Mapping BP – using open data to track Big Oil


The idea that huge corporations can secretly control the world is part of our Zeitgeist. Campaigners like the Occupy movement articulate it forcefully but it is far more widely present in the Culture. Just look at the number of Hollywood films that have mapped onto this meme in recent years: the Firm, Antitrust, Erin Brockovich. It’s now such a commonplace it is appearing in kids’ films – the arch villain of Lego Movie is Lord Business. These days it is only the political fringes, the committed Left or insurrectionary Right, who see it as an addressable problem and declare an intention to do anything about it. But the belief itself is quite mainstream. For most of us just muddling along I suspect it often seems like a fact of life, inert most of the time but always there. Climate change is inevitable. The Middle East will always be at war. England will never win the World Cup again. Big business rules the world.

But what if you could take that proposition out of its normal barroom habitat and test it under laboratory conditions? Chewing it over with our friends at OpenCorporates, we thought the idea that Big Business operates “invisibly” should be a testable proposition. After all, companies are required to file annual accounts and returns and other disclosures in almost all countries around the world. What happens if you are deliberately “naive” enough to try and put all that together?

To cut to the chase: in three weeks, using BP’s own public filings, we managed to put together a “power map” of BP which includes 1,180 company structures in 84 jurisdictions in chains of ownership up to 12 layers deep – you can play with and download the dataset at This essay describes that process, what the data set says and doesn’t say, and what we think it means for the potential to create a new way of tracking business in its dealings around the world.


Why BP? It could have been any massive oil corporation – but then BP is a massive oil corporation. It employs about 100,000 people and its turnover, about $350 billion in 2013, ranks it as a bigger economic concern than the GDP of Colombia or the Philippines. Although it began life very much as a British company, and in fact the British government once owned a majority stake in it, BP has expanded out across the globe and with its acquisition of Amoco in 1999 has an extensive presence in the United States.

Of course, it still has high global recognition because of the Deepwater Horizon spill in the Gulf of Mexico in 2010, but the company’s history goes back over a century. Born as “Anglo-Persian” in 1909, it was one of the famous “Seven Sisters”, the original oil companies who operated as a cartel for decades and carved up the fields of the Middle East. BP was part of the consortium which ran the Iraq Petroleum Company, for example, which in the mid-20th century decided how much oil Iraq was going to produce, keeping production deliberately low for years at a time to prop up global prices.

More recently it had had been led by John Browne who at one stage was so adept at re-inventing both the company and himself that he had been up there with Richard Branson as a rock star CEO. At the end of the 90s, BP was “Beyond Petroleum”, aggressively exploring renewables while thinking about getting out of the dirty – and thin-margined – business of oil refining, out front in the financialisation of oil markets, developing a trading arm which rivaled any major in the world, in the vanguard on transparency, suggesting it publish its contracts with Angola.

As a Brit, I have to say BP has also long held a particular interest. It is the national oil company we never had. When the North Sea was developed in the 60s and 70s, Norway built its own oil company, Statoil, to run its sector. It may have been suitably understated and self-effacing, as we have come to know and love from Scandiwegians, but Statoil was a very real assertion of what is now called “economic nationalism”. There was a brief moment in 1975 when the late Tony Benn, then Minister of Energy, created the British National Oil Corporation. But it never developed production capacity, was shunted aside under Margaret Thatcher – and later acquired by BP. Thirty years later, Norway has the Government Pension Fund Global, currently worth almost $900 billion while we learned during the Deepwater crisis of 2010 that some 18 million British pensions were connected to BP stock through insitutional investments, as though it were also the nearest thing we had to a sovereign wealth fund.

We have no animus towards or about BP as a company. If I am being honest I would have to admit that the knowledge that Deepwater Horizon still triggers enormous resentment certainly played in our selection as we thought it meant a guaranteed minimum of interest in the way this company, in particular, organises itself around the world. In the USA, even Republicans normally friendly to Big Oil love to hate BP as crass and, above all, foreign, as though it were a kind of corporate James Mason or Alan Rickman, the posh Brit baddie. And that story is far from over. There are ongoing lawsuits at state and federal level and BP’s share price is effectively hedged against liabilities which are still unknown.

But our experiment was in no way an attempt to single out BP. Our work revolves around the “Resource Curse”, the fact that digging out large amounts of valuable stuff often ends up being bad news for societies in an amazing variety of ways. But it rarely seems useful, or right, to make it personal. A good friend of mine had worked directly for Browne and I have come across several other BP “graduates” in our work in recent years.

I should also say that by definition we were not expecting to find any smoking guns. We were piecing together public filings, not rifling around in dustbins or sticking shredded documents back together. Whatever arrangements were made we had more respect for the small army of accountants and lawyers who were paid small fortunes to make them, than to expect to find anything flagrantly violating any laws.

At one level what you might expect to get out of such a mapping is fairly straightforward: a big network graph of companies criss-crossing the world. In the last couple of years, we have all become used to these kinds of visualisations, whether they are of the Iranian blogosphere, the famous Kevin Bacon paradigm in Hollywood movies (who’s acted in which films with who), or state corporations in China.

But the question is what actual use would a mapped network be, beyond the ooh-ah of seeing a huge mesh of dots and lines connecting them, or “nodes” and “edges” as our network analyst friends taught us to say.

In trying to keep track of oil and mining industries, there are two particular issues connected to corporate networks.

The first issue is the “Bad Guy Issue” – corrupt access to natural resources, when dodgy companies are able to muscle in on lucrative oil and mining concessions because they are politically connected. This has happened in many African countries, Nigeria and Angola for example. But such companies are normally identified because they lack any serious industry profile or access to capital. Whatever anyone might think about BP, it clearly does not fall into that category.

But there is a second issue, the “Sharp Guy Issue”: whether complex corporate structures allow multinationals to engage in “aggressive” tax planning and use a technique known as transfer pricing to shift profits from, countries where they are operating, to low or no tax jurisdictions such as Bermuda or the Cayman Islands. This might be completely legal or at least open to interpretation. The question here is more whether, if governments in say, Africa, were able to see the whole corporate chain of the companies operatingthey might adjust their own taxation policies in light of that, or subject billions of dollars of tax-deductible costs submitted to their tax authorities to audit, or at least more rigorous examination.

That might sound abstract but not long ago I spent a morning in the headquarters of a national mining agency of a small African country meeting a “tax task force” from several different ministries and government agencies. Their job was to see if they could get more tax out of existing mining contracts and one thing that puzzled them was why the royalties they were seeing from iron ore production were so low. Even on paper, the government was only due 3% of the value of the iron ore. But what confounded them was why the company was submitting payments which seemed to be considerably less than 3% of the world market price. Well who are they selling it to, we asked. It turned out to be a processing company in China they were affiliated to.

Oil, gas and mining operations need to buy in hundreds of millions of dollars worth of goods and services to make the oil flow. What if they buy services in high from affiliated companies? And then, once they have produced their commodities, sell them on low, to other affiliated companies? Any tax stream sitting in the middle, such as a profit tax imposed by a small African country, for example, will be squeezed because the base from which it is calculated has just been eroded.

These are the issues which lay behind the stories about Starbucks and Google not having paid any profit taxes in the United Kingdom because they haven’t earned any profits there, despite operating sizeable businesses for over a decade. In these cases it seems as though profits shifted to Ireland, with a 12 percent corporate tax rate, because the UK operations were paying extensively for intellectual property provided by Starbucks and Google affiliates in Ireland. The nature of goods and services being swapped between oil and mining companies might be very different but the principle is the same. Make as little profit as possible in high tax jurisdictions and as much as possible in low tax jurisdictions.

In this sense, the idea that we live in a global free market is a myth. According to the best estimates more than half of the goods and services that pass across international borders every day are between affiliated companies – who may or may not be swapping those goods at market prices. Because it is such a major feature of world trade, transfer pricing is supposed to be regulated and there are complex agreements and pre-agreements and systems in place around the world, sometimes down to the level of individual enterprises reaching a specific understanding with a tax authority on how it will report its own transfer pricing. But it was clear that morning in thatrainy West African capital that very little of the knowledge and techniques to capture exactly who was doing business with who was available even to senior civil servants.

And they are not alone. The auditor-general’s office of a southern African country conducting its first audit of national agencies and state-owned companies in the mining sector also did not know who they were doing business with – as in, the precise legal entities who held contracts. The line ministry of a Central Asian country wondered about how to read accounts for costs submitted by a multinational from dozens of different suppliers. If you want to nerd out on the complexity of oil cost accounting, an Ernst and Young audit commissioned by the government of Uganda made it on to the Internet. It lists over 200 local and international companies supplying Woodside Petroleum, the concessionaire. But in fact these accounts are atypically small and simple: the contracts were for very early stage costs in developing a field that was not even going to produce for at least another decade, and the sums under review were less than $100 million, peanuts for early stage oil and gas development these days.

Of these two issues, the Bad Guy issue and the Sharp Accounting issue, the Bad Guy issue is the best known and the most directly emotive.

It could in fact be called the founding paradigm of the transparency movement. It’s the African president’s son who walks into a branch of Riggs Bank in Washington DC with two million dollars in cash. Or who has 15 Rolls Royces in Paris or 17 apartments in Monaco registered in other people’s names. The people who, if there were an underworld edition of Hello!, would dominate the multi-page spreads.

The Sharp Guy issue on the other hand is grey and mundane – and involves much, much bigger sums of money. Raymond Baker, who founded the US watchdog Global Financial Integrity, estimates that illicit financial flows out of the developing countries into the rich world total perhaps a trillion dollars a year. But of this maybe only five percent is bribe money and corruption associated with government officials. Another 30 percent might be criminal funds. But the other two thirds is money earned in standard business operations to make, buy and sell standard goods. It’s just that the money is not reported, or not reported for what it is, capital flight hidden in the terms of trade. Kofi Annan’s Africa Progress Panel estimated that Africa missed out on about $40 billion in taxes from these flows each year – roughly equal to the in-flow of aid to the continent.

And Sharp Accounting is an issue we might expect to be connected to the legal structures of the multinationals. After all, at some level one assumes the point of corporate accountants, the high fliers at least, not the amiable plodders who approve your and my expenses or tell us off for our sloppy paperwork, but who only think in millions and can make money evanesce around the world faster than you can say Thomas Piketty is that they are sharp. And occasionally the mask slips. A colleague, for instance, was recently poring through job ads for corporate accounting (because that’s the kind of thing we do) and found one which listed as a job responsibility “maximise tax deductible cost recovery”. Not “maximise the accuracy of cost recovery” as in, make sure everyone gets receipts, but just make the figure as big as possible.

Not that we have any particular reason to think that BP in any way… and so on and so forth…

Just Because

But for us, as we spent weeks grubbing away at the filings, there was a reason more compelling than specific applications of this kind of data like tracking corruption or money flows. Just Because. Because legal systems around the world imposed these filing requirements on companies generations ago. And they did that because there was a broad consensus that the quid pro quo for being allowed to form a company was a certain amount of transparency. And the reason there was a quid pro quo was that forming a company has always been, legally speaking, a privilege given under state law, not an inalienable right.

Transparency is often seen as a new thing. Born in the Nineties, at the time of the Washington Consensus and the Blair Ascendancy, marching in lockstep with concepts like corporate social responsibility and Blood Diamonds. But long ago in Britain, you needed an act of parliament or a royal charter to found a company – the British East India Company or the South Seas Company. It was only in the mid-nineteenth century that Gladstone introduced the Joint Stock Companies Act to allow relatively straightforward registration of companies and a few years later that the cornerstone of modern corporate law, limited liability, was enshrined in a separate act. Shareholders in a company could now only be liable for losses incurred by the company for any reason to the extent of their shareholding. In return, they had to register and file information about their activities. It was broadly the same story, on broadly the same timeline, in various jurisdictions of the United States. Recognition of the need for corporate governance and transparency, in other words, has been with us since the beginning of the modern corporation. It is as old as capitalism.

What has happened since is that the system has got way out of kilter. Raymond Baker says that when he began his career in the 1960s there were four or five tax havens, whereas now there are about 60. Back then, calling a different continent, flying there and dealing with bank accounts there was Bond-like and glamorous. Today there are a whole bunch of jurisdictions where you can set up a company while sitting at your desk, in the time it takes to brew some coffee and drink it and at a cost at which you might take a couple of friends to a not very fancy lunch.

Which brings me to a final pre-process point. Because of the speed and ease of these movements, a fair few people have thought we are on a fool’s errand trying to map BP or invest any effort in building corporate maps out of public domain data:

* “You realise they can create special purpose vehicles far faster than you can ever track them?” – lawyer working out of Dubai.

* “You’ll never get close to the real action and the real players. And you will give them cover. This could actually be dangerous,” – leading campaigner.

* A friend of mine who cut his teeth as an offshore corporate lawyer in the Carribean just listens, raises his eyebrow, and smiles.

We will see!

How we went about it

The decision to try and map BP’s global network was based on the fact that the company makes many public filings. We didn’t really appreciate quite how many until we had finished.

In the UK jurisdiction alone, the BP network we constructed shows 182 affiliate companies. Of these, the 26 which were second-tier companies, that is to say owned directly by the mother company BP Plc, made 157 filings in 2012 and 143 in 2011. The group probably then submits something in the region of 600 to 800 filings a year in the United Kingdom alone. Then we were able to access filings in probably a dozen other jurisdictions.

All in all, the BP group probably averages between 50 and 100 A4 pages of public disclosures per day to public authorities around the world, under normal conditions. That is, not counting investigations, legal proceedings, or any other event-driven process.

We didn’t know that when we started because we were at the start of the learning curve. So we signed up to the Companies House filings service, at a pound a time, and started pulling records.

We had come to the project ever so slightly armed with some ideas of where to look. Because of its profile, there are stacks of books about BP. We set ourselves a BP Primer out of three of them: Tom Bergin’s Spills and Spin, John Browne’s autobiography Beyond Business, clearly the founding declaration of the man to rehabilitate his reputation after his 2007 downfall, and the Evolution of a Corporate Idealist, by Christine Bader, mostly an account of her time at BP working on corporate social responsibility.

Through these we understood broad trends within the company in the past twenty years which we hoped to find traces of in the corporate hierarchies. Bergin, who had held the energy beat at Thomson Reuters for years and had huge experience and clearly great access, offered several major themes. Browne’s rampant acquisition strategy of the 1990s and all the Amoco and Castrol entities that must have adhered to the group. Since BP had moved aggressively into commodity trading, that suggested there should be prominent corporate investment and trading vehicles. As to individual theatres of operations, would there be any affiliates clearly linked to BP’s play with TNK in Russia, for example, or its expanded presence in Libya post rehabilitation of Muammar Gaddafi in 2004-5? Also, given that many of the vertically integrated oil companies seemed to manage retail operations in very atomised fashion, would we see companies that were essentially a group of petrol stations in Dorset, and another one for Cumbria?

We quickly found a huge variation of size and activity between one company and another. Some seemed not to have any turnover at all, while others, such as BP Exploration Company, reported a profit of 840 million pounds in 2012. So we began to identify which affiliates seemed to be the “monsters” – BP Global Investments, BP International, BP Holdings North America – and home in on them.

Each of these companies were showing direct subsidiaries in a list at the end of their annual accounts. So we began to build lists of these affiliates and, if they were in a jurisdiction we could reach online, to pull their filings too, to see if they in turn had any subsidiaries. As we found each new company, we recorded its jurisdiction of incorporation, full legal name and number of incorporation, and then stored all documents in directories which combined these into one string.

This was a key understanding from – the need to identify individual company structures. Not “BP”, or “BP Global Investments” but “BP Global Investments Limited, incorporated in the UK on March 4th, 1932, company number 00263889”.

The entire play in the way multinationals operate is in the interplay between the group as a co-ordinated whole, making a decision about how to invest in exploration in the Arctic, how to react to the US shale gas boom, or how to allocate this year’s profits, on the one hand, and the fact that this unified strategy is played out across over a thousand affiliate companies who each exist as a separate legal “person”. The company naturally seeks to maximise advantage across jurisdictions by combining these different legal persons in the most profitable and least liable way for any given business problem. But even if the group does act with one mind, the price of being able to maintain the affiliate structure as separate legal persons is a bare minimum of autonomous reporting by each of them.

It was as if the BP group is a superorganism and its affiliates were the constituent organisms included in the whole, like individual ants or coral. None of those companies had any purpose or would even survive without being integrated into the colony. Nevertheless, each of them has a unique footprint and what we were doing was studying the traces of their uniqueness, their “genetic code”, to see if significant information was stored there which could tell us something about the internal functioning of the colony.

We kept tabs day by day on this grinding process of building the network by attrition. By the end of the first week we had found about 400 companies. That seemed like a lot of companies, although we knew from an earlier study that in 2010 BP was thought to have had about 1500 companies. So we thought, gosh, well since we understand which files to pull down and where to look within them, it might only take another two weeks, three weeks max.

The team was made up of smart people with no specific experience in accounting or the oil industry. Part necessity and part design. The necessity was that we were on a tight budget and couldn’t afford people who had experience of corporate accounting, for example, or international tax structures, even if we could have persuaded them to stare at annual accounts until they went cross-eyed. But it was also choice since this was an ideal group to test the proposition that anyone with some brains and determination to rub together could grok enough of the basics to build these corporate maps. As well as Anton Rühling and myself, who belong to OpenOil, Claire, Avner and Miguel were grad students based in Berlin. They were intrigued by the hypothesis, clearly had a solid basis of research skills and it didn’t harm that two of them came from countries where hydrocarbons were either already a major feature of the political economy (Venezuela) or might shortly be about to become one (Israel).

Then we had a couple of breakthroughs. The first was we found an Annual Return from BP Plc for 2010. We had looked at a couple of annual returns and not found anything particularly extensive. But the 2010 Annual Return contained, half way through, a 65-page list of subsidiaries which were nested in a hierarchy so that we could see which were the direct stakes of BP Plc, which were the companies they in turn owned, and so on.

This gave us what we needed to make an initial stab at the entire BP network. But there were two issues: first, since the list was four years old, what changes might have happened within the network since then. Second, the document was very obviously scanned from a printed document. How could we digitise the data so that we could integrate it into a data system?


The discovery of the relatively complete list of affiliates in the 2010 annual return led us to look all the way through the other returns, and we found another list in the 2013 annual return. This was less complete and did not contain the full hierarchy. But the PDF was clearly the result of a digitally native document which had been converted into PDF, rather than the 2010 list which was a scan of a print out. This meant it might be easier to scrape.

These might sound like technical and trivial questions. But in fact the difference between a data set you can render into a digital format and one which you can’t is the difference between whether you can actively work with that data set or not.

To give a simple example, even in a text document. At OpenOil, we have spent the last couple of years working on making oil contracts, the primary agreements between governments and big oil companies, more accessible to public analysis.

Some, but not many, oil contracts have been published, offering a chance for independent analysts to go through them, take a look at the terms and evaluate them for public policy purposes. All the published contracts appear on the Internet as PDFs. But most of them are scans of printed documents, which means there is no way to simply extract the text of these 200 page agreements. But a small number are PDFs that have been generated from office programs like Word or Open Office.

If you have the contracts as extractable text you can do things like dump a dozen contracts in one large text file and then run searches against particular terms like “environmental liability” or “best international industry practice”. This instantly shows you how those terms are used across a small library of contracts, a basis of comparison with the reference in whatever contract you are looking at. Or you can search against the titles of individual contract articles to instantly find half a dozen instances of “Stabilization Clause”, for example. It’s the difference between a public library which is sorted and indexed and a private library where books are piled higgledy piggledy with no organising principle.

With structured data, such as these lists of BP affiliates, the difference between what you can do with “native” digital and analog data becomes more acute.

The 2013 annual return showed each affiliate as a “child” company and then listed its immediate “parent”. If that structure was available digitally, and so could be cut and pasted and searched and so on, then we could re-create the entire hierarchy by doing what computer programmers call recursion.

Here’s how it goes. We know BP PLC is the mother ship, so let’s call it “first tier” – the top level company. Any company listed as a “child” in the list which has BP Plc as its parent must then be a “second tier” company. For example, BP Global Investments Limited. If you go through the list again, any company which lists BP Global Investments Limited – or any of the other 34 direct “children” of BP Plc as its direct parent – must be a “third tier” affiliate in the network. And so on. If you know the top level, and every company in the list has a direct parent, you can re-create the entire family tree.

We could copy the list ourselves of course, type it in. Which was what we had been doing the first week. Courtesy of the Shuttleworth Foundation, we had put aside the time and I had a budget to pay a small team to spend several weeks doing that. But that’s hardly a scalable and sustainable solution, and we were interested not just in what the BP data set had to offer, but in figuring out a process to run on any major corporate network. And yet the irony is that we weren’t exactly students at the US embassy in Tehran trying to piece back together the CIA’s internal cables that they shredded after the Iranian revolution. We were simply trying to make sense of documents that had been specifically been prepared to go into the public domain!

So digital scraping became a necessary part of the process. You might think that technologies to do with rendering digital content, text and pictures and numbers and so on, would be a hard science, perhaps difficult to learn but with rules and laws as fixed as gravity. In fact it is a dark art.

Like programs which convert speech into text, which need to learn different accents and inflections, to choose from a range of possibilities based on context, and which slowly “learn” from experience, or Google Translate, which gets better and better as the system absorbs millions of snippets of implicit information which tell the administrators from user reactions to the machine translation which translations have worked and which haven’t, scraping programs need to be constantly refined and are context specific. One program is great at interpreting grids of horizontal and vertical lines as structured tables, for example, but can’t read curves as well so delivers much poorer accuracy on some letters than others. One program can “speak French” – interpret acute and grave accents – and another can’t. One program likes Sans Cerif fonts more than Cerif. And so on.

Eventually on the tenth morning, having used one program to split out the pages we needed from the document as a whole and another to align them all in the same way, I put the processed file into a third program and pressed the button… and out came a file we could load into Excel which displayed the 2013 list as data in cells. It was a beautiful moment.

The resulting data still needed scrubbing: it didn’t like foreign characters, it had converted all combinations of the letters “ri” into “n”, and half a dozen other quirks. But after a couple of days more checking against the original document we had a list which, while still not error free, was “good enough”, in the open data way, to put out to the world. We expect to hoover up the last remaining errors as the data set gets used.

What we had got was 1,180 companies in 84 jurisdictions in layers of ownership going 12 deep. Some of its results confirmed our rough guesses going into the project. No surprise that BP, which for decades was an organic arm of British imperial power, has many UK affiliates. We found 182 in total, including 26 out of the 35 companies that BP Plc owns directly (the “second tier” companies). But in fact that is only just over half as many as BP’s US affiliate structures – 384. Many of these seem to be inheritances from the 1999 acquisition of Amoco. The rest of the top ten jurisdictions of the BP network reads like this: Germany (198), Australia (50), Netherlands (48), France (39), China (28), Canada (26), Brazil (21), Egypt (17).

The group owned 672 of these affiliate structures, or 58 percent of the network, 100 percent. It owned more than half (50 percent plus one) of the equity of another 134 entities, giving a total of 808 affiliates in which the Group holds a directly controlling stake.

Just over a third, or 381 companies, had direct entries in OpenCorporates because they were incorporated in jurisdictions which have put at least a rudimentary index of their corporate registers online.

  1. The Socrates Principle

A realisation had slowly dawned on us as we worked our way towards a reasonably comprehensive mapping of BP’s corporate entities. We had struggled through hundreds of documents and boned up on disclosure requirements in a dozen jurisdictions to find the data that could build the network. But now we had arrived at what we called the Socrates Principle: we knew that we knew nothing.

What the filings gave us were equity structures. What about debt and other forms of capitalisation?

In general terms, we know that the amount of money flying around the oil industry at all stages in the value chain has multiplied in dizzying fashion in the last two decades. In the upstream, exploration and development costs have increased massively as oil is found in smaller quantities and harder places. Wood Mackenzie, an industry analyst, estimates that the costs of discovering a barrel of oil rose from one dollar to three dollars in the first decade of the new century. Estimates for the total amount spent on exploration and upstream development rose from about $100 billion in 2000 to maybe $700 billion by 2010.

In the downstream too, trading of crude oil and fuel products was “financialised” to use the term of art. Trading in futures and derivatives related to Brent and other benchmark crudes began in the 1980s and by the early 1990s were perhaps double the value of sales of physical crude, as major consumers of oil such as transport companies or industrial plants used the new instruments to lock in future supplies insured against price and foreign currency fluctuations – the original purpose of financial options and futures. But by 2011 the ratio of “paper oil” had increased to 20, perhaps 30, times the volume of physical oil as speculators and hedge funds moved into commodity trading.

Given all that, the equity structure alone gives relatively little information and the data set raises far more questions than it answers. Who are the other equity partners in the 508 ventures we have been able to identify in which BP owns less than 100%?How are those equity stakes, BP’s or others, leveraged by debt? What about other forms of debt, either between the companies in the group or across the boundary to the world of finance outside?

And then there is the question, in terms of figuring out patterns of influence and control, of the other business relationships emanating out of this network and these companies. If BP as a group turned over $350 billion in 2013 and its profit was $23 billion, then the BP group spent $327 billion. But who are BPs suppliers around the world? What kinds of contracts represent what proportion of the BP Group’s own budgets, and what proportion of the turnover of the companies it was interacting with? Who arranged those deals and what, if any, is the interaction of management and directors between each party in each transaction?

Oil industries are at the top of the food chain wherever they go in the world, generating billions of dollars of supply contracts. If its Aberdeen, or Stavanger, or Houston, maybe that plays out as the build-out of new supply industries to meet their demand for ships, logistics, rigs and just about everything else. Erle Halliburton, for example, started his eponymous company in the 1920s which grew to employ 100,000 people today by being world leaders in the techniques of how to case oil wells with cement. Conrad Schlumberger started a business at about the same time, again risen to a global behemoth employing 126,000, more than BP itself, on the advance of using electrical conductivity in well logs to get a better read on how to drill faster, cheaper and more successfully. Neither company has ever, to my knowledge, owned an oil concession directly, or negotiated a production sharing contract with a government. They bill themselves simply as service providers.

The oil industry globally turns over something like three trillion dollars a year. Even with wide profit margins, and all kinds of processes adding value in the “midstream” (transporting and refining) and “downstream” (fuel products, petrochemicals and plastics and retail marketing) this means the “upstream” – discovering and producing crude oil and gas – is generating hundreds of billions of dollars of business a year for sub-contractors.

Development economists want to catalyse this big business sector for maximum advantage to the economy as a whole, and in development economics there is often eager talk of creating more service industries, and maybe even towns like Aberdeen or Stavanger, across Africa and the global South, where oil and mining industries represent the lion’s share of foreign investment. At the same time, it is clear that in many countries where governance was shaky to begin with that a lot of corruption is bound up in the service sector.

It may even be that the corruption action has moved more to the “service sector”, as it is known, in response to emerging norms around transparency in the oil and mining industries in a kind of bubble effect. A lot of governance mechanisms, such as the Extractives Industry Transparency Initiative or new disclosures legislated in the EU and the USA, focus on the primary relationships between concession holders such as BP or Shell or ExxonMobil, and governments. What the terms of the primary contracts are which govern access to the oil fields, how much taxes of what kinds are paid to which state agencies, and so on. The service layer is much harder to get to, partly because it usually involves relationships between two private companies, not one company and a state, and so there are fewer tools available to create flows of information and insight.

Our painstaking Equity Map for BP tells us relatively little about all this.

Nevertheless, we set out to see what starting points for stories we could find in the data that might, in combination with more and better information and more and better knowledge, could be developed to provide greater insight.

    1. Tales from the upstream: BP returns to Iraq

I have been coming and going from Iraq since 2007 working on transparency in the oil sector there. BP was part of a group of international oil companies who went back into Iraq in 2009 when it won the contract to develop Rumaila, one of the largest fields in the world. Despite Iraq’s war torn and delapidated infrastructure, which had led oil minister Hussein Shahristani to take the risky political step of structuring deals to allow the internationals back in, Rumaila was already producing over a million barrels a day when BP signed the deal and held an estimated 17 billion barrels of reserves – comparable to what is left in the entire UK North Sea sector. As part of its contract it committed to raise production to nearly three million barrels a day – more than the UK and Norway combined.

In 2010, BP tendered a service contract to support its own work in the Rumaila field which was reported to be worth more than $500 million. At least four companies – Schlumberger, Weatherford International, an Iraqi state-owned company called Iraq Drilling Company and a Chinese firm Daqing Drilling – took various pieces of it. Another two contracts worth another $100 million went to companies called Centrilift and Al-Khorayf Petroleum to provide submersible electrical pumps.

As discussed, none of these contracting relationships, worth $700 million that we know of, are likely to appear in an Equity Map. What, then, does show up?

The 2013 Annual Return does not name a single BP entity registered in the Iraqi jurisdiction. There are three companies with “Iraq” in the name and one BP company which used to have Iraq in the name but now doesn’t.

BP Iraq NV is registered in Belgium. It then goes through a four stage chain up to the mothership BP PLC through three intermediate companies all registered in the UK: BP Exploration Operating Company, BP Exploration Company, and BP International Limited. The 2013 Annual Return states that the first step up is 99.99% owned by BP Exploration Operating Company, not 100%.

Britannic Investments Iraq Limited is incorporated in the UK. It hops to the Mothership in three, through Britannic Strategies Limited and BP International Limited – all UK vehicles.

Iraq Petroleum Company Limited is a fascinating piece of corporate history. IPC was the Seven Sister joint venture company which dominated Iraq throughout the colonial period and for three decades after its nominal independence in 1932. The BP group has a 23.75% stake which jumps in four hops to BP Plc through BP Exploration Company (Middle East) Limited, BP Exploration Company Limited and BP International Limited.

BP West Aru II Limited now seems to operate in Indonesia. But the same company – with the same UK incorporation number – was called “BP Iraq Limited” between December 2, 2010 and April 15, 2011. Interestingly, before that it was called “BP Rumaila Limited” between July 3 2008 and July 1, 2009.


The Equity Map doesn’t tell us where the major money flows from Iraq are passing through the BP corporate structure. All we can do is inch forward, factlet by factlet using what we can find in company filings.

The two current UK companies whose current names include Iraq don’t appear to be handling any of the major financial traffic. The annual returns for Iraq Petroleum Company in fact appear to show a dormant company despite its momentous past. In the years 2010 through 2012 net assets were just over 500,000 pounds. Britannic Investments Iraq Limited showed some signs of life: director emoluments were listed at 90,000 pounds for 2012 with net assets of just over a million pounds. And its direct parent, Britannic Strategies, shows shareholder capital of 30 million pounds in 2012. But nothing like the scale you would expect to see for a company processing any significant part of BP’s share of revenues from a million-barrel-a-day field, whatever the terms of the contract.

BP West Aru II Limited shows assets of about 10 million pounds now under its new name, which appears to be connected to an offshore gas field in Indonesia, although no turnover. We could take a wild conjecture and say maybe the assets are some calculation against reserves proven in that field, or the cost of acquiring the exploration license, which is usually accounted for as an intangible asset. That would explain how the company’s assets jumped from nothing to 10 million without either revenue or capital coming in to the company in any other noted form. But in any case, in the years in which it was called BP Iraq Limited and BP Rumaila Limited, from 2008 to 2010, it was completely dormant according to the Annual Accounts: no turnover and no listed assets. It was audited by Ernst and Young but seemed to have no activity.

That leaves BP Iraq NV. Here we have even more indirect information. The NV suffix stands for “Naamloze Vennootschap”, roughly translated as “nameless partnership”. The Equity Map as a whole shows seven companies with this suffix, of which five are in Belgium (out of a total of nine companies in total in Belgium) and the other two in the Netherlands. Globally, this precise structure, or at least this precise term describing the structure, is only in use in the Dutch speaking world – the Netherlands, Belgium, and Suriname, and formerly also the Dutch East Indies. It stands in contrast to B.V., a more widely used company format. Both are limited liability companies but the N.V. allows what are known as “bearer shares” where the identity of the owner does not need to be recorded, whereas the B.V. only allows registered owners. In other words, all the people or companies who own shares in a B.V. can be known in principle, at least to the authorities, whereas with an N.V. ownership can also be anonymous. Various financial websites state that N.V. structures are used when a company may want to allow its shares to be traded on a stock exchange. The website for the Belgian company registry provides no further information in terms of either company filings themselves or a list of what filings have been made. But BP’s own website lists BP Iraq NV as the contact for its Iraq operations, including a long description of the development of Rumaila field, albeit with a phone number and company address care of the BP group headquarters at Sunbury on Thames, just outside London.

As usual, this search presents more questions than it answers. Three leap out at once.

First, it seems unlikely the anonymity feature of the structure is significant, at least at this stage. BP Plc after all has listed its own ownership of 99.99% of the shares in filings to UK Companies House, and publicises its connection to the company on its website.

But second, who owns the remaining 0.01%? And what are the reasons for this allocation?

Third, it would also seem that the trade-ability of the shares may not be significant either, at least in the case of BP Iraq NV, since it is almost entirely owned by the BP Plc group which is the only traded company in the Equity Map.

A bit more grubbing around allows us to add some flesh to BP’s network in Belgium.

A Belgian finance website shows annual accounts for BP Iraq NV which shows it jumps into high volume activity almost directly from its foundation which is listed as September 2009 – just at the time when negotiations with the Baghdad government were concluded and the deal for Rumaila was announced. It shows turnover of just under a billion euros in 2011 and just over a billion in 2012 and a Belgiangovernmentwebsite lists directors who include Michael Townshend. Townshend is mentioned in numerous media reports as president of BP Iraq. Clearly, then, unlike some of the other companies, BP Iraq NV is where at least some of the Rumaila action is.

But is that it? Maybe, maybe not. A little bit of triage on BP’s Belgian network leads us to another company.

BP IFC Belgium BVBA is only a distant relative of BP Iraq NV. If we were to use family relationship terms to describe these equity structures, and refer to direct owners as parents and direct subsidiaries as children, for example, BP IFC would be BP Iraq NV’s first cousin once removed – its “grandparent” would be BP International Limited, which is BP Iraq’s “great-grandparent”. And even this distant relation has to be put against the context that the common ancestor company, BP International Limited, is one of the second tier monster vehicles. The 2013 data show it with 49 direct subsidiaries in 11 jurisdictions. BP International Limited is like some tribal patriarch fathering untold children. Their second and third generation descendants can’t all be expected to be intimate with each other.

BP IFC Belgium BVBA shows a massive jump in assets during 2011 and 2012, just the time when the BP contract in Rumaila is kicking in, from about a billion euros in 2010 to 25 billion or so in 2011. Could it be related?

Probably not, in fact. We are probably making 2 and 2 = 37. The tricky thing with investigating is you want so badly to find something that you are in danger of seeing things which aren’t there. Now that we have understood how to access company filings in scale, and poke around in them, and structure them so that patterns can be seen across disparate data sets, maybe we are in danger of seeing how everything is connected to everything! To the man with the hammer, everything is a nail, as the Sufi saying has it.

In fact, BPIFCslisting on the Belgian government website states that it absorbed two other companies in 2012, BP Capital EURO V 0 F, and a company it lists as BP Capital AUD. These both have huge assets which would explain the bump in BP IFC, although it is not clear why the merging of these companies at the end of 2012 would have shown in the asset structures of BP IFC in 2011. It is also interesting that the same Belgian government website states that BP IFC has now been wound up, when it had an estimated capital of $35 billion.

But we have reached the end of our not very extensive knowledge of corporate accounting.

What, then, have we got, exactly? The precise identity of one company, BP Iraq NV, we can be reasonably sure is handling much of the financial traffic from BP’s strategic operations in the Rumaila field, producing a million barrels a day. And its corporate chain back to BP Plc through four hops. And the likelihood that this structure is geared around tax incentives. The Tax Justice Network’s Financial Secrecy Index for Belgium states, for example, that Belgian company structures are widely used as “internal banks” or corporate treasuries by multinationals including, in the oil sector, ExxonMobil and Total as well as BP.

Belgian companies get a 95% exemption on taxes on dividends of profits from their subsidiaries, wherever their subsidiaries operate. So if any BP Belgian entity received $1 million in profits from a subsidiary, it would only have to pay tax (of 33.99%) on $50,000, or an effective tax rate on dividends of 1.7%. Presumably, because Belgium also has an extensive network of double taxation treaties with other countries, BP may then be able to re-export the dividend income to BP Plc, and the 18 million UK pensioners who hold positions in its stock through funds, among others, with no further charges.

Could BP’s Iraq operation also be taking advantage of another feature of the Belgian system, tax relief on interest earned by lending to other companies in the group? If, for example, BP IFC lent to some its UK relatives, the UK relatives could discount the interest payments against their profits, while at the same time the BP IFC could get relief on the profits it made from the interest payments. Many multinationals borrow and lend to themselves in this way. In this case, that could mean, then, that some UK affiliates would borrow money from Belgian affiliates racking up interest charges that decreased the profits the UK companies pass up into the BP Plc group reporting. BP IFC and the other Belgian vehicles would discount some of the interest from their taxable profits using the provisions in Belgian law. But then once that was done, because Belgium and the UK have a double taxation treaty, BP IFC could repatriate profits including these interest charges back to the UK Plc group, which would be exempt from further taxation in the UK, so adding to the group’s overall profits.

So no smoking guns. And lots of unanswered questions. And more than a few red herrings. All about par for a data expedition based on partial information and exiguous domain knowledge. But perhaps, the beginnings of the shadow of a glimpse into how the money is flowing from southern Iraq into the BP group.

A few shorter tours round the Equity Map and we started to get the hang of reading the contours of affiliate structures. Each yielded nuggets of insight.


It’s a fact slowly receding into the mists of oil history that Indonesia was an early member of OPEC. The Dutch found significant quantities of oil there during the colonial period, which played a role in Japanese military strategy during the Second World War, and by the mid-1970s Indonesia was producing 1.5 million barrels a day, a level it held more or less until 2000 and its national oil company Pertamina for a time was right up there alongside the Middle Eastern national champions like Kuwait or Qatar. But over that period, the country’s soaring population and rising standard of living, even under Suharto, meant that its internal energy consumption rose more than seven times until it equalled the amount of oil it was producing. Then production began to fell as mature fields declined and companies did not make enough new finds to replace it. Energy demand continued to rise. By 2009, it was ten times what it had been in 1975, even though the population had only doubled. That same year, Indonesia left OPEC, having become a major energy importer.

Nevertheless, because of its size, proximity to Asian markets and its own rapid economic growth in more recent years, Indonesia remains a considerable energy prospect and most of the Big Oil pantheon is active there in production or exploration. Indonesia is technically interesting in oil industry terms because it was an early developer of Liquefied Natural Gas production and coal bed methane projects and much activity remains around the upstream sector.

The four BP affiliates registered as Indonesian companies showed different and multiple paths back to the mothership.

One starting with P T Amoco Mitsui PTA had a chain which jumped out of Indonesia at the first hop to a string of no less than six stages of ownership in the US. The first two up also have Amoco in the name, giving rise to the possibility that this is a corporate chain inherited from the BP acquisition of Amoco in 1999 and left unaltered since. A second chain starts with the Indonesian firm PT Cakrawala Tata Sentosa and then passes through three companies with “Castrol” in the name, ending with Burmah Castrol PLC which is directly owned by BP Plc. This suggests another chain leftover from BP’s acquisition of Castrol in 2000. The third chain starting in Indonesia with P T Jasatama Petroindo hops straight to BP Global Investments in the UK and then to BP Plc.

When a friend who has worked in the oil industry saw this he commented: “The company structures are records of corporate strategy. But the thing not to forget is that sometimes that record might be fossilised. So if BP takes over Amoco, maybe key holdings and strategic assets will be fully integrated and rationalised. But with this Amoco chain (in Indonesia) it might cost, who knows, $30 million to unwind all the different stages. Why would you do that unless there was a compelling business imperative?”

If that principle applied here, we would indeed be seeing a fossilised record. Amoco ran some operations in Indonesia and so did Castrol. When BP bought both companies in 1999 to 2000, it simply left these particular chains intact. Only the top level then reports into the pre-existing BP network of affiliates.

But the other thing to notice about Indonesia is that there are many BP companies active in the country which are not incorporated there.

BP’s flagship investment in Indonesia is the Tangguh liquefied natural gas project. An extension to the project announced in 2012 would cost an estimated $12 billion, just to add a third “train” liquefying natural gas to the two that already exist. There are other big foreign investors, notably the Chinese CNOOC and Mitsubishi from Japan, but the BP website states that BP owns the largest share, some 37.16%.

Tangguh was clearly a strategic investment. LNG projects were the next big thing in energy markets about a decade ago when gas had already gone a long way to replacing oil in the energy mix and the companies looked to developing LNG as a way of making it a globally transportable commodity like crude. The advent of the US shale revolution from 2005 on has challenged future global demand for LNG but the lead times on such projects are so long that in the case of Tangguh the investment was sunk, and irreversible. Production began in 2009 and is now set to increase, targeting presumably the East Asian market, where booming economies are stimulating demand, US shale gas production notwithstanding.

BP inherited Tangguh from another global acquisition, of Atlantic Richfield (Arco), in 2000, part of John Browne’s legacy of aggressive expansion. It was Arco which discovered the gas in the 1990s. Now BP’s structure reflects that. Its main vehicle for the project, BP Berau Limited, jumps straight to BP America Production Company. Then there are another four hops within the USA ending at the mammoth holding company B P Holdings North America Limited before jumping to the mothership. The interesting thing from our point of view is that BP Berau is itself a US company.

A second company, BP Wiriagar Ltd, exists separately within the Tangguh project. It is also owned by BP America Production Company and runs from there through the exact same chain up to BP Plc.

A third company, BP Muturi Holdings B V, is Dutch, owned by another Dutch company, then hops across to Spain before ending up at BP Global Investments in the UK and up to BP Plc.

All three chains attach to one physical installation in a remote part of Indonesia. One “project”, as it is known in industry circles. Twelve separate companies are involved in four jurisdictions. But none of them are Indonesian entities.

Then there is the coal bed methane project in East Kalimantan. BP’s interest is represented by three companies, VIC CBM Limited, Virginia Indonesia CBM Limited, and BP East Kalimantan CBM Limited, all of which jump through three hops to BP Plc – all British companies. Then there is a separate company just called East Kalimantan – registered in the Bahamas – which runs up through a company called Union Texas International Corporation which then picks up the same chain as the two US-related chains from the Tanguhh LNG project, with BP America Production Company.

Not forgetting upstream exploration. There are at least eight separate BP companies which appear to be involved in exploration in Indonesia – they have Indonesian placenames in their titles: BP West Aru I, BP West Aru II, BP West Papua I, BP West Papua III, BP Kapuas I, BP Kapuas III, BP Tanjung IV Limited and BP Exploration Indonesia Limited. All of them report straight up to the British company BP Exploration Operating Company Limited and then through an entirely British chain to BP Plc. And all eight are themselves British companies.

If you were to study BP’s networks in Indonesia starting from its affiliates that are actually Indonesian companies, you could build the initial network we saw – three chains of 15 entities across Indonesia, the USA, and the UK. But a more complete network such as we identified searching the entire network simply using the words “Indonesia” and known Indonesian placenames such as Aru and Kalimantan has another 24 companies – 39 companies in total – operating in another 15 chains, involving another three jurisdictions of which two, the Netherlands and the Bahamas, are high financial secrecy.

Starting from the assumption that all significant activity has a locally incorporated presence you would miss more than half the game.

It would be fascinating to know what level of visibility any of these companies, and BP’s overall networks, have to the Indonesian government and public, and how profits and taxes are calculated. Governments have regulations about foreign companies registering their presence, of course. But does the registration of a company as active in a country achieve the same visibility as an incorporation?

A couple of five minute glimpses:

The Cayman Islands

The interesting thing about BP’s network in the Cayman Islands is that all eight companies appear to be connected to the Caucasus: Azerbaijan, Georgia, and Baku-Tbilisi-Ceyhan pipeline. From the Caymans, they bounce straight up into the UK. And because of that, there is a heavy component of pipeline companies. We know that pipelines these days never cost less than billions, so presumably with large sums of money flowing there is some advantage to running financing through a low tax jurisdiction. But why wouldn’t there be other projects of this kind also running through the Caymans. A wild guess: many of the investment decisions for the BTC pipeline and the development of Azerbaijan – BP’s famous “Deal of the Century” – were made between the early 1990s and the early 2000s. Could the concentration of companies around these projects simply be a matter of a different generation of corporate policy? That for a time Cayman Islands registrations seemed like a good solution for set of business problems, and then after a time, perhaps after a change of guard, not any more?


Two things jump out from a brief (and roughly arranged) map of BP’s affiliate structures in Egypt. The first is that there are so many of them. In the mid-1990s Egypt did once approach a million barrels of oil a day production, and in more recent years has developed significant gas. But BP has as many affiliates incorporated in Egypt as in the rest of Africa put together. If you look at the names of the companies “North October”, “Temsah”, many refer to specific projects. We’ve seen this before in Indonesia but the difference is that in Indonesia they are exploration companies under British jurisdiction, whereas here they are Egyptian. A wild guess: the structure is imposed by the Egyptian legal regime, which was strongly influenced by Arab nationalism with its associated concern over strong national control of resources. A condition for developing a field might be to incorporate a separate vehicle within Egyptian jurisdiction. The second thing that is interesting is the Netherlands cluster: perhaps where the finance runs.

Sunbury Secretaries Limited

Some stories start to emerge from other information in public filings. Because our data display platform is at a rudimentary stage we have not yet integrated ways to visualise these data sets, so plain old words will have to do.

The filings of the UK companies, like many jurisdictions, also include the list of company officers, board directors and so on. Including the Company Secretary function. The company secretary is broadly speaking responsible for ensuring the legal compliance of the company.

On the first day of the data expedition, I downloaded a list of company officers for BP Exploration Company, one of the “monsters” in the network with a recorded profit of over two billion pounds in 2011. Intriguingly, the company secretary was not a person but a company, Sunbury Secretaries Limited. My lay understanding of the use of incorporation services was that you typically found companies performing these roles in relatively small shell operations. The classic Belgian dentist’s structure in the Caymans, for example, used to park a couple of million, might have one company performing the secretary role and another one in a director position. But how could that be the case for BP, and for an affiliate which turned over billions?

At first I thought the name was a clue. Sunbury on Thames is the headquarters of BP, a vast modern campus thoughtfully placed, connected to London by the Tube but with a very leafy Middlesex feel, big lawns, pleasing blocks of glass and steel and tasteful modern sculpture. Sunbury Secretaries Limited must itself be a BP vehicle which the group found more convenient to serve the role.

But no. Because Sunbury Secretaries Limited is itself a UK-registered company, we could pull the record. Founded in February 2010, its paid up capital is just one pound and its only shareholder is PriceWaterhouseCooper Legal LLP. Its business address in fact is just south of London Bridge in what appears to be a Price Waterhouse office building, and it sole listed company officer is Jonathan Gibson, the head of Price Waterhouse’s Global Entity Governance and Compliance Team.

A search on Sunbury Secretaries Limited on OpenCorporates yields 994 results, including both active and passive directorships. Not bad for a company which was founded four years ago!

We have not exhaustively checked all the entries, so I cannot say for sure that all its directorships are for BP affiliates. But that was certainly the case in the 30 or so records we accessed randomly and it would seem to make sense from the name that Sunbury Secretaries Limited is a vehicle to perform the company secretary role for the BP network.

Another intriguing thing was the dates of transfer of these directorships. BP Exploration Company’s took place – transferring away from a named person who presumably was a BP executive on July 1st 2010. A quick search found seven other companies whose secretary role had transferred the same day.

Immediately my journalistic antenna were buzzing! Deepwater Horizon!

July 1st was week 11 of the disaster. President Obama had by this time announced to a furious American public that he was going to “make BP pay”, and BP had been browbeaten into offering $20 billion to go into a compensation pot for the disaster. Tony Hayward was wondering when he would get his life back. And a whole bunch of affiliates had the company secretary position outsourced away from BP executives to a shell company with a paid up capital of one pound run by Price Waterhouse!

Was Sunbury Secretaries Limited used as part of the “Macondo Shield”?

The July 1st transfers were not Sunbury’s first outing as a company. In the week after it was registered itself the secretary role of three giant BP affiliates were transferred to it at the end of February 2010 – a few weeks before the Deepwater Horizon disaster. But there certainly seems to be a speed up of transfer of company secretary roles beginning July 1st, which has continued since.

Which brings another interesting facet: what is the response time to emergencies in legal entities across a multinational company? If the multiple transfers on July 1st 2010 were in response to Deepwater, we see the legal structure reacting within weeks.

Does that mean that in 2014, for example, the BP network will morph to take account of the evolving situation in Russia and Ukraine? The crisis broke out at the end of January. Of all the Western oil majors, BP is the most engaged in Russia through its partnership with Rosneft. But Rosneft’s CEO Igor Sechin was placed early onto a US sanctions list in response to the annexation of Crimea, a sign that Rosneft is seen as an integral part of Vladimir Putin’s power base and geopolitical strategy. Have BP’s legal structures already been fine tuned to minimise possible exposure to further sanctions against Russia if there is an escalation?

You’re only as good as your data

What these first stumblings show is you’re only as good as your data. And in terms of the Equity Map, and company filings, jurisdictions vary wildly.

If only, for instance, the world’s oil industry was based in New Zealand everything would be just hunky. The New Zealand equivalent of Companies House offers all documents filed by companies for free, with no firewall, on the site. Annual Returns, Annual Accounts, changes in directorship, changes of address, all posted as PDFs with no registration and faffing about with credit card payments. Also, the information inside the filings is extensive, containing related party transactions.

The USA meanwhile presented a conundrum. On the one hand, the Securities and Exchange Commission website is watched closely by corporate analysts around the world as the rules to be listed on US financial markets require exhaustive filings and documentation. But on the other hand, actual registration of companies happens at an individual state level – Texas, Illinois, New Mexico and so on – and the amount of information readily available from these state registries, for the majority of companies who are not publicly listed, is far more patchy. Within the United States, for example, the tiny north eastern state of Delaware offers relatively high secrecy and is favoured by multinationals for incorporation within the USA. We paid $40 for access to files offered by the state registry and got nothing we couldn’t have got by spending half an hour surfing the web.

High fees for individual records is a common feature in some unexpected places. The main Netherlands company registry, for example, charges 9.50 euros per document – but the information only appears for as long as you keep the browser open. Once you close it, as I discovered on our first attempt, the scant summary of filings on a company clean disappears – you can’t download it directly, and even though the site makes you register both yourself and a credit card, your account does not store past records retrieved as history. Click off the screen and your nine euros is gone! In Australia, you also have to pay through the nose: 37 Australian dollars per record, or 25 euros. That price is clearly prohibitive for extended public interest use. But the files do contain useful information when you download them.

In Nigeria, we paid a lawyer about 25 euros a time to physically walk into the Corporate Affairs Commission in Abuja and pull documents, because no part of the registry is available online. That was fascinating because lists of shareholders show extended families of Nigeria’s business elite, shell companies in the Cayman Islands, and so on. You could be absolutely sure that the CAC records did not run to the end of any chains of ownership and power. But it did seem like in some cases they formed the start of the chain.

Finally, there is the case of BRELA, the company registry in Tanzania. BRELA has put its company index online and you can search through some level of information about 70,000 companies registered there from the comfort of your desk. But full records require an in-person visit, and it took a couple of weeks for Tanzanian colleagues to work out a scenario to walk in and ask for the records of a few companies that they felt comfortable with. We got a few records with negligible information. And meanwhile, a colleague was able to take this picture of the physical files they were pulled from.

Clearly, we have some way to go before the bold new policy announcements about transparent business are uniformly adopted around the world.

  1. Flesh on the Bones – Project Economics

So the Equity Map itself can give leads on some topics, particularly when combined with other headline information in filings, such as the financials in the annual accounts. There may also be deeper data and analysis in filings if domain expertise is presented with them in the right format.

One example of this is information we found in the filings that could serve to enhance knowledge of what is known in the oil industry as “project-level economics” – reserves in oil fields, operating and exploration costs and so on.

Project economics is a hot topic among analysts of extractive industries. Because you can’t really understand the dynamics of the industry until you have a handle, at least approximately, on costs and assets. Each stage of advance in transparency has only confirmed this.

EITI at first focused on certifying the exchange of payments between governments and companies at an aggregated level. The first EITI reports from Azerbaijan simply stated the total amount in payments by all companies working in the country to the government in the year 2003 for royalties, the total for government entitlement to gas or oil, and so on. Then payment reports started to be disaggregated down to individual companies per country – so Shell in Nigeria, Total in Nigeria, and so on. But a company could still operate in many different concessions, each with different economics. Even a total of payments by revenue category by a company each year, then, such as royalties or corporate income tax, could represent multiple payments with no way of knowing how many payments of how much.

Then there was an increasing emphasis on publishing not just a couple of top level figures but also the contracts which governments had signed with companies. Companies claimed that contracts could easily not be compared and could do more harm than good in the public space. In a wide swathe of the oil producing world, both companies and governments said they would love to publish the contracts because they had nothing to hide – it was just that the other party did not want them to.

In retrospect each of these stages seems relatively modest but both have been highly contentious. And yet, all together they still don’t add up to the ability to really understand the economics of particular projects, or of an oil sector as a whole. In order to do that, you really need to get a handle on costs and assets, particularly, in the case of oil, reserves.

There is currently a certain inconsistency of thinking about this. On the one hand the BP group absolutely wants you to know its consolidated assets, the amount of booked reserves to which it has title and which play a key role in shoring up its share price. So in the 2013 annual report it lists over 5 billion barrels of oil equivalent, which if it were all crude oil would have a current market value of about $500 billion. Royal Dutch Shell was caught in a scandal in 2004 when it was forced to revise down the quoted number of “booked barrels” it claimed as assets, forcing its CEO to resign, while his counterpart at ExxonMobil, Lee Raymond, routinely ignored SEC rules for listing reserves by including assets in the Canadian tar sands in order to show Exxon’s replacement ratio was beyond production. At this global level it’s pretty simple. Replace reserves below production levels and you are effectively announcing to the world that your company is slowly dying.

But yet, when it comes to specific fields, and wells, these same figures of assets, and how much it actually costs to get them out of the ground, are a fiercely guarded secret.

The industry invokes two principles to seek limits on the amount of information disclosed. First is that profits have to be allocated on the risk-reward principle. Exploration is a risky and expensive business, and if investors are to be attracted in at all, they have to win big on the still rare occasions when exploration yields commercially significant quantities of oil extractable at a low enough cost. The implicit connection to keeping cost and asset data unpublished is the idea that public debate in a given country cannot be fully trusted to accept the risk-reward principle as applied to operations in that country. Why should Libya care that Chevron lost hundreds of millions of dollars in unsuccessful exploration in Chile and allow them higher profits because of it? The companies’ second argument is that cost and reserves data represent a competitive advantage – if other companies could access what costs per barrel were, and the precise level of proven reserves, they would be better able to compete in a number of ways.

Some people suspect similar concerns about the ability to use detailed information to reverse engineer company business models lie behind the oil industry’s attempts to prevent project-level disclosure of tax payments. In 2010, the United States passed the Dodd-Frank Act which specified that extractive companies listed on US financial markets – which include most of the world’s majors in one form or other, not just American companies, would be required to disclose “project level” payments of all taxes over $100,000 to any government anywhere in the world. The oil industry immediately reacted to it, suing the Securities and Exchange Commission which had been charged with issuing the regulations attached to the law and creating a legal impasse. Four years later the law has yet to be implemented. The main arguments advanced by the American Petroleum Institute on behalf of the companies were that the new requirements would impose an enormous reporting burden, and that in any case there was no clear consensus about what a “project” was.

If it seems unusual that companies turning over a hundred billion dollars a year might find themselves administratively challenged to break out separate views of their accounts which detailed tax payments of over $100,000 around the world, it’s worth considering that project economics may be the elephant in the room here.

Think about it. If you know what the tax rate is, you can calculate what the profit of the company must have been that it paid $245,000 corporate income tax on. And think about what happens when you add in volume of production. Production levels are also not often published officially on a field by field basis but they are easier for external observers to estimate to within a reasonable margin of error. In large quantities, after all, there is noticeable action on the ground. Storage tanks are filled. Supertankers dock and load. If you can guess roughly what production levels are… and if you know the approximate market value of that crude… and if you also know what company profit was calculated as being… and if you have access to the terms of the contract specifying production splits… then… you should be able to hazard a reasonable guess as to what costs were allocated inbetween the gross sales and the documented profits.

As always, there are still plenty of details to get stuck on – calculations of sales may be made against a formula price, not a market price, for example, and there may be complicated arrangements about oil or gas going into domestic use at specified prices. Those allowed costs include both “opex” – the costs of producing oil day in day out over the lifetime of the project – and “capex” – the huge costs that go in up front to develop the field in the first place. The sub-questions of how much of the costs in any given year are opex and how much are capex, and under what terms that capex is depreciated, all have an impact on the core economics. But, with these few facts or reasonable conjectures at your disposal, you would certainly be on the road to understanding the project economics of any given operation.

Similarly, with the amount of reserves and the rate of depletion of the field. Companies use proven reserves in fields where they are entitled to own some of the production to “book barrels” and shore up their market cap. The rate of depletion of reserves also tells the company – and anyone else who can see those figures – what future production levels, and therefore revenue streams under various price assumptions, are likely to be.

Given the potential for these numbers to allow powerful analysis of an oil field’s profitability, and an oil company’s business model, it is not surprising that commercial databases charge tens of thousands of dollars a year for subscriptions to services which seek to estimate or report field-level data. But what about the rest of us? Including governments and tax authorities in poorer countries who cannot afford access to the likes of Wood Mackenzie or IHS CERA?

The conventional wisdom is that it is impossible to find field-level information about big oil company operations because they group their operations across dozens of countries. It might be possible, for example, that a smaller company which was only active in one or two concessions could publish data which could be broken down into fields. But never a giant like BP.

Except… there is the case of Bunduq Company Limited, a joint venture between BP, Total and a third company called United Petroleum Development Company Limited (Japan). Registered in the UK in July 1970, Bunduq exists to produce oil out of one concession area in Abu Dhabi, the Bunduq field. Abu Dhabi is the major producer among the seven emirates which make up the United Arab Emirates and fabulously oil rich. Its sovereign wealth fund is among the largest in the world at perhaps $770 billion, but the details are hard to come by.

But perhaps because Bunduq exists only to exploit one field, its annual accounts filed every year in the UK show production year on year, depletion of reserves, royalties paid, and operating and development costs. If you pull the records going back over time, you can compile a fairly complete history of the field.

This allows someone with a basic grasp of oil economics to start to tease out some story lines in the life cycle of an obscure field in the heart of the Arabian Gulf.

So, for example, in the decade of filings from 2003 to 2012 inclusive, production dropped from 21,200 barrels a day to 13,500. Proven reserves depleted from 64 million to 23 million barrels over the same period. Clearly, then, Bunduq is a mature field in decline. But the interesting thing is that reserves figures are constantly oscillating up and down beyond simply the impact of production.

For example, in 2007, reserves suddenly jumped from their 2006 figure. In 2006, reserves had been quoted as 42.6 million barrels and the field produced a further 6.7 million barrels that year. Other things being equal, then, you might expect reserves in 2007 to have sunk to 35.9 million barrels. Instead, they rose to 50.1 million barrels. It seemed like some 14 million new barrels have been put in the proven reserves figure. How is that possible? Another figure in the annual accounts may give a clue. Production and Development costs are listed as a line item. In this ten year period, they average $27 million, but that overall figure hides a peak in 2005. In 2004, for example, these costs were $20 million, and in 2006 they were $18 million. But in the intervening year, 2005, they leapt to $70 million. This figure remained an outlier throughout the rest of the decade.

So what’s going on? It seems reasonable to assume that because the line item includes both production and development costs, the period covered by the 2005 filing saw a capital investment project of some kind to boost reserves. And in fact while the companywebsite does not contain any of these numbers, it mentions “since 2006, the associated sour gas has been injected for environmental protection and enhanced oil recovery”. Enhanced production is a term of art, meaning techniques which are used to increase the amount of oil that can be extracted from a field. The most prized oilfields are those where pressure in the field is maintained naturally for a long time. But typically pressure pushing the oil up out of the ground can drop fairly quickly from the time a field starts production, and the operating company then faces a series of decisions about what the possibilities are to enhance production. Some widely used methods range from installing submersible pumps to injecting the oilfield with water, natural gas which is often found along with oil in the same rock structures, or carbon dioxide.

The company’s own reference to injecting sour gas seems to talk to this. “Sour” gas contains a high level of sulphur meaning that it is both a high pollutant and hard to market, because it requires more processing to be useable in power stations, or liquefied for long distance transport. The reference to environmental protection also suggests the gas, which comes up with crude oil, was flared or even vented before – the gain to the environment would come from no longer exposing the gas to the atmosphere.

Given the lag times, it is possible that the spike in costs reported in 2005 relates to an investment made to re-inject the gas, which actually started happening in 2006.

The interesting thing in this case is that, if we could confirm this chain, we would be able to put some order of dimension figures on the gas injection project. Broadly speaking, 2005 appears to show a leap of about $50 million above normal costs, and to lead to an increase in proven reserves of about 15-18 million barrels. Brent crude was selling on the spot market for about $50 per barrel that year, so an investment of this kind would have generated an increase of about $750 million worth of oil – although it is worth bearing in mind that because the extra oil was going to be produced over many years, the future sales value of the oil would have to be discounted against “the cost of capital”. If, for example, the extra 15 million barrels were going to be produced out at an even rate of a million barrels a year for the next 15 years, and the oil price stayed at $50 per barrel, the “Net Present Value” of the $50 million capital investment would be about $300 million assuming a standard discount rate of 10%. Still a healthy return, but it is important to keep the long term perspective in view.

But another thing that the time series seems to show is that the increased reserves will lead to a longer production life rather than increased production. Production continues to decline after gas injection is introduced, from 16,000 barrels a day in 2007 to 13,500 barrels a day in 2012.

Another interesting factlet is that employee costs rise every year over the whole decade, from $5.8 million in 2003 to $14.4 million in 2012. That could be for particular reasons in the Bunduq field, or it could be a data point for another widely talked about phenomenon, the rise in staff costs in recent years as the oil industry faces a skills gap caused by experienced professional retiring without being replaced in full by a younger generation.

The filings also contain nuggets of interest related to tax, specifying figures paid in royalties to the government of Abu Dhabi at a specified rate of 12.5%. It should therefore be possible to calculate the average formal price of oil produced from Bunduq in any given year simply by multiplying that figure by eight to reach 100% and then dividing by the number of barrels known to have been produced. So if in the period reported in 2003, Bunduq paid $30 million in royalties, we can calculate gross value agreed in one way or another (either as sold on markets or by formula in the agreement) as $240 million, against 7.7 million barrels of oil produced, giving us an average price of $31.20. The corresponding figures for 2005 and 2008 are $56.62 and $99.92 respectively, reflecting the historic rise in prices during the Noughties. The royalty figures are puzzling in the last two years of the series, because average price seems to rise to $116 in 2011 and $118 in 2012, above market for the benchmarks.

This is another indication of the limitations of the data. In theory one explanation for the high 2011-12 figures could be that the Bunduq blend sells at a premium over Brent or Dubai Light because it is a higher quality. But in fact the earlier parts of the series show no premium to either of those blends. Could it be that royalty payments do not correspond neatly to that year’s production, so that there is some crossover in years, and the 2011-12 figures could be affected by earlier price spikes? We would have to have access to the contract – and we don’t – to make a determination about that.

The costs line also needs to be treated delicately in terms of absolute numbers. Across the ten year span, they average a mere $4.68 per barrel, which seems very low, especially for an offshore field. Are there other costs somewhere else? Confusingly, although BP’s own affiliates list says the company has 33% of Bunduq Company Limited, the Bunduq Company website says a Japanese consortium has a 97% participating interest in something called El Bunduq Oil Development Project and BP has the remaining 3%. How can these two figures be reconciled? Is it that our Bunduq Company Limited only has ten percent of the overall project, and BP’s 33% stake in this vehicle therefore translates into only three percent of the field?

As usual, the data throw up more questions than they give answers. Nevertheless, Bunduq offers a precious view of data at the project level, from within the BP network.

How many other Bunduq Company Limited’s are there lying around? Affiliate companies of huge multinationals that for one reason or another are effectively what might be called “single asset vehicles”, and which therefore might contain interesting data points around project economics?

In the BP network, as referred to earlier, we can see other single asset companies. The eight exploration vehicles operating in Indonesia, for example – West Aru, West Papua, Tanjung, and so on. The filings for these affiliates show bits and pieces of information here and there: intangible assets representing the cost of acquisition of licenses, for example. But there may be other production vehicles lurking in the network somewhere. It could be Bunduq’s status as a joint venture that determines a higher degree of disclosure, as well as the fact that the field stands at the maritime border between Abu Dhabi and Qatar and both countries are involved in developing the field.

But it might seem as though a pocket of information about a single concession area would be neither here nor there in the scheme of the oil industry at a global level. There are about a million oil wells operating around the world. Who cares if data from a clutch in the Gulf can be sifted for real numbers from the field? How does it scale?

This is another open question. If Bunduq turned out to be utterly unique, the reservation would have merit. But nobody has mined public filings to see if that is the case – certainly not anyone who is prepared to return their results to the public domain, anyway.

It is unlikely there will be hundreds of such companies. But it is important to understand that used in the right way, this data can project beyond geography. We only need a few to be able to start building public libraries of project economics in the oil industry which can seriously advance the amount of information available to everybody to better inform analysis of that risk-reward mantra.

Because of the paucity of data, many people creating models of revenue flows typically use proxy or placeholder values. Trying to estimate what the operating cost is for a field in an East African country? Well, data for the region vary between $6 and $13 per barrel, so let’s call it $10. What will happen to prices in the next few years? Use a scenario sketched by the International Energy Agency until 2035, or just apply a uniform year on year inflation factor. And so on. Often only the operating company has a really close understanding of the cost data. In theory governments normally have access to all cost data relating to concessions on their territory but they rarely have the capacity to analyse it independently.

The Bunduq data could therefore, when properly scrutinised, serve as proxy data for other areas and fields. Bunduq’s reserve depletion, the impact of enhanced recovery investment, and production and development costs can be used as one data series which with others can be classified as appropriate. Say you can accumulate similar data from half a dozen similar kinds of fields, shallow offshore, mature, high capacity local operating environment, and you can then construct composite proxies for those kinds of figures to use in models for places where you can’t access the actual field data. It’s what experts paid thousands of dollars a day actually already do a lot of the time in models for investors or governments, it’s just that they don’t advertise it too loudly, or that the approximation that goes into a model is lost by the time it or its main conclusions enter policy making.

    1. Can We Go Deeper? The Corporate Finance Game

In the eyes of the law, a corporation is a legal person. This means it can enter into contracts like a person, own things like a person, and possess rights like a person. Like a person though, a corporation can possess both a public persona and a private persona. Its glossy public persona consists of its logo and branding, its publicly quoted share price, its smiling CEO at the press conference and its advertising for motor oil. These though, are surface level images distracting us from the private persona beneath. In reality, BP is not a coherent, smiling, whole. It’s an elaborate network of subsidiaries presiding over a complex array of global assets, set up in such a way as to allow a range of rival stakeholders to extract value from them.

Who are these stakeholders? When BP sells oil, the revenues from that get split out, some being passed on to suppliers, some to employees in the form of salaries, some in the form of interest to those who’ve lent to the company, some as tax or royalties to the government, some in the form of bonuses to senior executives, and then some, finally, to shareholders in the form of dividends.

Much of the politics of corporations concerns who gets what. Does the government get a fair tax take? How much do higher management get in bonuses, relative to how much gets paid to shareholders as dividends, and how do both of those relate to how much ordinary workers get? In developing countries in particular, a key battleground will be how much local workers get out of oil contracts, versus how much international investors and foreign senior managers in London get.

Shareholders and bondholders

One of the key political dynamics, though, is between different types of financial stakeholders. BP, like most big companies, is financed by a combination of equity financing from shareholders and debt financing from bondholders or banks.

Shareholders tend to be seen as the ‘investors’ in BP – because they technically own the company and have voting rights – but bondholders who lend BP money are BP investors too. It’s just that they try to extract returns from it in a more indirect fashion. There is always a game being played within any corporate structure between equity investors and debt investors. Debt investors use shareholders as a buffer to protect themselves. Shareholders on the other hand, use the lenders to try leverage, or enhance, their own returns.

BP equity financing: Shareholders

Shareholders provide the baseline capital to fill the war-chest of the BP group. In return, they get to own the company and receive dividends. BP shares are publicly traded on the London Stock Exchange though, so the actual shareholders change as the shares of ownership change hands in the stockmarket.

Institutional investors: Mutual funds & insurance funds

That said, there are a certain core group of anchor shareholders that tend to stay put. These top shareholders are not individual people, but rather large institutional investors like big mutual funds, insurance funds and pension funds, who pool individuals’ money together and invest it on their behalf.

For example, in May 2014, the huge mutual fund company Blackrock owned 5.7% of BP, and Capital Group owned 4.03%. Legal & General Investment Management, who runs insurance funds and mutual funds, owned 3.31%. To put this in context, Blackrock’s 5.7% shareholding was valued at around £5.3 billion on 20th May 2014.

Sovereign wealth funds

BP shares are also popular with big sovereign wealth funds, institutions that invest the surplus reserves of particular countries. For example, according to Bloomberg data from May 2014, Norway holds a 2.28% chunk of BP via Norges Bank Investment Management. The Kuwait Investment Authority owns 1.78%, and the People’s Republic of China owns 2.11%. Other sovereign wealth funds include Singapore’s GIC, the Abu Dhabi Investment Authority and the Korea Investment Corporation.

Pension funds

Pension funds, who look after people’s retirement savings, have long been heavily invested in BP shares, apparently accounting for one in every six pension pounds invested in the UK. BP shares are owned by public pension funds such as the West Yorkshire Pension Fund, Teesside Pension Fund, and West Midlands Pension Fund, and by many corporate pension funds like BAE Systems Pension Fund, British Airways Pension Fund, and, not surprisingly, BP’s own pension fund. In North America, big holders include California Public Employees’ Retirement System, the New York State Common Retirement Fund, and the Canadian Pension Plan. It even turns up in Asia in the Malaysia Employees Provident Fund.

Hedge funds

There are also lots of shorter term investors in BP, such as hedge funds. Prominent hedge fund holders of BP shares (in May 2014) include David Einhorn’s Greenlight Capital, Mondrian Investment Partners, Pzena Investment Management, and Odey Asset Management.

American ADRs

A portion of BP shares are partitioned off to be traded in US stock markets. This is facilitated byJPMorgan, who buys them in the UK stock market and resells them to US investors on the New York Stock Exchange as American depository receipts (ADRs). Top US institutions holding ADRs include State Street and the Wellington Management Company, whose runs the legendary mutual fund called the Wellington Fund.

Another Top 10 US investor is the Bill and Melinda Gates Foundation. Indeed, many NGOs and charitable foundations have shares in BP. It’s such a large company that anyone running a standard UK investment strategy is expected to own some.

The BP Treasury: BP’s internal bank

This disparate shareholders described above are the ones that – technically speaking – hire the core senior executive management team to search for shareholder returns. This management team is in charge of searching out new oil investments around the world, but they’re also in charge of courting debt financiers, whose money can be used to leverage the returns shareholders get from those oil investments.

The BP treasury plays a central role in all of this. It is the financial nerve centre of the BP group, responsible for monitoring the finances and risks of the entire BP group. Its main job, in BP’s own words, is to act as “BP’s in-house global bank, ensuring it has the right funding, liquidity and currency in the right amount on the right day, and in the right place.” BP’s website says the treasury moves around up to $1 trillion every year, channelling external investors’ money through the elaborate network of BP group companies.

Short-term borrowing

One crucial day-to-day role of the treasury is short-term cash management, making sure subsidiaries have enough cash on hand to keep up with everyday operations. BP group companies might own big fixed assets, but they frequently don’t keep a lot of cash on hand. A middle-eastern subsidiary, for example, may own an oil rig that is destined to bring in a lot of money in a month or two, but in the short-term might be £275 000 short of the amount needed to buy a new repair boat, so may request the cash from the treasury.

The treasury team collects cash requests from around the group, and issues short-term ‘commercial paper’ in the ‘money markets’, a way of borrowing for periods as short as a week (or even overnight) from investors such as money market funds. They then transfer the money to the subsidiaries that need it. As the short-term loans come due, the treasury will either use incoming cash from operations to pay it off, or will re-issue more short-term debt to pay it off.

Longer term borrowing

The routine job of short-term borrowing can be contrasted to the more strategic job of long-term borrowing for big projects. BP does not rely very much on bank loans for this, and rather draws primarily on the bond markets, convincing large investors to lend them money via bonds for periods between 18 months to 10 years. The BP treasury issues these bonds centrally, getting the investors to lend to the BP group as a whole, rather than lending to particular BP country subsidiaries.

The treasury issues an annual $10-15 billion of debt, though in 2013 they issued a mere $8.6 billion. They draw on debt investors from around the world, issuing bonds in different national jurisdictions under different ‘issuance programmes’, such as their US and European issuance programmes. BP, for example, now also borrows in Chinese Renminbi, issuing what are colloquially known as ‘Dim sum’ bonds. They’ve also issued Japanese ‘Samurai’ bonds, Canadian ‘Maple bonds and Australia ‘Kangaroo’ bonds.

The decision to borrow in particular jurisdictions is partly guided by economic considerations – such as which countries have the deepest pools of surplus capital held by investors – and partly by strategic political positioning. BP has to keep investors in different countries on sides with them, showing consideration to local investors in new markets that they’re operating in.

BP Capital Markets PLC

When we say the ‘treasury’ borrows though, we are being inaccurate. The BP treasury has actually set up a corporate entity called BP Capital Markets PLC to do the borrowing from international investors. According to BP

BP Capital acts as a finance company issuing debt securities and commercial paper on behalf of the BP Group… BP Capital’s business is raising debt to be on-lent to the parent company and other members of the BP Group on a comparable basis. BP Capital is accordingly dependent on the parent company and other members of the BP Group to service its loans”. (in thisprospectus)

So the treasury uses BP Capital Markets to borrow from investors, and then gets it to lend that money to other members of the BP Group, all the while being guaranteed by BP PLC.

Here are some examples of bonds issued by BP Capital Markets PLC. Note the involvement of investment banks, who will help to put the bond together and help find the investors to lend.

February 2014

Issuance of $2.5 billion bond by BP Capital Markets, arranged by Barclays, BNP PARIBAS, Credit Suisse, Mitsubishi UFJ Securities, Morgan Stanley & RBS

February 2014

BP Capital Markets issues 2 billion Euros in two tranches, arranged by Credit Agricole CIB, Goldman Sachs, HSBC, Lloyds, Mitsubishi & Santander GBM

Nov 2013

BP Capital Markets issues bond for 450 million Canadian dollars, arranged by Bank of America Merrill Lynch, Canadian Imperial Bank of Commerce, and Toronto-Dominion Bank

Oct 2013

BP Capital Markets raises 1.2 billion Renminbi, arranged by Bank of China, HSBC, ICBC International & Standard Chartered Bank (B&D)

Sept 2013

BP Capital Markets issues 600 million euro bond, underwritten by Commerzbank, Deutsche Bank, Lloyds and Santander

Financial engineering

The bonds that BP issues are normally pretty straightforward, but occasionally it enters into more exotic deals. For example, in the midst of the Deepwater Horizon crisis in 2010 it issuedoilbackedbonds, which were backed by very specific BP assets rather than the BP group as a whole. The first was a $2.25bn bond backed by oil sales in the the Azeri-Chirag-Deepwater Gunashli field in Azerbaijan, and the other was a $2.5bn bond backed by BP Angola’s crude oil sales.

Who lends to BP via bonds?

Bondholders tend to be a lot more difficult to track down than shareholders. We can assume that many mutual funds specialising in corporate bonds will be lending, as well as big insurance funds and pension funds, but confirming this takes a fair amount of detective work. For example, in the Teesside pension fund’s 2013 annual report we can see that it bought into a bond called ‘BP Capital Markets 1.7% MTN 15/09/2014 CNY’. This is a ‘medium term note’ (MTN) issued by BP Capital Markets in Chinese Yuan, that has a coupon of 1.7% and that matures in September 2014.


The bondholder money that enters via BP Capital Markets PLC joins shareholder money that enters via BP PLC, and all of it ends up under the control of the BP treasury team. For most observers, this is where the story ends. Investor money enters the black box of the BP corporate structure and disappears, re-emerging at a later date in the form of interest paid out to bondholders and dividends paid out to shareholders.

But what happens in that black box? Remember that BP is not one company. It is a giant federation of linked subsidiaries owned via the holding company BP Plc. Investor funds are piped in, and in turn cascade down through the subsidiaries. The actual way this cascade works is via intercompany equity flows, and intercompany loans, as subsidiaries higher up the chain send equity or debt to subsidiaries further down the chain. In doing so, they set up intercompany channels via which dividends and interest will eventually come rolling back up the chain towards BP PLC and BP Capital Markets PLC, and back out to investors.

Intercompany equity flows leave a distinct trace, because a flow of equity is the act of one company purchasing ownership in another by ‘capitalising’ it (providing it with capital). A flow of equity thus leaves a trace in the form of two companies linked together as parent company and subsidiary, and BP in turn is required to report this. The BP company network we mapped is the current result of historic equity flows between different tiers of subsidiaries.

The traces left by intercompany loans are much harder to see. An intercompany loan doesn’t leave any official trace like a parent-subsidiary relationship. It leaves a creditor-debtor relationship between two BP companies, but these don’t often get reported. It’s possible that, more often than not, intercompany loans follow the same lines as intercompany equity, as parent companies lend to their subsidiaries, but this is not guaranteed. Mostly we glean intercompany loan information from small notes found in scattered reports, and anomalies in reporting requirements, such as the UK requirement to disclose ‘related party transactions’ with subsidiaries that are less than 100% owned by a parent.

What information you’d need to create complete picture of internal financing

While it is notoriously difficult to see inside the black box of intercompany financial flows, it helps if you know what information you’d ideally want to find. So, to do this, let’s reiterate the basic financial dynamic found in most corporations:

  1. Money flows into corporation from external investors
  2. That is then parcelled out to subsidiaries from the top
  3. Who in turn distribute it out into further subsidiaries or real assets
  4. They then harvest the returns from those assets, and
  5. …pass them back up to the top (after passing some to government in form of taxes)
  6. What is not passed back is re-invested as new equity

In reality of course, corporations are messier than this, and all sorts of flows back and forth are occurring concurrently, but logically speaking this order should roughly play out. It thus should be possible to describe each company in the BP group in terms of a logical financial circuit, comprised of incoming and outgoing financial flows, roughly as follows:

  • Incoming equity: The company is capitalised by a parent company
  • Incoming debt: The company receives a loan from another group company (possibly parent company)
  • Outgoing equity: The company uses that money to capitalise another subsidiary lower down the chain, or alternatively to purchase real assets like refineries or oil rigs
  • Outgoing debt: The company might also use that money to lend to its own subsidiary
  • Incoming dividends: The company subsequently receives dividends from the shares it owns, or income from real assets like refineries
  • Incoming interest: The company also might receive interest from the loans it’s given
  • Outgoing expenses: The company pays expenses to suppliers (who could be members of the BP group)
  • Outgoing interest: The company now pays interest to its lender (possibly parent company)
  • Outgoing tax: The company pays tax (such as corporation tax) to the government
  • Outgoing dividend: The company pays dividends to its parent company
  • Reinvested equity: Retained profit is reinvested

If you were able to find figures for all of these categories for each BP subsidiary for each year, you’d have a pretty comprehensive picture of intercompany financial flows.

Detecting flows by changes in balance sheet, and P&L statement

So how do we find these figures? The best way is to get hold of the annual report of each subsidiary, and to look at the balance sheet. That’s because each financial flow changes something in a balance sheet, and therefore can be inferred from those changes. In the case of BP subsidiaries:

  • Incoming equity manifests itself in changes to the ‘called up share capital’ section of the subsidiary’s balance sheet
  • Incoming debt manifests itself in changes to the ‘creditors’ section of the balance sheet
  • Outgoing equity manifests itself in the creation of new assets, which are recorded on the asset side of the balance sheet, normally under ‘fixed assets: investments’, or, in the case of real assets like oil rigs, ‘tangible assets’
  • Outgoing debt manifests itself in the creation of new debt assets, often recorded on the balance sheet as ‘current assets: debtors’
  • The subsequent annual income the company makes (for example, incoming dividends and incoming interest) can be found on the annual Profit & Loss statement. When offset against outgoing expenses, outgoing interest and outgoing tax, the result is the final profit for the year. This final profit is then split into an outgoing dividend, and retained profit, which manifests itself in changes to the ‘profit and loss account’ section in the shareholders’ funds section of the balance sheet.

A BP financial circuit: Baku-Tbilisi-Ceyhan company chain

Let’s use this basic schema to show the basic direction of financial flows in BP’s Baku-Tbilisi-Ceyhan pipeline project in Central Asia. We know which companies are involved, but we don’t have figures for all of them, especially considering that some are registered in the Cayman Islands. We do know though what we need to look for. Here’s the logical sequence that needs filling out with data:

  • BP Global Investments Ltd (UK) capitalises BTC Pipeline Holding Company Ltd (UK)
  • Who in turn capitalises BP Pipelines (BTC) Ltd (UK)
  • While BP International Ltd (UK) lends to BTC Finance BV (Netherlands)
  • Who in turn lends to BP Pipelines (BTC) Ltd
  • And BP Pipelines (BTC) Ltd, uses the money to capitalise three subsidiaries called The Baku-Tbilisi-Ceyhan Pipeline Company (Cayman Islands), BTC International Investment Co (Cayman Islands) and Baku-Tbilisi-Ceyhan Pipeline Holding BV (Netherlands)
  • And, as an interesting twist, it appears that Baku-Tbilisi-Ceyhan Pipeline Holding BV capitalises BTC Finance BV, which was the one lending earlier on
  • On the way back, dividends flow to BP Pipelines (BTC) Ltd from The Baku-Tbilisi-Ceyhan Pipeline Company, BTC International Investment Co and Baku-Tbilisi-Ceyhan Pipeline Holding BV
  • BP Pipelines (BTC) Ltd in turn pays interest to BTC Finance BV, who then pays interest to BP International Ltd
  • BP Pipelines (BTC) Ltd then pays dividends to BTC Pipeline Holding Co, who in turn passes that on the BP Global Investments (who passes it to BP PLC)
  • After this we have the reinvestment process, whereby profit that is not taken out as dividends is steered back as new equity

This is just one circuit among many, a tributary system within the much larger overall group flows of the whole BP group. Remember though, that the logical structure sketched out above is just a start. It’s a theoretical picture, given the company information we have. The next step is to start putting colour and detail into that structure, accounting for the anomalies, and explaining why the structure was set up in that way.

  1. The Call to Arms

So our first exploratory forays into the stories behind the Equity Structure might lead us to amend the Socrates Principle and to talk up the concept of open data mapping of companies a little more. And philosophise a little about where this could go. It is true that, big picture, in building the Equity Map we have come to know that we know nothing in a deep sense. But there are two important qualifications to that.

First, the network of affiliates is a public framework for building actual knowledge about the BP group – or any multinational enterprise. Think of it as a drawing by numbers picture where we have at least joined up all the dots. We just haven’t coloured them in with substance.

Available now, to anyone on the Internet, is the web of equity ownership that connects nearly 1,200 legal structures around the world, uniquely identified from an authoritative source – BP itself – on which they can hang deeper domain knowledge. About operations in the field, about contracts which make it into the public domain, project economics or other nuggets which can be mined from the filings, and all the associated financial reporting that comes with being technically a separate legal entity.

The basis of this is the namespacing approach developed by OpenCorporates which is as effective as it is simple. Full legal name + jurisdiction denominator + company incorporation number. A fingerprint for every company in the world created by the company registries of the world. Compare the robustness of that to the old data approach of Dun and Bradstreet, the incumbent closed source provider of information about companies, whose approach in the 1960s was to create the DUNS numbering system, a proprietary attempt to identify legal entities.

And, most significantly, the namespace is “dereferenceable”, as the geeks say. That means it is devolved from any particular organisation. Third parties do not need to go through OpenCorporates or OpenOil to exchange data about companies. Flows of information can be “many to many”, computer to computer, without fear of ambiguity as long as each piece of information is tagged by one or more company identifiers.

Also bear in mind that the volume of public domain around companies is increasing exponentially. While I have been writing this essay Companies House in the UK have announced that as of 2015 all data in their system will be freely available like New Zealand’s.

So what we have is the beginnings of an open data ecosystem that can track corporate behaviour.

Imagine what you can do when you add concerted attention from real domain experts, such as corporate accountants, ready to explain the 500 or so annual accounts of the affiliates we have structured into the network. Or when you can create such networks not for one multinational but for 500, for all of Big Oil and mining, or what the Australians charmingly call Big Dirt. Such systems can be scaled up through some of the technical automation principles offered by scraping, RSS feeds, APIs, and building a system that can at least partially update itself through watch lists and targeted flows from authoritative sources like company registries or companies themselves – we are talking to another multinational of the scale of BP about whether they would like to assist in their own mapping and win some reputational points.

Imagine what you can do when global information networks can be supplemented by local knowledge and networks. Right now we have a civil society partner sitting in Port Harcourt, deep in the heart of the Niger Delta, working on a similar network of all international and Nigerian oil companies and their interrelationships in the sector. Nigeria’s corporate register is not online but you can walk into the office in Abuja and request company files, and our friends have commissioned a lawyer there to do that. Because they are close to the ground, they are also able to make informed judgements about broadening the source base to some media reports, for example. In our view, you will always need tight editorial control on such networks to maintain credibility, one reason why all the automation techniques in the world won’t replace the human hand at the heart of credible open data systems, and one of the roles we see for ourselves going forward is good old fashioned editorial brand, providing assurance about the right balance in accessing the widest possible range of sources while maintaining quality control. But the norms and rules for editorial judgement can be both scaled and enhanced by knowledge of local context, which a network of local and national partners can provide. Imagine then, these networks maintained by the diligent crowd, individuals and institutions who can be shown objectively through their user histories to meet the quality standards needed to drive these systems forward.

As well as Nigeria, we are already in conversation with other partners on the ground to create corporate maps of the extractive sectors in Mozambique, Colombia, Mexico, and a couple of Arab countries.

We are talking to a group in Paris interested in creating a global mapping of Areva the French power company that recently held controversial negotiations with the government of Niger over renewal of its uranium mining rights. The issue of substance there is the claims of a company that turns over $14 billion a year and makes $2 billion profit about how much it can afford to pay the second poorest country in the world for raw material to power millions of homes in France. But there is also an interesting issue of process: Areva is 80 percent owned by the French state. This means in theory that freedom of information legislation should come into play. Since four euros out of every five is owned by the French people and managed for them by the French government, surely Areva’s operations and legal structures around the world are a legitimate matter of interest for the French taxpayer, who is effectively the majority shareholder in a way that does not apply to BP. And, then again, the 20 percent that is privately held is actually traded on the Paris stock exchange where, like most financial markets, higher rules of disclosure kick in than for non-traded companies.

We are also beginning to map the holdings of a sovereign wealth fund. It turns out there are many issues of public interest in the way a wealth fund operates where you can only really get a reading if you build system-wide data. What is the fund’s exposure to various kinds of risk, for example, and how does that change over time? Where does it have joint investors or other kinds of business partners, who are they and are they appropriate? Can we see patterns in large data sets around investment in particular sectors, through particular channels, or around issues such as liquidity and currencies? And how does all that relate to the Santiago Principles which are the agreed best practice internationally for sovereign wealth funds. It is possible that this project could make the authorities responsible for administering the fund a little jumpy. But such a mapping, developed together with policy analysts inside the country, could be the first real tool to help gain a systematic view of the fund that is available to the public. It might even improve its management once it has settled down and become embedded in the system.

In all this, there are the data and then there are the visualisations of the data. Both are needed, form and content, and both are still in their infancy, and as a result there is often an elision or confusion of concepts.

As the total amount of information stored digitally increases exponentially, we see more and more visualisations which essentially convey the same story: look what a complex web this (and, conceptually, everything) is.

This is the story our first visualisation on Google Fusion gave us with the BP network. This is all very well, and at a primal level there is no denying the tickling ooh-ah pleasure in seeing a bunch of nodes and edges jumping out of your hard won data set and tootling around on the screen, animated as if brought alive by Google. Cultural historians may well look back on system complexity and network analysis as the scientific paradigm of this age, with its attendant concepts of clustering and emergent properties, recursive and iterative patterns, much as monodimensional linearity and brute quantification were the scientific paradigm for most of the twentieth century.

But what does our Google Fusion visualisation actually do, beyond representing to us what we already know? How can we, or any of the people we would like to engage, actually use the data? What does it change?

We wouldn’t pretend to have even formulated these questions well, let alone answered them. But in our stumblings along a path of investigation, we have tripped over a few useful heuristics.

First, you need great flexibility to slice and dice the large data sets. Too much information is just as unhelpful as too little. This presents technical challenges at two levels. There is the back end, the scale of computation and complexity involved in traversing networks, because permutations of “edges”, or connections, are in non-linear relationship to data sets, which are themselves increasing exponentially. To ask a data set which contains even just a few thousand entities the shortest path from person A to B across a variety of different possible relationships involves a fair bit of processing. Traditional relational databases may not be as well-suited to do this as new graph database structures, but this is an early stage technology.

And on the “front end”, to be able to distil down all the various walk-throughs such a networked data set can offer into a user experience that is easy to navigate is very challenging, and assumes a close interaction with real users and their use cases, in order to privilege a few very commonly needed kinds of queries over huge numbers of other theoretical but practically useless ones. We can begin to speculate now what kinds of queries might be useful to help people seeking to understand corporate behaviours – give me a list of companies that are not known to be connected by any equity or debt relations but do share all of: a business address, an accountant, and an auditor – for example. But any system needs to quickly learn from its own use where the preponderance of user interest actually is, and adapt to it.

Second, to be useful, the visualisation needs what we are calling “Click to Check”. This is key to building credibility. You can zoom in to any entity (node – dot) or relationship (edge – line) and understand what our assertion is. But how can you know that it is true? Every atom of data in the network, the dots representing companies and lines connecting them, should never be more than a click away from a source document. Don’t take our word that BP Global Investments owns 100% of Tootle Pip Refining Limited – here is the company filing that states it if you need that.

“Click to check” is one reason we are working only with public domain documents. Another is we believe the value of such data have not been harvested. But the third and most important point of all is physical safety. Addressing the question of who owns what can get you killed in many parts of the world, even in countries where press freedom exists more broadly in the sense of reporting on government policy. It is more primal than any policy level general issue. It’s the Politovskaya Scenario. Not so much for us sitting in Berlin but for if we do want to work with colleagues in the heart of the Niger Delta, or Mexico, or Maputo. Working solely with public domain data doesn’t by itself guarantee safety, of course. It’s not foolproof of course. Just because we restrict ourselves to using public domain data, which companies have themselves already agreed to supply, it doesn’t follow that they or the powerful people who often stand behind them will accept our bona fides or our right to do that. What such people do when they aren’t happy varies very widely around the world of course. Poor Politovskaya is obviously an example at one end of the spectrum but there are plenty of less violent but equally malicious paths pursued by others. Broadly speaking, we make a rough equivalence in terms of evaluating the danger level of any given project or data set between Britain’s punitive anti-libel laws and the risk of our partners getting killed in more lawless countries. If you can’t be sued for it in London, you are unlikely to be killed for it in Uttar Pradesh.

And then of course there are very serious ethical issues. Chief among them what to do about people. In the BP project we accessed lists of past and present company officers across the network of affiliates which is far from complete but counts over 1,500 individuals. We could publish that list right now – it is, after all, made up entirely of data which meets the right of consent criteria, data those individuals agreed to release into the public domain when they became directors of this or that subsidiary. We could start to correlate it with other lists, and publish who has served with who on boards across the entire sector, or who are connected in some identifiable way to “Politically Exposed People”, individuals who serve or have served in high or relevant public office. For now we have chosen not to do that, simply to establish the principle of public domain mapping without undue emphasis on individuals in the first stage.

But at some stage not very far down the line, you run the risk of being coy. When I explained to a roomful of Nigerian colleagues that a pilot initiative by a government agency there to track chains of ownership would not, in the case of Shell and Western oil companies, lead to many individuals, because once you arrive up at the mothership of a company like BP Plc or Royal Dutch Shell you don’t find many individuals owning significant percentages of the stock, their faces visibly fell. People in Nigeria feel – with great justification – that they have been robbed blind through the agency of the oil industry over the last 40 years and they want to know which people are controlling these deals and how those people are connected.

The term of art in investigation for this is “beneficial ownership”. as campaigning groups like Global Witness and others have shown, it is precisely because individuals are able to hide behind corporate structures that they can salt away billions. The reality that Nigerians and billions of other people around the planet live is that formal institutions are almost irrelevant. Even the nominal owners of a company may be stooges, put there by a powerful relative or connection who stays in the shadows. In this view, the world is viewed by “Big Men” and everyone knows who they are because they flaunt it with cars, boats, mansions, women – bling. But if you think you can find their traces in proper structures and formal paperwork you are a fool.

It would be foolish to underestimate the role individuals play in creating, managing and subverting formal structures, and the need to find ways to track this through open data systems.

But how far do you go? In Nigeria, for example, a four-minute Internet search around a prominent and controversial individual, worth tens to hundreds of millions of dollars, revealed his Facebook network of 150 friends. What do you do with that?

Our working approach is to err on the side of caution. While that network must contain some people who are connected to his business interests, it must also contain many individuals – and there is no prior consent.

This must be our approach precisely because we subscribe to the view that the surveillance state is a real danger. We won’t play that game. Our job is purely to focus on prior consent public domain information and see what it can tell us if it is harvested.

A final reason for us to limit ourselves to public domain information is around the politics of investigations. We need to build an open data community around corporate mapping. The fact is that, even if they are conducting an investigation in the public interest, investigative journalists and campaigning organisations are intensely competitive environments and in practice are reluctant to share data. When I was a Reuters correspondent, like most of my colleagues, I operated a number of “bilateral trade agreements” with specific journalists based on trust, liking, mutual benefit and whether they were direct or indirect competitors. The idea that I would contribute any of my data to an unlimited open space, the creative commons, would have seemed bizarre and challenging. What about my scoops? Even within Reuters bureaux the office contacts book was distinctly thinner and more outdated than the individual correspondents personal ones. That’s how it has always been. You have to respect the social engineering.

At the same time, that model of all parts of all investigations remaining in the files of the investigator is a tremendous waste of potential, and no longer necessary. Most parts of most investigations are in fact based on public domain information, and we now have the systems, and a broad enough understanding of how to use systems, to allow a clear differentiation in the way that different kinds of data are handled within an investigation. So if the Jazeera network has spent three years investigating the Lockerbie bombings, or a year investigating Goldman Sachs, to make a series of documentaries, all the public domain information components of that investigation could do into the public space at a suitable time – after the films have gone to air. There is no need for those thousands of hours of research to stay locked on a hard disk. Meanwhile, the private parts of the investigation could remain off limits forever, to respect confidentiality, or competitive edge or any other need of the investigator.

The predominance of public domain information is as true, by the way, in the world of corporate intelligence as it is for journalists or campaigners. Plenty of businesses lay down plenty of money for “inside” information only to have well-dressed twenty something graduates Google their targets for days at a time. Not always – but more often than the directors of such enterprises might feel comfortable admitting to their clients.

Such an approach of course will require a lot of slow consensus building among people and institutions unused to open data ways. But a key part of whether we can build that buy in rests in whether people think that open data systems around corporate information are doable. Whether they can be broad enough, and reliable enough, and easy enough to use, and updated enough, to offer genuine value, and so be integrated into existing work flows.

We believe they can, and that the BP experiment is a baby step in the right direction. But that the scaling of data and the development of a community of genuinely open corporate analysis can happen fast.

It need not be restricted to the oil industry, of course. Following are two suggestions for other corporate mappings of areas of lively public interest. The first, from my fellow Shuttleworth Foundation Fellow Jesse von Doom, who runs a non-profit called Cash Music which seeks to help musicians use open source software to improve their chances of making a living out of their art. His idea is that corporate maps of the global music industry could help establish who is really profiting by how much, and allow the broader community of artists to make up own their minds what they thought about that.

Music is often dismissed as mere entertainment for off-hours or in the background of our work. In truth, it brings context to our lives. The arts show us our world from a new perspective, challenge our perceptions, and help mold our opinions. We’ve seen the power of music in counter culture and we’ve seen it sell products in the mainstream.

By it’s very nature music plays to a global audience, with international IP rights and distribution at the heart of the business surrounding it. The businesses built on music generally fall into two categories: independents (labels, promoters, managers, etc) and the majors, multinationals whose arms reach into all areas of entertainment and mass marketing. Unsurprisingly, the true structure of the majors is hard to determine. There are relationships to movie and television studios, lobbying and legal trade groups, and other media companies. The majors have wholly owned label imprints and partial ownership of independent labels — blurring the distinction between major and independent.

Major labels use their back catalogs and legal prowess to force deals with startups in music and beyond. A great example is Spotify, which gave ownership to each of the majors as part of a catalog licensing deal. The terms are secret under NDA, but those ownership stakes could incentivize major labels to agree to lesser licensing payouts, taking their money back as earnings while forcing down royalties to true independents based on most favored nations clauses the licensing deals. Other abuses have been seen recently, forcing SoundCloud to give up ownership against threats of litigation due to DMCA/IP issues.

Because music plays a critical role in advertising, political campaigns, movies, retail sales, and every other facet of our lives it’s important we have an idea of what we’re dealing with. There will be no level playing field for artists without understanding just how far reaching the influence of the majors truly is. Without a level playing field for artists our ideas and culture are open to manipulation at all levels. This isn’t a matter of abolishing multinational entertainment conglomerates, but of exposing them and giving artists the tools and power to dismantle them.

– Jesse von Doom

CASH Music

Another colleague, George Turner at the Tax Justice Network, has already started a corporate map of UK soccer clubs, seeking to lay bare what has become a multi-billion pound business to millions of fans who love the game, but suspect they are being gamed.

What human endeavour better demonstrates the importance of fair competition than sport. Without it sport is meaningless; merely a spectacle, and an extremely poor one at that.

Sports regulators since the beginning have fought a constant battle against new innovations which might give a team or a competitor and unfair advantage, be these tactical, mechanical or narcotic. Only recently, in Europe at least, have they began to look at financial and corporate structures which some teams may employ to buy advantage.

The financial prizes in sport can be enormous, and so too the temptation to buy success. In England’s football league clubs have crashed in spectacular style after they were loaded up with debt and failed to perform. Often their owners, hiding behind secrecy jurisdictions can float away, lacking any accountability.

And it isn’t just about competition either. Modern sports competitions are awash with money, and often than money ends up being washed, laundered through clubs and sponsorship into the pockets of a small elite. The losers are always the fans, who see an increasing amount of their income taken and used poorly.

In the inimitable words of Massimo Cellino, owner of Leeds United: “They been ripping off the fucking fans everyone. With Compass, with the stadium, with the training centre, with insurance, with the fucking Virgin Island company, Cayman Island company”

All of this is unacceptable when we consider that sports clubs, although technically just limited companies like any other, are in reality much more than that. They carry the hopes and dreams of millions. As David Boyle former CEO of supporters direct said, no one ever had their ashes scattered in the aisle of Tescos.

That is why the Tax Justice Network has launched a project to track and analyse the ownership of every club. We want to give fans the tools to hold their clubs to account, to help them understand who owns their dreams, and what they are doing with them.

 George Turner

Tax Justice Network

I also want to reiterate here what I said at the start. This is not a politically radical project. We believe open corporate analysis can significantly increase competition. If certain business interests were to maintain that it were, well, that would only be the latest example of corporatism, and the vested interests of incumbents posing as defenders of the free market. It will just be the new normal benefiting not just governments and their publics, but investors, and anyone running a business who has nothing to hide.

In Mexico and Ecuador, in fact, we have been in early stage discussions with players interested in developing applications as a means of facilitating entry by international companies into the oil sector. The industry has been so politicised for so long, and there is so much mistrust around foreign multinationals, that an open systemic view of the way those companies operate could be a critical factor in persuading governments that they can, in turn, make a case to their publics that Big Business can be monitored effectively.

The idea that we can systematically map big business is not anti-business. It is good business.








One Response to “Mapping BP – using open data to track Big Oil”
  1. Martin says:


    Really interesting read and depth of information.
    Do you have any more recent information about BP (2016-2017) – affiliates, org structure, etc?
    Is there any cost for getting updated information?



Leave A Comment