Multinational enterprises, financial centres and their implications for external imbalances: a euro area perspective

Prepared by Virginia Di Nino, Maurizio Michael Habib and Martin Schmitz

Published as part of the ECB Economic Bulletin, Issue 2/2020.

This article analyses how the operations of large multinational enterprises (MNEs) affect the external account of the euro area and, in general, financial centres. The increased ease of moving intangible assets, profits and headquarters across borders poses challenges to the current framework of international statistics and economic analysis. First, the article shows how MNE operations are recorded in cross-border statistics, as well as the challenges in measuring such data. Second, the article highlights evidence of the impact that MNEs have on the external account of the euro area – this is most evident in current account balances and foreign direct investment in euro area financial centres, often involving special-purpose entities (SPEs). Third, the article looks at the tendency of financial centres to report current account surpluses that may be tentatively attributed, in part, to the activity of MNEs. Multilateral initiatives could help to improve the transparency of MNE operations and ensure an exchange of information across borders for statistical and tax purposes.

1 Introduction

The rise of large, profitable, global firms and the mobility of intangible assets^[1] have increased the relevance of firms’ profit-shifting activities, posing challenges to the current framework of international statistics. The balance sheets of large multinational enterprises (MNEs)^[2] have become very sizeable. The assets of the largest listed companies in major advanced economies, amounting to a value of several hundred billions of US dollars, are roughly equal to the gross domestic product of many small open economies. In order to reduce their tax burden, MNEs carry out a range of activities: these include shifting profits to low-tax jurisdictions by manipulating transfer pricing^[3] and shifting intra-company positions – often this involves complex financial structures and the creation of SPEs in low-tax, or no-tax, jurisdictions. These activities are extremely difficult to track. The novelty of some activities – in particular the growth in intellectual property products and improved opportunities to strategically choose their location – poses significant challenges for the existing framework of national and international statistics, which is based on the concept of residence^[4].

International tax avoidance by MNEs is not a novel phenomenon but its rapid growth increasingly attracts the attention of academics and policy makers.^[5] Global firms respond to tax incentives when recording worldwide income among affiliates. A recent survey of this literature finds that a decrease by one percentage point in the statutory corporate tax rate translates into a 1% expansion of before-tax income for global firms.^[6] Importantly, this study shows that the estimated impact appears to be increasing over time. Transfer pricing and licensing seem to be the main channels of tax avoidance – these appear to be more important than financial planning.^[7] International taxation may also alter the geography of foreign direct investment (FDI): a higher statutory tax rate in a target investment country discourages the acquisition of firms in that country, while lower tax burdens may attract FDI related to profit-shifting activities.^[8] Another area of research focuses on the implications of these tax-avoidance activities for the measurement of the external wealth of nations and the diminished ability of governments when it comes to taxing the corporate profits of global firms.^[9]

A number of policy initiatives at the international level have been launched to counteract the intensification of tax avoidance. The Organisation for Economic Co-operation and Development (OECD) estimates that 240 billion US dollars in tax revenues are lost globally every year as a result of tax avoidance by MNEs. As a result, the OECD and the G20 sponsored the Base Erosion and Profit Shifting (BEPS) Project, including an action plan that identifies 15 actions intended to limit international tax avoidance.^[10] This initiative currently involves over 135 countries, including the European Union (EU) Member States. The EU built on the BEPS Project’s recommendations by adopting two Anti-Tax Avoidance Directives, which entered into force between 2019 and 2020. The EU reform package includes concrete measures to reduce tax avoidance, boost tax transparency and move towards a level playing field for all businesses in the EU, but also new requirements for MNE financial reporting (see Box 1).^[11]

The traces of MNE operations are particularly apparent in the external statistics of financial centres. Since the euro area hosts some significant financial centres, this article discusses the dynamics of their external accounts. We adopt a standard operational definition of financial centres on the basis of the size of their stock of foreign liabilities relative to GDP. These are therefore economies where financial activities tend to dominate domestic economic activity. In particular, financial centres are defined as the ten advanced economies with the largest ratios of foreign liabilities to GDP in a large sample of more than 60 countries. These ten financial centres include six euro area economies (Belgium, Cyprus, Ireland, Luxembourg, Malta and the Netherlands) and four non-euro area economies (Hong Kong SAR, Singapore, Switzerland and the United Kingdom).^[14] Chart 1 shows the ratio of foreign liabilities to GDP for three groups of countries: advanced economies (excluding financial centres), financial centres and emerging market economies. In contrast to the effect it had on other advanced economies, the global financial crisis in 2008 does not appear to have dented the rise in the international financial integration of financial centres. In financial centres the median value of foreign liabilities increased, from around seven times GDP before the global financial crisis, to almost 11 times GDP at the end of 2018; the dispersion of the distribution of this statistic – foreign liability to GDP – markedly increased over the same period.

The importance of MNEs within the global economy has increased over time – as has the role of financial centres. It is worth considering whether this has an impact on current account imbalances, particularly on those of large financial centres. Financial centres tend to record large current account surpluses: eight out of the ten financial centres, as defined in this article, had a current account surplus over the past two decades on average. However, each one has its own business model, which is reflected in the diverse composition of their current accounts. Chart 2 shows the breakdown of the current accounts of these economies into their main subcomponents since 2010, when the stock of FDI liabilities started to grow rapidly. For the first group of economies – Singapore, Switzerland, the Netherlands and Ireland – the current account surplus is mainly the outcome of a large surplus in the balance of goods. As explained in Section 2 and Section 3, the activities of MNEs (such as merchanting and contract manufacturing) may boost the goods balance of financial centres. For a second group of economies – Luxembourg, Malta and Hong Kong – the surplus is mostly due to the service balance, in turn driven by the financial services sector.

The correct measurement of external statistics, such as those discussed in this article, is important for central banks. Large external imbalances may raise concerns about the sustainability of economic growth and about financial stability, which can affect monetary policy and macroprudential policies. For instance, central banks monitor external accounts to assess the equilibrium value of exchange rates, while noting potential misalignments – this is because abrupt and significant corrections in exchange rates may influence inflation developments. A distorted representation of aggregate current account imbalances could provide flawed signals to policy makers.

This article is structured as follows. Section 2 explains how typical operations by MNEs are recorded in balance of payments (b.o.p.) and international investment position (i.i.p.) statistics; it also highlights relevant challenges faced when measuring these statistics. Section 3 aims to gauge the quantitative relevance of MNE operations for the external accounts of euro area countries, in particular distinguishing financial centres from other euro area economies, and focusing on aspects of trade and the composition of euro area FDI. Section 4 summarises and concludes the article.

2 Recording multinational enterprise operations in balance of payments statistics

2.1 The origins of measurement challenges

The operations of large MNEs affect national accounts statistics and, in particular, external accounts, thus creating challenges for statistical compilation and economic analysis.^[15] This section reviews how typical MNE operations are captured in b.o.p. and i.i.p. statistics; it also highlights some of the associated measurement challenges. MNE tax planning strategies mainly affect b.o.p. data in three ways: (i) by shifting profits to affiliates in low-tax jurisdictions, which can involve moving IPPs or manipulation of transfer prices on intra-firm trade; (ii) by shifting intra-firm debt obligations and capital linkages; (iii) by redomiciling headquarters and legal incorporations to financial centres with favourable tax arrangements. This section also shows why these activities have different implications for the current account and i.i.p. of countries hosting MNEs and their affiliates.

Measurement challenges are caused by friction between residence-based national statistics methodologies and the global activities and ownership structures of large MNEs. B.o.p. and national accounts statistics, and their associated data collection processes, are based on the residency concept, according to which each institutional unit^[16] is resident of one economic territory: the place where they have their centre of predominant economic interest. However, MNEs tend to organise their production chains and corporate structures across the globe involving numerous legal entities, including SPEs (see Box 2).^[17] Data on these entities are recorded in the national b.o.p. statistics for the economy of the country where they reside. Consequently these data are not consolidated across borders with the home country of their parent MNE.^[18]

Measurement challenges are exacerbated by digitalisation and the increasing importance of IPPs, which are particularly relevant for financial centres. Over time the corporate structures of MNEs have become increasingly dynamic as a result of the redomiciling of headquarters and the increased relevance of intangible assets (such as patents and copyrights), which can be moved across borders with greater ease than physical assets, such as factories. These phenomena can have large effects in terms of magnitude and volatility of statistical indicators, which become especially visible in those economies where MNE transactions and balance sheets are large relative to the size of the domestic economy.

2.2 MNEs and current account balances

To trace the impact that MNE operations have on external accounts, various components of the b.o.p. need to be looked at separately.^[22] According to the b.o.p. identity, it holds that

C A + K A + E O = F A (1)

where CA stands for the current account balance, KA for the capital account balance (comprising mainly transfers of capital and non-produced non-financial assets), EO for errors and omissions (capturing any statistical discrepancy), and FA for the financial account balance.^[23]

MNE operations affect various items of a country’s current account balance, the key variable measuring trade, and income and transfer flows vis-à-vis non-residents. The current account consists of the trade balances in goods and services as well as cross-border factor income (primary income) and transfers (secondary income), with the first three being directly affected by the actions of MNEs:

C A = G o o d s + S e r v i c e s + P r i m a r y I n c o m e + S e c o n d a r y I n c o m e (2)

Cross-border production arrangements and merchanting activities related to MNEs can affect the trade-in-goods component of the current account. This might involve foreign subsidiaries of MNEs (in what is known as offshoring) or an unrelated foreign company (i.e. outsourcing). B.o.p. statistics are based on the concept of change in economic ownership. Which means, in contrast to international trade statistics that measure all goods crossing a country’s border, trade in goods recorded in b.o.p. statistics also includes contract manufacturing and merchanting. In contract manufacturing, an MNE hires a foreign company to produce a good. During the production process, the ownership of the inputs remains with the MNE and hence no trade flows are recorded in the b.o.p. (with the exception of an import by the MNE of manufacturing services from the foreign company that is producing the good). However, the b.o.p. does include the sale of the final products to third countries, which is consistent with the change in ownership principle. Merchanting is the process whereby a company purchases a good from an entity resident abroad, and subsequently sells it to a buyer in a third country without the good crossing the border of the country where the merchant is based.^[24] If such transactions involve foreign entities belonging to the same group, their pricing has a decisive impact on the amount and location of profits booked, which is in line with the well-established concept of transfer pricing.^[25]

MNE business operations affect trade in services, reflecting the rise of the knowledge economy and digitalisation. As IPPs can often be easily moved across borders within an MNE group, possibly involving SPEs, they affect exports and imports. ^[26] IPPs are hard to value at market prices and, therefore, MNEs may use them to avoid taxation. For example, one entity of an MNE might own the group’s IPP assets, while other entities in the same group pay licence fees and royalties for its use.

The primary income balance, which is dominated by investment income flows, is another component of the current account affected by MNE operations.^[27] Investment income reflects the receipts and payments generated by an economy’s external assets and liabilities (such as dividends and interest), and can be further decomposed into functional categories of the b.o.p. (FDI, portfolio investment, other investment and reserve assets).

MNE operations are particularly visible in FDI income.^[28] Income on FDI comes from its equity and debt components. Equity income can be further decomposed into dividends (profits distributed to the direct investor) and reinvested earnings (profits retained in the foreign affiliate). Crucially, the direct investor’s decision to reinvest earnings (i.e. to keep them in a foreign subsidiary) is recorded twice in offsetting ways in the b.o.p. – once as income on FDI, and once as a reinvestment of equal size in the financial account. In practice, MNEs can use complex corporate structures to optimise their tax burden – for example, by concentrating reinvested earnings in certain jurisdictions and by organising intragroup debt obligations. Apart from FDI, the cross-border ownership of MNEs may also affect portfolio investment in equity. In portfolio investment equity, only dividend payments are recorded in the income account, while non-distributed profits are not included.^[29]

The MNE operations described in this article mainly affect the composition of a country’s current account balance, while leaving the level of the current account balance unchanged. For instance, let’s first assume that a company residing in “country A” manufactures a pharmaceutical product and exports it to “country B”. This will generate a trade surplus in “country A” and a trade deficit in “country B”. Now, assume that the company resident in “country A” decides to move production offshore to a subsidiary, which is resident in “country C” (a financial centre economy) and subsequently the goods are sold to “country B”. This implies, all other things being equal, that the current account of “country A” records a profit – from the subsidiary in “country C” – equal in size to the net exports recorded before the decision to move production offshore. Thus, the value of the current account balance of “country A” is the same in either scenario, but the composition is altered in the second scenario because an investment income surplus replaces a trade surplus.

In contrast, MNE redomiciliation strategies – i.e. relocating their headquarters to another country – may have a significant impact on headline current account balances.^[30] Even if the redomiciliation of an MNE is not associated with additional economic activity in the economy of residency, the current account balance may be affected in several ways (e.g. due to attribution of net exports resulting from contract manufacturing or IPP related services trade). Primary income may be affected due to the differing treatment of reinvested earnings in FDI and portfolio equity. The country hosting the redomiciled global firm will record an improvement in the net FDI position and deterioration in the net portfolio equity position, to the extent that its shareholders are located outside the economy that hosts the new headquarters, which is typically the case for a small FDI hub. However, these two offsetting positions produce two different income streams. Reinvested earnings from foreign subsidiaries are recorded as income receipts and boost the recorded current account balance, whereas profits payments to foreign MNE shareholders are only recorded if they are distributed as dividends (in portfolio investment).

2.3 MNEs and cross-border financial and national accounts

Mirroring the current account, MNE operations also affect the financial account of the b.o.p. and external assets and liabilities. Changes to a country’s net i.i.p. can be broken down into net financial transactions as captured in the financial account (FA), revaluations due to changes in exchange rates and other asset prices (REV) and other volume changes (OVC).^[31]

I I P t - I I P t - 1 = F A t + R E V t + O V C t (3)

MNEs have a particularly large impact on FDI, both in the i.i.p. and the financial account. All FDI transactions (such as withdrawals of equity and reinvestment of earnings) are recorded in the financial account and hence affect the i.i.p. as shown in equation (3). Redomiciliations, which imply cross-border movements of MNE balance sheets, may give rise to OVC as defined in equation (3) and can thereby substantially change a country’s i.i.p.

Finally, it should be noted that MNE activities not only impact cross-border statistics but also affect the broader national accounts. A case in point is Ireland, where investment income flows, related to redomiciled MNEs, the depreciation of IPPs and aircraft leasing, had a large impact on Irish GDP and GNI.^[32] As a result, Ireland’s Central Statistics Office publishes a number of modified economic indicators (such as GNI* and a modified current account CA*) that exclude these phenomena and thereby provide a more focused view of domestic economic developments.

3 How do multinational enterprise activities affect the euro area balance of payments?

MNE operations affect the external accounts of the euro area, though their impact varies markedly across the 19 euro area countries. The aggregate b.o.p. of the euro area masks the varied impact of MNE activities on the external statistics of each individual country. Euro area countries can be classified into two groups, which present marked differences in their external accounts: six economies that are specialised in providing financial services^[33] and another 13 economies that are not.

The size, composition and volatility of the current account and financial account balances of euro area financial centres are significantly affected by MNE transactions. Section 3.1 presents stylised facts on the euro area b.o.p. related to the activity of specialised subsidiaries, such as SPEs in financial centres, whose location is primarily determined by tax-related, financial and regulatory considerations. Section 3.2 then focuses on the impact that SPEs have on FDI.

3.1 Euro area current account

When comparing the composition and size of the current accounts of financial centres with those of other economies in the euro area, five key features stand out.

First, financial centres in the euro area share a similar current account composition: they exhibit large trade surpluses that are partly counterbalanced by income deficits. This is shown in Chart 3 and corroborated by the empirical evidence in Box 3 based on a larger sample of the top ten global financial centres. The trade surpluses of financial centres often reflect exports with large value added, such as those related to licences in the field of information and communications technology. The literature on global value chains (GVCs) has established that value added is mainly created in very upstream activities (e.g. research and development, design and financial services) or very downstream activities (e.g. merchanting, logistics, royalties from licences, branding and marketing) – financial centres appear to have comparative advantages in several of these activities.^[34] If production is fragmented across borders, the allocation of value added across the firm’s network may result in financial centres appropriating a significant part of the value added on a global level. Income deficits can also reflect the practice of booking profits in financial centres.

Second, the negative income balances recorded by euro area financial centres partly reflect the redistribution of profits to foreign shareholders. The sum of the income deficits in financial centres was 5% of their cumulated GDP in 2018, whereas the primary surplus in the other euro area economies stood at 1.6% of GDP. The global value added retained in financial centres is ultimately owned by foreign investors that receive an after tax profit which is recorded as income deficits. In practice, however, while aggregate income deficits are very common in euro area financial centres, not all arise from FDI income. They may also be driven by portfolio income, as in the case of Luxembourg and Cyprus. Heterogeneity in income balance composition reflects specific business models, i.e. different net direct investment and portfolio investment asset positions, as well as their position in the global capital network and in relation to other financial centres.

Third, the practice of moving value added to low-tax euro area jurisdictions may also inflate their trade surpluses, while producing the opposite effect in higher-tax economies. This is suggested by the different scale of the vertical axes in Chart 3. MNEs pursue several strategies aimed at avoiding taxes that, while vested differently, ultimately boil down to value added being shifted across borders; these strategies affect the trade balances of euro area countries.

Available evidence shows that, as a result, the trade surplus of euro area financial centres stood at 13% of their combined GDP at the end of 2018. As shown in Chart 3, this contrasts with a surplus of less than 3% in the average of other euro area economies. Moreover, the surplus recorded by financial centres has tripled over the past decade, mirroring the growth in FDI recorded in the financial account of the b.o.p.

Fourth, contract manufacturing and merchanting conducted by entities resident in financial centres have generated a growing discrepancy between b.o.p. statistics and international trade statistics for euro area financial centres. Different concepts underlying the compilation of b.o.p. data with that of international trade statistics lead to some differences (see Section 2.2). In the euro area the gap between these two sources has been growing over time, in particular since 2015 (see Chart 4). Among euro area countries, financial centres account for the bulk of the growing discrepancy, whereas the discrepancy has remained stable for the other economies. This may be partly driven by MNE practices such as change of domicile and outsourcing of merchanting activities to specialised subsidiaries located in financial centres.

Fifth, the trade surplus of financial centres is mainly driven by value added that is produced elsewhere (i.e. foreign value added) and then re-exported. This contrasts with the group of other euro area economies, whose cumulated trade surplus primarily reflects domestic value added that is traded with final consumers. For a more detailed discussion of this feature, see Box 4.

3.2 Euro area foreign direct investment

FDI is a very significant component of the euro area’s financial account. In recent years it has gained prominence as a result of the striking expansion of gross transactions channelled by euro area financial centres (see Chart 5). The increase in gross FDI flows in turn reflects MNE activities, as discussed in this subsection.

The size of gross FDI flows going through financial centres is so large that they drive the aggregate developments of gross FDI in the euro area as a whole. FDI transiting through financial centres is, on average, between two and three times higher than that recorded by the other euro area economies. It is also three times more volatile. On a net basis, however, the FDI flows of the other euro area economies are more important in determining the aggregate net external position of the euro area (see Chart 5).

As a result of MNE activity, gross FDI transactions in the euro area have become less stable and less predictable compared with when FDI mostly consisted of mergers and acquisitions and greenfield investment.^[36] Furthermore, the volatility of gross FDI flows in the euro area, once considered a stable source of external financing, rose above that of other financial flows in the post-crisis period (see Chart 5). Conversely, over the same period the volatility of gross FDI flows in the other euro area economies declined compared to pre-crisis values.

Another defining feature of FDI is the strong positive correlation between gross assets and liabilities, especially in financial centres. The very large degree of co-movement of FDI inflows and outflows is determined by capital passing through financial centres en route to other destinations (Chart 5).^[37] Complex international investment schemes have been engineered to take advantage of favourable corporate tax and legal conditions; this makes financial centres highly interconnected while also allowing them to preserve their own business models.

The bulk of FDI transactions in financial centres are carried out by financial subsidiaries or holding companies of MNEs, including SPEs. In fact, other financial institutions’ transactions (which include these entities) dominate the size and dynamics of FDI in financial centres, whereas NFCs drive gross asset and liabilities flows in the other euro area economies (see Chart 6). According to the dedicated IMF Task Force (see Box 2), SPEs are set up by MNEs specifically to access capital markets or sophisticated financial services; isolate owner(s) from financial risks; and/or reduce regulatory and tax burden; and/or safeguard confidentiality of their transactions and owner(s).^[38] Euro area financial centres offer many of these advantages. In particular, they have developed sophisticated financial instruments, such as securitised products. The SPEs located in euro area financial centres typically hold MNE equities, manage corporate MNE debt-issuance, and allocate financing across parent and subsidiaries.^[39]

SPEs channel European and global capital around the world, also involving securitisation schemes. Some SPEs operate by pooling parent company debts and often transferring asset backed securities to a third subsidiary entity that is legally separate and possibly resident in another financial centre within or outside the euro area. This set of within-group financial transactions accounts for part of the earnings of SPEs and other subsidiaries in financial centres and represents another potential profit-shifting channel. Finally, to the extent that these securitisation schemes consist of within-group financial operations, neither the assets nor the risk underlying the securitised assets are shifted off the balance sheet consolidated at group level.

MNEs not only exert a significant impact on the size of gross FDI flows, but can also be a source of asymmetries in the measurement of bilateral external positions. These asymmetries are particularly pronounced for bilateral FDI income recorded in US and euro area b.o.p. (see Box 5).

4 Conclusions

This article analysed how the operations of large multinational enterprises (MNEs) are affecting the external accounts of the euro area and, in general, financial centres. First, the article presented how MNE operations are recorded in cross-border statistics, as well as the related measurement challenges. Second, this article showed the impact of MNEs on the external accounts of the euro area, which is most evident in the current account balances and in foreign direct investment of euro area financial centres, often involving special-purpose entities. Third, financial centre economies generally report current account surpluses that may be attributed, in part, to the activity of MNEs.

Multilateral initiatives to improve the transparency of MNE operations are necessary to ensure exchanges of information across borders both for tax and statistical purposes. Such initiatives should help national authorities to take action against tax avoidance. Moreover, close international cooperation between statistical compilers – including sharing of potentially confidential information – would help to ensure consistent cross-border recording of MNE activities, thereby improving the quality and consistency of macroeconomic statistics. In particular, such initiatives could help to ensure clarity by disentangling the transactions conducted by SPEs in the context of FDI in the b.o.p.