Articles
A global assemblage of tax haven clusters: profit shifting, tax dodging and money laundering
A global assemblage of tax haven clusters: profit shifting, tax dodging and money laundering
Investigaciones Regionales - Journal of Regional Research, núm. 60, pp. 23-37, 2024
Asociación Española de Ciencia Regional

Recepción: 22 Diciembre 2023
Aprobación: 20 Mayo 2024
Abstract: In this paper, we examine tax haven clusters considering them as diverse but recognizable examples of about fifty phenomena worldwide rarely studied from an economic geography perspective. Three canonical but diverse examples of tax haven clusters are used for the analysis: Wilmington in Delaware (United States), Ireland (European Union), and Gibraltar (former European Union, now post-Brexit British Overseas Territory). The objects of study are not treated as being almost perfect expressions of the classical, canonical cluster configuration. On the contrary, we focus on the 'agentic' impulses of 'desire' motivating human action and use 'pattern recognition' to identify the features explaining these clusters and the 'agentic' actors motivating them. It is concluded that the characteristics of this type of clusters – tax haven clusters - require more recognition of the interactive and often innovative communication networks through which cluster members interact globally, given they exist within a global system akin to an 'assemblage'. Unlike other types of clusters, we confirm their extreme dependence on the legislation on which they are based as well as their fragility in the face of legislative changes imposed from the outside.
Keywords: Clusters, assemblages, tax havens, agency, desire.
Resumen: En este artículo, examinamos los clusters de paraísos fiscales, considerándolos como ejemplos diversos pero reconocibles de alrededor de cincuenta fenómenos en todo el mundo, raramente estudiados desde una perspectiva de geografía económica. Para el análisis se utilizan tres ejemplos canónicos pero diversos de clusters de paraísos fiscales: Wilmington en Delaware (Estados Unidos), Irlanda (Unión Europea) y Gibraltar (antigua Unión Europea, ahora Territorio Británico de Ultramar post-Brexit). Los objetos de estudio no son tratados como expresiones casi perfectas de la configuración clásica y canónica del cluster. Por el contrario, nos centramos en los impulsos "agenticos" del "deseo" que motivan la acción humana y utilizamos el "reconocimiento de patrones" para identificar las características que explican estos grupos y los actores "agenticos" que los motivan. Se concluye que las características de este tipo de clusters - clusters de paraísos fiscales- requieren un mayor reconocimiento de las redes de comunicación interactivas y a menudo innovadoras a través de las cuales los miembros de los clusters interactúan globalmente, dado que existen dentro de un sistema global similar a un ‘ensamblaje’. A diferencia de otro tipo de clusters, confirmamos su extrema dependencia de la legislación en la que se basan, así como su fragilidad ante los cambios legislativos impuestos desde el exterior.
Palabras clave: Clusters, ensamblajes, paraísos fiscales, agencia, deseo.
1. Introduction
This is an unusual paper because it focuses on an unusual unit of analysis: tax haven clusters. In order to understand the characteristics of tax haven clusters, we seek to perform three main tasks.
Our first focus is on a widely present but mostly under-studied subject in the broad panoply of cluster analytics (Cooke, 2001; 2002; Lazzeretti, Capone, F. and Boix, R. (2012); Hervas Oliver et al. (2019), namely tax haven clusters. Prefiguratively canonical clusters of the kind so ably interrogated by Belussi (2018) consist of agglomeration in space, untraded as well as market dependencies, predominantly small and medium-sized enterprises, associative or agentic relationships, interaction with large firms inside or outside the cluster and knowledge acquisition, advisory, consultancy and knowledge translation services from research and innovation intermediaries - public and/or private. Tax haven clusters fit the bill almost perfectly, although their nature also confers on them unique characteristics. But until recently, although there are at least fifty such clusters worldwide research on them was fairly scant in the spatial science literature (Cooke, 2023).
The particularities of tax haven clusters requires a point of view alternative to the usual used for canonical cluster analysis. The second focus of our paper is to utilise and assess the value of ‘assemblage theory’ and its methodologies for future geographical cluster, cluster platform and wider global cluster network analysis (De Landa, 2019; Deleuze and Guattari, 2004; Buchanan, 2021). We do this by referencing key literature, reviewing relevant research, and applying it in this present contribution to new spatial science, spotlighting three tax haven clusters, namely Delaware, Ireland and Gibraltar.
In passing we wish to define the cluster-types we designate as geographical cluster, cluster platform and global cluster network. The first category (geographical cluster) is outlined above as canonical cluster. The cluster platform is characterised by overlapping clusters that form a ‘platform’ of ‘related variety’ clusters in proximity. This occurs, canonically, in Silicon Valley, which combines information and communication technology (ICT), biotechnology, digital media, cybersecurity, cleantech and electric automotive businesses in a horizontal array of technology clusters. These have their overlaps from relying on computation, for example, as the common ‘general purpose technology’ (e.g. ChatGPT for artificial intelligence [AI]; Jovanovic and Rousseau, 2005) while constructing distinctive relational ‘sectoral assemblages’. The third category, the global cluster assemblage is like a global supply chain with one or more clusters embedded within it (Belussi, 2006). An example, drawn typically from biotechnology, would be the Cambridge (UK) biocluster which contains numerous Cambridge University spin-outs and more mature, evolved biotechs, incoming firms originating in distant clusters such as Cambridge, Massachusetts and global suppliers of diagnostics or contract drug manufacturing organisations (CDMOs) from, for example United States (US) (Catalent) or Korea (Celltrion; Jeong et al., 2023) with clients like Moderna, Johnson & Johnson and Novo Nordisk.
This paper is constructed as follows. The first main section covers three things. The first of these is to outline the ‘pattern recognition’ methodology, which is abductive and constructs explanation of a qualitative nature, based on interrogating truth claims, according to binary psychological traits. The second step is theoretical and uses critical assemblage theory modified because of its professed tolerance of ‘amorality’. Given the nature of this enquiry this is an unacceptable form of discourse, in which the cause of ‘amorality’ is precisely the issue. So, its take on ‘assemblage’ is moderated with reference to ‘thirdspace’ reasoning which is tolerant of inclusivity rather than exclusivity (Bhaba, 1994; Soja, 1996; Appadurai, 1996). Finally, the third step is to reveal the deeper structure of motivation which it is shown lies in what may be called ‘dark’ as compared to ‘lighter’ psychological motivations for human action. In the second and third main sections accounts are given of the structure and mechanisms of our two lengthy accounts of representative tax haven clusters, including their ‘agentic’ actors and processes by which they are constrained and enabled. Then we anatomise anew the Gibraltar ‘gaming’ cluster. This leads to the final section which contains discussion, conclusions and implications of such cluster practices as are revealed by the evidence.
2. Methodology: a novel approach to understanding clusters
Tax havens grounded in cluster-like structures are quite complex artifacts. Their study requires more complex methodological approaches than those based on traditional geographical clusters of regional science, business, and international business theories. In this section we beginbriefly to explain the composite method we follow to disentangle those artifacts. First, we introduce an ontology based on agency and assemblage theory. Second, based on that ontology we apply abductive approach and pattern recognition to trace the elements on which tax haven clusters are grounded.
2.1. Philosophical approach: the assemblage theory
2.1.1. Assemblages territorialization
We introduce the philosophical notion of ‘assemblage’ to analyse the ontological distribution of agency across an arrangement resulting from a combination of things. The notion of assemblage finds its roots in Deleuze and Guattari (2004a,b) and describes a complex, dynamic, and non-hierarchical arrangement of various elements, both human and non-human, that come together to form a temporary and contingent whole, where the whole is more than the sum of the parts[1].
Two concepts are described which interact as drivers for the theory of ‘assemblage’:
a) The first is ‘desire’ which motivates all action in the human, animal and even the mechanical worlds. It is stronger than, for example, even the profit motive or arguably the quest for power though it can be tightly bound up with such impulses: ‘there is no desiring-machine capable of being assembled without demolishing entire social sectors’ (Deleuze and Guattari, 1983, 118).
b) The second is the ‘abstract machine’ which is the ‘interlocking system of things’ by which desire is expressed in the world (Buchanan, 2021). Typically, such relational ‘things’ take the form of markets (e.g. for tax avoidance; Jones et al, 2018); power generation systems (Cooke,2024); cities or regions (Rizzo et al., 2024); or research laboratories (Latour and Woolgar, 1979). ‘Free money’ is the key motivator for tax avoidance. Much depends on abstract but purposeful machines in this second dimension of the evolved ‘assemblage’.
Assemblages are related to ‘places’ through movements of ‘territorialisation’ and ‘deterritorialization’. Territorialisation is an imbricated spatial form of the assemblage complex, which also has its ‘rhizomatic’ (non-hierarchized) underground and overground networks – obviously regarding power grids, cities and regions and tax haven ‘profit shifting’ offline and online equipment, whose intimate communication relays scientific knowledge and information over, by satellite and fibre, or under the sea by Internet cable.
c) Accompanying the two conceptual components of ‘desire’ and ‘machinic abstraction’ are ‘agency’ (or the ‘agentic’ impulse) activating the desire and the ‘real’ or ‘material’ product or outcome.
These are the process elements of the assemblage. In evolutionary terms each complex of active concepts and their products transforms territory by ‘de-territorialisation’ or ‘re-territorialisation’. Capital flows de-territorialise from their places or origin (that is one of the characteristics of capitalism) but then re-territorialize. Tax havens are extremely well geared for this necessity. In the words of Smith and Protevi (2018):
‘Capitalism's command is utterly simple: connect deterritorialized flows of labor and capital and extract a surplus from that connection. Thus capitalism sets loose an enormous productive charge—connect those flows! Faster, faster!-the surpluses of which the institutions of private property try to register as belonging to individuals.’
2.1.2. Tax haven clusters as bodies without organs
A powerful metaphor in ‘assemblage’ theory is the ‘body without organs,’ which is not a human but a ‘corporate body’ that comprises the ‘agentic’ motivator of desire, with the ‘abstract machine’ whose interlocking system parts - together comprise the ‘assemblage,’ set in its territory.
This ‘machinic’ descriptor is referred to by certain ‘assemblage’ theorists as ‘New Materialism’ because they adopt the ‘assemblage’ approach solely for the interlocking parts of the system and downgrade the ‘desire’ sub-system (De Landa, 2019). However, they are criticised by those who argue for the completeness of Deleuze and Guattari’s dualistic version of ‘assemblage’ seeing large complex entities as an interesting, dynamic and evolving ‘actualité’ or real set of phenomena, but uninteresting unless imbricated in the creative, innovative or replicative motivating ‘desire’ elements.
Such critique is a fortiori extended further in the dismissal of a further variant of ‘New Materialism’ in the form of ‘Vital Materialism’ which dispenses totally with the ‘desire’ element, analysing only the ‘interlocking parts’ system as the ‘assemblage’, which even exerts its own motivating impulse back on the ‘arrows of desire’ as the poet William Blake (Maidment, 2001) called them, activated by ‘my bow of burning fire’. Thus, as the critic Buchanan (2021) explains there is research which, for example, endows the ‘abstract machine’ of geographical space with the origins of Sartre and De Beauvoir’s discovery after 1941 of existentialism in Café de Flore in Paris, ascribing it to its vital ‘assemblage’ vibe, rather than to any human agentic intent, which seems to mistake an effect for a cause.
Finally, two significant second order concepts deserve mention in our hybrid theory. First, we return to Deleuze and Guattari’s (2004) reference to assemblage’s ‘rhizomatic’ character, which emphasises the underground/overground aspects of the complex – easily understood in the context of power generation and distribution systems with their underground gas pipelines and overground cabling ‘filaments’ which echo features expressive of mycology – but especially regarding the dark arts of ‘profit shifting’, money laundering and tax avoidance (legal) or evasion (illegal) among tax haven clusters.
2.1.3. Thirdspaces
And, finally, we wish to emphasise the hybridity of our version of assemblage theory’. In our analysis ‘assemblage’ is moderated away from its acceptance of ‘amorality’ by combining it with the geographical notion of ‘thirdspace’ which includes marginalised interests in its perspective.
Thirdspaces are social and cultural spaces that exist beyond traditional binary categorizations. These spaces are neither fully public nor private, urban nor rural, physical nor virtual. Instead, they are hybrid, overlapping zones where diverse social interactions and identities can flourish (Soja, 1996). Thirdspaces, unlike ‘assemblages’, treats exclusion (social, racial, historical) neutrally as a consequence of ‘amoral’ evolution (Soja, 1996; Bhaba, 1994; Appadurai, 1996; Thelen and Mahoney, 2010).
Understandably, tax havens involve significantly immoral relational and ‘agentic’ cluster activity. Tax havens are a type of hierarchical cluster in which the peak ‘agentic’ actors (the tax shifters) are absent, but which exercise a kind of ‘quantum’ space by being both present and absent simultaneously. Global intermediary agents represent them, for example, the Big 3 consultants, the Big 4 auditors and the Banks and, in turn, by their affiliates and nominees in the cluster as such ‘agents’ diminish to local cluster scale.
2.2. The method: abduction, pattern recognition and cross validation
Based on the previously announced philosophical ontology, we use an abductive methodological approach on which we position a pattern recognition process. Different from deductive and inductive approaches, the abductive approach is a form of reasoning that involves generating the best explanation or hypothesis for a set of observed facts (Peirce, 1994; Magnani, 2009). Abduction is often used when faced with incomplete or ambiguous information. For example: rain fell in the night because there are puddles in the street.
Pattern recognition is a process through which we make sense of observed data and information by identifying underlying patterns or regularities (Bishop, 2006). This combined method seeks the underlying lineaments beneath the visible or contextual surface. The data sources typically used in this kind of research are more qualitative than quantitative drawing upon ‘grey literature’ and on financial newsprint, websites, official government and consultancy reports and other ‘fugitive’ sources, as well as specialist academic literature, publications, reports, and books.
We apply this approach to three cases of study of tax haven cluster assemblage. Regarding the first two (Delaware and Ireland), we depart from an early elaboration in Cooke (2023). The third is Gibraltar, which constitutes another interesting case due to its pre- and post-Brexit circumstances.
3. Delaware tax-haven cluster: de-territorialization of banking, black boxes, and interest caps
3.1. Background
Weitzman (2022) explains Delaware’s unique role as a state and tax haven, hosting many large U.S. corporate registrations and over a million smaller companies, while being the smallest state after Rhode Island. Wayne (2012) highlighted Wilmington’s 1209 North Orange Street as one of the major corporate clusters in the World, with 285,000 businesses registered there. The Corporation Trust, part of Dutch company Wolters Kluwer, is central to this hub. Major companies like Ford, GMC, Coca-Cola, Intel, Google, and others have legal addresses there. These include specialised trusts, special purpose entities (SPEs) or special purpose vehicles (SPVs) crafting or processing Cayman Island credit default options (CDOs) that caused the Great Financial Crash. Despite the high-profile names, the office had only eighty low-paid clerical workers, not high-fee lawyers (Shaxson, 2012). Weitzman (2022) showed how Delaware, especially the Wilmington cluster, provided a haven not only for legal activities but also for criminal activities (money launderers, kleptocrats, etc.).
What are the main mechanisms derived from agency through which this assemblage internalizes benefits? The first is Delaware’s business-formation industry, called 'the Franchise', that charged a $300 fee for firm registrations, generating $1.6 billion by 2020, accounting for 40% of the state’s budget and keeping taxes low. The Franchise gives Delaware significant power, as it writes incorporation laws for the U.S. and many other countries[2]. A consequence is that this mechanism diverted funds from some of the poorest Americans.
The second is 'escheatment', which refers to unspent balances on expired gift cards from Delaware-incorporated companies, which revert to the state (Cooke 2024a)[3]. Since 1980, Delaware’s corporate tax revenue declined from 50% to 2.5% of total revenue. In 2020, Delaware earned only $245 million from corporate income tax, making the Franchise tax the state's primary income source. However, 'escheatment' became the third largest source of funds, generating $444 million in 2020.
Another mechanism is Delaware’s legal system. Many firms are incorporated there, leading to a high volume of litigation, bankruptcies, company registrations, and mergers and acquisitions, which provide income through high legal fees and related expenses (e.g. accommodation for lawyers and their staff):
‘Attorney fees, lawyer fees in Delaware are the most expensive in the country, on average, more expensive than New York, more expensive than California. And Washington, D.C., they’re the most expensive hourly rates’ (Sadek, 2022).
3.2. The formation of the assemblage: de-territorialization of the de-regulated banking and transformation of Delaware in tax haven
The state of Delaware, and Wilmington particularly, was historically a maritime, shipbuilding, foundry and railway stop on the New York City to Washigton DC rail line. Wilmington was a heavy industry port, which in 1802 was founded as a gunpowder manufacturer and grew to become headquarters of the Du Pont (now Chemours) chemical empire, benefiting from the US Civil War and subsequent conflicts.
Heavy industry dominated until post-World War 2 reconstruction, the arrival of the I-95 highway and urban renewal cleared the inner city, which led to serious race rioting in 1968. Subsequently, by 1981 it began ‘de-territorializing’ by capturing parts of the de-regulated banking, especially the credit-card industry, augmenting its growing Franchise ‘business formation’ (incorporation) and asset management industries. This had followed long after early tightening of business incorporation rules, which had diverted that business from New Jersey in the early 1910s.
New Jersey had been known as the ‘traitor state’ for its loose corporate legislation, for which then-Governor Woodrow Wilson had sought to cleanse it of its murky reputation. In mid-century, it promoted ‘shell companies’ to foreign investors who sought such ‘black box’ anonymity for often nefarious motives. Specialisation in incorporation of businesses led to Wilmington and its District of Delaware attracting the United States Bankruptcy Court, which became the busiest of the 94 federal bankruptcy courts in the US. Among the largest Wilmington employers is M&T Bank (Wilmington Trust Corporation) a provider of third-party trustee ‘special purpose vehicle’ (SPV) and administrative services to hedge funds, investment management, and private banking. The firm was founded as a banking, trust, and safe deposit company by paternalist Du Pont’s ‘agency’ in 1903 then acquired by M&T bank in 2010, employing 1,900 by 2019; Blackrock Capital Management Inc, (which Varoufakis, 2023, 111, notes as the world’s largest hedge fund and asset manager ahead of Vanguard and State Street, forming another ‘Big 3’ with shares in 90% of firms listed in the NYSE), employed 834 in 2019; and WSFS Financial Corporation included commercial banking, retail banking, cash management, trust and wealth management services that in 2019 employed 801.
There is an opaque episode that further explains the transformation of Delaware into a significant global tax haven, the origin of which lies in the early issuance of credit cards. It involves the decision in 1978 by the state of Nebraska to enable the First National Bank of Omaha to break the biblical-era ‘usury’ sanction for out-of-state cardholder interest rates. In the founding case, the target was consumers in neighbouring Minnesota. Could the Nebraska interest rate (18%) be charged to residents of Minnesota (capped at 12%)? The US Supreme Court ruled it could. Delaware’s governor Pierre Du Pont, realised it could do the same as ‘first-mover’ South Dakota. This state had subsequently removed all interest caps. Doing so would enable dissemination of local credit cards countrywide on condition all major US banks were incorporated in Delaware.
Its Financial Center Development Act (FCDA) of 1981 facilitated that. Next the phenomenon of proto-cluster clientilism occurred at the venerable Wilmington Whist Club. Here local bankers expressed the fear that non-local oligopolists would out-compete them while the oligopolists, in turn, simply needed to operate in a lower-tax environment than New York. A compromise was reached in the 1981 FCDA that, while legally innovative, was morally indefensible. Delaware’s tax ladder would be ‘regressive’, meaning local banks would be protected by a bank franchise interest cap of 8% on income below $20 million, 6% on $20-$25 million and so on, up to peak incomes of oligopolist subsidiaries incorporated in Delaware that paid only 1.7%. Oligopoly banks, subsidiaries and intermediaries inundated Wilmington with local incorporations and its credit card business boomed. This in fact heralded and coincided with the rise to prominence of ‘Reaganomics’ after the 1980 US election. It also promulgated the debt-ridden complexities and catastrophes of the 2008-2009 Global Financial Crisis (Shaxson, 2012).
3.3. Geographical assemblage and thirdspace
Accordingly, it is possible to assess the resulting financial ‘assemblage’ from the qualitative data and interpretations made of the results. A simplified ‘layering’ of this ‘thirdspace’ (Bhaba, 1994; Soja, 1996; Thelen and Mahoney, 2010; De Landa, 2019) ‘assemblage’ and its community hubs indicate the following key nodes and networks:
a) First, at regional paradigm level, the ‘reversibility’ of the Delaware Valley up to Philadelphia and even Pittsburgh identifies a varied metropolitan ‘carbonscape’ composed of coal, steel, and shipyards. Delaware shared some of that, though Wilmington as its main city ‘inverted’ from a ‘munitionscape’ to a company town ‘chemoscape’ under the patriarchal, ‘agentic’ Du Pont interest.
b) At the regional regime level inter-state rivalry and the ‘dark’ reputation of, in particular, New Jersey led to regulated ‘reversibility’ that undermined New Jersey’s lax and predatory business incorporation advantage. In De Landa (2019) ‘amorality’ as a cognitive value is an allowable sentiment in ‘assemblage’. As a traditionally weak ‘agentic’ parastatal actor, Delaware nevertheless usurped New Jersey, also shaping a financial agency ‘singularity’ out of its ‘agentic’ Du Pont ‘chemoscape’ monoculture. A Du Pont heir – as Delaware state-governor - subsequently innovated legalised US ‘anti-usury’ conventions in the 1980s.
c) Deindustrialisation of Wilmington led to further collective ‘agentic’ initiative of a negative kind by the ‘creative destruction’ of worker and black residential districts with hegemonic 1960s ‘carbonscape’ transportation-led, urban renewal initiatives. More amorally but good for growth after the 1980s, the new century transitioned pre-feudal ‘usury’ reforms, near feudal ‘Franchise’ and ‘escheatment’ revenue-raising and business tax relief incentives. These favoured core banking and ancillary financial regulatory and jurisdictional ‘singularity.’ Alongside waterfront development, designation of the Financial Quarter and of federal institutions like the US Bankruptcy Court these anchored the new cyclical profile.
d) Finally, these consolidated into the Wilmington ‘hub’ that also sustains annually 6,000 University of Delaware graduates as a fintech talent pool, many of whom are cognitively and physically proximate through personal and transactional contacts to core commercial law, venturing, regulatory, political, and financial services markets. Accordingly, the assemblage had been ‘territorialized’, ‘de-territorialized’ then ‘re-territorialized’ during three evolutionary cycles.
To reflect on the geographical 'assemblage' cycles mentioned, the 'rhizomatic' (De Landa, 2019) character of its non-linear ‘everywhere’ node and network spatial relations requires examination. Their 'fungal' nature implies both 'underground' and 'overground' aspects. The 'underground' aspects are captured in the darker elements like the amoral interactions of 'the Franchise' and 'escheatment,' involving business manipulation and that can even reach crimes tried in the Delaware Court of Chancery (its Federal Bankruptcy Court) such as fraudulence, embezzlement, tax avoidance, money-laundering, conflict of interest or corruption. The 'lighter' spatial relations are seen in Wilmington's ‘agentic’ history, from the Du Pont legacy to its colonial past as Fort Christina (the capital of New Sweden from 1638-1655). More recently, the replacement of the state’s carbonscape legacy and transition to fintech embedded new nodes in major trader training at the University of Delaware, networks of fintech start-ups and new urban entrants to the Wilmington Financial Quarter.
4. Ireland tax-haven cluster: Schrödinger’s Cat and anther Quantum Effect
The case of Ireland retains some of the core elements described in Delaware but fed with different fiscal transfer mechanisms. Unusual for ‘institutional’ analysis, each of O’Toole (2021) and O’Boyle & Allen (2021) refer to the ‘agentic’ architect(s) of this economic governance organisation as politician and shortly to be Prime Minister of Ireland Séan Lemass (1959-1966) and civil servant T.K. (Ken) Whitaker, secretary to the Irish Department of Finance. Advised initially in 1952 by US consultancy Stacy May to emulate Puerto Rico by becoming a tax haven, Whitaker led a small team of officials to write a 250-page document entitled Economic Development.
In 1959, Lemass succeeded Éamon De Valera as President. Lemass, who had long disagreed with De Valera's import substitution policy, supported Whitaker's alternative. The government endorsed the Economic Development analysis and approved the white paper Programme for Economic Expansion, aiming to double economic growth to 2%, develop export-oriented industrial production, attract foreign capital, reduce trade barriers, and improve workforce skills and technology. This policy was adopted in 1959, with Ireland’s Foreign Investment Agency (IDA) as the main facilitator. However, O’Boyle & Allen (2021) note that the shift towards becoming a tax haven really began earlier, with the Export Profits Tax Relief in 1956 that had pointed the way towards ‘Tax Haven Ireland’. The Shannon Airport Free Trade Zone established in 1959, symbolized the IDA’s success in attracting US investors to ‘double your after-tax profits’. Its generous tax regime rose for capital allowances from 20% in 1956 to 100% in 1978 with capital depreciation varying from 10% to 25% annually and 50 per cent tax relief for five years on exports (Donnelly, 2013).
By the 1980s, Ireland's tax haven strategy had succeeded, attracting significant US tech foreign direct investment (FDI). The IDA promoted university city growth outside Dublin as a successful geographical translation of the state’s Programme for Economic Expansion, particularly benefiting the rural West. Cork (University College) was a key beneficiary of FDI attraction to outer growth locations. For instance, the Apple plant, opened in 1980, was situated near Cork airport as overseas manufacturer of the successful Apple Mac PC. Similarly, near UC Galway, the first overseas Digital Equipment (DEC) plant was established in 1971, though it closed in 1998 after being acquired by Compaq, which was later absorbed by Hewlett Packard in 2002.
What caused this FDI roller-coaster? According to O’Boyle & Allen (2021), Apple’s Hollyhill site was among the last opened before the European Union (EU) abolished Ireland’s Export Tax Relief Scheme in 1981, which had guaranteed zero tax for firms exporting products. Those arriving before 1981 continued to benefit for a decade. But by 1984, the Fine Gael government sought ways for US investors to avoid taxes in Ireland. In 1991, when Apple faced financial issues, it hinted at leaving Ireland due to high recorded profits. Apple argued this was due to intellectual property and branding, not manufacturing. They proposed reducing reported net profit from $270 million to $30-40 million, with Ireland helping to 'disappear' the remaining $230 million, reminiscent of Schrödinger’s Cat—both present and absent, alive and dead. O’Toole (2021) ventriloquises the meeting with state’s agents accepting the following:
“...we’ll say that the profit we make in Ireland is $30-$40 million, you will agree to collect the tax on that amount, and we will both be happy. All you need to do is to accept that the rest of the money exists in that place that has always been so central to the Irish imagination: elsewhere” (O’Toole, 2021, 495).
So, Apple established two Irish incorporated companies as part of the Apple holding: Apple Sales International and Apple Operations Europe. The profits recorded by these two offices were assigned to a ‘head office’ which did not exist. Accordingly, they were not subject to tax in any country and the after-tax $30-40 million (to the Irish Revenue) residue (of $230 million) had ‘disappeared’, or in O’Toole’s (2021) words:
“... a new kind of unknown known...about profit rates and companies that were not companies and magical money that could be everywhere and nowhere” (O’Toole, 2021, 496).
From the late 1980s until its dismantling in 2014, the 'Double Irish' tax scheme was the largest tax avoidance tool in history, shielding $100 billion annually by 2010 in US multinational foreign profits from taxation. Prime Minister Haughey boosted Ireland's tax haven status by establishing the Irish Financial Services Centre (IFSC), transforming it into a conduit for tax haven activities rather than just an FDI platform. Despite claims of the Irish state that the country was not a tax haven because incentives built businesses and created jobs in Ireland, capital flow data from the IFSC contradicted this. The IFSC culture included illegal schemes like the Ansbacher Cayman scheme and allowed investors to incorporate without becoming residents (à la Delaware), relying on competitive global tax regimes. O’Boyle & Allen (2021) highlighted how intellectual property rights were used to exaggerate transfer pricing between corporations in different jurisdictions. In 2016, the EU fined Apple €13 billion for shielding €111 billion in profits from US and Irish taxes between 2004 and 2014, the largest tax fine in history. The Irish state opposed this fine.
Coming back to the Apple example, the company created valuable intellectual property rights (IPR) which it then sold to an Irish incorporated firm run out of, say, the Cayman Islands. For the US Internal Revenue Service, the Cayman entity is actually bona fide Irish thus a Controlled Foreign Company (CFC) which avoids US taxation. Apple then contacts the Irish Revenue that sees Apple as a CFC controlled from Cayman therefore tax resident in the West Indies. The Cayman controlled entity then licenses the IPRs to a second Irish company (hence ‘Double Irish’). This one, being in the EU, can trade the IPR there tariff-free at an accepted profit of, say, 95%. This returns to the second Irish firm, having paid Corporate Profit Tax (CPT) of, say, 33% on the 5% owing, a negligible sum. In Ireland CPT at 12.5% is avoided by the Irish entity for the Cayman entity’s IPR. This sum is taxed in Cayman at 0% meaning Apple retains the 95%. Because of the Irish Revenue’s withholding tax of 20% for transfers to known tax havens, the Irish entity sends the sum to an EU (possibly Dutch) firm that is not liable for such charges. Thus, the two Irish entities represent the ‘Double Irish’ tax avoidance trick and the Dutch client acts as the ‘Dutch Sandwich’. This enabled Apple to become the world’s greatest hoarder of avoided taxation, estimated at $246 billion in 2017 starting from 1991 with Apple’s variant dating from the early 1980s (O’Boyle & Allen, 2021).
But the list is not limited to Apple. Major Irish accounting and law firms in the IFSC, such as Matheson, Arthur Cox, and Goodbody, have created and managed significant tax avoidance schemes, including a €1 billion per year variant of the 'Double Irish' for Microsoft. Irish tax lawyers developed international tax tools like the 'Single Malt,' using Malta as a counterpart tax haven instead of Bermuda. This also included a further Irish bilateral tax treaty with the United Arab Emirates (UAE) which still pertains.
By 2010, the 'Dutch Sandwich' was eliminated under the Organization for Economic Co-operation and Development’s (OECD) BEPS (base erosion profit shifting) anti-transfer pricing rules. The Irish Finance Minister then promoted the 'Green Jersey' scheme involving Jersey as another tax haven. By 2017, the 'Single Malt' was replacing the 'Double Irish' scheme. Major companies like Apple, Google, Facebook, and Pfizer transitioned to new tax avoidance tools like Capital Allowances for Intangible Assets (CAIA), Section 110 Special Purpose Vehicles (SPV), Qualifying Investor Alternative Investment Fund (QIAIF), and Irish Collective Asset-management Vehicle (ICAV), reinforcing Ireland's tax haven status. A University of Amsterdam study identified the IFSC as one of the world's largest conduits Offshore Financial Centres for corporate tax avoidance (Zucman, 2015).
5. A validation-case: Gibraltar, between ‘the Rock’ and a Not So Hard Place
5.1. Background
Our third case of study is Gibraltar. If our first cluster under study (Delaware) was specialized in finance and the second (Ireland) in tech companies, our third case study has an unusual specialization that makes it particularly interesting: online gambling.
Gibraltar is a small British Overseas Territory of less than 7 km2, (2.6 miles) and with only 33,000 residents, embedded in the southern part of the Spanish region of Andalucia. It was originally part of Spain but was captured by the Anglo-Dutch fleet during the War of the Spanish Succession and ceded to Great Britain in the Treaty of Utrecht (1713).
Gibraltar's transformation into a tax haven is linked to its status as a British Overseas Territory with a legal system based on British common law but a distinct tax system. Historically claimed by the Spanish state as part of its territory, from 1946, Gibraltar was listed as a territory without self-governance. In 1960, the United Nations (UN) declared its right to be a colony.
To counteract the increasing demand for self-determination and independence of Gibraltar and the pressure of Spain to recover control of the territory under international law, Great Britain increased Gibraltar’s rights and autonomy in the constitutions of 1964 and in 1969. Gibraltar was recognised as a British colony with an autonomous government: the parliament (House of Assembly) is locally elected and is responsible for all domestic matters, including financial services, taxation, company law and commerce, whereas the United Kingdom (UK) remained responsible for internal security, defence and foreign affairs. In 1970 the UN recommended no longer considering Gibraltar a colony although by then Gibraltar had already developed a more autonomous position.
With Gibraltar holding a high degree of autonomy in legislative and economic matters, the Income Tax Ordinance was introduced in 1968. This ordinance aimed to attract international business and foreign investment by offering a favourable tax regime based on low corporate taxes and tax exemptions, while sovereignty under British law provided legal stability and predictability to business and investors. Corporate tax rates were considerably lower than those of many other jurisdictions (including the UK). At the same time, the Ordinance introduced various tax exemptions that increased over time: exemption subsequently concerned value-added tax, capital gains tax, wealth tax, inheritance and donation tax and estate duty. In addition, these advantages also applied for offshore incomes of companies with a permanent presence in Gibraltar.
This made Gibraltar appealing to business engaged in international trade and financial services and helped in attracting business looking to minimize their tax liabilities. Accession in 1973 by the UK to what subsequently became the European Union gave Gibraltar full access to the European market so that by establishing a company in Gibraltar a firm acquired passport rights: it did not need to establish subsidiaries in other EU countries to set up branch operations and provide services to other European member countries, while also enjoying the advantages of Gibraltar legislation and taxation.
In 1989 the Gibraltar government passed Financial Services Commission Act. The Act established the Gibraltar Financial Services Commission (GFSC), a body of supervision and regulation of financial services. It is the regulatory authority responsible for overseeing financial services in the territory and compliance with international standards and regulations (Schembri, 2018). The GFSC nowadays supervises a broad range of firms including: auditors, banks, company managers, e-money institutions, professional trustees, payment services providers, funds and fund service providers, insurance companies, managers and intermediaries, investment firms, and insolvency practitioners. The activities of the GFSC subsequently facilitated the attraction of financial and gambling companies.
In 2011, Gibraltar increased tax advantages for companies, taxing only income generated locally, which was minimal. Following the Great Crash, this move acknowledged that globalization, financialization, and digital trading had intensified tax competition for mobile capital. The corporate tax rate was cut to 10%, less than half the EU average of 21%, attracting funds, online gambling, and offshore companies with no significant local economic activity. In 2013, Spain accused Gibraltar of dishonest tax practices and included it in its list of tax havens due to its tax advantages as well as tobacco and drug trafficking issues.
Further disruptions to global investment led multilateral bodies like the UN, EU, and OECD to increase controls on corporate tax rates and ‘profit shifting’ to tax havens, raising EU corporate taxes to 12.5%. Brexit in 2019 changed Gibraltar’s status, with the EU designating it a British colony in the visa regulation. Post-Brexit agreements with the EU now require negotiation between the UK and Spain. The EU Council stated that any UK-EU agreements on Gibraltar need Spain's prior approval (European Council meeting 25 November 2019). Brexit memorandums aimed to curb tax avoidance by ensuring Gibraltar-based companies and individuals with false residency are taxed according to EU regulations. Gibraltar must also exchange tax information with Spain and other EU countries.
As of 2024, Gibraltar remains on Spain's list of tax havens. Post-Brexit, regulatory changes ended Gibraltar's passporting rights with the EU from the end of 2023, limiting service provision without separate licenses and reducing double taxation agreements. International pressure from the OECD, EU, and Spain suggests that Gibraltar’s banking secrecy could be affected if it fails to meet agreements with Spain. However, this had not occurred, and Spain continues to list Gibraltar as a tax haven.
4.2. The formation of the assemblage: from colony to gambling tax haven
The clustering of online betting and gambling companies in Gibraltar began in 1961 with the UK's Betting and Gaming Act 1960, which legalized betting shops. This permitted licensed bookmakers to establish betting shops nationwide. Gibraltar, as a British Overseas Territory, followed suit and permitted the establishment of betting shops. Over time, Gibraltar became a hub for gambling companies due to two key factors.
First, Gibraltar’s stable legal environment has made it attractive for company ‘offshoring’ registrations. In 1998, the Gibraltar Regulatory Authority (GRA) was established, later extending to online gambling. The Gibraltar Gambling Act 2005 regulates gambling with codes of practice, including anti-money laundering and data protection controls. The Minister, advised by the Gibraltar Gambling Commissioner (GGC), acts as Licensing Authority. This has made Gibraltar appealing to gaming and betting companies, and in particular for the UK sportsbooks (fixed-odds betting e.g. Bet365; Entain [Ladbrokes/Coral]; Betfair; Betfred, and others; Betfair merged with Paddy Power and is now listed as ‘Flutter’ in Dublin).
Second, Gibraltar's low taxes are attractive: companies pay 12.5% corporation tax (compared to 15% in the UK or 25% for the rest of activities), and Pay As You Earn (PAYE) income tax is only on Gibraltar-derived income for firms and non-residents. Additionally, there are no taxes on VAT or capital gains, and only 0.15% tax on Gross Gambling Yield (GGY).
These advantages encouraged UK gambling firms to migrate to Gibraltar. Victor Chandler (BetVictor) was one of the first, followed by others like William Hill and Ladbrokes (now Entain) to save on taxes. Qualified technical and professional services, such as software and internet providers, also moved to Gibraltar, creating a supportive cluster infrastructure for online gambling. Crucial was the technological upgrading, including high-speed internet and advanced telecommunications systems essential for online platforms. Over time, Gibraltar developed a pool of experienced English-speaking professionals in legal, IT, accounting, and other services related to the online gambling industry, reminiscent in some aspects of the Delaware case.
Hence, these conditions attracted numerous online gambling companies so that the sector could account for about 40% of Gibraltar’s GDP and more than 3,800 jobs (about 13% of Gibraltar’s jobs). According to Gibraltar government official data, in 2023 there were 39 companies licensed as remote gambling (including casinos), betting or bookmaker operators in Gibraltar, although some of them hold their main operations and headquarters in other places[4]. Some of those firms include businesses also specialising in gaming and online casinos[5]. It is clear that typical cluster support activities are present to the predominant gambling, gaming and casino incumbents. However global regulatory and jurisdictional impulses affect the fluidity of ownership, specialisation and location as indicated in the acquisition tags and movements towards the US market and the perceived advantages of location inside or outside the EU.
Gibraltar's journey to today’s British Overseas Territory from a military colony to a tax haven cluster has been shaped by its unique geopolitical and historical context. The territory's low-tax regime, regulatory framework, legal structure, and access to the EU, have attracted businesses and investors seeking to optimize their tax liabilities (EC, 2017). However, Gibraltar's tax haven status is not without challenges. Some of the main pieces of the assemblage have been changing after Brexit, favouring a process of deterritorialization, in which lightweight companies that want to preserve their status and benefits have started to move to other tax havens and tax spots, as for example Malta or the Spanish autonomous cities of Ceuta and Melilla in the North of Africa, which enjoy special taxation regimes.
The two main changes for the future – shared with other clouds on the tax haven horizon are: international regulatory discourse against tax evasion, money laundering (penalties and fines are growing exponentially (e.g. Entain’s – 2023 - £615 million [ $777 million)] UK fine for bribery cf. earlier fines of £20 million for money-laundering). OECD’s Common Reporting Standard (CRS) and Base Erosion and Profit Shifting (BEPS) initiative influencing Gibraltar/UK compliance. And, again, Gibraltar is affected by UK’s Brexit decision that weakened Gibraltar’s specific tax haven status (Tatham, 2017). Accordingly, with alternative aspirant locations waiting, since 2018 some gambling operators have moved their headquarters or main operations to Malta: Bet365, 888 Holdings, William Hill, (Dublin HQ for Flutter Entertainment; Malta for International Operations Pokerstars, Singular and Paddy Power Betfair) and Betclic Everest.
6. Discussion and Conclusions: Deconstructing a Cluster
The three case studies of Delaware, Ireland and Gibraltar draw a unique type of clusters: tax haven clusters. The theories developed around the canonical cluster model pose limitations for studying other types of different clusters. In our case, we have used a theoretical framework based on agency, assemblages and thirdspaces. This leads us to conclude that there are cases of clusters that differ from the canonical model of geographic cluster to which the literature usually refers.
Secondly, through this framework, we have been able to conclude that at least three elements are related to the creation of those tax haven clusters that can be found in many other tax haven clusters. The first is its colonial past, and it is from legal autonomy that the transformation begins. This shift is part of the long-term tax haven timeline, similar to other post-colonial outposts of former empires, as described by Buchanan (2021) as the ultimate functioning 'assemblage'. Thus recall Delaware was once a colonial outpost of the Swedish, then the British, crown before American independence. Ireland was a colonial ‘foodscape’ based on its largely agricultural assemblage, then after independence, developed its FDI ‘technoscape’ led by Apple and other IT giants, until today it has been re-territorialised, notably in its Dublin financial cluster as a ‘offshorescape’ for profit shifting tax avoidance. Finally, Gibraltar de-territorialised from a colonial ‘navyscape’ to a post-colonial ‘resortscape’ and presently its re-territorialised status is as a leading ‘gamingscape’ located in its Europort cluster.
The largest question, which we cannot resolve but may suggest some leads towards understanding, is how ‘colonialism’ and its post-colonial aftermath leads so commonly to tax haven cluster status. One important reason is that the colonial experience is socially, politically, and culturally so traumatic that the colony seeks to relieve the trauma by a desperate quest for ontological security through risky financialisation as it aspires/desires to achieve relative affluence to insure against its endemic lack of resilience, or fragility (Taleb, 2012).
The desire for colonies is often summarized as 'God, Gold, and Glory,' but it is more complex. Religion was a significant motivation for European colonizers, who saw themselves as civilizing native peoples. Different religions played a role, such as the Catholic and Protestant motivations behind European colonization. For instance, New England was colonized by Puritans rejecting the wealth and corruption of the Catholic Church, while the Spanish conquistadors sought both gold and the salvation of native souls. Non-European powers like the Japanese also pursued territorial expansion for gold and glory rather than religious reasons.
So, the motivations are both complex and contradictory, but they come together in the desire ultimately for ‘control’ and its corollary for the colonised of ‘freedom’ from control. Deleuze and Guattari (2004a) are clear states are necessarily oppressive, representing an ‘abstract interest’ in resisting domination from the ‘other’. Power thus takes ‘active’ and ‘reactive’ forms. That is ‘power over’ and ‘power to’. Both state and market (capitalism) are ultimately regressive towards everyone, but some have a greater interest in their continuing subordination i.e. the financially favoured (Appadurai, 1996; Kapoor, 2004).
The second element is the existence of predatory/maximizing agents, companies or individuals, and the amoral/optimizing behaviour they exhibit in the use of tax havens. The third element is the appropriation of functions and the cannibalization of rents from functionally or geographically connected territories. In this game there are two winners and many losers. Tax haven clusters and the companies/actors that use them make a profit from their residents at the expense of other states (most of whom do not use the services of tax havens).
Funding
This research is framed within the research project PID2021-128878NB-I00 financed by the Agencia Estatal de Investigación, Ministerio de Ciencia, Innovación y Universidades, Gobierno de España.
References
Appadurai, A. (1996). Modernity at Large: Cultural Dimensions of Globalisation. University of Minnesota Press.
Belussi, F. (2006). In search of a theory of spatial clustering: Agglomeration vs active clustering, in B. Asheim, P. Cooke and R. Martin (eds) Clusters in Regional Development, (pp. 69-89). Routledge.
Belussi, F. (2018). New perspectives on the evolution of clusters. European Planning Studies, 26(9), 1796-1814.
Bhaba, H. (1994). The Location of Culture. Routledge.
Bishop, C. (2006). Pattern Recognition and Machine Learning. Springer.
Buchanan, I. (2021). Assemblage Theory and Method. Bloomsbury.
Cooke, P. (2001). Regional innovation systems, clusters and the knowledge economy. Industrial & Corporate Change, 10, 945-974.
Cooke, P. (2002). Regional innovation systems: general findings and some new evidence from biotechnology clusters. Journal of Technology Transfer, 27, 133-145.
Cooke, P. (2023). ‘Agentic’ cluster aggression: KIBS auditors & law firms as key tax haven drivers. Competitiveness Review. https://doi.org/10.1108/CR-07-2023-0159
Cooke, P. (2024). Rogue auditors: Dark motivations of the Big 4 accountants in regional sustainability and the creative economy. European Planning Studies, 32, https://doi.org/10.1080/09654313.2023.2220379
De Landa, M. (2019). A New Philosophy of Society: Assemblage Theory and Social Complexity. London.
Deleuze, G., and Guattari, F. (1983). Capitalism and Schizophrenia. University of Minnesota Press.
Deleuze, G., and Guattari, F. (2004a). A Thousand Plateaus. University of Minnesota Press.
Deleuze, G., and Guattari, F. (2004b). Anti-Oedipus: Capitalism and Schizophrenia. Penguin.
Donnelly, P. (2013). How foreign firms transformed Ireland’s domestic economy. Irish Times, November 13, https://www.irishtimes.com/business
European Commission. (2017). Study on the application of the anti-money laundering directive to the gambling sector.
Gibraltar Government. (2023a). Online gambling. Retrieved from Gibraltar Government (2023) Regulatory authority for gambling. Retrieved from https://www.gibraltar.gov.gi/regulatory- authority-for-gambling
Gibraltar Government. (2023b). Regulatory authority for gambling. Retrieved from https://www.gibraltar.gov.gi/regulatory-authority-for-gambling
H2 Gambling Capital (2020). Gibraltar: Online gambling overview.
Hervas-Oliver, J-L., Estelles-Miguel, S., Mallol-Gasch, G., and Boix-Palomero, J. (2019). A place-based policy for promoting Industry 4.0: the case of the Castellon ceramic tile district. European Planning Studies, 27(9) 1838-1856.
Jeong, H., Song, C., Kim, S., and Shin, K. (2023). Quantum jump in biopharmaceutical industry: a case of Korea’s catching up with Europe and US. European Planning Studies, 31(5) 885-904.
Jones, C., Temouri, Y., and Cobham, A. (2018). Tax haven networks and the role of the Big 4 accountancy firms. Journal of World Business, 53(2), 177-193.
Jovanovic, B., and Rousseau, P. (2005). Chapter18 – General Purpose Technologies in P. Aghion, and S. Durlauf (eds.) Handbook of Economic Growth, 1 (Part B) (pp. 1181-1224).
Kapoor, I. (2004). Hyper self-reflexive development: Spivak on representing the third world ‘other’. Third World Quarterly, 25(4), 627-647.
Latour, B., and Woolgar, S. (1979). Laboratory Life: The Social Construction of Scientific Facts. Sage.
Lazzeretti, L., Capone, F., and Boix, R. (2012). Reasons for Clustering of Creative Industries in Italy and Spain. European Planning Studies, 20(8), 1243-1262.
Magnani, L. (2009). Abductive Cognition: The Epistemological and Eco-Cognitive Dimensions of Hypothetical Reasoning. Springer-Verlag.
Maidment, B. (2001). Reading Popular Prints, 1790–1870. Manchester University Press.
O’Boyle, B., and Allen, K. (2022). Tax Haven Ireland. Pluto Press.
O’Toole, F. (2021). We Don’t Know Ourselves. The Bodley Head.
Peirce, C. (1994). The Collected Papers of Charles Sanders Peirce. Harvard University Press.
Perera, J. (2021). An Overseas Territory in Europe: Gibraltar as a test case for discussing the non-colonial. The Commonwealth Journal of International, 110(3), 333-346.
Rizzo, A., Sjöholm, J., and Luciani, A. (2024). Smart(en)ing the Arctic city? The cases of Kiruna and Malmberget in Sweden. European Planning Studies, 32, https://doi.org:/10.1080/09654313.2023.2217850
Sadek, N. (2022). ‘Delaware is Everywhere’: How a little-known tax haven made the rules for corporate America, June 27, ICIC. https://www.icij.org/inside-icij/2022/06/delaware-is- everywhere
Schembri, P. (2018). Gibraltar: A tax haven or a low-tax regime? SSRN Electronic Journal, 305-332.
Shaxson, N. (2011). Treasure Islands. Vintage.
Smith, D., and Protevi, J. (2018). ‘Gilles Deleuze’, The Stanford Encyclopedia of Philosophy. E. Zalta (ed.). Stanford, Stanford University Press.
Soja, E. (1996). Thirdspace: Journeys to Los Angeles and Other Real-and-Imagined Places. Blackwell.
Taleb, N. (2012). Antifragile: Things That Gain from Disorder. Random House Penguin.
Tatham, C. (2017). Betting against the house: Gibraltar, Brexit, and gambling firms. The Guardian, 21 January.
Thelen, K., and Mahoney, J. (2010). Explaining Institutional Change: Ambiguity, Agency, and Power. Cambridge University Press.
Varoufakis, Y. (2023). Technofeudalism: What Killed Capitalism. The Bodley Head.
Wayne, L. (2012). How Delaware Thrives as a Corporate Tax Haven, The New York Times, June 30. Visited 31 January, 2023. https://www.nytimes.com/2012/07/01/business/how-delaware.
Weitzman, H. (2022). What’s the Matter with Delaware?How the First State Has Favoured the Rich, Powerful and Criminal — and How It Costs Us All.Princeton University Press.
Zucman, G. (2015). The Hidden Wealth of Nations: The Scourge of Tax Havens. Chicago University Press.
Información adicional
JEL Classification: R1; K4; F2