1Illicit financial flows (IFFs) are found to be a key obstacle to the development of low- and lower-middle-income countries, depriving them of billions of dollars in tax revenues (Brandt, 2023). Yet beyond a shared moral concern about this loss of revenue, views differ about IFF concepts and definitions (Musselli and Bürgi Bonanomi, 2020), and even more about appropriate and effective measures to curb IFFs. This holds particularly for IFFs related to foreign direct investment and trade, where transfer mispricing, trade mispricing and profit shifting play an important role (Crivelli, De Mooij and Keen, 2015; Wier and Zucman, 2022; Mosquera Valderrama, Lesage and Lips, 2018). Which policy measures are seen as effective and should be deployed is a highly contested point. This is unsurprising as the question involves distributional conflict: more for those deprived of revenues now means less for those who currently benefit.
2There are many possible regulatory responses and policy innovations on the table that might curb trade-related IFFs. These range from increasing transparency in trade and financial transactions to strengthening capacity-constrained tax administrations in low-income countries to fundamentally overhauling the existing international taxation system.
3While measures proposed to address IFFs are highly diverse, the complexity and interconnectedness of the topic require, in most cases, coordination between jurisdictions and careful consideration of the broader institutional and regulatory context to avoid new loopholes and ensure policy measures are effective. International organisations play an essential role in this coordination. Since each international organisations has its particular mandate and decision-making procedures to represent the interests of its members, the question of where interests are coordinated and where international rules are negotiated matters. For example, in global tax governance the role of the Organisation for Economic Co-operation and Development (OECD)—whose members are the 36 most advanced economies—as the principal forum for tax negotiation has increasingly been challenged by non-member countries, all of them low-and middle-income countries, calling for UN-based negotiations where they have a vote (Brugger and Engebretsen, 2020; Christensen and Hearson, 2019; Hearson, Christensen and Randriamanalina, 2023; Seabrooke and Wigan, 2016; Teo, 2023).
4Given the complexity of the topic, the plethora of proposals to curb IFFs, and the numerous actors at different levels involved, the question emerges as to what policy space is available to advance reform. The concept of ‘policy space’ delineates the limits to, and opportunities for, policy actors’ choices or actions. Factors that can constrain or expand a policy space include institutional frameworks such as legal and regulatory structures at national and international levels, pressures from other states or from international organisations or interest groups, the financial resources available and technological advances. Another critical aspect determining the policy space available is the dominant political climate. The combined effects of the agenda of policymakers, the balance of political power, the strength of interest groups and public opinion can influence the acceptance of reform proposals among different stakeholder groups, and hence the range of policies considered. This is particularly important for highly contested policy issues where the stakes for the various actors are high. Understanding the areas where views converge can reveal the room for manoeuvre that a government, organisation or other decision-making body has when formulating and implementing policy reforms, whether at the domestic or the international level.
5Taking the case of commodity trade–related IFFs, we explore the policy space available for reform by eliciting policy experts’ preferences for measures to curb IFFs. Study participants include public officials, and representatives from international organisations, the private sector, policy think tanks and academia. Although our focus is on commodity trade and commodity trade mispricing more specifically, several of the policy measures considered are broader in scope, targeting different types of IFFs, whether tax-related, corruption-related, or originating from transnational criminal activity.
6We proceed in two steps. First, building on existing research by Musselli and Bürgi Bonanomi (2021) into IFF policy measures, we provide a conceptual framework delineating the range and scope of different policy options to address commodity trade–related IFFs. On this basis, we distil a diverse set of policy options to elicit stakeholder preferences through a Q-methodology study to identify the policy space that intersects stakeholder groups.
7We find that views on the most appropriate policies to curb IFFs differ widely. Critical tensions run along preferences for short-term fixes versus more comprehensive structural reforms regarding extending legal liability, and whether host countries should be encouraged to use or expand their legislative sovereignty to introduce (unilateral) measures for curbing IFFs. Policy measures to increase targeted transparency are the most likely to garner approval from all stakeholders.
8Our conceptual framework builds on the comprehensive compilation and analysis of policy options to address commodity trade–related IFFs of Musselli and Bürgi Bonanomi (2021). The inventory captures the current policy debate related to commodity trade–related IFFs; however, despite striving for completeness, this is an evolving field and new policy options continue to emerge (see Table 6.1). The inventory is organised along a three-by-three matrix (see Figure 6.1). The x-axis considers the temporal dimensions and sorts the policies into short-, medium-, and long-term interventions. The key distinguishing criterion here is the complexity of decision-making and implementation. Measures that can be implemented in the short term build on existing regulations and do not require legal changes; some can even be adopted by a single or just a few actors. Short-term measures are relatively low-hanging fruit, but their potential to substantially curb IFFs is often limited. Medium-term measures, in contrast, require additional political and organisational effort and depend on the cooperation and coordination of multiple actors across countries; typically, they also involve developing and adopting a new legal basis. Long-term policy measures aim for more profound structural reform, which is only possible with concerted action at the multilateral level.
9The y-axis distinguishes between the implementing jurisdictions. The home country jurisdiction is where the multinational enterprise (MNE) is headquartered or incorporated. It is typically where the firm’s central operations and management are located and where the parent company is subject to laws, regulations, and economic policies. The home country may provide support and incentives to companies headquartered in its jurisdiction, and in turn benefits from their global success through taxes, employment, and economic growth. The host country, meanwhile, is a foreign jurisdiction where the multinational company operates or conducts business activities and establishes a subsidiary company. In resource extraction and commodity trade, host countries are typically low- and lower-middle-income countries. While the host country has its legal framework regulating resource access, environmental protection, labour laws and fiscal regime, legislative sovereignty is often limited by international agreements that the country has adopted. For example, the investments and operation of foreign-controlled firms are usually subject to bilateral investment treaties (BITs). Most of the roughly 2,200 BITs in force (UNCTAD, 2023) are based on either the OECD or UN Model Tax Convention (OECD, 2017b; UN, 2017), which both require tax administrations of host countries to follow the OECD Transfer Pricing Guidelines (OECD, 2017c; Soong Johnston, 2017). The current transfer pricing guidelines have been identified as conducive for IFFs (Durst, 2010; Faccio and Picciotto, 2017; Picciotto, 2018).
10As sovereign jurisdictions, the home and host countries can take unilateral measures if international agreements like the BITs or relevant multilateral conventions do not restrict them from doing so. In this case, bilateral or multilateral negotiations or global cooperation is required to effect changes to the existing regulatory framework. Such negotiations are typically protracted but provide an opportunity for more fundamental reform. That said, the distinction between whether the measures can be taken by the host or the home country is not always clear-cut in practice. For some policies, it needs cooperation between the two to negotiate and implement particular measures.
Table 6.1 Inventory of policy measures to curb commodity trade–related IFFs.
AA1 Introduce the use of smart containers.
AA2 Introduce systematic scanning of containers.
AA3 Introduce the use of blockchain to manage trade transactions.
AA4 Improve the flow of information and data matching between customs authorities, tax authorities and banks, within and between countries.
AB1 Build capacity and funding for independent assaying capacity in producer countries (laboratories, inspection firms).
AB2 In state-owned enterprises, ringfence the buyer selection process from political interference.
AB3 Make information related to all stages of the buyer selection processes in state-owned enterprises publicly available.
AC1 Offshore wealth centres shall spontaneously inform low-income countries on accounts and amounts held by their residents in a de-identified manner.
AC2 Offshore wealth centres shall publish the number of accounts and amounts from low-income countries in a de-identified manner.
AC3 Trading and financial hubs shall supply information – automatically or on request – to low-income countries on a non-reciprocal basis.
AC4 Sharing tax information received from partner countries between public entities within the receiving state shall be allowed.
AC5 Trading and financial hubs shall coach low-income countries to request and interpret tax information beyond the Tax Inspectors without Borders initiative.
AC6 For gold, establish granular statistical categories under the Harmonized System for different gold purity levels to better select reference prices.
AC7 In trade statistics, introduce statistical keys to identify related-party transactions.
AC8 Add the origin of gold received in custom statistics based on refinery information.
AC9 Support efforts to establish geo-forensic methods for the identification of the origin of gold.
AC10 Publish granular merchant trade data on a commodity-by-commodity basis (instead of net receipts only).
BA1 Producer countries shall legislate the use of reference prices to determine the tax value of commodity sales.
BA2 Producer countries shall put limits to deductible taxpayer costs in intra-group transactions.
BA3 Producer countries shall introduce profit allocation rules for subsidiaries (e.g. minimum profit margins or fixed margins).
BB1 Make professional service firms criminally liable when they fail to prevent their employees or agents from facilitating criminal tax evasion.
BB2 Require legal professionals, accountants and service providers to report suspicious transactions to the anti-money laundering authority.
BB3 Legislate penalties for the designers, marketers or facilitators of abusive tax arrangements.
BB4 Legislate mandatory disclosure of cross-border tax arrangements when they have certain predefined features that might signal aggressive tax avoidance.
BB5 Introduce national beneficial ownership registries and verification mechanisms covering all types of legal entities.
BB6 Introduce international beneficial ownership registries that are interoperable with national registries.
BB7 Require the parent company to make the Country-by-country report public.
BB8 Require the parent company to send the Country-by-country report to its subsidiaries for local filing.
BB9 Allow the use of Country-by-country data for tax adjustment purposes.
BB10 Financial centres and trading hubs shall legislate unilateral information exchange in lieu of existing reciprocal standards.
BB11 Offshore centres shall assist other countries in recovering undeclared wealth through withholding taxes on interest and dividend payments to non-residents.
BB12 Payments-to-governments disclosure requirements shall be extended beyond extraction to include disaggregated information on trading.
BB13 Offshore centres shall introduce effective protection frameworks for whistle-blowers compliant with the OECD Anti-Bribery Convention.
BB14 Home countries shall influence the conduct of their companies abroad through legal means, including through mandatory supply chain due diligence, requirements on institutional investors and financers, listing requirements, and redefinition of directors’ fiduciary duties.
BB15 Make taxes paid at source in low-income countries tax deductible against the tax due in residence countries.
BB16 Negotiate development-friendly Double Taxation Agreements with low-income countries based on the UN Double Tax Treaty template.
CA1 Introduce unitary taxation with formulary apportionment as global tax standard.
CB1 Combine transparency approaches into a Global Asset Registry linked with legal entities and taxpayer IDs covering securities and other financial assets (trusts, funds, real estate, luxury items).
CC1 Make Environmental, Social and Governance (ESG) disclosure mandatory within a multilaterally agreed framework.
CC2 Exclude companies convicted of tax evasion, bribery, or money laundering from finance, insurance coverage, and equity investment.
|
Source: the authors, based on Musselli and Bürgi Bonanomi (2021).
Figure 6.1 Conceptual framework for organising policy proposals to address trade-related IFFs.
|
|
Implementation timeframe
|
|
|
Short term
|
Medium term
|
Long term
|
Implementing jurisdiction
|
Cross-border/ global
|
AC1-4
|
--
|
CA1-CC2
|
Host country
|
AB1-3
|
BB1-3
|
--
|
Home country
|
AA1-10
|
BA1-16
|
--
|
Note: The information in the cells refers to the policy statements derived to build the Q-set as discussed in Section 3 and listed in Table 6.1.
Source: the authors, based on Musselli and Bürgi Bonanomi (2021).
11Q-methodology is a research methodology that combines quantitative and qualitative techniques to systematically study human subjectivity (Brown, 1980; McKeown and Thomas, 2013; Watts and Stenner, 2012; 2005). It is a valuable tool for understanding how small-n-groups (as in our case) perceive a specific issue. Q-studies have found application across various disciplines and have become increasingly popular in policy analysis, offering insights for political decision-making (Alderson et al., 2018; Brown, 2019; Molenveld, 2020; Sardo and Sinnett, 2020; Zabala, 2014).
Figure 6.2 Q-sorting grid.
Source: the authors.
12During Q-studies, participants rank predefined statements that broadly represent the spectrum of viewpoints on the topic of interest, representing the ‘concourse’, in Q-parlance (Watts and Stenner, 2012; 2005). These statements are displayed on cards, and each participant sorts them into a forced normal distribution based on his or her preferences (see Figure 6.2), resulting in one Q-sort per participant. Patterns of each Q-sort are then established. By employing by-person factor analysis, similar ranking patterns in participants’ Q-sorts (i.e. Q-sorts with similar significant loadings) are combined into so-called factors (see Figure 6.3), whereby each factor represents a shared viewpoint or perspective of study participants (Alderson et al., 2018; Brown, 1980; Schuijff, De Jong and Dijkstra, 2021; Stephenson, 1993; Watts and Stenner, 2012).
Figure 6.3 By-person factor analysis.
This by-person factor analysis analyses how different preferences are expressed within each individual, helping to find groups of individuals who have sorted the statements in a similar way, indicating shared perspectives or viewpoints.
Source: the authors, based on Baker (2016).
13In this study, the concourse from which the statements are derived pertains to the spectrum of regulatory measures and policy responses to curb commodity trade–related IFFs. Musselli and Bürgi Bonanomi (2021) have presented a comprehensive stocktaking of the regulatory and policy options from the academic literature and the policy discourse, which we use as our starting point for formulating the Q-statements (collectively called the Q-set). To generate the Q-statements, we summarise each policy proposal in a one-sentence statement (see Table 6.1). We reconfirmed with the authors that each single-sentence statement accurately represents the intention of the respective policy proposal. This results in a Q-set of n=40 Q-statements. The policy measures (and hence the Q-statements) are organised in the three-by-three matrix (see Figure 6.1) that distinguishes short-, medium-, and long-term responses, home country, host country, and cross-border or global measures. Of the nine resulting fields, policy options were identified for six: short-term, home country (AA); short-term, host country (AB); short-term, cross-border/global (AC); medium-term, home country (BA); medium-term, host country (BB); and long-term options at the global/multilateral level (CA-CC).
14Study participants were purposefully selected based on their broad classification as relevant policy stakeholders and their general familiarity with the topic. In our selection of participants we also sought to include those with diverse backgrounds and professional roles to ensure a heterogeneous distribution (Watts and Stenner, 2012; 2005). In total, a sample of 20 participants (P = 20) were recruited, resulting in a 2:1 Q-set-participant ratio, which is good practice in Q-studies (Webler, Danielson and Tuler, 2009). Participants comprise representatives of private sector entities active in the commodity sector (five), representatives of think tanks and interest groups covering the broad spectrum of political positions and interests (seven), academics working on IFFs (six), and regulators (two). The low number of government officials is the result of a low response rate. As several bureaucrats who declined to participate argued, this reluctance is due to the sensitivity of the topic. How severe is their absence from the study? Such professionals are tasked with developing policy. It is not they, however, who will ultimately set policy. Rather, they are the agents of politically appointed principals (the minister and the parliament) who have the decision-making power. This does not mean that bureaucrats have no agency or room for manoeuvre. However, in politically contested topics such as IFF, an administration’s independence should not be overestimated as it is closely watched by its political constituencies. The fear of being perceived as acting politically explains government officials’ reluctance to participate. We therefore argue that the low representation of bureaucrats does not reduce the existing diversity of policy stances in our sample. Elected parliamentarians, in turn, engage in policy work in consultation with their preferred interest group organisations and think tanks since they combine technical expertise and the preferred political ideology. Interest groups and think tanks actively engage with the policy process and advise policymakers. Therefore, to represent the different policy stances, the diversity of interest group organisations and think tanks is as relevant, which also explains why this is the largest group from which we recruited study participants.
15Data were collected in the form of participants’ sorted rankings of statements (Q-sorts) as well as their written comments. Before the study, selected participants were contacted via e-mail and provided with information about the study’s background and purpose. Participants were assured of anonymity, and their involvement was based on prior informed consent. The Q-sorts and participants’ comments were collected remotely between September and October 2021 using the cloud-based application QMethod Software (Lutfallah and Buchanan, 2019).
16For the Q-sort administration, participants were required to rank the 40 policy statements from the most to the least representative given their personal views. In the first step, participants were asked to read the statements carefully and sort the cards into three initial piles: most representative of their views, least representative, and neutral. Then, participants were requested to sort all statements onto the Q-sorting grid, from the most representative (+5) to the least representative (-5) (see Figure 6.2). Participants were free to change their sorted rankings of statements until satisfied with the final distribution.
17In the next step, we asked participants to explain their ranking choices, especially the statements ranked most representative (+5) and least representative (-5). These explanations helped with the factor interpretation and allowed us to gain a better understanding of stakeholder viewpoints and preferences. Subsequently, participants were given the opportunity to provide any further comments.
18The analysis included 20 Q-sorts, corresponding to the number of study participants. Eight factors were initially extracted, and three were selected for rotation, employing a varimax rotation method. Factor selection is based on a combination of statistical criteria (e.g. the Kaiser–Guttman criterion on eigenvalue, minimum and cumulative variance) and theoretical considerations (Armatas et al., 2014; Brown, 1980; Molenveld, 2020; Nguyen et al., 2018; Watts and Stenner, 2012; Webler, Danielson and Tuler, 2009; Zabala, Sandbrook and Mukherjee, 2018). In our case, the three factors retained (F1, F2, F3) explain 43 per cent of the study variance (18%, 14%, and 11%) with no significant correlation between factor scores.
19Out of 20 Q-sorts, 19 were found to load significantly on one or more of the three factors (factor loading of 0.41 or higher at the p>0.01 level), indicating a strong relation between each participant’s Q-sort and the factor (Zabala, Sandbrook and Mukherjee, 2018). One Q-sort was confounded, showing significantly positive loadings on both F1 and F3, and was excluded (Watts and Stenner, 2012; Webler, Danielson and Tuler, 2009). As a result, F1, F2, and F3 are defined by six, five and seven variables (i.e. Q-sorts), respectively, whereas F1 and F2 are bipolar as both factors include significantly negative loading Q-sorts. Bipolar factors in Q-methodology mean that the Q-sorts associated with this factor take on diverging, tentatively opposite viewpoints. The negative loading has nothing to do with a wrong or negative position; the two represent just the views from the opposite poles (Watts and Stenner, 2012).
20For interpretation, the bipolar factors F1 and F2 were each split into two separate factors, F1a and F1b, and F2a and F2b, resulting in the construction of four-factor arrays (defined by five, one, three and two variables, respectively), whereby F1b is the Q-sort of a single participant. Detailed statistics are provided in the appendix.
21We interpret the factor arrays mainly by analysing the extreme statements (±5, ±4). In some situations, we also consider statements that received more moderate ratings (±3). We first assess the extent to which the policy proposals ranked under each factor are coherent and cluster around one or more policy approaches to curbing IFFs. We then ask where policy preferences espoused by the respective factors are located on the conceptual matrix introduced earlier (see Figure 6.1). Taken together, this information outlines the position of the ‘policy community’ represented by a given factor. Superimposing the three positions reveals the policy space in which preferences might overlap and where reforms are most likely to gain majority support.
22We label the perspective that is represented by Factor 1 the compliance view (see Figure 6.4) because the policies that this view espouses pivot around extending liability for curbing IFFs to actors that are currently not held legally accountable. This also entails expanding legal consequences in the case of wrongdoing.
23The compliance view subjects professional service firms (BB1) and designers, marketers and facilitators of tax evasion schemes (BB3) to judicial prosecution if they assist in facilitating IFFs. It increases the consequences for perpetrators by excluding companies convicted of tax evasion, bribery or money laundering from accessing financial services, particularly project and trade finance, insurance coverage and equity investment (CC2). The focus placed on reining in those facilitating and benefiting from IFFs is combined with measures to reduce opportunities to legalise IFFs, in particular by the preference for an interlinked system of national and international beneficial ownership registries (BB5, BB6, CB1) and the introduction of unitary taxation (CA1).
24Under the compliance view, such reforms should not be sidetracked by technocratic measures such as more effective container management (AA1, AA2), traceability of goods through blockchain technology (AA3), or geo-forensic methods to establish the provenience of high-value commodities such as gold (AC9). These measures attempt to change how trade operations work but are seen as short-term technical fixes and are the least preferable policy options from the compliance perspective.
25While the topic of transparency receives less attention than does legal compliance, the position taken under Factor 1 mirrors the same preference: only targeted transparency measures are welcome, while pure statistical transparency is seen as less helpful. In concrete terms, providing more granular commodity-by-commodity merchant data in customs statistics (AC6, AC10), publishing the number of accounts and amounts in offshore wealth centres (AC2) or adding statistical keys to identify related-party transactions (AC7) are seen significantly less favourable than taking measures that give tax authorities in low-income countries access to information that allows for direct and targeted follow-up and investigation. The latter include mandatory publication of country-by-country tax reports and critical inter-company tax arrangements (BB7, BB8, BB4) and unilateral information exchange by financial centres and trading hubs (BB10), but the protection of whistle-blowers (BB13) also belongs to this cluster of measures.
26Noteworthy and somewhat inconsistent with the tenets of this perspective is the negative rating of the proposal that capacity-constrained jurisdictions should be allowed to use country-by-country data for tax adjustment purposes (BB9). The scepticism with regard to directly empowering host country tax administrations is, however, not unique. Other policy proposals that would embolden host country tax administrations—including the use of reference prices to determine the tax value of commodity sales (BA1), limits to deductible taxpayer cost in intra-group transactions (BA2), and the introduction of profit allocation rules for subsidiaries in the form of, for example, minimum profit margins or fixed margins (BA3)—are also met with reservations.
27In sum, the compliance view favours mid- to long-term reforms over short-term fixes and puts obligations on host countries but is less consequential on proposals that would empower host countries to expand and use their sovereign policy space.
28The opposing perspective to the compliance view—represented in Factor 1b—shows the opposite preferences (see Figure 6.5). It displays reliance on all those technological measures that are disregarded by the compliance view (e.g. AA2, AB2, AC6, AC9) and shows the lowest rankings for setting up beneficial ownership registries and extending legal liability to service providers and other actors, with one notable exception: this view strongly supports the proposal that companies convicted of financial wrongdoing should be excluded from financial services (CC2), which is assumed to work as a powerful legal threat and deterrent.
Figure 6.4 Composite Q-sort for Factor 1a.
Source: the authors.
Figure 6.5 Composite Q-sort for Factor 1b (opposing view to Factor 1a).
Source: the authors.
29Factor 2 presents a different mix of proposals that it envisages as most effective in reducing IFFs. We call it the enigma of empowerment view because it takes an indecisive and contradictory stance on the extent to which host countries should be empowered to curb illicit financial outflows (see Figure 6.6). Three elements stand out: First, the enigma of empowerment view shows equally strong support for extending legal liability to enablers and marketers of schemes that facilitate IFFs (BB1-3) as the compliance view, but without the explicit support for beneficial ownership registries. Rather, extending legal liability is complemented by a second set of policy measures that proposes strengthening tax authorities in capacity-constrained jurisdictions, directly empowering local authorities. It requests the parent company of an MNE send the Country-by-country report to its subsidiaries for local filing (BB8), a measure that provides relevant information to tax authorities in host countries. Yet the empowerment element goes a step further, suggesting that the local tax authorities should be allowed to use country-by-country data for tax adjustment purposes (BB9, the proposal that also has the highest composite ranking under Factor 2). These medium-term reform proposals are complemented by a short-term policy measure consistent with the idea of empowering local tax authorities that can be realised within the existing regulatory set-up. It suggests requiring data matching and improving the flow of information between customs authorities, tax authorities and banks within and between countries (AA4). Different documents—including sales and purchase documents, customs declarations, and commercial, payment and transport documents—shall be cross-matched to spot discrepancies that may point to misinvoicing in specific trade transactions.
30The policy proposals rejected partly contradict the profile of the empowerment view that started to emerge from the most favoured policy proposals. Foremost, the policy option that producer countries might unilaterally introduce profit allocation rules for subsidiaries (e.g. minimum profit margins or fixed margins) in the medium term is firmly rejected (BA3). Such a proposal would significantly empower host countries’ taxing capacity but would challenge the basic tenets of the current transfer pricing doctrine. Further, even modest measures to empower low-income tax authorities are firmly rejected, including the idea that trading and financial hubs shall supply information—automatically or on request—to low-income countries on a non-reciprocal basis (AC3) and the proposal that trading and financial hubs shall significantly expand the coaching of low-income countries to request and interpret tax information (AC4). It is important to note that the enigma of empowerment view also rejects extending incentives to multinational firms by making taxes paid at source in low-income countries deductible against the tax due in residence countries (BB15).
31It is its internal contradictions that led us to label Factor 2 the enigma of empowerment. It brings to the fore a tension that is rarely made explicit in discussions about global tax policy and the question of tackling IFFs: To what extent should the capacity, autonomy and agency of host countries, mainly low-income countries, be strengthened with regard to establishing and enforcing rules, and to what extent should home countries, those benefiting from the existing rules, be tasked with curbing IFFs? While the official discourse, as exemplified through the Inclusive Framework at the OECD, emphasises cooperation and joint responsibility (OECD, 2017a), research finds a reluctance to move towards an equal partnership and negotiation at eye level (Brugger and Engebretsen, 2020; Christians and van Apeldoorn, 2018; Hearson, Christensen and Randriamanalina, 2023). The enigma of empowerment view nicely illustrates the unease in policy circles with regard to the role some are willing to concede to host countries.
32The (partially) opposing factor F2b is presented in Figure 6.7. It provides consistent rejection of the proposals favoured by the enigma of empowerment view, particularly rejection of extending liability to service providers (BB1-3) and allowing the use of country-by-country reports for tax adjustment purposes (BB9). The most preferred policy option of the opposing stance to the enigma of empowerment is sstrong support for making taxes paid in low-income countries deductible in home countries (BB15). The other provisions show a less consistent picture, combining targeted transparency measures like international beneficial ownership registries (BB6) with improving the flow of information and data matching between customs authorities, tax authorities and banks, within and between countries (AC4) with the obligation for offshore centres to assist other countries in recovering undeclared wealth through withholding taxes on interest and dividend payments to non-residents (BB11). Empowerment-focused proposals, such as producer countries limiting deductible taxpayer costs in intra-group transactions (AC4), receive second-order support only.
Figure 6.6 Composite Q-sort for Factor 2a.
Source: the authors.
Figure 6.7 Composite Q-sort for Factor 2b (opposing view on Factor 2a).
Source: the authors.
33The most supported policy proposals under the third perspective represented by Factor 3 pivot around the idea of targeted disclosure (Figure 6.8). This involves proposals that enhance the exchange of information and data among customs authorities, tax agencies and financial institutions within and across nations (AA4), extend payment-to-government disclosure beyond extractive industries to explicitly include disaggregated information on trading (BB12), and introduce beneficial ownership registries at the international and national levels (BB5, BB6).
34These measures are complemented by the mandatory requirement for legal professionals, accountants and service providers to report suspicious transactions to the anti-money laundering (AML) authority (BB2) and the legal protection of whistle-blowers in offshore centres. Yet support for targeted disclosure is not unrestricted, as other proposals that would advance targeted disclosure are among those rejected, including the option that financial centres and trading hubs shall legislate unilateral information exchange instead of existing reciprocal standards (BB10).
35The least supported policy proposals include, on the one hand (and similar to the compliance view), short-term measures targeting the day-to-day operations of trade, such as improved container management (AA1, AA2), the use of blockchain for traceability, and publishing more granular merchant trade data on a commodity-by-commodity basis instead of publishing net receipts only (AC10). On the other hand, the targeted disclosure perspective firmly rejects sanctioning companies by excluding those convicted of tax evasion, bribery or money laundering from finance, insurance coverage and equity investment (CC2). It also rejects policies that strengthen host countries’ means of limiting outflows of IFFs, including, for example, the policy proposal that producer countries shall introduce profit allocation rules for subsidiaries, such as minimum profit margins or fixed margins (BA3), and the option that producer countries shall legislate the use of reference prices to determine the tax value of commodity sales (BA1).
36In sum, the targeted disclosure view is more reticent than the first two perspectives to support policy options that involve more profound shifts, such as extending legal liability or empowering host countries. Rather, the targeted disclosure view assumes that the most effective way of curbing IFFs is a combination of short- and medium-term reforms that gradually strengthen the existing system of safeguards and expand access to relevant information. Focusing on targeted disclosure and supporting advanced transparency measures that are contested, this view is more ambitious than calling for statistical transparency only, which finds much broader acceptance.
Figure 6.8 Composite Q-sort for Factor 3.
Source: the authors.
37The analysis of subjective policy preferences shows a broad range of views on how best to curb trade-related IFFs. Comparing the compliance, the enigma of empowerment, and the targeted transparency views, at least three overarching tensions can be identified that ripple through the policy discourse on IFFs. The first tension arises between support for short-term measures versus medium- or longer-term measures. It seems that these are not seen as consecutive steps nor as a virtuous cycle. Rather, they represent different levels of ambition, and some stakeholders fear that focusing too much on short-term fixes could undermine the willingness to engage in more substantial reform. For example, the explicit rejection of most short-term measures in the compliance view (and vice versa in the polarised complement) also signals a certain level of ambition, and that focusing on technical fixes in the short term might prevent more ambitious reforms (see Figure 6.8). While this tension between short-term and long-term reforms is clearly visible in all perspectives, the majority of highly ranked policy options focus on medium-term reforms. There seems to be an agreement that short-term fixes will not be enough to address the problem, and that more needs to be done, but within limits: profound long-term reforms find even less support.
38The second tension is over the role that legal obligations should play for those involved in facilitating IFFs and what the legal consequences should be for companies convicted of financial misconduct—so, tax evasion, bribery, or money laundering. The softest form of extending legal responsibility—that is, requiring legal professionals, accountants and service providers to report suspicious transactions to the AML authority—is the only proposal that seems acceptable more broadly, but still not unanimously.
39The third tension is about whether, and if so, how much, autonomy in rule-making should be conceded to host countries, which are often capacity-constrained low-income countries. While policy proposals in this regard appear prominently in some views, they do not receive cross-cutting support. Rather, a robust majority prefers home countries to drive reforms (see Figure 6.8 and 6.9). Whether this is because of home countries having a moral responsibility to rein in their MNEs or out of a desire to keep control over the reform process (and outcome), or a combination thereof, is beyond the scope of this study. The tension over the role and involvement of host countries in defining tax policies has grown recently, with some host countries increasingly resisting following the OECD-brokered consensus. In November 2023, the UN General Assembly adopted a resolution requiring the United Nations to start an intergovernmental process on international tax cooperation (UN, 2023). The resolution is intended to pave the way for a UN convention on taxation and a new global tax body that challenges the role of the OECD, which has dominated international tax policy for the last six decades. This development illustrates that the actors seeking to empower host countries are gaining momentum at the global level.
Figure 6.9 Policy options with significant support for each perspective: the five (eight) highest-ranked statements.
|
|
Implementation time frame
|
|
|
Short term
|
Medium term
|
Long term
|
Implementing jurisdiction
|
Cross-border / global
|
AA1-4
Compliance 0 (0) / 1 (1)
Empowerment 1 (1) / 0 (0)
Transparency 1 (1)
|
--
|
CA1-CC2
Compliance 1 (3) / 1 (1)
Empowerment 0 (1) / 0 (0)
Transparency 0 (0)
|
Host country
|
AB1-3
Compliance 0 (0) / 1 (2)
Empowerment 0 (0) / 1 (1)
Transparency 1 (1)
|
BA1-3
Compliance 0 (0) / 0 (0)
Empowerment 0 (0) / 0 (1)
Transparency 0 (0)
|
--
|
Home country
|
AC1-10
Compliance 0 (1) / 2 (2)
Empowerment 0 (0) / 1 (2)
Transparency 0 (0)
|
BB1-16
Compliance 4 (4) / 0 (2)
Empowerment 4 (6) / 3 (4)
Transparency 3 (6)
|
--
|
Note: The second pair of figures in the compliance and empowerment views refers to the bipolar complement.
Source: the authors.
40The different stances in the three areas of tension are firm, and there is little convergence recognisable. The policy proposals that are potentially agreeable across positions are those that can be summarised under the label of targeted transparency. The targeted transparency view is indicative of this. The perspective has no polar opposite view, and the targeted transparency policy measures also find significant support in the other perspectives. We call this set of transparency policies targeted transparency because these measures release information that is directly relevant and actionable to tax authorities in host countries. Targeted transparency measures include policies such as improving the flow of information and data matching between customs authorities, tax authorities and banks, within and between countries, but also unilateral information exchange, whereby home countries and financial centres adopt legal provisions that allow information to be provided to host countries without requiring a bilateral exchange treaty.
41Research into the promotion of social accountability shows that greater access to information alone is not enough to trigger change (Fung, Graham and Weil, 2007; Kosack and Fung, 2014; McGee and Gaventa, 2010). The assumption that information about misdeeds, such as corrupt practices or embezzlement, would create a public outcry and put pressure on those in charge to improve the situation has proven unreliable. In the extractive sector, this has been observed for data released to the public through the Extractive Industries Transparency Initiative (EITI) (Lujala, Brunnschweiler and Edjekumhene, 2020; Öge, 2016; Rathinam et al., 2019; Sovacool, 2020). Transparency measures that are targeted and actionable prove to be more effective (Fung, Graham and Weil, 2007; Kosack and Fung, 2014; McGee and Gaventa, 2010). In that perspective, the results from the Q-study are interesting as participants support targeted transparency measures that are directly actionable by tax administrations and do not deem additional generic information, such as more detailed statistics, useful; nor do they see this as the start of a virtuous cycle. While many targeted transparency proposals find broad support across perspectives, this also has its limits. For example, the unilateral information exchange proposal, which goes further than most other measures, does not find broad support.
42Our study has several limitations. First, Q-studies provide a general overview of relevant standpoints on a subject, revealing shared and diverging perspectives among the stakeholder representatives included. By design, Q-methodology does not aim to produce generalisable findings for an entire (stakeholder) population as in a representative survey. Hence, while being highly relevant for the overall discourse on policy options for curbing IFFs, the identified perspectives and areas of consensus (or dissent) do not necessarily extend to all stakeholder representatives at all levels. Second, caveats to factor interpretation apply. Since not all Q-statements receive high Q-scores, the identified perspectives do not cover all policy proposals included in the Q-sample. Low Q-scores do not necessarily indicate that stakeholders do not attach importance to an issue, as low-salience rankings by individual sorters may primarily indicate a stronger preference for other policies. Another potential caveat concerns the level of expertise of study participants. Given the highly diverse set of policy options, to some extent participants’ ranking choices also reflect their degree of professional understanding of the presumed pros and cons of specific policy instruments. Low or neutral rankings may indicate a limited understanding of a given policy instrument rather than rejection or indifference, as partially indicated by qualitative comments. Given the complexity of the topic, this is, however, reflective of a general characteristic of the wider policy debate on IFFs. In real-life policy debates, policymakers have imperfect information on some or all options, which puts this limitation of our study into perspective.
43Further, when drawing on the study findings, it must be kept in mind that all participants (i.e. their Q-sorts) receive the same weight in factor construction and interpretation. However, not all study participants (and stakeholders) necessarily carry the same weight in policy processes. Finally, as discussed earlier, our selection of study participants attempted to include the most diverse range of positions we could identify in the policy debate. However, all participants are from home countries in the global north, and there is no direct participation of experts from host countries. This limitation is only partially compensated by including experts from organisations with the explicit mission to promote/strengthen the role of low-income countries and represent their interests in policymaking in the global North. This limitation also pertains to the fact that in commodity trade, there are transit countries through which commodities pass. Currently, they are not treated separately.
44The application of Q-methodology to public policy is still in its infancy. Our study demonstrates the usefulness of Q-methodology for understanding policy preferences and informing policy processes. At the research level, our study advances the analysis of complex policy processes via a standardised semi-quantitative methodology suited for expert and small-n groups. Further research is needed to expand the study beyond participants from home countries to include stakeholders from host countries, increase the representation of policymakers and government officials, and consider including transit countries as a separate category. Future research could also dig deeper into the positions within stakeholder groups. Other relevant stakeholder groups could include the trade policy community, the financial services community, and the tax community, among others. Finally, future research could also take an interdisciplinary approach and explore the cross-fertilisation of Q-methodology with research into the agency of different stakeholders.
45The results of this research reveal that the distributive struggle that lies beneath debates over policy measures is not only about the distribution of financial gains but as much about the distribution of agency and decision-making power: who should be reined in and who should be empowered to lead anti-IFF policy measures; the outcome might well set a precedence for other policy areas. Bringing these more fundamental questions to the surface and into the policy discourse is important if we are not to lose sight of the bigger picture and the question of what the architecture of cooperation in tax matters and the combatting of IFFs shall look like.
46The IFF and tax policy debates are unlikely to become less contentious, even less so with the upcoming negotiations on a UN tax convention. A tax policy observatory conducting Q-studies at regular intervals could be a cost-effective means to provide near real-time information on ongoing policy processes that are highly dynamic. In light of the growing pressure to curb IFFs, the contestation over policy solutions combined with increasing debate over the representation of host countries and the appropriate forum for negotiation, such an observatory might be a useful academic contribution to inform policy processes.
Figure 6.A-1 Factors F1, F2 and F3 together explain 43% of the study variance (18%, 14% and 11%).
Source: the authors.
Figure 6.A-2 Distribution of participants on factors, with factor loadings.
Source: the authors.
Figure 6.A-3 Descriptive statistics on factors and factor relations.
Source: the authors.
|