StuW Sonderheft NeSt 2025

StuW Sonderheft NeSt 2025

Abhandlungen – Steuerwissenschaften

S70

Koch / Scheider – The Known Unknown: Tax Avoidance by European Multinationals

best possible manner. One potential concern with our metho dology is how we separate the effects from the location of sub sidiaries and profit differences across subsidiaries. In this re spect, we either assume that, in the absence of profit shifting, all affiliates of a business group have the same profit or that the profits are allocated in accordance with local GDP of the affili ates ’ countries. Our measures of profit shifting, thus, reflect both differences in the size and in the profitability of affiliates. The accuracy of this step could be improved if we could merge our data with other available information, particularly from firms ’ private country-by-country reports. Access to this infor mation would allow us to observe the allocation of firms ’ in vestments across countries more precisely. In general, the exist ing problems in answering the research questions addressed in this paper illustrate the great need for improving the tax data infrastructure (see in this respect also Blouin & Robinson , 2020; Bilicka , 2019; Bradbury et al., 2018). In this sense, we strongly support and appreciate the great efforts of NeSt (Netzwerk em pirische Steuerforschung – Network for empirical tax research) of the German Federal Ministry of Finance to improve the tax data infrastructure in Germany. As a second contribution, our analysis is focused on profit shifting by European multinationals. By contrast, most other recent studies are based on US data (e.g., Clausing , 2016, 2020; Garcia-Bernardo et al., 2023) or use global samples with only a small share of European multinationals (e.g., Johansson et al., 2017a; Tørsløv et al., 2022; Delis et al., 2025). The findings of prior research let us assume that European multinationals and US multinationals differ in their attitude towards and the use of profit shifting. Having results for EU data is thus important for European policymakers in order to evaluate the necessity of further regulation against BEPS. The remainder of this paper is organized as follows. In Sec tion 2, we discuss the findings of related literature. We describe the employed data and our methodology in Section 3 and re port the results in Section 4. Section 5 concludes the paper. 2. Related Literature Several studies estimate the scale of international profit shifting and the total revenue losses associated with such tax planning schemes. An OECD working paper by Johansson et al. (2017a) assesses tax rate elasticities of profits based on a global panel of unconsolidated financial statements from ORBIS over the peri od 2000 to 2010. They find that net revenue losses from inter national profit shifting amount to 4 to 10 percent of global cor porate tax payments. However, their results do not distinguish between different locations of multinational headquarters. Tørsløv et al. (2022) use a different methodology and data to answer this question. They combine country-level national ac counts data and foreign affiliate statistics for the year 2015 as their primary data source. They compare the profitability of af filiates of foreign multinationals with that of purely domestic firms and find that multinational affiliates are more profitable in several low-tax countries, which they attribute to the use of profit shifting. Overall, they estimate that 36 percent of global multinational profits are shifted to tax havens. U.S. multina tionals shift twice as much of their profits to tax havens as mul tinationals from other countries. Jansky & Palansky (2019) as sess the extent of profit shifting for a large number of countries

based on bilateral foreign direct investment data from the IMF. The 79 countries in their sample lose $125 billion per year in tax revenue due to profit shifting. According to their results, the tax revenues of OECD countries are the least affected, while those of low- and lower-middle-income countries are the most affected. Other studies estimate revenue losses from tax avoidance for individual countries. Clausing (2016) uses a series of BEA data on U.S. multinationals from 1983 to 2012 to assess the amount of profits shifted out of the United States. She estimates that the total revenue loss from profit shifting in 2012 is in the range of $77 to $111 billion in the U.S. and about $280 billion globally. Clausing (2020) provides a similar analysis based on 2017 U.S. country-by-country reporting data and estimates a revenue loss of about $100 billion for the U.S., or about one third of U.S. federal corporate income tax revenue. Fuest et al. (2022) analyse profit shifting by German multinationals. Using private CBCR data for 2016 and 2017, they find that – although 82 percent of these multinationals have at least one tax haven affiliate – only 9 percent of profits, 3 percent of employees and 4 percent of fixed assets of these multinationals are located in tax havens. According to their estimates, Germany loses 5.7 bil lion euros a year in tax revenue due to corporate profit shifting. Despite the existence of these studies, the true magnitude of revenue losses due to corporate profit shifting remains a mys tery. In particular, there are concerns about the accuracy of the data sources used. Studies based on unconsolidated financial information from ORBIS may suffer from a lack of representa tiveness of the firms included in the database. OECD (2020) shows that firms in ORBIS are on average larger, older and more productive than the population as a whole. In addition, the coverage of unconsolidated financial information for affili ates of multinational enterprises varies widely across countries. In particular, information on affiliates from North America ( Arndt , 2023) and tax havens ( Fuest et al., 2022) is known to be underrepresented in ORBIS. Measurement error can also arise from inconsistencies between taxable income and accounting profits, as shown by Bilicka (2019). Using confidential tax re turn data for UK firms, she shows that estimates of profit shift ing based on financial accounting data significantly underesti mate the true effect. Dyreng & Hanlon (2021) question some of the estimates of profit shifting based on macroeconomic data as implausibly large. Another common concern with this type of study is the poten tial bias due to the so-called double counting problem of for eign affiliate profits ( Blouin & Robinson , 2020). Simply put, the double-counting problem describes the problem of including lower-tier profits as equity income (dividends) in the profits of higher-tier affiliates, which may bias profitability of these high er-tier affiliates upward or their effective tax rates downward. The following example illustrates this issue: A German multinational owns a U.S. subsidiary indirectly through an intermediate holding company in Luxembourg. The U.S. subsidiary ’ s after-tax profits of $10 million are distributed to the German head office via the intermediate holding. No divi dend or withholding taxes are incurred in any of the three coun tries. As a result, the pre-tax profitability of the Luxembourg holding company and the German headquarter is biased up wards and their effective tax rate is biased downwards.