Data journalism in a time of epic data leaks
Hamish Boland-Rudder and Will Fitzgibbon
“Interested in data?”
These three words, sent from an anonymous whistleblower to a reporter in Germany, would ultimately lead to the biggest leak of financial data the world had ever seen.
Hidden within a trove of 11.5 million documents were the financial secrets of a wide range of politicians, celebrities, businesspeople, criminals and other individuals willing to use tax havens and shell companies to launder money, hide riches and commit crime behind a veil of anonymity.
These three words would also spark the largest investigative journalism collaboration in history. When the Panama Papers broke on April 3, 2016, its impact was both global and immediate. Within hours, the Twitter hashtag was trending worldwide, and four years later, it still attracts a steady stream of tweets.
Over the course of the first 24 hours after publication, tens of thousands of protesters filled the streets in Iceland, Malta, Pakistan, the United Kingdom and beyond. Within 48 hours, the prime minister of Iceland was forced to resign over undeclared links to an offshore company. He would not be the last head of state to lose his job over the ensuing scandal.
By the end of that first week, lawmakers were raging in parliaments around the world, governments were forming committees to draft legislative recommendations and authorities began making arrests and raiding law firms and businesses to seize evidence of wrongdoing.
From investigations to resignations to criminal prosecutions to legal reforms; for almost fou r years (at the time of writing), this cache of secret files has produced a steady stream of impacts. At the time of writing, more than $US1.2 billion have been recouped, thousands of investigations launched and multiple laws changed. The Panama Papers remains unprecedented.
However, for many of the 376 reporters who spent 12 months producing stories from the dataset, this project was part of a much longer data journey they had embarked upon many years previously. Over the course of the past decade, technological advances have empowered whistleblowers to share ever-growing tranches of inside information with journalists. This is both a blessing and a curse: as the amount of data has grown, so too have the challenges of receiving, ingesting, processing and organising it and, ultimately, turning oceans of complex data into important, compelling stories (Reich and Barnoy 2016).
From the early days of WikiLeaks, to Edward Snowdens NSA Files, to the ongoing leaks of financial data, including Offshore Leaks, Luxembourg Leaks and the Panama Papers, journalists (and whistleblowers) have employed very different methodologies to understand these large datasets and to wring stories from them. While every dataset has its own challenges, there are certain constants common to the way they should be approached. Journalists are trained in cross-referencing to check stories, in cultivating sources and relying on experts when it comes to handling complex data. Leaks of this size demand another, more unfamiliar dimension: collaboration and trust with peers. As ICIJ’s Gerard Ryle puts it:
This is an era of big leaks, so that is a bonanza for journalists. . . .
This is also an era of easy and cheap communication. It would not have been possible for us to do this investigation 5 or 10 years ago. And because this is the era of big data, there has never been a better moment to bring hundreds of reporters together to get their collective eyes on things.
Although this may appear intuitive, news organisations are competitive in nature, and such collaboration requires high-level buy-in. Taking into account the journalistic and technological foundations laid in the lead-up to the Panama Papers, this chapter will explore the recent history of data journalism and the emergence of massive data leaks and show how journalism and journalists have adapted to exploit these invaluable sources of information.
A new paradigm
As technology expands the size of the playing field, the players too have expanded their abilities to deal with an unprecedented amount of information from various sensitive sources. Whistleblower platform WikiLeaks is a prime example. Launched in 2006 by enigmatic Australian hacker-activist Julian Assange, WikiLeaks drew heavily on the legacy of Daniel Ellsberg and the Pentagon Papers for inspiration, pitching itself as an “anonymous global avenue for disseminating documents the public should see” (WikiLeaks 2007).
WikiLeaks has since tested numerous publishing models, including high-profile partnerships with major media outlets (Leigh and Harding 2011, p. 61). Since its inception, the site has existed somewhere in a grey zone between legitimate journalistic enterprise and a home for radical transparency activists. But its role in popularising technology' as a key tool for both transmitting and publishing leaks is undeniable.
It first came to global prominence in 2010 when it captured world headlines not only because of the sensational material it published but also on account of the backstory of the whistleblower at the heart of the leaks. United States Army Private Bradley Manning (now Chelsea Manning) had not yet turned 22 when she first reached out to WikiLeaks. An intelligence analyst stationed at Camp Hammer in Iraq in 2009, Manning had broad and easy access to top secret information. She increasingly believed that the way the United States was conducting its operations in Iraq was profoundly wrong. Armed with a rewritable CD (disguised as a Lady Gaga album), Manning began extracting secret files and searching for a conduit that could help put this information where she thought it belonged, that is, in the public domain.
“It’s public data. It belongs in the public domain. Information should be free,” he wrote in a chat session with a confidant. “If its out in the open, it should be a public good, rather than some slimy intel collector.”
(Zetter and Poulsen 2010)
WikiLeaks had caught Mannings attention when it published a leak of what was probably National Security' Agency data. After making contact and verifying Assange’s online identity', Manning used encryption and secure connections to transfer her collection of secret files (Leigh and Harding 2011, pp. 31,75).
Over the course of 2010, in collaboration with major news outlets, including The Guardian, the New York Times, and Der Spiegel, WikiLeaks would go on to publish three major releases of this information. This included a video, editorially' titled “Collateral Murder”, which showed U.S. forces killing 12 people, including two Reuters journalists, in Baghdad. It also contained a tranche of top-secret military' logs related to the U.S. war in Afghanistan and a cache of confidential diplomatic cables, amounting to hundreds of thousands of documents.
Apart from the effect of the content, the very' act of leaking brought about a new era, one in which even the most precious of state secrets could be divulged en masse for transmission and publication to the world, sometimes after careful vetting and reporting by' journalists, sometimes not. Regardless of the arguments over ownership, technology enabled the collection of this data and also enabled its dissemination.
It was in this new paradigm of massive data leaks and ever-advancing technology' that a whistleblower, known only as John Doe, gained access to a new cache of information from a Panama-based law firm called Mossack Fonseca.
After years of reporting on a massive fraud scheme that was swindling Australian investors and using tax havens to hide the proceeds, in 2011, an Irish-Australian journalist, Gerard Ryle, received a hard drive by post.
On the drive was 260 gigabytes of data - about 2.5 million files — of secret financial information from two offshore financial service providers. It was the first large-scale leak from an insider in a shadowy industry that helps clients hide money in tax havens such as the British Virgin Islands, Singapore and the Cook Islands. Rather than keep the information to himself, Ryle took the data with him to a new job: director of the International Consortium of Investigative Journalists (ICIJ) based in Washington, D.C.
ICIJ was formed in 1997 as the international arm of an American non-profit investigative media outlet, the Center for Public Integrity. The consortium was made up of invited journalists considered among the best investigative reporters from countries around the world. The idea behind the organisation was simple: through this trusted network of members, reporters would help each other chase leads across borders and collaborate on global investigations.
Through ICIJ s network, Ryle was able to have the data processed and shared with 86 journalists in 46 countries who worked together for 12 months to produce an unprecedented global investigation, published in April 2013, that quickly became known as Offshore Leaks. Offshore Leaks would be the first in a series of ground-breaking ICIJ exposés that revealed how the rich and powerful were able to buy into a secretive parallel economy — a world where financial rules could be bent, taxes dodged and shady business deals hidden from scrutiny.
In 2014, ICIJ s international team of collaborators published an investigation based on another trove of leaked documents, this time from the offices of Price Waterhouse Coopers’ Luxembourg office, exposing a system that allowed multinational companies to slash their tax bills dramatically by routing money through the tiny European country. Five years after this investigation was published, Luxembourg Leaks, or LuxLeaks, is still referenced regularly in European Parliament communiqués and legislative actions related to corporate tax dodging.
In 2015, yet another leak of financial data revealed the secret account details of tax-dodging and sanction-evading customers of HSBC’s private banking arm in Switzerland. This prompted numerous government inquiries as well as a full apology from the bank itself.
This truncated history of ICIJ’s investigations in the years leading up to the Panama Papers is worth recounting. The leaked material at the heart of each of these investigations was the result of earned trust and increased visibility.
ICIJ had developed an enviable track record of cracking global stories and building cohesion among erstwhile competitive news organisations. Its reporting was proof that this team of journalists was trustworthy, diligent and able to produce stories that had results. Governments, companies, individuals and their enablers that chose to operate in this netherworld now had to look over their shoulders. This was an attractive proposition for would-be whistleblowers.
Among the collaborative cohort were the German reporters who would go on to receive the Panama Papers leak: Bastian Obermayer and Frederik Obermaier.
John Doe’s path to the leak — a ‘concerned citizen’
We do not know how John Doe came by the data, but there was a lot of it. “More than anything you have ever seen,” he would eventually tell German reporter Bastian Obermayer (Obermayer and Obermaier 2016, p. 9).
As we have seen, Obermayer and Obermaier had been part of the international team working together on ICIJ’s previous leaks of financial information, but why did John Doe choose them as the best conduits for his information? The answer likely lies in one very specific story about the raid on a German bank.
On February 24, 2015, more than 150 prosecutors, tax investigators and detectives swooped on offices and bank branches across Germany as part of a massive investigation into tax evasion and money laundering focused on the country’s second-largest bank: Commerzbank. The German authorities were acting on information from leaked data purchased for almost 1 million Euros two years previously. The data came from inside a Panama-based law firm, Mossack Fonseca, and related mostly to offshore companies linked to the tiny European tax haven of Luxembourg.
Obermayer and Obermaier had received an inside tip that the raids were imminent. But they also had another weapon in their armoury - they, too, had obtained a more comprehensive version of the same dataset. Following on from ICIJ’s Luxembourg Leaks investigation (published in 2014), a source had reached out to the LuxLeaks team and had given a copy of the Mossack Fonseca material to 1CIJ and the German reporters (Bernstein 2017, pp. 206—208).
When they learned that Commerzbank would be raided over the data, Obermayer and Obermaier used the dataset in their possession to prepare a three-page spread for their newspaper, detailing the authorities’ investigations and providing background information about the Panamanian lawyers at the heart of the case.
Their initial trove of data on Mossack Fonseca was about to be rendered insignificant compared to a new source with far more to offer. Shortly after the publication of their stories on the Commerzbank raids, John Doe sent his initial messages to Obermayer, with the first tantalising snippets of what would become the 2.6 terabytes that constitute the Panama Papers (Bernstein 2017, pp. 206—208).
Dealing with the data first
In many ways, journalists have to treat data like any other source of information. Reporters have to establish credibility and reliability. Data needs to be verified. Each piece of data requires assessment before it can be relied upon for a story; it must be treated with the same level of scepticism and corroborated with the same level of care as any other source.
If the identity of the source is known, this makes it easier. Can the existence of this type of data be proven? Is it likely that that person has access to this type of data? Can they be trusted? What are their motivations?
However, as was the case with the anonymous John Doe, how best to check? It might be possible to discern whether this sort of data exists and whether its format matches the possible origins of the data. This reveals nothing about the trustworthiness of the source and/or their access to the material. Therefore, attention must focus on the data itself.
In this case, authentication was helped by the datasets consistency with the first set of Mossack Fonseca material obtained prior to the Commerzbank raids. Other sources were also employed, such as matching details from court records, company registries and more.
Finding the structure
Every dataset has its clues as to where to start, and most of the time, this begins with the metadata - the data about the data. The bigger the dataset, the more important understanding the metadata becomes. Good metadata can be a guide to the data, in the same way that knowing a source’s background, qualifications, position, age or language can help a reporter prepare more informed questions and better understand the answers.
The Panama Papers was a collection of diverse files, mostly containing unstructured information that could not be easily categorised or sorted. Among the 11.5 million files, there were 4.8 million emails, 2.1 million PDF files, 1.1 million images and more than 320,000 text documents. Furthermore, of the 3 million structured files (spreadsheets, databases), it quickly became clear that the dataset did not reflect Mossack Fonsecas database in its original format. The structure was a mess, but, crucially, there were some clues hidden in the metadata that could help clean it up (Cabra 2016).
Through analysing the folder structure of the leak, for instance, it became clear that the information was sorted by client codes that could be discerned from file- and folder-naming conventions. These client codes and other hints at structure made it possible to reverse-engineer the client database to create a structure around the data that could be used as a map.
The difference between structured and unstructured data, email messages, PDF files and Excel spreadsheets can be lost on even the best investigative reporters. Regardless of the size of the dataset, data needs to be presented in a way that makes it possible to compile stories.
This will always be a challenge. The format of the data itself may present obstacles, but reporters must also be constantly aware that the dataset will likely be incomplete. So, while it may present vital evidence to back up a story,
DATA JOURNALISM IN A TIME OF DATA LEAKS traditional methods of interviewing and cross-referencing other sources must also be followed.
The first step with the Panama Papers was to make everything searchable. This might sound obvious, but with 11.5 million files, it was a monumental task. If an individual were to spend a minute reading each document, it would have taken about 22 years to go through all the files, and even with dozens of journalists, it would have taken far too long. Obviously, this was not an option.
Instead, ICIJ turned to technology. Original hard-copy files (scanned documents, images) were digitised using optical character recognition (OCR), where a computer program recognises each letter. Then each file needs to be “read” by a computer to ingest all the words. All of this information is then combined into an index that can be queried by a journalist. Again, even with the latest technology, size still mattered.
Processing this much data required more power than any individual computer could muster, so ICIJ turned to the power of the cloud. At its peak, ICIJ was running more than 30 servers simultaneously to index the information contained in each file in the dataset.
Building and tailoring tools
The next step was to make these documents searchable by hundreds ofjournal-ists with wide disparities in technical competence.
The ICIJ data team built a Google-like search engine based on Blacklight, open-source software originally designed for librarians. Ongoing consultation with reporters allowed the team to prioritise additional functionality', such as filters to search by date and file type, for example, PDFs or Excel spreadsheets. The team then began ticking off the boxes in the reporters’wish list, executing functionality' such as the ability to link files in various folders and step through email chains message by message.
In time, the data team extended the system’s capability' still further to executing lists of searches through the system to find certain categories of individuals, such as elected officials. In time, journalists would also be able to use a bespoke batch-searching tool to return a tailored spreadsheet of results based on input search terms.
This two-pronged assault on the dataset enabled the development of entirely new tools that would help build a deeper understanding of the data.
When chasing money flows across jurisdictions, reporters often become deeply enmeshed in labyrinthine structures comprising anonymous owners of Russian doll-like shell companies registered by a single agent. To better understand these complex relationships, the ICIJ used a unique database format, Neo4j, combined with a third-party' visual enhancement tool, Linkurious, that allows journalists to search for individuals and see at a glance the companies and addresses linked to them in a network diagram. Moreover, a simpledouble-click expands the network and unveils other people with interests in the same companies or using the same addresses.
Bringing the team together
Collaboration is key to the ICIJ journalism model - combining skills, experience and resources helps reporters tell better stories than they would if working alone. Large datasets spanning interests across the world crystallise this realisation (Hamilton 2015).
There are several ways to think about a collaborative journalism investigation, but a core rationale prevails. The best collaborations are mutually beneficial — there should be a clear benefit to teamwork compared to working alone (Alfter 2016, p. 304). Leaked data can be an extremely valuable piece of this equation: the most skilled and sceptical of journalists can be persuaded to put natural competitiveness aside as the price to be paid for unprecedented access to secret documents.
Of course, competitiveness can only be sacrificed temporarily and for a specific purpose - something akin to the best club players coming together to represent their countries in pursuit of a world trophy.
In the case of the Panama Papers, the prize was big enough to entice many of the best reporters in the world. Aside from the bigger picture and the kudos involved, there would also be more than enough stories to go around - stories that would likely never have been told otherwise.
The financial secrecy at the centre of the Panama Papers required journalists with specialist financial knowledge, including politics, corruption, moneylaundering and tax evasion. The global nature of the dataset also made location a key consideration. Rather than parachuting foreign journalists into unfamiliar jurisdictions, the ICIJ model is based on sourcing local reporters with invaluable local knowledge (Alfter 2018, pp. 45—46).
For example, in this case, the team needed a reputable and trustworthy reporter in Iceland to help tell the story of the Icelandic Prime Minister’s connections to an offshore company. Like many stories, ICIJ reporters combed through the documents with a view to providing possible interesting angles before their partners came on board.
Other promising leads had also piqued curiosity. ICIJ required reporters in Russia and Eastern Europe to help unwind Vladimir Putin’s network of allies and track the money that flowed from country to country, bank to bank, between his associates’ accounts and businesses.
Partners in the United Kingdom and United States would help uncover details in key secrecy jurisdictions (from Nevada to the British Virgin Islands) and ensure the story would eventually make its way to the desks of policymakers and regulators who had the authority to change the sharp practices — even if they were technically legal.
ICIJ also needed a diversity of skills to accomplish its ambitious task. Even with the most user-friendly tools, there was still a mountain of data work and detailed research necessary to penetrate the complex corporate structures and uncover patterns of company ownership that could make or break a story.
The collaboration needed reporters skilled in complex database querying, as well as journalists whose contact books and interviewing skills could provide access to key sources. The team also required linguistic assistance — who better to chase down Swiss prosecutors than seasoned Swiss court reporters? Who better to unravel the dynamics of Pakistan’s ruling family than an experienced Pakistani political reporter?
In the end, ICIJ put together a team of more than 370 reporters from 80 countries working across 25 different languages (plus a handful of coding languages, too).
At a practical level, the collaborative process began with an in-person meeting hosted in Washington, D.C., in June, 2015, where a few' dozen reporters traded initial findings, swapped reporting plans and, most importantly, formed or renewed relationships with each other that would form the foundation of this collaboration. These bonds would set the tone for any journalists who would join later.
Secure communication was essential to ensuring the investigation would not be compromised. ICIJ employed two-factor authentication (where users needed both a password and a time-sensitive six-digit code to log in) and protected sensitive emails using PGP encryption.
Finding the stories inside and outside the data
The leaked data provided tantalising signposts, but it would never prove conclusive - in the end, the Panama Papers represents the marriage of data and traditional journalism.
In some cases, the documents included low-hanging fruit that almost guaranteed a story: the passport of Syrian president Bashar Al-Assad s cousin, official letterheads from Nigerian senator David Mark and a barely legible fax that referred to possible proceeds of an infamous UK gold heist.
But even these documents required significant shoe-leather reporting to unpack the full story. Individuals who choose to hide their wealth using offshore companies tend to be devious (or have devious individuals in their employ). It would be foolish to expect to find a smoking gun email that reads “I want to pay a bribe.” None were found.
Journalists matched addresses on emails, contracts and corporate filings with public property and company registries, including Open Corporates. They cross-checked the names of shell companies with details from court records -both public and confidential. Journalists verified spellings and alternative surnames of clients who used multiple aliases.
To understand the “why" question at the heart of the Panama Papers, journalists called on finance, taxation and legal experts to explain the significance of complex inter-company loan agreements, consultancy contacts and other financial deals.
The power of collaborative publishing
The days and weeks leading up to April 3, 2016, were tense for the journalists involved. Letters were sent to the targets of the investigation — including Vladimir Putin, which prompted a pre-emptive campaign to discredit the investigation. Final interviews and doorstop calls were conducted. The team held firm as a major political bribery scandal began to unfold in Brazil and as authorities started to indict FIFA officials over corruption allegations — both of which were stories that the Panama Papers held new details on but which couldn’t be published until the entire investigation was ready.
The ICIJ is often asked how more than 370 reporters managed to keep the investigation secret for a year and agree to simultaneous publication across the world. The answer is simple: every reporter that agrees to join the investigation appreciates the value of unity. It is in nobody’s interest to go on a solo run: working closely together builds strong bonds and a realisation that a single deadline creates maximum worldwide impact (The Listening Post 2016).
A single news organisation publishing an in-depth investigation in one country can cause a stir — but 100 media outlets publishing various, localised stories using the same dataset on the same day makes it impossible to ignore. The ICIJ had already proven this time and again.
On Sunday, April 3, 2016, at 2 p.m. Eastern Time in the United States, 100 news organisations collectively signalled to the world that the issue of financial secrecy — and the corruption and inequality that it enables - was worthy of global attention.
Simultaneous publication also provided a measure of protection for reporters operating in regions with restricted press freedoms. In the wake of the Panama Papers, journalists in Panama, Turkey, Niger and elsewhere were harassed and threatened over their stories; journalists in Venezuela and Hong Kong were sacked; media companies in Spain and Mongolia were sued; journalists in Finland were warned that their homes might be raided. Tragically, 18 months after publication, a journalist working on a related investigation in Malta, Daphne Caruana Galizia, was assassinated.
Ever}' time a member of the Panama Papers team faced a backlash, the whole network stood up in support. It wasn’t always enough to stop sackings or litigation, but it certainly ensured that no reporter could be singled out without facing the international solidarity (and publishing power) of hundreds of their colleagues.
On May 9, 2016, a little more than a month after the first stories generated headlines around the world, ICIJ was ready to make another splash - by publishing the underlying structured data from the leak.
Information on more than 214,000 offshore entities (companies, trusts, shareholders) from 21 jurisdictions were added to a revamped interactive application called the Offshore Leaks Database. The database already held data from ICIJ’s Offshore Leaks and China Leaks investigations from 2013 and 2014; the addition of the Panama Papers meant users had access to more than 360,000 names of people and companies behind secret offshore structures.
The Offshore Leaks Database provides an interactive, explorable map to parts of this system and acts as a repository of downloadable data that can be compared against other datasets. The data has been used by individual investigators as well as government authorities to augment their probes into the wrongdoing enabled by offshore finance, to confront those who may have broken the law and to push for reforms.
Maintaining integrity and ethics
The data ICIJ chose to publish included the sort of information that could typically be found in a searchable corporate register - such as Companies House in the United Kingdom. However, most of the companies in the Panama Papers were registered in secrecy jurisdictions - places where registries of information were either limited or non-existent.
Of equal importance is the data the ICIJ chose not to publish. ICIJ did not publish any personal information or private account details en masse: the Offshore Leaks Database does not comprise the full tranche of 11.5 million documents that formed the Panama Papers cache.
This quickly became a point of contention. On social media, some users cried hypocrisy. “If you censor more than 99% of the documents you are engaged in 1% journalism by definition,” WikiLeaks tweeted, one of a string of criticisms the organisation levelled at ICIJ in the wake of the Panama Papers (WikiLeaks 2016).
For ICIJ, the equation was - and is — very different. As journalists, bound by a code of ethics that dictates responsible, fair reporting, ICIJ is only comfortable releasing documents that have been 100% verified and reported out, where subjects have been given the chance to respond, and where there’s a strong public interest justification for publishing the information (Greenberg 2016; Walker Guevara 2016).
Most of the 11.5 million documents in the Panama Papers leak related to private, personal matters, issues that would not shed light on important figures or entities, documents that would not add anything to the understanding of this shadow}' underworld. No organisation in the world could responsibly sift through these documents to determine which would pass a simple public interest test. It was simply not possible ethically to publish all the files.
Instead, ICIJ endeavoured to publish as many primary documents as possible. This decision was driven by ICIJ’s mission to bring more transparency to society. But of equal importance is the credibility that it adds to reporting by allowing readers to scrutinise the process and the data directly.
The truism that change is the only constant certainly pertains to data journalism, where the pace is rapid and relentless. Although cutting edge at the time, much of the technology employed to produce the Panama Papers is already obsolete. However, the foundational theories and processes driving massive data investigations remain constant.
This idea that good journalism begets good sources begets more good journalism played out directly and specifically in the case of the Panama Papers. This continues to hold true. ICI J’s next investigation, the Paradise Papers, published 18 months later, was based on another massive trove of leaked data that revealed the questionable conduct and wrongdoing enabled by tax havens and financial secrecy.
This dataset was leaked to the same German reporters, who then brought the information to ICIJ — validation for the impact of the Panama Papers. It worked for the whistleblower and the news organisation that shared it (Allsop 2017).
Data momentum has continued at breakneck speed. In the years since the publication of the Panama Papers, ICIJ has published several cross-border investigations based on leaked material whose authors never envisaged it would be held up to public scrutiny. More material continues to arrive on a regular basis - some financial and some more general information, ranging from human rights abuses to kleptocracy.
Developing cutting-edge technology requires specialist knowledge and skills - skills that many journalists do not possess. In the lead-up to the Panama Papers, ICIJ had already learned the importance of taking the time to process data before wandering around blind in a mountain of unstructured data. ICIJ also understands the benefits of forging carefully vetted technology partnerships to better parse the data it has at its disposal.
Ultimately, ICIJ realises the stories it aims to tell cannot be told to maximum effect without the unparalleled local knowledge of its partners. Similarly, these partners appreciate an inability to achieve similar impact should they publish these stories on their own (even if they were capable of doing so). This symbiotic relationship underpins the ICIJ ethos and will continue to do so. In summary, the whole is immeasurably greater than the parts.
Alfter, B. (2016). ‘Cross-border collaborative journalism: Why journalists and scholars should talk about an emerging method’. In Journal of Applied Journalism & Media Studies, vol. 5, no. 2, pp. 297-311.
Alfter, B. (2018). ‘New method, new skill, new position? Editorial coordinators in cross-border collaborative teams’. In Sambrook, R. (ed.), Clobal teamwork: The rise of collaboration in investigative journalism. Oxford: Reuters Institute for the Study ofjournalism.
Allsop, J. (2017). The quiet impact of the Paradise Papers. [Online]. Columbia Journalism Review. Available from: www.cjr.org/watchdog/paradise-papers-icij-tax-havens.php [Accessed: 18 May 2020]
Bernstein, J. (2017). Secrecy world: Inside the Panama Papers investigation of illicit money networks and the global elite. New York: Henry Holt and Company.
Cabra, M. (2016). Wrangling 2.6TB of data: The people and the technology' behind the Panama Papers. [Online]. ICIJ.org. Available from: www.icij.org/blog/2016/04/ data-tech-team-icij/ |Accessed: 18 May 2020]
Greenberg, A. (2016). How reporters pulled off the Panama Papers, the biggest leak in whistleblower history. [Online]. IVired.com. Available from: www.wired.com/2016/04/ reporters-pulled-off-panama-papers-biggest-leak-whistleblower-history/ [Accessed: 18 May 2020]
Hamilton, M. (2015). All together now: News partnerships increase in digital age. [Online], American Journalism Review. Available from: https://ajr.org/2015/05/18/all-together-now-news-partnerships-increase-in-digital-age/ [Accessed: 18 May 2020]
Leigh, D. and Harding, L. (2011). WikiLeaks: Inside Julian Assange’s war on secrecy. London: Guardian Books.
Listening Post. (2016). Behind the Panama Papers. [Online], Al Jazeera. Available from: www.aljazeera.com/programmes/listeningpost/2016/04/panama-papers-media-l 60409 024928741.html [Accessed: 18 May 2020]
Obermayer, B. and Obermaier, F. (2016). The Panama Papers: Breaking the story of how the rich and powerful hide their money. London: Oneworld Publications.
Reich, Z. and Barnoy, B. (2016). The anatomy of leaking in the age of megaleaks: New triggers, old news practices. In Digital Journalism, vol. 4, no. 7, pp. 886-898.
Ryle, G. (2016). In Raab, B. Behind the Panama Papers: A Q&A with international consortium of investigative journalists director Gerard Ryle. [Online], Ford Foundation. Available from: www.fordfoundation.org/ideas/equals-change-blog/posts/behind-the-panama-papers-a-qa-with-international-consortium-of-investigative-journalists-director-gerard-ryle/ [Accessed: 18 May 2020]
Walker Guevara, M. (2016). Frequently asked questions about ICIJ and the Panama Papers. [Online]. ICIJ.org. Available from: www.icij.org/blog/2016/04/faqs/ [Accessed: 18 May 2020]
WikiLeaks. (2007). Website homepage. [Online]. Available from: http://web.archive.org/ web/20070202025339/http://wikileaks.org/ [Accessed: 18 May 2020]
WikiLeaks. (2016). Post on Twitter. [Online]. Available from: https://twitter.com/wikile aks/status/717753531483168768 [Accessed: 18 May 2020]
Zetter, K. and Poulsen, K. (2010). ‘I can’t believe what I’m confessing to you’: The WikiLeaks chats. [Online]. IVired.com. Available from: www.wired.com/2010/06/wikileaks-chat/ [Accessed: 18 May 2020]