis, 2006). The consolidation made use of various Offshore Jurisdictions, with companies in Bermuda, British Virgin Islands and Cayman Islands acting as significant centres of M&A activity, allowing corporate combinations with minimum tax and regulatory burden (Guo et al., 2021). The consolidation was mostly horizontal, rather than vertical, and was accompanied by the emergence of large, specialised international oilfield services companies, mostly of US origin, including Schlumberger, Haliburton and Baker Hughes (Guo et al., 2021). Both M&As and fundraising for the oil industry represented major profits for banks (mainly of US origin) arranging these deals, and slowly gaining expertise in the industry (Wójcik, 2012).

A second major development came from the very beating heart of financial markets – exchanges. In 1983 the New York Mercantile Exchange, where the main futures contract for decades was Maine potatoes, introduced futures for crude oil based on the West Texas Intermediate (WTI) grade of oil. The innovation was first dismissed as ‘a way for dentists to lose money’ (Yergin, 2008: 707), but soon industry actors realised that it was ‘cheaper to explore for oil on the floor of the NYSE’ than ‘under the topsoil of Texas or the seabed of the Gulf of Mexico’ (Yergin, 2008: 708). The International Petroleum Exchange in London introduced crude oil futures the same year, using the Brent Field grade of oil, but were suspended due to a geographical mismatch between the existing oil trade in the North Sea and the contract design. ‘[. . .] the futures contract was for North Sea oil to be delivered from or in tanks in Rotterdam or Amsterdam. The crude oil trade is, however, used to dealing in North Sea oil free on board at Sullum Voe, in the Shetland Islands’ (Ottino, 1985: 10). The contract was revised and reintroduced successfully in 1988 (Horsnell and Mabro, 1993).

The 1980s thus marked a major milestone in the evolution of oil markets and its prices. The world now had two major crude oil price benchmarks, WTI and Brent and markets for oil have become more globally standardised and integrated (Rodriguez and Williams, 1993). From the near-monopoly of Standard Oil, through the oligopolies of Seven Sisters, and OPEC, a more market-based system emerged, with prices co-determined on trading floors in the world’s leading Financial Centres (Bridge and Le Billon, 2017). Price reporting agencies continue to play a big part by collecting, assessing and reporting prices from physical oil transactions.

The age of peak trading – 2000 to present

While trading has been vital to the oil GPFN since its inception, by the turn of the millennium deregulation set the stage for its acceleration in both financial and physical forms. The Commodity Futures Modernization Act of 2000 opened US commodity markets more widely to global investors and allowed banks and other financial institutions to trade commodities on a larger scale. In the 2000s commodity trading volumes and values have multiplied, as did commodity prices until mid-2008. Oil led the charge in this growth, with the price of crude more than quadrupling between 2000 and mid-2008 (see Figure 4). Low inflation (despite high oil prices) and low interest rates (at least until 2022) have facilitated the age of trading in oil making credit and leverage (trading with borrowed money) cheap and abundant. Meanwhile, the growing geographical distribution of oil production (see Figure 3 above) and consumption (driven by the economic growth of emerging markets) has fuelled further globalisation of the oil GPFN.

During this time the structure of trading changed. Before 2000 trading was dominated by buyers and sellers of commodities hedging their costs and revenues (e.g. power generation companies or airlines), but post-deregulation trading has been dominated by commodity traders, working for specialised trading firms, banks, oil companies themselves and other financial institutions (Cheng and Xiong, 2014; Hanieh, 2021). Investment banks, led by Goldman Sachs, Morgan Stanley and JPMorgan (all headquartered in the Financial Centre of New York), traded futures and other oil-based financial instruments, but also engaged in the physical trading, transport and storage of oil, and were referred to as Wall Street refiners ‘dancing between the physical and financial worlds to profitable effect’ (Blas and Farchy, 2021: 111). They mostly withdrew from the oil GPN after the global financial crisis and the 2010 Dodd Frank Act, which constrained their dealings. While this increased the share of specialised commodity trading firms in the market, commodity trading has become more regulated, particularly in the US and the EU. In the 2010s the US government has also imposed multi-billion-dollar fines on US and foreign commodity traders and their financiers for bypassing US-led financial sanctions and bribing foreign officials (Blas and Farchy, 2021). This illustrates the extra-territorial power of the US as a World Government in the oil GPFN. Meanwhile, the oil industry remains a major source of revenues for large commercial banks acting as lenders to oil companies and commodity trading firms (Bridge and Le Billon, 2017; Urban and Wójcik, 2019).

Data on who trades oil (and their location) is opaque (Nesvetailova et al., 2021), but the best available estimate shows Switzerland with 35%, London 25%, New York and Houston (combined) 20%, Singapore 15% and other centres 5% (Goldthau and Hughes, 2021). In 2018, Switzerland hosted 282 specialised trading firms employing over 3000 people. Over 50% of these firms, and 65% of jobs, were in Geneva, the headquarters of Vitol and Gunvor, the biggest firms in this sector (Swiss Statistical Office, 2018). The second biggest centre was the Zug Canton, the base of Glencore, with diversified trading activities, and oil exploration in Africa. Though landlocked and without oil extraction, Switzerland is well positioned in the oil GFN given its long tradition of commodity trading including when traders in the 19th century used Swiss neutrality to arrange politically inconvenient exchanges between British and French merchants (Guex, 1998). Moreover, having access to the US market, but not within its territory provides advantages, such as when the US banned Iranian oil in 1979. Very low corporate income taxes (particularly in Zug and Geneva), a private banking industry, and a high level of financial secrecy, usefully combines features of Financial Centres and Offshore Jurisdictions.

IOCs have also increased their share of oil trading. Shell is considered the world’s largest oil trader, ahead of Vitol and Glencore (Blas and Farchy, 2021). BP’s trading arm is staffed by about 3000 people in London, Chicago, Houston and Singapore (Blas and Farchy, 2021). Both Shell and BP trade more oil every day than the daily production of Saudi Arabia, and 5–10 times more than their own production (Blas and Farchy, 2021). NOCs have entered the realm of oil trading as well, including the trading arm of Aramco, the largest oil company in the world. Offshore Jurisdictions are heavily enrolled in trading to bypass taxes and other advantages. For example, with just 36 traders in Nassau, Shell reported profits in the Bahamas of $848 million in 2019 and did not pay a single dollar in taxes (Blas and Farchy, 2021: 5). While it would be prohibitively expensive to shift oil physically through Offshore Jurisdictions, routeing trading profits through them is easy.

The ascendancy of IOC trading is part of more general corporate restructuring driven by the continued dominance of NOCs in oil production, and the challenge of transition towards lower carbon economies. This restructuring includes M&As, diversifying into gas and renewables, and lowering costs through outsourcing, offshoring and more flexible labour arrangements. This work has been facilitated by Offshore Jurisdictions in the US, with ExxonMobil incorporated in New Jersey, and Chevron in Delaware. As a result, the organisation of IOCs became more functional than geographical, resulting in more power concentrated in corporate headquarters, where financial expertise and trading reign supreme (Labban, 2014).

The rising role of Financial and Business Services in the oil GFN and GPN also manifests in the ownership and control of the GPN. The leading nine owners of US IOCs are: Goldman Sachs, Morgan Stanley, JPMorgan (investment banks); Vanguard, State Street and Blackrock (asset managers); Blackstone, Carlyle and Riverstone (private equity firms), with six of the nine headquartered in New York (Hanieh, 2021). In contrast, the NOCs from oil producing countries exist in a close relationship with sovereign wealth funds (Haberly and Wójcik, 2017). NOCs use Offshore Jurisdictions less than IOCs, and much less than specialised commodity traders (Nesvetailova et al., 2021). Offshore Jurisdictions, however, remain useful for them, as illustrated by the top three Chinese NOCs registering companies in the British Virgin Islands and Cayman Islands (Wójcik and Camilleri, 2015).

China has become a significant oil producer and its main consumer and importer. In addition to NOCs, it hosts Zhuhai Zhenrong, a state-owned oil trading company, and the largest trader of Iranian crude oil (Blas and Farchy, 2021). By one measure, the number of commodity futures contracts, China contains the world’s three largest commodity exchanges, Dalian, Shanghai and Zhengzhou (The Economist, 2021). This does not, however, translate into a central position of China in the oil GPFN. First, Chinese commodity trading focuses on agricultural commodities, and is dominated by retail investors, resulting in large volumes but small values. It is also poorly integrated with foreign markets. Chinese exchanges have no warehouses or storage tanks where commodities could be delivered outside of China, while foreign exchanges are not allowed warehouses in China. While Argus and Platts (now part of S&P Global) have presence in China, foreign investors’ access to the Chinese market is restricted to selected companies and contracts, traded mainly on the Shanghai International Energy Exchange (The Economist, 2021). In addition, Chinese authorities have a history of intervening in the market, with exchange executives moving between exchanges and government departments. To be sure, China does not lack ambition regarding the oil GPFN. ‘For officials in Beijing, the next step in the development of China’s commodity markets is to turn the country’s benchmarks into global standards’. (The Economist, 2021: 76). Meanwhile, the gradual re-routeing of major oil exports from the Gulf countries to East Asia has increased the role of Dubai as a Financial Centre trading oil and setting price benchmarks (Fattouh, 2011; Jacobs, 2014).

Pricing anomalies and the Global Production and Financial Network

Building from this historical contextualisation we now focus on the ongoing transformation of the oil GPFN via an analysis of the world leading oil price benchmarks (WTI and Brent), as the critical points where production networks meet financial networks. Through this we seek to capture ‘how oil moves across space from the messiness and environmental externalities of extraction sites, through complex pipelines and tankers, mediated by the global financial sphere and national polities, to consumption sites’ (Haarstad and Wanvik, 2017: 441). In particular, to understand oil price anomalies, we must consider midstream, upstream and downstream networks. ‘It is tempting to see the incredible complexity and embeddedness of networks of pipelines, rail lines, trucks and gas stations as a testament to the stability and deep territorialization of carbon society. Yet they are also highly vulnerable, contested and exposed, and thereby unstable’ (Haarstad and Wanvik, 2017: 443).

Since the late 1980s the WTI and Brent indices tracked each other closely (Figure 5). Whenever the spread became sizable it was profitable for market actors to ship crude oil across the GPN and close the gap in pricing; confirming market expectations of arbitrage and the law of one price. In short, it was an exemplar of a well-functioning global market and integrated GPFN.

OPEN IN VIEWER

Figure 5.

Crude oil prices, Brent and WTI from 1988 to 2021.

However, after more than two decades of almost perfect co-movement of prices, the situation changed in 2010 (see Heidorn et al., 2015). The magnitude of the spread (climbing above $10 in early 2011 and peaking at $30 later that year) and duration (continuing for almost 5 years) was unprecedented. The prevalent market wisdom assumed a swift arbitrage driving prices back to equilibrium. What was missing, however, was a full understanding of the role of the oil GPFN in shifting WTI from a global to a local price indicator.

Financial actors did not fully appreciate changes in the North American oil GPN that fundamentally affected arbitrage. New techniques for oil extraction – horizontal drilling and multistage hydraulic fracturing (‘fracking’) – made oil in shale rock formations economically accessible. As a result, oil production in the US and Canada increased by 46% from 2008 to 2013 (from 7.5 million to 11 million barrels per day) (US Energy Information Administration (USEIA), 2016). This rapid growth of production of Bakken crude (and Canadian tar sands, see Pasqualetti, 2009) ran counter to the design of the ‘midcontinent pipeline system [. . .] configured to deliver crude oil imported to the U.S. Gulf Coast and domestic production from West Texas to the refineries in the Midwest via Cushing, Oklahoma’ (USEIA, 2013).

The GPN of WTI has been rooted in the central southwestern region of the US, which emerged as a key site of petroleum extraction around the start of the 20th century, including major discoveries at Cushing, Oklahoma in 1912 resulting in storage tanks and related infrastructure (Kimes, 2005). Standard Oil further developed pipeline networks around Cushing to connect the region to the rest of the US and this logistical advantage helped maintain Cushing as a key node in the oil GPFN (Simpson, 2019) including the inflows of imported oil (c.f. the Seaway pipeline from Freeport, TX to Cushing constructed in 1976, see Figure 6).

OPEN IN VIEWER

Figure 6.

Major crude oil pipelines.

This material infrastructure also made Cushing an ideal delivery point for sales of WTI oil and led to the centrality of Cushing for GFN centred around the WTI futures market developed by NYMEX. This helped establish the Cushing based WTI index as a key indicator for crude oil pricing in the GPFN given it was a high-volume market with many buyers (the minimum WTI contract is for 1000 barrels) in a regulatory environment that made price manipulation more difficult relative to other indexes (Fattouh, 2011). In contrast to WTI, the Brent index is organised around a seaborne crude which can be delivered at several ports.

The emerging geographic mismatch in the US GPN depressed prices in Cushing and by extension widened the spread shown in Figure 5. The resulting bottleneck at Cushing (with already large existing supplies in storage tanks) made it difficult to arbitrage away price differences via the GFN and encouraged those with access to tank capacity to hold oil until prices recovered as alternative transportation, truck or train were already in high demand and more expensive (Simpson, 2019).

The continued function of the WTI futures market in the GPFN during this time is remarkable: Despite being a defunct index for both the global and US oil markets for about 5 years, WTI futures contracts survived relatively unchanged and trading remained constant between 2011 and 2015 (see Martén and Jiménez, 2015) for U.S. prices, even though crude oil imports (at higher Brent indexed prices) continued throughout this time (see USEIA, 2017). This continued construction of the US oil market largely around Cushing, and by this the stickiness of the oil GPFN is puzzling. Market participants trading WTI futures in the Financial Centre of New York did not even change the delivery point from Cushing to, say, Port Arthur on the Texas coast. The inertia for Cushing as a delivery point for contracts is tied to the specialised material infrastructure of crude oil delivery required for commodity trading within the GPFN (Simpson, 2019). It might be replicated elsewhere but neither quickly nor cheaply. Second, and more importantly, WTI traders in the Financial Centres developed practices around Cushing’s deep, liquid financial ecosystem that was also a source of profit for NYMEX and associated banks and brokers. This network effect made it attractive for buyers and sellers to remain and was especially important for larger offers seeking not to impact prices with their own order. Changing the delivery specification in a contract would shift traders to a less liquid market ecosystem, at least initially, and so it made sense to stay with a potentially misleading indicator rather than to abandon it for a new one (see Christophers, 2015).

Simpson (2019) shows that storage capacity in Cushing strongly expanded starting in March 2011, efforts attributed to holders and speculators seeking to profit from later price increases. Ultimately the materiality of the GPN and market was reassembled by reversing the flow of the Seaway pipeline (see Figure 6) in 2012. This change meant that Cushing could readily remain the delivery point for WTI oil contracts as was preferred by futures traders. However, just getting oil to the coast was not a sufficient reconstruction of the WTI market structures given that US crude oil had not been legally integrated in the oil GPFN. US policy originating from the 1970s oil crisis banned crude oil exports. Accordingly, prior to fracking the oil market was constructed around the US’s role of a net importer and many refineries in the US prefer imports to domestically-produced crude oil (due to differences in oil qualities) creating a GPN mismatch between the new domestic supply from fracking and customer demand (Johnson, 2015). This territorially-defined, regulatory constraint on the market ultimately disappeared when the export ban was lifted by the US government in December 2015, making it possible to sell crude oil from fracking for international export. While U.S. producers had long advocated eliminating the export ban, the sustained low prices for oil in the US’ WTI market provided a useful argument for reconstructing the GPFN so that US producers could profit from higher international prices. As a result, in 2015 the long-running spread between the WTI and Brent indices ended, and the ‘global’ oil market was in sync with itself again.

However, this was not the last time the localised physical infrastructure of the GPN played a key part in the oil GPFN: Even more dramatic is the short, but extremely sharp, divergence between WTI and Brent on April 20, 2020 when WTI was priced at −$37.63, $55 lower than Brent. This unprecedented negative pricing resulted in considerable media attention and caught many investors by surprise - for many, it was simply inconceivable that the price of oil could go below zero. To understand how negative oil prices occurred first requires a closer look into the oil GFN, defining exactly what turned negative: that is, the price for a WTI contract traded on NYMEX to be delivered in Cushing, Oklahoma in May 2020. This particular contract was expiring and as a result financial trading was shifting to later contracts, that is, June 2020 and beyond. This meant that on April 20, market liquidity in paper oil with a May delivery was thinning with few buyers and sellers left in the GFN who were primarily interested in delivering or receiving physical oil. It is noteworthy that WTI futures with June and later expiration dates (as well as all Brent futures) never became negative. In other words, the ‘negative price of oil’ seen within the GFN that turned markets upside down, was an extremely specific and localised portion of the oil GPN.

The larger context for negative oil prices stems from anti-coronavirus measures and the depressed economic climate worldwide resulting in reduced demand for oil (Kumar, 2020, see Gardner et al., 2020), concurrent with an increased supply from most global oil producers (Hanieh, 2021). In the GPN, Cushing, Oklahoma experienced increasing supply and a dwindling demand for actual physical oil resulting in storage filling up at record speed (Flowers, 2020), a frantic search for storage (Bousso et al., 2020) and little spare capacity in mid-April 2020. Adding to these constraints was the April 13th decision by the pipeline operator from Cushing to the Gulf coast, to partially reverse the pipeline to let shippers move crude from Houston to Cushing and access the storage capacity there (Marino and Tobben, 2020). In short, the materiality of market infrastructure was contributing to a bottleneck in the oil GPN centred around Cushing.

Attracted by the apparently very cheap oil prices, traders bought May 2020 WTI futures despite low market liquidity and the extreme localness of the so-called ‘global’ oil market at that moment, that is, the need to take physical delivery of at least 1000 barrels of crude oil in Cushing. Many of these traders were new: the trading platform Robinhood reported an increase from 8000 users in February to 222,000 users in April holding USO, an oil-related fund (Bouchouev, 2020a; Schwartz, 2020). Likewise, Chinese retail investors bought into an Exchange Traded Fund (ETF) called ‘Yuan You Bao’ (crude oil treasure) focused on WTI futures and marketed with the slogan of oil being ‘cheaper than water’ (Bouchouev, 2020b). As a result, on April 20th the oil GFN contributed to creating a market full of retail investors without the capacity to receive and handle thousands of barrels of physical oil and relatively little access to local information (see Clark et al., 2006 for a similar setting). The price dropped to zero and further to −$40 and settled at 2.30 pm at −$37.63. This fall was likely reinforced by margin calls (Schwartz, 2020), that is, traders needing to cover potential losses, which would further contribute to a sell-off. As the thin liquidity of the market became more apparent the remaining owners of futures sold at any prices. Meanwhile, many commodity trading firms, including Glencore, bought large amounts of US oil and put it in tankers in the Gulf of Mexico, which then waited offshore, mostly in Asia, for the oil price to recover (Blas and Farchy, 2021).

In summary, the drop of April 20 was primarily fuelled by retail investors investing in WTI contracts in March and April to ‘buy the dip’. To some buyers in the GFN oil prices looked cheap, but only to those who were unaware of the material and localised constraints of the oil GPN in Cushing. In short, it is the trading infrastructure built on the interrelation between the oil GFN and GPN that led to this moment of crisis in the oil market. If investors had displayed a fuller appreciation of the oil market’s GPFN – taking explicitly into account the very local nature of delivery in Cushing, local storage capabilities and pipeline directions as described above – they could have avoided heavy losses driven by negative oil prices.

Conclusions

In this paper we proposed to combine the GPN and GFN frameworks into a GPFN and applied the combination to oil as one of the world’s most important commodities. While the GPFN is not a new theory but a specific articulation of GPNs and GFNs, it helped us identify the key actors and geographies of oil networks and analyse their relationships and evolution over time. Our analysis overcomes the physical-financial dichotomy in the studies of oil and shows that oil networks have always been financialised. As a commodity trading executive said: ‘oil is just a form of money’ (quoted in Blas and Farchy, 2021: 16), so it was for John D. Rockefeller. Just as Bridge and Le Billon (2017: 3) argue that ‘the political character of oil is normal and continuous state of affairs and not an aberration or interrupting event’, so is the financial character of oil. The latter has evolved but oil has always been financial, just as it has always been political.

Our analysis highlights the stickiness of the oil GPFN. An oil strike from more than 100 years ago continues to influence the spatial configuration of the US oil infrastructure. Likewise, 160 years after the first wells were drilled in Pennsylvania, the oil GPFN is still populated by descendants of Standard Oil, from the Financial Centres of New York and London, with the US as the key World Government. At one level this stickiness reflects geographical and historical inertia with new geographies (additional oil reserves), actors (NOCs taking over from the ‘seven sisters’ oligopoly), and processes (trading emerging as the industry shifted from vertical to horizontal organisation) integrating with existing systems rather than forming completely new structures. China, with its NOCs and sovereign wealth funds added to Russia’s and OPEC’s; Shanghai added to New York, London and Geneva; Delaware and Switzerland added to New Jersey as Offshore Jurisdictions. Incumbent actors adjust to new entrants as well as political and regulatory changes as they seek (not always successfully) to retain power. Investment banks, for example, evolved from acting solely as financiers and advisers to the oil industry to becoming financial and physical traders of oil, but have since largely retreated from physical trading.

Our analysis of the development of financial and physical oil trading through our generalised framework has implications beyond oil. Using oil as an example, we show that commodity prices or prices in general can be analysed as a meeting point of GPN and GFN in the GPFN. When we consider the GFN and GPN jointly, a price is no longer the product of supply and demand operating everywhere and nowhere at the same time or like angels dancing on the head of a pin. The network focus of the GPN and GFN obliges one to identify specific nodes and connections, including pipelines and trading rooms, with accounting, law, and finance all intertwined with each other and with the physical and virtual flows of goods and data in a specific spatial and temporal context. This offers a quintessentially geographical approach to the topic of price formation as a ‘holy grail’ of economics.

The inseparable relationship between production and finance is key to understanding the transition from carbon-intensive fossil fuels to renewables. Just as the GFN is married to oil, it is married to fossil fuels in general. Financial and Business Services, Financial Centres, many Offshore Jurisdictions and much of public finance are addicted to fossil fuel revenues (Cojoianu et al., 2023). As regulatory and market pressures and incentives push the GFN slowly away from fossil fuels, renewable energy is becoming subject to similar forces as oil, with large-scale land acquisitions, investments in solar and wind farms, hedging and speculation on energy prices. With decarbonisation, GFN players, particularly Financial and Business Services, including commodity trading firms, can turn to other commodities, such as minerals to produce electric vehicle batteries. Oil companies are developing renewable energy capabilities too. Thus, there is a case for studying the emerging GPFNs of renewable energy.

As we write these words, we have seen another period of high volatility in commodity markets. Following the Russian invasion of Ukraine in February 2022, prices of oil, wheat, and many other commodities went up by 50% or more within a few months, unleashing high inflation, which has continued into 2023, despite commodity prices returning to 2021 levels. High inflation in turn, has triggered interest rate rises, which threaten debtors and banks that became dependant on cheap money. The war and sanctions on Russia are leading to a major shift in the oil GPN, with Russia doubling on its connections with Asian importers, with China and India in the lead. Oil companies and commodity trading firms are enjoying record-breaking profits, with benefits for both Financial Centres and Offshore Jurisdictions. The GPFN can shed light on these developments. To advance it, more collaboration is necessary among the fields of economic and financial geography, as well as geopolitics and economic history.