Jump to ContentJump to Main Navigation
Estimating Illicit Financial FlowsA Critical Guide to the Data, Methodologies, and Findings$

Alex Cobham and Petr Janský

Print publication date: 2020

Print ISBN-13: 9780198854418

Published to Oxford Scholarship Online: March 2020

DOI: 10.1093/oso/9780198854418.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 21 October 2021

New Proposals for IFF Indicators in the Sustainable Development Goals

New Proposals for IFF Indicators in the Sustainable Development Goals

Chapter:
(p.144) 6 New Proposals for IFF Indicators in the Sustainable Development Goals
Source:
Estimating Illicit Financial Flows
Author(s):

Alex Cobham

Petr Janský

Publisher:
Oxford University Press
DOI:10.1093/oso/9780198854418.003.0007

Abstract and Keywords

The search for indicators for the UN target to reduce illicit financial flows (IFF) is complicated by data issues and the difficulties of estimating what is deliberately hidden. In this chapter, we propose our two preferred indicators, that address instead the measurable consequences of IFF. In this chapter we lay out the two measures, and evaluate their strengths and weaknesses including in respect of data availability and potential work-arounds at national level. Both measures use relatively newly available data to establish scale: in the first case, the scale of the misalignment between where multinational companies carry out their economic activity, and where they are ultimately able to declare the resulting profit; and in the second case, the scale of offshore wealth which is undeclared to tax authorities.

Keywords:   Illicit financial flows, IFFs, offshore wealth, profit shifting, tax evasion, tax avoidance, SDGs

As discussed in chapter 1, the illicit flows target in the Sustainable Development Goals (16.4) leaves much to be desired. Politically, it made sense to maintain the overall framing that the Mbeki panel had set. But technically, the combination of the quite different issues under the IFF umbrella creates a challenge. That challenge is exacerbated by the proposed, single IFF indicator:

16.4.1 Total value of inward and outward illicit financial flows (in current USD)

The various types and channels of IFF have different implications in terms of tax revenue loss, damage to governance and political representation, and effective market functioning. They also create different patterns of inflow and outflow from a given jurisdiction. Summing across every type, and treating inflows and outflows as equivalent, will provide at best a blunt measure of the scale of the underlying phenomena.

The target is, however, politically significant. It reflects political leadership from the global South, and the African region in particular. It would be unfortunate if the UN system were to fail to deliver meaningful progress—beginning with agreed indicators.

Looking ahead, the target is of substantive importance. While there is uncertainty associated with many individual estimates, there are literally no results for the scale of any type or channel of illicit financial flow which support the view that the phenomena should not be considered a sizeable obstacle to development. The international nature of IFF requires, by definition, international coordination as well as national action; the SDG target represents the best opportunity to drive progress.

If IFF were readily separable by type and/or by channel, it would be possible to create sub-indicators that summed to the proposed indicator 16.4.1. For example, given separability by IFF type and robust estimates of each, 16.4.1 could be the sum of 16.4.1.1: market-abusive IFF; 16.4.1.2: tax-abusive IFF; 16.4.1.3: abuse of power IFF; and 16.4.1.4: IFF due to laundering of the (p.145) proceeds of crime. An equivalent but more complex structure could be envisaged on the basis of the IFF channels in table 1.1.

In theory, either of these would offer not only a comprehensive indicator for monitoring and accountability purposes, but also a disaggregable basis for policy prioritisation at any and all of the global, regional and national levels. In practice, however, as we have seen in the foregoing chapters, current estimates provide neither the robustness nor the separability necessary to support such an approach. The one broad IFF approach that could be relatively straightforwardly generated on a global basis is that of Global Financial Integrity. But neither the capital account component, nor the trade component of this approach, appear robust; and nor, then, are any conclusions about the relative importance of trade-based IFF.

The risk approaches outlined in chapter 5 do provide a more granular basis for policy prioritisation, and sidestep the problems of estimates by relying instead on proxy measures that can be directly calculated. They do not, however, offer indicators of scale of the type required by SDG 16.4.1.

To meet the constraints of the SDG process, we have therefore proposed two measures—measures, rather than estimates—that capture the scale of two key outcomes of IFF. These reflect, first, the volume of profit shifting; and second, the extent of undeclared offshore assets (Cobham & Janský, 2017d, 2018b). These form part of basis for national pilots in the process led by UNCTAD and the UN Economic Commission for Africa to develop and test SDG indicators of tax-related illicit flows (UNODC and the UN Economic Commission for Latin America and the Caribbean are leading work on illegal market IFFs).

Rather than inevitably imprecise estimates of the scale of illicit flows, our two preferred indicators (or indicator components) address instead the measurable consequences. In this chapter we lay out the two measures, evaluate their strengths and weaknesses including in respect of data availability and potential work-arounds at national level, and consider the scope to combine the measures, if necessary, into a single indicator 16.4.1.

6.1. Profit Shifting: SDG 16.4.1a

6.1.1. Overview

‘Another governance dimension of IFFs relates to the unequal burden of citizenship imposed on other sectors of society, both in terms of tax fairness and (p.146) ‘free-riding’. When large companies, particularly multinational corporations, engage in base erosion and profit-shifting activities, the bulk of the tax burden as a result falls on small and medium-scale enterprises and individual taxpayers. This runs counter to the idea of progressive taxation, in which those who earn more income contribute a larger percentage of tax revenues. Just as pernicious to governance is the ‘free-riding’ that results when entities evade or avoid taxes where they undertake substantial economic activities and yet benefit from the physical and social infrastructure, most of which is still provided by the public sector in Africa.’

  • African Union/Economic Commission for Africa, 2015, Report of the High Level Panel on Illicit Financial Flows from Africa, (p.52).

‘We will make sure that all companies, including multinationals, pay taxes to the Governments of countries where economic activity occurs and value is created, in accordance with national and international laws and policies.’

  • United Nations, 2015, Addis Ababa Action Agenda of the Third International Conference on Financing for Development (p.12).

‘The G20 finance ministers called on the OECD to develop an action plan to address specific BEPS issues in a coordinated and comprehensive manner. Specifically, this Action Plan should provide countries with domestic and international instruments that will better align rights to tax with economic activity.’

  • OECD, 2013a, Action Plan on Base Erosion and Profit Shifting (p.11).

‘Developed countries… have special responsibilities in ensuring that there can be no safe haven for illicit capital and the proceeds of corruption, and that multinational companies pay taxes fairly in the countries in which they operate.’

  • United Nations, 2013, Report of the High Level Panel of Eminent Persons on the Post-2015 Development Agenda (p.11).

As seen in chapter 4, estimates of multinational companies’ profit shifting may be the robust of all the IFF areas. The revenue losses associated with these tax abuses may very well be the largest. But for a consistent basis of (p.147) annual monitoring and accountability, with near-global coverage, these estimates too remain lacking.

For the SDG indicator, we therefore propose an approach which differs from most of the literature. Rather than a necessarily imperfect estimate of the profit shifting flow, we construct a more precise measure of a somewhat broader phenomenon: the ultimately achieved misalignment of profits with the underlying real economic activity.

As the quotations illustrate, reduction of this misalignment is the now well established and unique aim of international attempts to combat these abuses. As discussed in chapter 1, the exact nature of profit misalignment means that this measure will necessarily include a degree of licit activity. Recall that profit shifting is made up of lawful and unlawful avoidance, along with criminal evasion. Profit misalignment—the phenomenon that can be measured, rather than estimated—is a broader term, including these three elements but also misalignment that may arise simply from the fact that national and international tax rules do not explicitly seek alignment.

The result is that, since some divergence from full alignment might therefore be expected even in the absence of tax-motivated shifting, the value of the indicator consistent with IFF elimination need not be zero. There is, however, no reason to expect any systematic change over time in the overall, global degree of non-tax-motivated misalignment. There is also no reason to expect that individual jurisdictions would experience particular swings in non-tax-motivated misalignment over time. On this basis, we favour tracking a relatively precise measure which includes some noise, in preference to more uncertain and imprecise estimates of a more closely defined phenomenon.

6.1.2. Data

To construct a global measure of profit misalignment requires data on profits declared and real economic activity at the country level, plus ideally data on tax paid to understand the likely motivation.

Now for a company operating in a single jurisdiction, as was the case for all companies at the time when corporate law and accounting norms began to emerge, most of this information will be contained in the annual accounts. Those annual accounts in many jurisdictions, have long been required to be placed in the public domain. This reflects a crucial decision in the development of entrepreneurship, by which governments allowed the liability of those running companies to be capped—so that commercial activity was not (p.148) held back, for example, by the risk that business failure would also mean the loss of one’s family home. While having sporadic use across millennia, it was only from the early 19th century that limited liability companies were the subject of formal legislation followed by widespread use.

The effective quid pro quo for this protection was the publication of company accounts, signed off by an approved auditor. Where limited liability socialises (some of) the private risks of business failure, the publication of audited accounts provides transparency to allow external stakeholders and investors to manage their own exposure to those risks.

In the 20th century, the growing emergence of business groups operating transnationally necessitated major changes to national regulatory frameworks that had hitherto been purely domestically focused. Most obviously, this process saw the League of Nations take leading role in establishing the basis for international tax rules that first governed the imperial interactions in the multinational tax sphere, and were later taken up by the OECD.

Perhaps unsurprisingly, compared to tax, there was less pressure to ensure transparency regulations were adapted for the globalising world. With most multinationals headquartered in and owned from current or former imperial powers, these OECD country governments were largely able to ensure domestic regulatory compliance and to access any data they required to ensure appropriate tax was paid—in their own jurisdictions.

And so it fell, eventually to the G77 group of countries to force the question of greater corporate disclosure (Ruffing & Hamdani, 2015; Meinzer & Trautvetter, 2018). Following growing anger at the apparent impunity of multinationals operating in lower-income countries, and after lengthy negotiations at the United Nations, the Center for Transnational Corporations (UNCTC) was established in 1975. The Center in turn convened a Group of Experts on International Standards of Accounting and Reporting (GEISAR), to increase the financial transparency of multinationals and their global networks, including proposals for publication of the accounts of each entity in each country of operation.

While the work of GEISAR was eventually blocked through the mobilisation of business lobbyists, major accounting firms and OECD member states, and the UNCTC shut down in 1992, the issues remained unaddressed. The International Accounting Standards Board in London, and in the US the Federal Accounting Standards Board, allowed during some periods for geographic segment reporting—but typically this did not break out more than a handful of individual countries of operation, if that, and left the rest aggregated by broad region.

(p.149) Then in the early 2000s, the lack of jurisdiction-level reporting by multinationals become the subject of discussions among a small expert group in what would soon become the Tax Justice Network. And so it was that some months before the network was formally established, the first ever draft accounting standard for country-by-country reporting was published (Murphy, 2003). This set out the basis for public data to ensure that multinationals, too, would provide effective disclosure about their activities and risks at the jurisdiction level. Although swiftly taken up by civil society transparency advocates, initially focusing on the extractive sector and subsequently looking at tax avoidance more broadly, the proposals were consistently resisted at the International Accounting Standards Board and at the OECD.

In just ten years, however, the powerful G20 group of countries had required the OECD to put aside any misgivings and deliver a standard for country-by-country reporting, to apply to all multinationals in the world over a certain size threshold. The eventual standard in most technical respects hewed closely to the proposals developed by the Tax Justice Network, but with one, crucial difference: the OECD data was not to be made public, but provided privately only to home country tax authorities.

This limitation, despite the complex information sharing arrangements constructed since, largely defeats the purpose of the proposal. Above all, it prevents any public scrutiny by stakeholders including investors, labour, and people in the communities where companies’ activities take place. But on top of that, the OECD approach manages to take a measure designed to level the playing field between the tax authorities of lower- and higher-income jurisdictions, and instead to exacerbate the inequalities faced. By construction, the information sharing arrangements result in systematically worse access to information in those (low- and middle-income) countries that suffer the most intense revenue losses (Knobel & Cobham, 2016).

Aside from the question of access, the data itself is not perfect. Table 6.1 compares the OECD standard with civil society proposals, and a range of other existing requirements: CRD IV (limited country-by-country reporting for EU financial institutions, under the fourth Capital Requirements Directive); the now-repealed Dodd-Frank requirement for US-listed extractive sector firms; the Canadian and EU equivalents; and the standard of the Extractive Industries Transparency Initiative (EITI). Indeed, EITI and the extractive industries sector has been discussed recently in comparison with other reporting requirements for extractive industries (Porsch et al., 2018) and within a broader discourse on illicit financial flows in extractive industries (Lemaître, 2018). While most of the key variables are included in the OECD (p.150) (p.152) standard, there are important absences: notably, of economic activity indicators (employee remuneration and tangible assets), and of intra-group transactions including interest and royalties.

Table 6.1. Comparison of data fields in CBCR standards

Civil Society Proposal

OECD CBCR

CRD IV

Dodd Frank

Canada

EITI

EU

Identity

Group name

Group name

Group name

Group name

Payee name

Payee name

Group name

Countries

Countries

Countries

Countries

Countries

Legal and institutional framework

Countries

Nature of activities

Nature of activities

Nature of activities

Projects (as in: by contract)

Same data required per project as well as per country

Allocation of contracts and licenses

Projects (as in: by contract)

Names of constituent companies

Names of constituent companies

Receiving body in government

Subsidiaries if qualifying reporting entities

Exploration and production

Activity

Third party sales

Third party sales

Social and economic spending

Turnover

By the process of addition

Turnover

Number of employees FTE

Number of employees FTE

Number of employees

Total employee pay

Tangible assets

Intra-group transactions

Intra-group sales

Intra-group sales

Intra-group purchases

Intra-group royalties rec’d

(p.151) Intra-group royalties paid

Intra-group interest rec’d

Intra-group interest paid

Key financials

Profit or loss before tax

Profit or loss before tax

Profit or loss before tax

Payments to/from governments

Tax accrued

Tax accrued

Tax paid

Tax paid

Tax paid

Income taxes paid

Tax paid

Profits taxes

taxes levied on the income, production or profits of companies

Any public subsidies received

Any public subsidies received

Source: Cobham, Gray, & Murphy (2017)

In addition, the OECD standard has faced criticism over the failure to require that the data be reconciled with the published, global consolidated accounts of multinationals; and that it currently applies only to the largest 10−15 per cent of multinationals, those with a turnover above three quarters of a billion euro. The Global Reporting Initiative (GRI) has now launched on a technical standard for voluntary public reporting which addresses this question and a number of others, and sets the basis for future improvements to the OECD standard also. Vodafone, a participant in the GRI working group, have also become the first major multinational to commit to publish their OECD standard reporting, from 2019.

Notwithstanding its weaknesses, the OECD standard marks a turning point in the debate. Previous arguments that such data was not held by companies, or would be prohibitively expensive to collate, have been eliminated. This in turn means that the opportunity is there for the data to be made available.

The OECD BEPS Action 11 team, tasked with generating consistent measures to track progress in reducing base erosion and profit shifting by multinationals, soon recognized that existing estimates cannot provide such measures—and, moreover, that this can only be done with country-by-country reporting data. But BEPS Action 13, which includes the responsibility to introduce country-by-country reporting, had already been the subject of energetic lobbying. Major multinational lobby groups, the big four professional services firms and certain OECD member states (Meinzer & Trautvetter, 2018) were successful in having country-by-country reporting data designated as confidential, with removal of access for states that violate this.

For this reason, it took until the OECD Secretary-General’s report to G20 finance ministers of July 2018—some five years after the start of the BEPS project and three years after its formal end—for the BEPS 11 team to obtain agreement that it would collate and publish the partially aggregated country-by-country reporting received by each relevant home jurisdiction tax authority. Even then, the commitment is only to publish data by end-2019. But this is potentially a great step forward delivering not transparency of individual multinationals, but the ultimately more important accountability of individual jurisdictions for their role in promoting and/or tackling profit shifting.

The difficulty, of course is that BEPS Action 11 can only publish the data that states provide—and this will include at least some regional aggregation to (p.153) accommodate confidentiality concerns. It remains to be seen to what extent any further aggregation is imposed, and to what extent it limits the value of the data for the proposal made here.

Consider a country with few multinationals above the reporting size threshold (annual turnover of $750 million), such as the Czech Republic. Here, tax authorities may find that data aggregated to the country level (e.g. the total employment of all reporting Czech multinationals in France) could threaten confidentiality.

For a major headquarters jurisdiction like the US, with a large number of reporting multinationals, there may be no obstacle. But with a range of (e.g. lower-income) countries in which only a smaller number of US-headquartered, reporting multinationals operate, the problem may still arise. In fact, the current Bureau of Economic Analysis survey of all US multinationals has many data suppressions for this reason (Cobham & Janský, 2019).

Before that data is scheduled to be available in early 2020, three other channels are under exploration. One is the voluntary route. There are potential champions here—Vodafone, for example, has committed to publish its OECD standard reporting from 2019, and its fellow members of the ‘B Team’ alliance have indicated some interest. The Global Reporting Initiative’s new standard is likely to see broad take-up from 2020. But voluntary approaches are difficult, since the data will inevitably focus attention on the absolute levels of a given multinational’s profit misalignment—rather than any relative superiority to less transparent rivals.

A second channel is that of unilateral requirement for publication. The UK parliament has already legislated to allow publication, but the government has not yet chosen to impose the requirement. The French parliament had passed a measure mandating publication, before the previous government reversed this with an archaic, technical manoeuvre. In the absence of multilateral agreement at the OECD, pressure will continue for others such as the EU to take a lead—albeit that Germany has now emerged as the key blocker.

The third channel is for the issue of corporate disclosure to return to the UN system (on which see Cobham, Janský, & Meinzer, 2018). One possibility here would be for ISAR, the successor to GEISAR, to develop a mandatory public standard. Another would be for the requirement to be embedded within the draft treaty on multinationals and human rights. Perhaps the most obvious channel, however, given the organisation’s central role in analysing data on the investment (and more recently, profit shifting) behaviour of multinationals, would be for UNCTAD to become the repository for (p.154) country-by-country reporting data, and the guardian of a strong standard delivering the data to underpin an indicator of profit misalignment for the SDGs.

Only this latter channel, of course, has the potential to generate consistent data of sufficient coverage to support the proposed SDG indicator, should the OECD release prove insufficient. An additional channel, for countries involved in the initial pilot studies at least, is to construct nationally-relevant equivalent measures by requiring local filing of OECD standard data, and/or by creating work-around approaches using nationally filed tax returns and additional data (public and privately filed) on the global operations of each multinational. Figure 6.1 compares the current OECD arrangements to access the data, with an alternative proposal put forward by the Tax Justice Network.

New Proposals for IFF Indicators in the Sustainable Development Goals

Figure 6.1. Approaches to sharing country-by-country (CbC) data.

Source: Knobel & Cobham (2016).

For the moment, we may optimistically assume that states providing data to the OECD as now agreed will not spuriously use confidentiality arguments to downgrade the quality of data supplied. Were this the case, however, it is possible that a delegated UN body such as UNCTAD could—in tandem with the OECD or quite separately—obtain additional data direct from member states’ tax authorities, with the guarantee of protecting confidentiality of individual reporting multinationals.

Even in a more pessimistic view, however, the overall transparency provided is likely to far exceed any current public information about the activities and profit declaration of multinationals. Unnecessary aggregation aside, the quality of the data required is high. Although not currently required to be audited and consolidated to global accounts, the basis in reporting by individual multinationals to their home tax authorities sets the likely standard well above any current alternative—from the limited, publicly available company balance sheet data, to elements of national accounts data or bilateral aggregates. Misreporting faces potentially criminal consequences.

In terms of coverage, the capturing of multinationals above the threshold is expected to be complete—and therefore global in terms of their operations.

Finally, in the discussion of data, we note that reliance on country-by-country reporting would reflect a central policy position of the High Level Panel on Illicit Financial Flows from Africa (AU/UNECA, 2015), whose work underpins the global agreement on SDG 16.4.

“We were encouraged by the emergence of discussions on country-by-country reporting of employees, profits, sales and taxes as a means of ensuring transparency in cross-border transactions. Country-by-country reporting, publicly available, will help to show where substantial activity is taking place and the (p.155) (p.156) relative profits generated and taxes paid. In the absence of a universal tax administration, country-by-country reporting will enable tax and law enforcement agencies to gain a full picture of a company’s activities and encourage companies to be transparent in their dealings with African countries.”

(p.45)

“African States should require multinational corporations operating in their countries to provide the transfer pricing units with a comprehensive report showing their disaggregated financial reporting on a country-by-country or subsidiary-by-subsidiary basis. African governments could also consider developing a format for this reporting that would be acceptable to multiple African revenue authorities.”

(p.81)

“The Panel calls for partner countries to require publicly available disaggregated, country-by-country reporting of financial information for multinational companies incorporated, organized or regulated in their jurisdictions.”

(p.85)

These policy positions provide further confirmation, of course, that those responsible for this globally leading work saw multinational profit shifting as a major component under the IFF umbrella. At the same time, the approach proposed for SDG indicators allows full separation of the profit shifting component, to support the distinct policy approaches needed.

6.1.3. Methodology

The misaligned profit indicator is defined as the value of profits reported by multinationals in countries, for which there is no proportionate economic activity of MNEs. It is defined for each jurisdiction and it can be summed across some or all countries. For each jurisdiction we define the misaligned profit as:

χi=πiωiΠ
(1)

where:

πi is the value of all multinationals’ gross profits declared in jurisdiction i;

ωi is the share of all multinationals’ economic activity in jurisdiction i; and

Π is the global, gross profits of all multinationals, such that Π=i=1nπi.

We propose to capture economic activity as the simple average of single indicators of production (the share of full-time equivalent employees in a jurisdiction, ιi) and consumption (share of final sales within each jurisdiction, γi). We define, for all i:

(p.157)

ωi=12(ιi+γi)

It follows that the global sum of misaligned profits, X, is equal to zero:

Χ=i=1nχi=0

We propose that the profit misalignment indicator for use in SDG target 16.4 is the global sum of positively misaligned profits—that is, the total excess profits declared in jurisdictions with a greater share of profits than would be aligned with their share of economic activity. Equivalently, this can be calculated as half the sum of the absolute values of misaligned profit:

SDG16.4.1a=12i=1n|χi|
(2)

Note that the SDG indicator as defined in the current framework is expressed as the sum of inward and outward IFF, so the sum of absolute profit misalignment could be used; this seems inelegant at best. Note also that the underlying country-level misalignment measures provide monitoring and accountability for individual states seeking to reduce the (negative) misalignment suffered—for example, to demonstrate to citizens and domestic businesses that multinationals are being fairly taxed; and for states that benefit from profit-shifting at the expense of others, an accountability mechanism to demonstrate their own commitment to global progress.

Although not proposed as an element of the SDG framework, such a profit misalignment indicator can also be constructed at the firm level. Equation (2), calculated with the data from a single multinational group, will provide a measure of the value of profit that is misaligned. The χi for individual jurisdictions will show the group-specific pattern of misalignment, including which jurisdictions are suffering and benefiting.

In addition, a further indicator can be constructed to allow easier comparison across multinationals and of a given multinational over time. This is simply the level of the group’s global profit misalignment, per Equation (2), expressed not in currency terms but as a proportion of total group profits: in other words, a comparable measure of the intensity of profit misalignment at multinational group j:

μj=i=1n|χi|2Πj
(3)

(p.158) Note that, perhaps counter-intuitively, this intensity ratio can easily exceed 100 per cent. This is because the sum of positive profits declared at jurisdiction level (and so potentially able to be misaligned) will exceed the global profit total if there are also losses declared in some jurisdictions.

We can explore the methodology at the firm level by applying it to one of multinationals currently leading the way in publishing some form of country-by-country reporting. Vodafone, who for many years were subject to campaigning by tax justice activists in the UK over a questionable deal with the tax authority, have become the first multinational to commit to publish their OECD standard country-by-country reporting, from 2019. In the meantime, they publish data on a roughly equivalent basis (Vodafone, 2018) which we use here for illustrative purposes.

In Vodafone’s case, for 2016–17 data, the sum of positive, declared profits at the jurisdiction level is €4.128bn. This is 221 per cent of the overall, (net) global profit of €1.867bn. Applying the approach in equations (1) and (2) reveals misaligned profit of €3.574bn, or an intensity of profit misalignment, per equation (3), of 191 per cent of global profit.

Figure 6.2 shows the extent of misaligned profit and the effective tax rate paid, for the ten jurisdictions in which more than 1 per cent of the global profit is declared and where there is positive profit misalignment. Of the nearly €1.5bn declared in Luxembourg, more than 99.5 per cent is not aligned with the real economic activity taking place there (as captured by employment and sales). The effective tax rate on these profits, according to Vodafone’s data, is around 0.3 per cent.

New Proposals for IFF Indicators in the Sustainable Development Goals

Figure 6.2. Major jurisdictions in Vodafone’s profit misalignment.

Notes: Figure shows ten jurisdictions for which declared profit exceeds 1 per cent of Vodafone’s ultimate, global profit. The size of the bubbles is in proportion to the absolute value of misaligned profit (the largest, Luxembourg, is equivalent to €1.44bn). An eleventh jurisdiction with 1 per cent of Vodafone’s global profit, Romania, is excluded, since it sees negative profit misalignment (i.e. lower profit than its share of economic activity indicates).

Source: authors’ calculations, using Vodafone data for 2016−17. We are grateful to Tommaso Faccio for providing the data.

On the other hand, South Africa appears to benefit from a 33 per cent effective tax rate on declared profit of around €1.08bn, of which nearly 90 per cent is misaligned. Egypt, Kenya and at a lower effective rate, Italy, appear to follow a similar pattern; while Malta sees misalignment on an effective rate around 6 per cent. Not shown are the range of EU states and the US and Australia, where losses are declared in significant, long-established markets with major activity.

In a series of basic simulations, using constrained, randomly generated values for activity and profit in the jurisdictions of hypothetical multinational groups, we confirm that while misalignment intensities in excess of 100 per cent are not necessarily common, they are certainly plausible—especially, as noted, when a group records losses in a number of jurisdictions.

The same simulations confirm, as would be expected, that aggregation of the country-by-country reporting of multiple multinationals will tend to result in substantially lower, overall profit misalignment. For example, aggregating (p.159) across 10 simulated multinationals operating in the same 26 jurisdictions, with misalignment intensity varying between 55 per cent (44 per cent) and 135 per cent (129 per cent), we obtained an overall misalignment of 22 per cent (17 per cent) of global profits.

The limitations of these basic simulations aside, we would expect actual misalignment to be higher, however, if the patterns of positive misalignment are not in fact random—if, for example, multinationals in general are likely to use a particular set of jurisdictions for tax-motivated profit shifting. The methodology here is broadly comparable to the approach of Cobham & Janský (2019) surveyed in section 4.6, in which we use data on US multinationals and find a level of aggregate misalignment that rises from 5−10 per cent of global profits in the 1990s, to 25−30 per cent by the early 2010s. Tørsløv, Wier, & Zucman (2018) use a different but not unrelated methodology, evaluated in section 4.9, to reach a finding that 40 per cent of the profits declared by foreign affiliates are misaligned (note: this is substantially lower than 40 per cent of multinationals’ global profits, since it excludes home jurisdiction entities).

(p.160) The data available for banks under CRD IV (limited country-by-country reporting for EU financial institutions, under the fourth Capital Requirements Directive) provide an opportunity for a case study. Banks in the European Union recently started publicly reporting data on profit, number of employees, turnover and tax on a country-by-country basis. Janský (2018) introduces the largest, hand-collected, public data set of its kind, which covers almost 50 banks for up to 5 years between 2013 and 2017. he identifies the main locations of European bank’s profits, which include the largest European economies as well as tax havens. He focuses on answering the question of how geographically aligned these profits are with economic activity. He finds that some of the tax havens have maintained high shares of profits in contrast with their much lower shares of employees. Figure 6.3 below illustrates this point for Ireland and Luxembourg, for which there are ample data and both of which are important locations of profit. Janský (2018) concludes that his results indicate that banks are likely shifting their profits to tax havens, but for the profit shifting to be directly observed, regulators will need to ask banks to publish even better data.

(p.161)

New Proposals for IFF Indicators in the Sustainable Development Goals

Figure 6.3. Relative misalignment between profits and employees and turnover, respectively (% of gross profits), 2013–2017 mean for countries with at least 1000 million euro in profits reported in at least one of the years.

Source: Janský (2018).

6.1.4. Conclusions

The proposed SDG indicator is a measure, rather than an estimate, of the global total of country-level profit misalignment of multinationals with annual turnover in excess of $750 million. There are three potential criticisms, and one area of significant uncertainty.

The first criticism would take the form of a claim that some estimate, specifically of profit shifting, could more closely capture the scale of the problem. As discussed in chapter 4, we do not believe that there is currently an estimate which combines a robust methodology with sufficiently high quality data of near-global coverage, that could justify being put forward as part of the SDG framework.

A second criticism relates to just what is being measured. Given that profit misalignment is broader than profit shifting, is a direct measure of the former really preferable to an estimate of the latter—which is the IFF concern? While a direct measure of the specific phenomenon would of course be preferable, it is also unlikely—if the IFF were directly observable, it would not represent such a threat and would likely tend toward zero. The choice is therefore between a measure of a related phenomenon, and a less precise estimate of the specific phenomenon.

As figure 1.2 shows, profit shifting IFF are made up of the cross-border components of corporate tax evasion; unlawful tax avoidance; and lawful tax avoidance. Profit misalignment comprises these three elements, plus that of misalignment which is not tax-related—which, in other words, reflects simply that the current international tax rules do not have alignment as their goal or inevitable outcome, absent any tax-motivated behaviour.

Cobham & Janský (2019) find that by the 2010s, around 25−30 per cent of the global profit of US multinationals was misaligned. Tørsløv, Wier, & Zucman (2018) estimate that around 40 per cent of the profits of multinationals’ foreign affiliates is misaligned. In both cases, tax motivations drive the misalignment, and would support the assumption that overall, the non-tax-related component is likely to be small and should not be expected to demonstrate any systematic trend over time, or even across countries.

The extent of misalignment captured by the indicator in error (that is, not related to profit shifting) is therefore likely to be both small and random. As such, the indicator should provide a broadly consistent indicator of the scale of the IFF in question, allowing meaningful comparison across countries and of progress over time.

(p.162) The third criticism is related and something of a truism: namely, that the measure is not a fair evaluation of the current international tax rules, but instead a measure of how far they deviate from a unitary tax approach. The current tax rules rely on the arm’s length principle: namely, that entities within a multinational group should transact with each other at genuine market prices (assuming these exist for the goods or services in question), and the resulting distribution of taxable profits will be the ‘right’ one. The OECD’s separate accounting approach therefore takes each entity within the group as individually profit maximising.

A unitary approach, in contrast, identifies the unit of profit maximisation as the group itself—recognising that it may be in the group’s interest for some entities to make no profit, or even a loss, on paper. The total global profit is then allocated as potential tax base between jurisdictions where the group’s activity takes place, according to some formula. That formula could reflect, for example, the shares of a group’s tangible assets, sales, employee numbers and remuneration in each country (as the EU’s proposed Common Consolidated Corporate Tax Base does); or, say, sales and employee numbers (as the formula for apportionment between Canadian provinces does).

In this way, a unitary approach allows precisely the alignment of economic activity and profit that is sought, according to the global consensus reflected in the quotations in section 6.1.1. However, the indicator here simply reflects that consensus on the need to reduce misalignment—it does not set unitary tax as a global goal. Policymakers looking to move beyond the failed arm’s length principle might, of course, reflect on the alignment potential of unitary approaches.

Finally, this approach faces remaining uncertainty around the availability of data. It is unclear, first, whether the partially aggregated country-by-country reporting data provided to the OECD by national governments, for publication in early 2020 and annually thereafter, will be sufficiently consistent and of high quality and coverage to support the approach proposed here. Second, it is unclear whether a UN agency such as UNCTAD could or would step in to ensure better data if the OECD is unable to deliver. A further possibility would be to establish the range of necessary variables as part of the system of national accounts, and so to ensure their publication for the longer term. Third, it is uncertain to what extent tax authorities can generate their own national analyses—especially in lower-income countries which have least access to company reporting, and may be more likely to see their data suppressed in host countries’ reporting to the OECD. Immediate opportunities will depend on the quality of submitted tax returns and publicly available, global consolidated accounts; and a willingness to follow non-OECD countries (p.163) such as China, India and Uruguay (Knobel, 2018b) in requiring direct filing of country-by-country data.

6.2. Undeclared Offshore Assets: SDG 16.4.1b

6.2.1. Overview

With multinational profit shifting addressed by the proposed indicator SDG 16.4.1a, Table 6.2 shows the range of illicit flows outstanding. It is immediately clear that the range is wide indeed. Recall, too, that the underlying channels as set out in table 1.1 show greater detail and variety.

While each of the behaviours are unlawful, the main split is between the illicit use of legally generated funds, on the one hand, and the flow of criminal funds on the other. In the first category, the most iconic IFF behaviour is that of outright, cross-border tax evasion: the use of secrecy jurisdictions to hide, and to hold, undeclared assets and income streams resulting from legitimate business activity. We also find here illicit transfers of licit income, for example to circumvent capital controls; and licit transfers for illicit purposes, for example the financing of terrorist activity.

Table 6.2. A simple outline of illicit financial flows, excluding multinational profit shifting

Legal category

Origin of assets

Behaviour type

Unlawful

Legally generated profits, capital gains and income

Market/regulatory abuse

Criminal

Illicitly transferred, and/or

transferred for illicit purposes

Tax evasion

Proceeds of corruption

Bribery; Grand corruption; Illicit enrichment; Embezzlement

Proceeds of theft/related crime

Theft; Extortion; Kidnapping; Fraud; Bankruptcy

Proceeds of illegal markets

Drug trafficking; Counterfeiting; Firearms trafficking; Trafficking in persons; Smuggling of migrants; Wildlife trafficking

Source: Extract from table 1.2.

The third channel of illicit flows of legitimate income, however, is perhaps the largest and only in recent years has begun to receive greater public and policymaker attention. This is the use of anonymous ownership vehicles (p.164) to circumvent market regulations—for example, anti-monopoly limits on ownership concentration, or to hide potential conflicts of interest—for example, policymakers’ financial interests in regulated entities or in companies benefiting from political discussions such as the granting of mining rights, or telecoms licenses, or tax incentives. Most famously, in 2016 Iceland’s Prime Minister Sigmundur Davíð Gunnlaugsson stepped down after the Panama Papers revealed an anonymous company owned with his wife. The pivotal revelation was that the company held bonds in three major Icelandic banks—the value of which, after the financial crisis, was largely dependent on decisions taken by Gunnlaugsson’s government.

The second category includes three types of flows of illegally earned funds: the proceeds of corruption, the proceeds of theft, kidnapping and related crimes; and the proceeds of criminal markets including trafficking in narcotics, humans and firearms, illegal wildlife, waste, and illegal logging and fishing. As explained in chapter 1, this book has not sought to survey the literature on criminal market IFFs, surveyed recently in the crime-focused World Atlas of Illicit Flows (Nellemann, Stock, & Shaw, 2018). While work continues at organisations like UNODC to refine the approaches to estimation for specific, individual markets in individual countries, the gap is great indeed to build from these to credible, robust measures with broad coverage both of countries and of markets.

Overall, the sheer range of IFF types here raise problems for measurement or estimation. In addition, none of the approaches surveyed has suggested a comprehensive approach. The GFI and Ndikumana & Boyce approaches do aim to cover both capital account- and trade-based IFF, but even if perfect would not necessarily capture e.g. payments made offshore for trafficked goods or people; or hidden ownership through anonymous ‘foreign’ investment.

These IFF channels do, however, have a common element: the creation of undeclared offshore assets and/or income streams of domestic taxpayers. This varies in its centrality to each IFF type. For tax evasion, the creation of undeclared offshore assets is the essence of the IFF. For regulation-circumventing anonymous ownership, undeclared offshore assets are almost a byproduct of the process, which aims to hold domestic assets. For IFF relating to the proceeds of illegal markets, undeclared offshore wealth is a result that it is often unwanted, with further laundering used in an attempt to overcome it.

The proposed indicator takes the sum of undeclared assets as a potentially measurable proxy for the scale of IFF other than multinational profit shifting. In this way it collates the range of quite different IFF into a single indicator of scale, of the type envisaged by the SDG drafters.

(p.165)

6.2.2. Data

As with the proposed profit-shifting indicator, so too in the area of undeclared offshore wealth there is a new possibility due to the recent adoption at global level of a key tax justice proposal. In this case, it relates to the ‘A’ of the Tax Justice Network’s ABC of tax transparency: the automatic exchange of tax information. This measure requires jurisdictions that are signatories to the OECD Common Reporting Standard to provide bilaterally to other jurisdictions, detailed reporting on financial assets of the other’s citizens—for example, for Switzerland to report to Germany the Swiss bank holdings of German citizens.

This policy measure is intended above all to address offshore tax evasion by individuals. The category of undeclared assets, however—and hence the proposed indicator—should include the results of the great majority of illicit flows as set out in Table 6.2. With only certain exceptions, maintaining the success of the illicit flow will require continuing not to declare ownership of the results offshore assets to the home authorities.

More than 100 of the leading financial centres are committed to exchange financial information under the CRS, starting either in September 2017 or September 2018, and annually thereafter. Unfortunately, the OECD has allowed jurisdictions to breach the originally understood commitment to exchange automatically with all other CRS signatories, leading Switzerland and others to restrict their detailed reporting to only economically and politically powerful states. But as with 16.4.1a, the proposal here does not require full access to the detailed data.

Since financial institutions are required for CRS effectiveness to confirm the citizenship of accountholders, reporting of aggregate data is straightforward—that is, not the data on individual German citizens with Swiss bank accounts, but on the totality of their holdings. At the same time, to participate in the CRS requires tax authorities to organise their own data on citizens’ self-declaration in an equivalent manner. This therefore makes it reasonable to publish aggregate data on the totality of holdings in each other jurisdiction—e.g. of German citizens in Switzerland, in France, in Austria, and so on.

The major financial secrecy jurisdiction that has not committed to the CRS is the United States, which ranks second in the Tax Justice Network’s Financial Secrecy Index 2018. The major financial secrecy jurisdiction that has committed to the CRS, but used the bilateral ‘dating’ approach to ensure that it only provides data to a small number of fellow signatories, is Switzerland, ranked first in the Financial Secrecy Index. But both Switzerland and the US publish (p.166) aggregate data which is broadly equivalent to what would be required in the form of aggregate CRS reporting: the liabilities to foreigners, by jurisdiction, of financial firms in the reporting country (Knobel, 2018c).

The Tax Justice Network (Knobel, 2018a, Meinzer & Knobel, 2017) has published proposals for aggregate CRS reporting, to provide consistent data in support of global monitoring and accountability—just as the SDG indicator should ideally do. The complexity of the CRS approach, hinted at in figure 6.4, means that full accountability requires consistent data on the assets and income of the whole range of both reported and non-reported accounts, by jurisdiction of account-holder.

New Proposals for IFF Indicators in the Sustainable Development Goals

Figure 6.4. Proposed aggregate CRS statistics.

Source: Meinzer & Knobel (2017).

A particular concern relates to the ongoing activity to create non-reportable asset classes such as insurance ‘wrappers’ that may allow circumvention of the CRS. In this sense, automatic information exchange can be thought of as a form of capital control, and in common with all such measures will require ongoing strengthening as financial institutions and others ‘innovate’ to avoid (in this case) transparency. The standard should be expected to evolve over time for this reason, potentially raising issues for comparability in the longer term. In addition, the somewhat narrow range of financial assets currently covered makes leakage inevitable; but there is no serious alternative in terms of data quality with wider range.

(p.167) The benefits of having all major financial centres bar the US collate data for exchange on a consistent basis should not be understated. The quality of the data required is expected to be generally high.

In terms of coverage, comprehensive data would require some improvement in implementation. Some implementing jurisdictions have sought to limit the data gathered by requiring their reporting financial institutions only to collect data for the jurisdictions with which they will initially be exchanging information—ensuring that no data is available, even in the aggregate, on the assets of citizens of other jurisdictions. In addition, the current systematic exclusion of lower-income jurisdictions must be addressed—either through pressure on the OECD to implement the fully multilateral instrument that the G20 first sought, or through an alternative UN measure to require it.

The most consistent data currently available is that from the Bank of International Settlements (BIS). For many years, the BIS published bilateral data only on a consolidated basis. This consolidation of bank branches around the world up to the jurisdiction of their parent produced largely unhelpful data. For example, the sums held by an Ethiopian account-holder at Credit Suisse in Addis Ababa would be shown in the consolidated statistics as a Swiss-Ethiopian stock.

Following civil society pressure, the BIS now also publishes data on a locational basis—so that the Swiss-Ethiopian stock shows only the funds of Ethiopian account-holders at Credit Suisse branches in Switzerland—or indeed those of HSBC, etc.

In order to complement international reporting on bilateral funds held by financial institutions, domestic data are required on the extent of declared assets and income streams. For various reasons, some of which underpin non-reportable categories in the CRS and some of which reflect decisions not to seek the relevant information for tax purposes at home, not all accounts held offshore need always be declared to the home tax authority. But good practice suggests that, at a minimum, any decision not to collect all such data should be made clear.

The High Level Panel on Illicit Financial Flows from Africa (pp.65–66) is again clear in its findings on the importance of automatic exchange, and of engagement by lower-income countries:

Transparency is key to achieving success in the fight against IFFs. The admonition of the late Justice Louis D. Brandeis of the United States mentioned earlier that ‘sunlight is the best disinfectant’ is especially pertinent in this regard. The importance of transparency is evident in ongoing approaches to tackle IFFs, (p.168) whether through the automatic exchange of information, country-by-country reporting …

Policy implication: The policy implication of increased transparency is that it should ensure access to such information and the right to obtain it. While various countries and regions are developing mechanisms for information sharing, there is a need to move to a common global mechanism. African countries in turn need to show commitment to the various voluntary and mandatory initiatives by joining them and mainstreaming their requirements nationally and regionally, including through legislation and adoption of common standards. They also need to develop the capacity to request, process and use the information that they obtain.

In keeping with this, the construction of an IFF indicator that requires the collation of tax authorities’ own data on declarations, and may over time require a wider range of declarations, will have benefits above and beyond the indicator itself.

6.2.3. Methodology

The undeclared offshore assets indicator is defined as the excess of the value of citizens’ assets declared by participating jurisdictions under the CRS, over the value declared by citizens themselves for tax purposes. For each jurisdiction we define the undeclared assets as:

ϕi=j=1nβj,iαi
(4)

where:

αi is the sum of assets declared by citizens of jurisdiction i as being held in jurisdictions j =1, …,n where ji; and

βj,i is the sum of assets of citizens of jurisdiction i reported as being held in jurisdiction j.

We propose that the undeclared offshore assets indicator for use in SDG target 16.4 is the global sum of jurisdiction-level undeclared assets:

SDG16.4.1b=i=1nϕi
(5)

Again, the underlying jurisdiction-level measures will allow monitoring and accountability in a number of ways. Individual states seeking to reduce the (p.169) under-declaration suffered, for example, will be able to demonstrate to taxpayers that the economic elites who make disproportionate use of ‘tax havens’ are being fairly taxed. For example, the underlying data would allow France’s tax authority to show progress in closing the undeclared assets gap, thereby bolstering revenues and also confidence in the system, with wider benefits for tax morale and compliance.

For states that benefit from providing financial secrecy at the expense of others, the measures offer an accountability mechanism to demonstrate their own commitment to global progress. This would allow Switzerland, for example, to be held to account over the number of countries and the volume of assets for which it still refuses to provide automatic information exchange.

6.2.4. Results

Without systematic publication by tax authorities of the jurisdiction-level aggregate data from declarations of taxpayers’ offshore holdings, it is not yet possible for independent researchers to construct the proposed indicator. It is, however, possible to see how it would work.

Knobel (2018c) considers the example of Argentina:

[T]he Argentine media outlets reported that Argentina, an early adopter of the CRS, in 2017 received information about 35,000 foreign accounts, mainly in Belgium, Bermuda, Cayman, France, Isle of Man, Luxembourg, the Netherlands, Spain, and the UK. Interestingly, the US and Switzerland aren’t mentioned—because Argentina has no agreement to automatically exchange banking information with the US, and exchanges with Switzerland will only start in 2019.

But because of the data those two countries publish, Argentine authorities still have cards to play. On the one hand, they can check if the total amount of money Argentines declared in US banks matches what the US Treasury reports as belonging to Argentines. If Argentines declared less, then authorities can start investigating who has failed to declare their holdings, or make specific requests for information. The same goes for Switzerland.

… If all or most countries (or at least all major financial centres) published these details, a world of new patterns and knowledge would emerge, upon which citizens and responsible governments could act. Now imagine if all CRS adopting countries published this data not just at the legal owner level—as the US and Switzerland currently do (still allowing individuals to hide behind e.g. a company that is holding the bank account), but if countries also published this data at the (p.170) beneficial owner level (identifying also the individuals who may be hiding behind a shell company that holds the bank account). The CRS framework already requires beneficial ownership information to be collected and exchanged, so countries are already in a position to publish this aggregated banking data (both at the legal and beneficial ownership level) at no extra cost. A lot more data would emerge, no confidentiality would be breached, and a host of benefits could flow.

An engaged tax authority could construct a range of measures, comparing the gaps between taxpayer declarations of offshore holdings with public BIS data on the bilateral position of financial institutions elsewhere, and with received CRS data. Public, aggregate CRS data would complete the picture.

6.2.5. Conclusions

Like SDG 16.4.1a, this second proposed SDG indicator is a measure, rather than an estimate. In this case the indicator is a measure of the global total of undeclared offshore financial assets, which we take as one result of, and therefore a broad proxy for, the scale of illicit financial flows excluding multinationals’ profit shifting.

This second indicator is more ambitious, in two ways. First, it relies on financial centres being willing (or being required) to publish aggregate CRS data. The alternative, of potentially less consistent data from unilateral reporting (as e.g. the US and Switzerland currently do), or from reporting to the Bank for International Settlements, may provide an acceptable alternative in the meantime. Second, the indicator relies on tax authorities being able and willing to collate data on offshore assets declared by taxpayers, in order to demonstrate the gap vis-à-vis CRS reported totals.

6.3. Combining the Two Components

Overall, we believe that the two proposed indicators have the potential to provide global measures, respectively, of the scale of multinationals’ profit shifting and of undeclared offshore assets (a key result of all other illicit financial flows). In addition, both indicators are fully decomposable to support jurisdiction-level accountability for those that procure the underlying illicit flows, and those that suffer them—in some cases at least, without (p.171) sufficient challenge. As such, the adoption of these indicators into the UN SDG framework, after the county pilots, could go a long way to ensuring policy focus and eventually progress against illicit flows.

A final issue to consider is that the proposed measure for undeclared assets is not of a form consistent with the profit shifting indicator. To align the two—so that, for example, they could be added to form a single number as the currently framed UN target envisages—would require a conversion of undeclared asset (stock) into undeclared income (flow).

There are two possible approaches. One would be to assume some rate of return on the measured stock, to estimate the associated annual income flow—much as Henry (2012) and Zucman (2013) do. The fact that those two studies reach similar values for global revenues lost to tax evasion, despite a threefold difference in the asset base, speaks to the sensitivity of such approaches to the assumed rates of return.

The other approach would be to track the growth in total undeclared assets, year to year, as effectively a net flow of undeclared assets. This approach could also be adapted, à la Henry (2012), to allow for given rates of consumption of assets.

A weakness here is that if CRS circumvention through asset class innovation is effective, the data are likely to show falling flows even as the stock of undeclared assets (in non-CRS classes) grows more strongly. This is, of course, also ultimately a weakness of the approach in equation (4)—and so continuing tightening of the asset class definitions will be needed to maintain effectiveness. Using the potentially less sensitive BIS data to provide backup measures of overall scale may be valuable.

Since either conversion of the undeclared assets measure to a flow approach would introduce complications, we would propose to report separately on 16.4.1a and 16.4.1b, rather than seeking to combine the two—but as indicated, conversion and combination are possible if deemed strictly necessary.

In combination, the two measures respond to the main components of illicit flows as presented, most simply, in Table 1.2. Proposed indicator 16.4.1a captures the level of achieved profit shifting by multinationals. Proposed indicator 16.4.1b captures the level of achieved creation of illicit (undeclared) assets offshore, an important result of the remaining IFF types.