A Case of Non-Ideal Guidance: Tackling Tax Competition

1 Introduction

In the global justice literature, growing attention has been given to problems particular to a globalised economy such as tax competition and illicit financial flows. Political philosophers have started to reflect on how these problems intersect with theories of global justice (Brock and Pogge 2014). This paper explores the idea that action-guiding principles of justice can only be formulated at such intersections. This is the starting point from which I develop a ‘non-ideal theory’ of global justice. The methodology of this theory posits that principles of justice are formulated according to the practice they are intended to regulate. Individual practices provide insights about the formulation of principles, for the non-ideal circumstances that prevent the realisation of justice are only revealed through the interpretation of each practice. With regard to the content of principles, I reject the notion that non-ideal theory is applied ideal theory. I offer instead an overview of the main features of a conception of justice for a non-ideal world based on the ideas of compliance, fact-sensitivity, feasibility and path-dependence.

The objective of this paper is essentially meta-theoretical: it is to explain the contribution of a non-ideal approach to our theorising about global politics and to explore the action-guiding potential of non-ideal theory by identifying its defining features. The contribution of this paper is twofold: to provide the conceptual framework for an action-guiding non-ideal theory of global justice and to show why this theory is well-suited to address problems of a globalised economy, such as tax competition.

More precisely, this paper will argue for three claims: (i) ideal theory is not analytically prior to non-ideal theory, (ii) a theory of global justice is best advanced in a pluralist fashion, in which different sets of principles are proposed to regulate different practices and (iii) determining our duties of justice in a non-ideal world is not a question of non-compliance with ideal principles, but rather an inquiry about compliance with non-ideal principles. It is by seeking compliance with non-ideal principles that we formulate an action-guiding theory of justice. Claims (i) and (ii) are argumentative in nature and they relate to the first contribution of this paper: they aim at showing how the conception of non-ideal theory championed in this thesis overcomes central problems in the global justice literature and contributes to our theorising about global politics. Claim (iii) is analytical and relates to the second contribution: it explores the internal structure of an action guiding theory of justice.

The paper proceeds as follows. Section 2 introduces the problem of tax competition. Section 3 argues that cosmopolitan theories of global justice have failed to make room for the specificities of social relations in the formulations of principles of justice. The subsequent two sections focus on the methodology of non-ideal theory. Section 4 covers the second step of the argument: clarifying that social relations are practice-mediated and therefore principles should target individual practices. Section 5 highlights the contribution that can be made by a theory that takes practice as its starting-point: such a theory stands to give past and ongoing injustices the normative weight they require. Collectively, the arguments of these sections support claims (i) and (ii). Section 6 explore the key features structuring the content of non-ideal theory and revealing its potential for guiding action (claim iii). Section 7 returns to the problem of tax competition in more detail: it shows how a non-ideal conception of justice targeting this problem delineates the elements that have to be addressed for global justice to progress today.

To avoid confusion, I take all theories of justice, ideal and non-ideal alike, to be aiming at reforming a feature of the world and to be making idealisations. I do not wish to characterise ideal theories as completely separated from real world problems. Ideal theory refers to political theories that aim at the formulation of principles of justice and do so by abstracting from or idealising about features of the current institutional order and institutions in particular. Ideal theories need not to be completely fact-insensitive. They need only to idealise certain features of society and institutions, such as, but not exclusively, by assuming general compliance. Ideal theories do not aim at the regulation of institutions and do not aim at addressing the non-ideal circumstances that explain, among other things, non-compliance. That is, instead of relying on a categorical distinction between ideal and non-ideal theory, I opt to qualify ideal theories based on their function. The objective is to open the door to an understating of non-ideal theory that does not presuppose ideal theory, one that carries out a different task, which proceeds from ongoing practices and the potential injustices they perpetrate. Ideal theories do rely more heavily on idealisations, but what is particular to these theories is a special attention to the definition of our values and principles independently from the functioning of actual practices and institutions. Rawls’s theory is not completely idealised and fact-insensitive. It is idealised in the sense it aimed at defining a theory of ‘perfect justice’ that assumed ‘full compliance’ of agents in society; and because it takes the site of justice to be the basic structure of society, instead of deriving its principles from the functioning of the different practices and institutions that composed the basic structure. Contrastingly, non-ideal theories are concerned with addressing the practices that are the causes of injustices in the first place.

2 The problem of tax competition

According to the methodology of non-ideal theory, our first task is to provide an interpretation of the targeted practice. Providing a full analysis of the problem of tax competition is beyond the scope of this paper (Avi-Yonah 2000; TJN 2005; Clausing 2016; Dietsch and Rixen 2016 provide a thorough and insightful account of the phenomenon). I will thus focus only on questions such as fact-sensitivity, path-dependence and compliance, which allow me to prove claims (i), (ii), and (iii).

Today’s globalised economy is characterized by high capital mobility but largely domestic tax policy. In this context, individuals and corporations can select different tax regimes. Tax competition is broadly understood as the phenomenon of countries lowering tax rates in order to attract capital, corporate investments and labour (Rixen 2011b). Tax competition occurs when tax systems are sensitive to tax differentials. Three forms of tax competition can be distinguished in the areas of: portfolio capital, paper profits, and foreign direct investment (Dietsch and Rixen 2012, p. 2).

So-called ‘tax havens’ are an example of tax competition in the area of portfolio capital. Tax havens offer low or zero tax rates to attract capital. They also offer bank secrecy rules in order to hide ownership of bank accounts from external governments. This procedure is illegal when conducted for the purposes of tax evasion. In this case, individual hold their capital gains ‘offshore’. The curtain of secrecy surrounding this form of tax competition indeed allows many people to evade tax (Dietsch 2016, p. 233). Notably, 50 % the wealth of Latin American countries and up to 70 % of the wealth of Middle Eastern countries is held in tax havens (Dietsch 2015, p. 3).

In the area of paper profits, we observe different practices. One of the most common phenomena is called ‘transfer mispricing’. One way through which multinational enterprises (MNEs) manipulate transfer prices is by assigning their profits made in high-tax jurisdictions to their subsidiaries in low-tax jurisdictions. This constitutes a legal form of tax avoidance. Multinational firms manipulate prices on intra-firm transactions (Avi-Yonah 2016; Clausing 2016, p. 34). No money actually circulates between these affiliate companies. ^[1] Consider the case of a company that sells office supplies in the US and declares very low profits, because it manipulate prices so that is appears to buy its pens, printer paper and laptop bags from its subsidiary in the Cayman Islands, a tax heaven, at an extremely high price. ^[2] The Cayman Islands company in fact register this transaction as if it has bought these from another subsidiary company, a manufacturer in Honduras. The Honduras subsidiary appears to sell the goods to the Cayman Islands subsidiary at a very low price, again declaring very little profit. This way, the US company and the Honduras subsidiary do not declare important profits. However, the Cayman Islands subsidiary makes a gigantic profit, for it apparently bought pens, printer paper and laptop bags at a low price from Honduras but sold them at an exorbitant price to the multinational office furniture company in the US. MNEs do not need to actually buy and sell goods from its subsidiaries. Most of this is only written and calculated for tax purposes. The overpriced items will produce revenue in subsidiaries located in low tax countries and reduce them for subsidiaries in high tax countries. This is one example of how transactions are manipulated so that the bulk of profits of multinationals is declared in low-tax jurisdictions. Other examples in the area of paper profits are what is today known as the ‘Double Irish’ tax structure, used by Google and many other multinationals. Also known as ‘earnings-stripping’, it involves taking out loans from foreign affiliates and paying tax-deductible interests (Dietsch 2016, pp. 233–234). As opposed to tax evasion, most of these strategies in the area of paper profits are legal. To get an idea of the magnitude of the problem, without claiming that all of it is transfer mispricing, the OECD in 2002 has shown that 60 % of all international trade is done intra-firm (Dietsch 2016, p. 234). In the first two forms of tax competition, governments ‘poach’ capital from the potential tax revenue of the governments where individual and corporations are actually located.

The third form of tax competition occurs in the area of foreign direct investment (FDI). This is the competition to attract companies. MNEs choose the location of their company according to various factors such as labour qualification and quality of local infrastructure. Yet, one important factor is the tax burden in the chosen country (Clausing 2016). Governments wishing to attract FDI can for example lower their general tax rates or design preferential tax regimes for particular sectors of activity. One example that comes to mind in this case is Ireland. With its low corporate tax rates, Ireland attracted 25 % of US companies’ FDI in Europe between 1990 and 2004 (Economist 2004). The direct investment stock in Ireland is greater than in Germany and France combined (Times 2015). This form of tax competition involves a real relocation of economic activity. A government that engages in this third form of tax competition ‘lures’ the tax base of foreign governments to make it a legitimate part of their tax base. Whether the luring government will actually tax these revenues is an open question. The available data (see below) indicates that these taxes will be considerably lower at the ‘luring’ destination.

Indeed, it is helpful in this context to distinguish between the impact of the phenomenon on richer and poorer countries. Empirical studies have shown that nominal corporate tax rates in OECD countries decreased from an average of 50 % in 1975 to an average of 25.7 % in 2010 and the nominal top personal income tax rates decreased from 70 % to 41.4 % over the same period, according to the OECD tax data base. Since the tax base in these countries expanded over this period, overall revenues as percentage of GDP remained stable (Dietsch and Rixen 2012, p. 5). What studies show is that, internally, the burden shifted among tax payers. Bigger corporations benefitted more from the decrease of tax rates than did small and medium companies. Also, the tax burden shifted from capital to labour. It has been observed that, if richer countries wished to prevent important revenue losses, they could to adopt regressive fiscal policies, notably by shifting the tax burden from capital to labour, and from taxation on revenue to indirect taxation of consumption (Dietsch 2015, pp. 47–8).

In contrast, in developing countries, we do observe losses in revenues. The losses of revenue have not been compensated by a broadening of tax bases (Clausing 2016). Corporations and corrupt officials take advantage of weak or non-existing rules governing financial transactions today. They not only become rich, they also weaken the institutions that are meant to sustain the jobs, living conditions and overall rights of the world’s poor (IBA 2013). Tax abuses deprive agents, many of them poor countries, of the capabilities to create or strengthen the institutions that uphold political, cultural and socio-economic rights. And they augment countries’ dependence on foreign assistance and thereby diminish their financial autonomy. In sum, tax competition creates important distributive problems in developed countries and considerable revenue losses in developing countries. In absolute numbers, in 2008, income shifting by multinationals deprived the US government of an amount somewhere between $60–$90 billion. Although data in a context of bank secrecy is hard to obtain, globally, estimates are that about $8 trillion (a lower-bound estimate) is located in tax heavens, which represents about 8 % of total wealth (Zucman 2013, 2014).

The empirical analysis of the previous sections will inform the normative analysis to follow. In lowering their tax rates to attract investments and capital, countries exert pressure on other countries to lower their tax rates in order not to lose capital and investments. Two points are worth mentioning. Internally, agents benefit unequally from it: corporations more than individuals and big corporations more than small ones. Globally, developing and poor countries are deprived of significant revenues. Although countries maintain their capacity to set the tax rates as they wish, they lose their ‘de facto fiscal self-determination’ (Dietsch and Rixen 2012). Without regulation, there are risks of this phenomenon being exacerbated and countries being dramatically deprived of their capacities to act upon justice internally. A state’s capacity to implement justice within its borders can thus be undermined. Whether this practice leads to a race to the bottom – all states entering competition and lowering their tax rates to a minimum – is debatable (Ronzoni 2009, 2014). What the data suggests is that the race to the bottom has been prevented in richer countries precisely because the tax burden shifted, exerting more pressure on immobile factors like labour and smaller companies. If the race did not end completely at the bottom, injustices have nonetheless been created. This has heavy consequences. Illicit financial flows and tax competition infringe state sovereignty in forcing states to forego a welfare system, or at least in reducing significantly its chances to implement a welfare system. Undermining their internal capacity to act might make states lose power to control their economy and therefore monitor external interference. Empirical evidence seems to point in this direction (TJN 2005; Avi-Yonah 2007; Dietsch and Rixen 2012; OECD 2014). If countries wish to prevent revenue losses, they have to adopt regressive fiscal measures.

This is only a brief summary of the phenomenon of tax competition, but it indicates we have reasons to believe tax competition raises a problem from a justice standpoint. As mentioned, the task of this paper is less to formulate original principles of justice to tackle this problem, than to show the normative elements that must be considered by any theory of global justice wishing to address it. The presentation of the problem in this section should help us in observing in the following sections that the complexity of the problem of tax competition reveals normative elements that theories of justice must consider.

3 The global justice debate

The global justice literature today is less centred on the debate between cosmopolitans and statists. We see more sectorial studies about the pressing moral problems of a globalised economy. I wish to point out to one aspect of the global justice debate that has gone largely unnoticed and has obscured the locus of non-ideal research in the field.

This discussion was originally structured around one question, the question of extension: ‘should principles of global justice be the extension of principles of domestic justice?’ Cosmopolitans answer it by the affirmative. (Pogge 1992; Tan 2004; Caney 2005; Pogge 2008, pp. 48–49). Statists answer the question of extension in the negative. They argue that egalitarian principles are of concern only within self-contained political communities. (Walzer 1987; Rawls 1999a; Blake 2002; Nagel 2005).

A new set of theories has emerged in recent years in the global justice debate, aiming to surmount the problems of cosmopolitanism and statism. These theories have been described as ‘relational’ or ‘practice-dependent’ (Cohen and Sabel 2006; Sangiovanni 2007, 2008; Valentini 2011). However, since what characterises these approaches is their methodology, the relational approach has let is proponents to both cosmopolitan and statist conclusions. Relational approaches maintain that equal respect applies only between people who stand in a specific relation to one another (Valentini 2011 explores relational aspects of cosmopolitans and statists positions).

Now, consider again the question of extension, around which much of the global justice debate has revolved: should principles of global justice be the extension of principles of domestic justice? What both cosmopolitans and statists fail to see is that there are in fact two steps we must take in answering this question. First, we must know whether there are considerations of justice that are global in scope. Second, we must determine whether the principles of justice that apply globally are simply the same principles that apply domestically. These are two analytically distinct steps. A relational perspective should help us see to this, because it makes room for answering the first question positively and the second negatively. That is, we can claim that the scope of justice is global because there are relations at the global level that raise justice-related concerns, without having to claim that the same principles will apply at both levels. I will clarify below that relational theory and non-ideal theory intersect methodologically: the methodology of non-ideal theory is relational. The task of non-ideal theory is precisely to understand the specifics of these relations, why they raise justice-related concerns, and how a conception of justice can guide action in order to address these concerns (James 2012 offers a powerful account about the moral importance of considering the particularities of relations).

Cosmopolitans wrongly presuppose that all problems of global justice are of the same kind. However, this is not the case: for instance, the problem of tax competition has the form of an ‘asymmetric prisoner’s dilemma’, whereas the problem of climate change is one of collective action but where the attraction to act in one’s self-interest is much stronger. Sharing a common humanity does not entail advocating for one single set of principles for the plurality of problems of global injustice. The global practices are what guides us best in the formulation of principles of justice, not the fact that we share a common humanity. The causes of global injustice – e. g. tax competition, colonialism, migration, borrowing privileges, resources privileges, trade agreements on intellectual property rights, climate change – are profoundly heterogeneous.

4 The methodology of non-ideal theory

The second contribution of this paper is to explore the idea of action-guidance by examining the defining features of non-ideal theory. In order to do this, further scrutiny about the first contribution is warranted. That is, we should try to clarify why a non-ideal methodology is well suited to address global justice problems such as tax competition. It will be easier to explore the content of non-ideal theory if first we define the methodology of non-ideal theory. To this end, I will offer support for claims (i) and (ii) in this section and the next. ^[3] To recap, claim (i) states that ideal theory does not have analytical priority over non-ideal theory, and claim (ii) states that different principles regulate different practices at the global level. Broadly taken, these two objectives bring us closer to narrowing the gap between the formulation of principles and the practices that constrain the advancement of justice in the world today. More specifically, they claim that it is the understanding of these practices, not the abstract interpretation of the principles, which allow us to bridge this gap. The second claim aims to overcome the problem of cosmopolitanism, that theories of global justice seem incapable of formulating principles that speak to the causes of injustices in the global realm.

The first claim must be established first, however: I must show that there are good reasons to proceed from the causes of injustice and to see what we can learn about what to do here and now by studying these unjust practices. The first step of this demonstration is showing that there is value in proceeding from the interpretation of the practices, which is the question the previous section left us with: how exactly does a theory that proceeds from the interpretation of practices contribute to our theorising about justice?

I will offer an answer to this question in the next section. First, though, we must define the notion of practice and the role of practices in the formulation of principles of justice. Sangiovanni claims that ‘all relational views share the idea that principles of distributive justice cannot be formulated or justified independently of the practices they are intended to regulate’ (Sangiovanni 2007, p. 5). He states that relational accounts will differ according to the identification of the relevant ‘practice-mediated relations’ (Sangiovanni 2007, p. 5).

I have so far employed the ideas of relations and practices quite indistinctively. Henceforth, by relations, we should understand things like domination, exploitation, reciprocity and coercion, or, more positively, reciprocity or participation. Practices are more precise. They mediate relations, and can therefore perpetuate unjust relations. To a certain extent, my definition of practice follows Sangiovanni’s definition of institutional systems. Practices should be understood as a set of formal or informal norms, rules and decision-making procedures that regulate a political or social activity (Sangiovanni 2008, p. 142). Social and political activities regulate access to goods, the division of opportunities and patterns of political authority. Practices can include political organisations or regimes; political organisations are structured around a set of rules, most often formal rules, and act as social actors exercising authority over its members. A political organisations exercises de facto authority when it claims to impose duties and confer rights, and exercises de jure authority when this clam to authority is legitimised by its members. Regimes are sets of formal or informal rules that regulate a sphere of activity. Regimes can be regulated in part through political organisations, but can also be totally unregulated and still have an impact on individuals’ lives.

Examples of political organisations are states or international institutions such as the World Bank, the UN or the EU. Examples of formal regimes are the practice of human rights or the Agreement on Trade Related Aspects of Intellectual Property Rights (the TRIPS agreement) under the World Trade Organisation (WTO 1994), which controls access to medicines. Informal regimes are unregulated phenomena such as illicit financial flows, global tax competition or the escalation of trade tariffs.

Relations have different meanings according to the theoretical contexts and levels of analysis where they are used, as we have seen. Social, cultural and institutional practices affect and often shape the relations individuals can bear to one another. There are relations of reciprocity, sharing, participation, domination and exclusion, all of which are social relations. Relational approaches to justice want agents to be able to engage fairly in social relations and we want to prevent the rise and spread of unjust relations. But these relations are not the objects of regulations; they are ‘practice-mediated’. I want to suggest here that we cannot formulate principles to regulate relations as such in non-ideal theory. We can formulate principles to regulate formal or informal practices in order to obtain a society that prevents the rise and perpetuation of unjust relations, and also distribute social goods to enable individuals to engage fairly in social relations. Most relations are mediated through practices. For instance, discrimination can be the consequence of bad laws. Exclusion can result from inappropriate social safety net measures. Domination can be the consequence of missing working regulations (missing regulations would still be a practice in the sense of this thesis, since they are informal regimes that call to be formalized). Once regulated, informal regimes become at least de facto formal regimes. Of course, regulations do not immediately or solely prevent unjust relations. The social or cultural implementation of rules through time and contingencies will condition the possibility of preventing unjust relations. I assume henceforth that it is plausible for principles to target practices, in the sense defined above, rather than relations.

In other words, we can say that international practices are the site of global justice.

The site of justice is not the same as its scope: the site of justice refers to the kinds of objects (individuals’ actions, individuals’ character, rules, or institutions, and so on) appropriately governed by principles of justice, that is, to which the principles of justice rightly apply, whereas the scope refers to the range of persons who have claims upon and responsibilities to each other arising from considerations of justice. (Abizadeh 2007, p. 324).

Take Rawls’s theory of justice. For Rawls, the site of justice is the basic structure of society, whereas the scope of justice is limited to the national borders of each constitutional democracy (or any decent society). Indeed, Rawls and Scanlon explicitly conceive different principles for different domains of justice. In Justice as Fairness, Rawls states that the principles that are formulated for the basic structure of society are not fit for all subjects. These principles do not apply to universities or churches, and neither do they apply to the law of peoples (Rawls 1996; Rawls and Freeman 1999, pp. 522–523).

I believe it would be a mistake to assume that by identifying a given site of justice we would be determining the scope of justice. According to Rawls, it is the different structure of social frameworks, the role of its various components, and the way they all fit together that explains why principles apply to different subjects (or sites). That is, I am arguing that the sites of justice are individual practices, whereas the scope of justice is global. A non-ideal methodology therefore allows us to address the practices that cause the problems of global background injustice.

5 Proceeding from the practice

For Rawls, the main characteristics of ideal theory are the formulation of a theory of perfect justice, under the idealized assumptions of full-compliance and favourable circumstances (Rawls 1999, pp. 8–9). For Rawls, this inquiry has to be undertaken prior to the development of any non-ideal theory, so that it can provide a systematic grasp of non-ideal issues. To prove (i) – the claim that ideal theory does not have analytical priority over non-ideal theory – is to contradict Rawls on this point. It is to challenge the notion that non-ideal theory is applied ideal theory.

Given (i), principles of non-ideal theory need not to be modelled according to an ideal standard. This opens the door to developing non-ideal theory independently of ideal theory and, simultaneously, to think of ideal theory as something less demanding than perfect justice theory. Amartya Sen argued that the feature of perfect justice – understood as a situation above which no principles of justice of a higher order could be formulated – made the two theories analytically disjointed (Sen 2009). My point is that even if we relax this constraint (as we should) and accept that an ideal theory can be less demanding than perfect justice (say a theory of basic rights), non-ideal theory would still not be modelled according to this standard. ^[4] Claim (i) thus means that we do not need an ideal theory of any kind (as a theory of perfect justice, basic justice or some other form) to undertake the work in non-ideal theory.

To prove (i), it would be helpful to clarify that relational positions do not deny moral equality of persons, but they justify their principles by appealing to further moral considerations, which limit the scope of principle (Valentini 2010, pp. 399–400). The scope of the principles will be limited by the boundaries of the practices that can be identified following the justificatory element of justice. This proves that relational approaches and non-ideal theory, at the level justification and scope of justice, intersect with regard to their methodology. A relational theory is a theory that is non-ideal with regard to its methodology. ^[5] The methodology of non-ideal theory is relational. It follows that, although this approach proceeds from a practice that is the cause of structural or immediate problems of justice, it would only show concern for equality indirectly, or in time. Principles that regulate tax competition will probably target states, financial institutions and corporations. Such principles might not directly reflect normative individualism.

Now, we may rightly ask what justificatory element triggers the formulation of principles of justice. Although we might not have a definitive answer to the questions ‘what global justice requires’ we could nonetheless answer the question ‘what normative element allows us to say that a given practice is prima facie unjust and should trigger a non-ideal theoretical analysis?’ Whilst a non-ideal theory can be pluralist about values and accommodate a specific interpretation about why each individual practice is unjust, there could be one sufficiently weighty normative element that would trigger the consideration of the practice from a justice standpoint.

One possibility is that any institutional system or practice that has a significant, pervasive or profound impact on individual opportunities is subject to considerations of justice. One of Rawls’s reasons for choosing the basic structure as the site of justice is that the basic structure has a profound impact in individual opportunities (Cohen 1997; Abizadeh 2007). If we ground this view on the notion of pervasive impact alone, any practice that has a pervasive impact on individuals’ lives would be a subject of justice, regardless of its function in coordinating social cooperation.

As opposed to a view based on social cooperation, a view based on the idea of pervasive impact is not limited to the participants in a system of cooperation but rather to all those significantly affected by the impacts of an institutional system. That would notably include people that might only feel the negative impact over institutions of which they have no control. It would also include people that are only affected by the externalities of an institutional system without ever being subject to its policies. In all such cases, agents (like states and individuals in the case of tax competition) cannot escape from the impact of these practices. This view is thus a good candidate to provide further normative grounds to a non-ideal approach to global justice (adding to the strengths of a normative theory that considers the particularities of individual practices).

Without saying anything about the content of principles of justice, a practice that has a pervasive impact on individual opportunities triggers the formulation of principles of justice. In this case, a formal or informal practice that has a profound effect on individuals’ lives has to be a pre-existing condition for justice concerns to arise. Thus, pervasive impact is a necessary condition for the formulation of principles of justice. A pre-existing condition is a necessary condition but not all necessary conditions need to be pre-existing. Also, pervasive impact is a sufficient condition for considerations of justice to arise. That is because although the necessary institutions, political authority or policy tools to address a problem of justice might not be in place, the fact that there is a problem is a sufficient reason to look to establish the adequate structure to address the problem. As Ronzoni argues, we have a duty to create the background of interactions that would allow for the implementation of ideal principles (Ronzoni 2009). Similarly, I argue that we have a duty to create the adequate institutional scheme or reform if the existing mechanisms are insufficient to address problems of justice. Note that pervasive impact is a trigger to formulate principles of justice. The appropriate notions required to formulate principles will be given by the content of non-ideal theory and an interpretation of the relevant practice, as I will show below.

In sum, pervasive impact on individual opportunities is the criterion based on which we identify practices that are subject to considerations of justice. An analysis based on the notion of pervasive impact is one way to cement the methodology of a non-ideal relational account of justice. Therefore, if we accept the conclusion of the last section, that international practices are the site of justice and that all practices that have a pervasive impact on individuals’ life prospects trigger the formulation of principles of justice, we have established claim (i). Indeed, following Section 4, we can claim that the scope of justice is global, but that different principles apply to the domestic and the global realm. Section 5 provided solid reasons to consider practices as the site of justice. And to the extent that these practices have an impact on individuals’ life prospects and thereby cause ongoing problems of justice, we have reasons to address them directly and thereby contribute to the realisation of justice. This shows ideal theory is not analytically prior to non-ideal theory.

In the remainder of this section, I will reply to one objection that might seem to threaten both (i) and (ii). The objection challenges the project of a non-ideal theory of justice in both its content and methodology:

To assign any greater role to institutions and practices – to allow them, as I have said, to influence the formulation and justification of first principles of justice – is a fundamental mistake: constraining the content of justice by whatever social and political arrangements we happen to share gives undue normative weight to what is, at best, merely the product of arbitrary historical contingency or, at worst, the result of past injustice itself. (Sangiovanni 2008, p. 137. Sangiovanni rejects the objection, but not for the same reasons I provide here.)

‘First principles’ of justice are principles that are not themselves derived by applying other more fundamental principles to particular circumstances (Sangiovanni 2008, p. 137 n.1). They can be derived from facts about human nature or from moral values (such as fairness or respect), but not from other principles. Non-ideal theories including the one I am defending claim that first principles should apply to practices individually. ^[6] This implies that the nature of the practice will influence the formulation of the principles. The objection rejects this methodology, claiming that the nature of practices should not play any role in the formulation of first principles of justice.

The reply to this objection is central to the project of non-ideal theory. I argue that it is wrong to claim that the consideration of practices at the level of justification implies giving undue normative weight to contingent products of history. It is quite the opposite. By doing so, we give ongoing unjust practices and past injustices the normative weight they require. Colonialism, access to medicine or tax abuses are all past or ongoing practices that should be subject to considerations of justice. Addressing them as such is to show equal respect to all those that suffer from the injustices propelled by these practices. Responding to particular injustices and making sure our principles reflect the specificities of these injustices is perhaps to give them the weight they command in order that they may be prevented in the future and for compensation to be given for the past.

In light of this, it is odd that the burden of proof about analytic priority is taken to lie on those who try to address these injustices directly and individually. We may pursue this reasoning and look for other reasons to tackle these practices individually. A good candidate is agency. Colonialism deals with present agents trying to compensate for the wrongs of past agents. Climate justice wants present agents to compensate for the wrongs of past agents in order not to undermine the life prospects of future agents. Tax competition involves international institutions, states and corporate agents, that may have nothing to do with the two problems just mentioned. These examples support claim (ii): that we should formulate different principles for different practices. Indeed, the exact formulation of the principles of global justice will more likely, and perhaps will only, be determined by an analysis of these practices individually. This reinforces the argument against the analytical priority of ideal theory – claim (i) – and surely suggests that the set of principles that will target each practice will be formulated for these practices individually – claim (ii). Although we had sufficient practical reason to address these practices individually, such as to give past injustices sufficient consideration, we also have major conceptual reasons, like agency, to do it. Exploring other reasons is beyond the scope of this thesis. Agency is a complex enough case to support claim (ii) on its own, for it involves: the temporal action of agents (past, present, future), the nature of agents (individuals, states, transnational institutions) and the nature of actions (harm reduction, distribution, background regulation, compensation) among other problems. Moreover, as mentioned above, the interpretation of the practices might reveal the different nature of the problems, when for instance we observe that tax competition is best described as an ‘asymmetric prisoner’s dilemma’ case and the problem of climate change another type of collective action problem, one where self-interest plays a larger role. This provides further support for (ii).

Claims (i) and (ii) consolidate the methodology of non-ideal theory: non-ideal theorising proceeds from existing practices, addresses them individually in a sectorial approach and does not presuppose an ideal theory.

6 The content of non-ideal theory: a theory of action-guidance

The recent debate on ideal and non-ideal theory revealed ways in which we can interpret action guidance paradox of ideal theory and how it is possible to respond to it from the standpoint of ideal theory (Robeyns 2008; Swift 2008; Valentini 2009; Hamlin and Stemplowska 2012). This section provides insights about how non-ideal theory responds to it, thereby exploring the content of non-ideal theory. For Swift, action guidance means ‘knowing what options are feasible, over what time scale, with what probabilities, given where we are now’ (Swift 2008, p. 374). These are indeed relevant issues, each one deserving careful examination. I distinguish in this section and the next between the notions of compliance, feasibility, fact-sensitivity and path-dependence of a given institutional proposal. These are four components of action-guidance in a non-ideal theory of institutional design.

7 Tackling tax competition

This section presents insights about the ways non-ideal theorising contributes to solving the problem of tax competition. I will begin by quickly reviewing a few issues with ideal theorising that are avoided by an adequate non-ideal theory. One idealisation avoided by non-ideal reasoning is the one of thinking about global taxation by abstracting from the world as divided into states, as the division of states would presumably consist in a morally arbitrary division which would itself be the cause of much injustice. With all that has been said so far about action-guidance, I do not need to dwell much on this point. But it is important to remind ourselves here that making such an idealisation would make us mischaracterise the very subject of our inquiry. Whether or not the world would be more just without it being divided by national borders is speculative and irrelevant. What is pertinent to the discussion is to know that the problem of tax competition arises because capital mobility is global and tax policies are determined domestically. Not taking seriously the existence of states, with their interests and prerogatives, capabilities and submission to private interests, is to misunderstand the problem. Whatever regulation, especially of international coordination, has to be sensitive to the powers and limits of nation-states today. A second potentially harmful ideal principle of institutional design is the harmonisation of global tax rates as a simple solution to tax competition. Preventing countries from adjusting their tax rates in moments of economic growth and recession is probably dangerous economic policy (Ronzoni 2014; Dietsch 2015, p. 118). Governments require flexibility to raise taxes in moments of great expenditure and to diminish taxes in times of unemployment. Arguing for the harmonisation of global tax rates as a simple solution is potentially damaging. ^[9]

Moving to the level of non-ideal theory, recall we distinguished the three forms of tax competition – in relation to portfolio capital, paper profits and foreign direct investments (FDI) – and two ways via which the autonomy of states is undermined: the ‘poaching’ (when the capital owner does not follow the investment) and the ‘luring’ (when countries attract FDI by making investments part of their tax base) of taxes. For the large affluent countries, if governments wish to prevent important revenue losses, they have to adopt regressive fiscal policies, notably by shifting the tax burden from capital to labour, and from taxation on revenue to indirect taxation of consumption (Dietsch 2015, pp. 47–8). For less affluent countries, the result was more directly a net and massive loss of revenues.

It follows from the assessment of the practice from a justice standpoint that the involuntary deprivation of resources needed to act upon justice internally, and its consequences on aggregate welfare, is a sufficient reason to regulate this practice. States should not suffer from undue external interference, and thereby lose their capacity to express a collective self-determination that should have a say both on the size of their public budget and on their redistributive capacities.

Dietsch and Rixen formulate two principles to regulate the three different kinds of tax competition: the membership principle and the fiscal policy constraint (Dietsch and Rixen 2012; Dietsch 2015, p. 80). The membership principle targets competition for portfolio capital and paper profits. It requires that natural and legal persons pay taxes in the countries in which they benefit from the existing infrastructure. The membership principle is associated with a demand for transparency. Individuals and corporations must make their income data available to tax authorities. The consequence of eliminating the competition for portfolio capital – recall the substitution effect mentioned in 2 – is that it would increase the competition for real investments. This is when the second principle comes into play. The fiscal policy constraint forbids tax policies that are simultaneously unjust and strategically motivated. It states that:

A tax policy is legitimate if it does not produce a collectively suboptimal outcome. A collectively suboptimal outcome is here defined as one where the aggregate extent of fiscal self-determination of states is reduced (Dietsch and Rixen 2012, p. 13).

Respecting states’ fiscal self-determination is morally required for states to act upon justice domestically and for states to be able interact freely globally. Following what has just been said, a practice that infringes upon the fiscal determination of states is a practice that constrains states’ capabilities to implement programmes and provide public goods its society desires. The intuition behind regulating tax competition is to promote justice by preventing injustice in the first place. Its objective is to allow for better national distributive capabilities, including the capability to maintain the overall size of a national budget and the capability to determine the internal distribution within this budget.

It is clear now that the methodology of non-ideal theory is effective in pressing us to address the problem of tax competition directly and individually. We have sufficient reasons to believe tax competition is a cause of injustice in the world today. No ideal theory is required for us to address this problem. And we know enough about the specificities of the practice to guide us to the formulation of principles (much more than the thin lines exposed in this paper). I wish now to point out how the content of a non-ideal theory, including the principles formulated above, needs to be fact-sensitive and target compliant.

First and foremost, the principle of states’ self-determination needs to be balanced against the principles of transparency and information exchange. Values of transparency and self-determination cannot be balanced against each other abstractly, in ideal theory. For instance, one idea put forward by the OECD consisted in providing tax ‘information exchange on request’ (OECD 1998). Yet, for a country to gain access to other countries’ information on individual cases it needs to provide initial proof of tax evasion. Until recently, such proof was very hard to build. More robust measures based on routine verification and multilateral exchange of information are necessary (Dietsch and Rixen 2012, p. 21). This shows why addressing non-compliance is a constitutive part of a non-ideal theory of tax competition. Asking agents to report on audited financial statements and to favour fostering banking transparency is central to the formulation of this conception of justice. That is, the precise design of principles necessary to enhance states’ distributive capabilities requires the balancing of their self-determination and their engagement for financial transparency. Today, many countries have agreed that automatic exchange of information (AEOI) should be part of the new standard in global tax matters (OECD 2016). ^[10] Mechanisms of compliance are constitutive of a non-ideal theory of tax competition.

An example hinging on feasibility, fact-sensitivity, path-dependence and compliance, to which I promised to return, is the ‘arm’s length principle’. In the example about transfer mispricing in section 3, the office supply company based in the US declared low profits in the country. It also manipulated transactions so as to appear to have sold its goods and made a huge profit – tax-free – by ‘paying’ its subsidiary in the Cayman Islands an exorbitant price for pens, paper and laptop bags. In principle, such transactions should not have been permitted because of the arm’s length principle, according to which transactions have to be conducted as they would have been with an arbitrary third party. This has as a consequence that companies can only buy and sell at the fair market price. The OECD transfer pricing guidelines are based on this principle (OECD 2011). This is a very specific principle, based on impartial fairness in transaction that needs to be enforced as part of a general proposal for curbing tax competition. We will need to be sensitive to facts and non-compliance, not only to make sure the OECD guidelines provide effective means for countries and courts to assess transfer pricing practices of multinationals, but to make sure more complex cases can be addressed. This principle is indeed particularly complex to monitor and to enforce (Avi-Yonah 2016). It does not effectively prevent companies from engaging in profit shifting. Consider the case of intellectual property rights that challenge the feasibility of a successful implementation of this principle. A company could sell coffees in the US and write in its books that it paid its subsidiary in Ireland a high price for the intellectual property rights on its famous ‘Makkachino’. Intellectual property rights are more difficultly assessed under the arm’s length principle. These are questions that a non-ideal theory of tax competition needs to discus in order to guide action in the non-ideal world.

Relying on the arm’s-length principle does not allow for the required path to be paved. It has been observed that the OECD’s (2013) efforts to repair the arm’s length standard have in fact contributed to a cat-and-mouse game, where the regulators are always one step behind (Avi-Yonah 2016; Dietsch and Rixen 2016). Path-dependence is crucial for the formulation of an action-guiding theory and it is best determined by non-ideal level theorising. Path-dependence is observed not only in the choice of a system of regulation, but also on the very way to approach the governance of international tax interactions. ‘Global tax governance thus consists of the set of institutions governing issues of taxation that involve cross-border transactions or have other international implications’ (Dietsch and Rixen 2016, p. 3). Governance involves hard choices such as the choice between limiting the mobility of capital by shifting power to an international body and finding mechanisms of global governance that enhance the capacity of states to tax mobile capital. Exactly what level or what dimensions of a nation’s power to tax need to be shifted to the international level is a question beyond the scope of this paper. What is central for the purposes of this thesis is to observe that the way to organize collective action in the case of tax competition can only be given in non-ideal theory.

This allows me to return to claim (iii) and the other elements that define the content of non-ideal theory. Clearly, in order to prevent injustices caused by tax competition, the role of non-ideal theory is not to determine what to do when agents do not comply with ideal principles, but how to make agents comply with non-ideal principles. The question of compliance in non-ideal theory is motivated by actual non-compliance as the core of the problem. It is not a question that can be avoided by non-ideal theory, but one that should structure it. Moreover, for regulations to effectively target compliance, they have to be fact-sensitive. It is in relation to real world facts that we assess the desirability of proposals such as the harmonisation of tax rates. Also, for principles of justice that target tax competition (such as Dietsch and Rixen’s two principles) to be feasible, mechanisms that make the implementation of principles feasible have to be enforced. The agreements on AEOI of the OECD have to be in place for membership taxation to be monitored. This is the kind of non-ideal theory feasibility issues that we must keep in mind. Finally, it is crucial to understand why a principle such as the ‘arm’s length principle’ is ineffective. In today’s context, too much is based on a principle that has very little chance of succeeding and too much attention is given to a coordination response to a problem that requires global governance. A question for a non-ideal theorist is whether to build mechanisms around this principle to make it effective or find an alternative principle. These are all questions for a non-ideal theory of global tax competition.

8 Conclusion

Considerations of justice aim at enabling individuals to enjoy their right to freedom. Whilst freedom is mostly realized at the domestic level, within the boundaries of a political community, international interactions and institutions very often undermine individuals’ capacities to pursue their plans of life. For such institutions to stop interfering negatively would already have immense positive effects. Regulating the problems that are particular to the global level should have a great positive impact on individual opportunities in different nations. The redistributive duties between nations would be significantly diminished in an international context where individual practices were effectively regulated.

The main task of this paper was to provide an account of a non-ideal theory of global justice, detailing its methodology and content. It has opened the door to a conception of justice based on the ideas of non-compliance, fact-sensitivity, feasibility and path-dependence. It did so by arguing for three claims. We saw the theoretical and practical value of a theory that proceeds from a specific practice that is the very cause of injustice. This paper provided a case for a non-ideal theory of justice that was not aimed at implementing ideal principles, thereby suggesting that ideal theory is not analytically prior to non-ideal theory, i. e. claim (i). Sections 4 and 5 supported the case for (ii): global justice is a pluralist theory pursued by sectorial contributions. We saw in section 5 that an adequate interpretation of the practices provided insights to the formulation of principles (transparency and information exchange in the case of tax competition; incentives of the pharmaceutics industry for the case of agreements on intellectual property). Clearly, global justice theories should be open to pluralism at the level of principles and perhaps even at the level of values. The complex nature of individual practices require principles tailored to address these complexities (think about the overcoming the deficiencies of the arm’s length principle to curb transfer mispricing). Section 7 provided support to claim (iii): determining our duties of justice in a non-ideal world is not a question about non-compliance with ideal principles but rather about compliance with non-ideal principles. This shift from the standard interpretation is warranted. Non-compliance is at the very core of non-ideal theorising. To avoid the question of compliance by asking whether our duties are affected in cases of partial compliance is to misunderstand the role of non-ideal theories. The non-ideal theorist should aim to bridge the gap between the values underpinning our moral assessment of contemporary problems of justice and the development of regulations that can provide genuine guidance to agents living in the real world as it is today.