This paper examines how the choice between book profits and taxable income affects profit shifting estimation. Using administrative tax data from France (2014-2022) covering the universe of firms with matched tax returns and financial statements, we document substantial book-tax differences, with book profits exceeding taxable income by a factor of 3 to 4. Book profits explain only 23.3% of taxable income variation, challenging the assumption that financial statement data can reliably proxy for tax bases. We demonstrate that measurement choices systematically bias profit shifting estimates. The profitability gap between multinational enterprises and domestic firms is substantially larger using taxable income than book profits. Semi-elasticity estimates vary dramatically by profit measures: pre-tax profit increase by 1.1 percent when the foreign tax rate increases by 1 percentage point, while taxable income increases by 1.9 percent. Tax haven analysis reveals missing profits of €17.7-38.9 billion over 2014-2022, depending on the measure used. These findings suggest studies using financial statement data may substantially underestimate profit shifting and have important implications for policies like the OECD’s Global Minimum Tax.

Over the past decade, more than 100 jurisdictions have signed automatic exchange of financial information agreements (AEoI) in an effort to fight cross-border tax evasion. This paper studies the effectiveness and coverage of these agreements using account data leaked from an Isle of Man bank with a large customer base in countries participating to AEoI. We establish three sets of results. First, we find that the design of the governing AEoI agreement absolved the bank from reviewing and reporting a very large share (81%) of all the wealth owned by tax residents of AEoI participating countries, and instead the responsibility passed to smaller entities with weaker incentives to comply. Second, out of the wealth that fell under the bank’s reporting responsibility, foreign tax authorities only received reports covering 50% of what their tax residents held at the bank. We estimate that a further 32% went unreported due to loopholes in rule design. The rest of the accounts did not appear to have been reported, although through the information available in the leak we classified them as reportable. Third, we find evidence that bank clients who were more at risk of being reported on preemptively closed their accounts, potentially circumventing the AEoI reporting process. This paper provides new evidence on the potential limits of these agreements and how sophisticated individuals can ultimately avoid the AEoI transparency shock.

Cryptocurrencies pose substantial challenges to tax enforcement due to their anonymous and decentralized properties, undermining conventional regulatory practices. We study the impact of an ambitious new enforcement initiative aimed at addressing these challenges: domestic third-party reporting of crypto income. We estimate tax compliance and behavioral responses to this new policy by combining unique Danish microdata from domestic crypto platforms, administrative tax records, and cross-border bank transfers. Despite the introduction of domestic third-party reporting, over 90% of crypto investors do not declare crypto income. Moreover, we identify a significant and persistent evasion response to the policy as investors shift trading activity from domestic platforms, subject to third-party reporting, to foreign platforms outside regulatory reach. Our findings underscore the limits of domestic enforcement strategies in addressing tax evasion for decentralized, borderless assets like cryptocurrencies, highlighting the need for international coordination.

Using a quasi-experimental setting, we document that corporations decrease declared profits and corporate income taxes in response to an increase in the VAT rate. In an attempt to raise tax revenue during the Greek economic crisis, a 16% VAT rate, which existed for historicopolitical reasons in Greek islands, was harmonised to the national 24% rate. We combine tax filings with Orbis and ICAP data that enable us to geolocate corporations and to construct comparable groups based on locations in or out of the preferential rate. Counteracting the reform’s intended effect, declared profits decreased by 28% and corporate income taxes by 34% on a permanent basis. Macroeconomic factors and a fall in reported revenue cannot fully explain this decrease. Pervasive tax evasion in the Greek islands, where corporations might have an opportunity to adjust profits, offers a plausible explanation of the magnitude of responses.

Country-by-Country Reports (CbCRs) have emerged as a unique public source of information to track the country-by-country activities of multinational corporations. However, concerns about double counting and comparability have raised questions about the reliability of these reports for economic analyses. In this paper, we conduct a benchmark analysis focusing on publicly available CbCRs to assess the reliability of CbCR information compared to respective consolidated financial information. Our findings suggest only limited double counting issues. Most CbCR information matches well with the consolidated information, with only a few exceptions. Nonetheless, we document differences in the definition of variables and in the scope of the reports that may complicate comparisons across multinational corporations. We subsequently discuss the implications of our findings for the use of CbCRs as a source of information in economic analyses. In addition, we provide recommendations for improving the reliability and comparability of CbCR information.

In this paper, we analyse a sample of voluntarily published country-by-country reports (CbCRs) of 35 multinational enterprises (MNEs). We assess the value added and the limitations of qualitative and quantitative information provided in the reports based on a comparison to individual MNEs’ annual financial reports and aggregate CbCR data provided by the OECD. In terms of data quality, we find that the inclusion of intra company dividends and equity-accounted profits are a minor concern on average but that for individual MNEs corrections might be substantial. Our sample MNEs seem to pay higher effective tax rates than the global average and many of them report relatively little profit in tax havens. We only find a very weak correlation of the location of profits and effective tax rates. This might indicate that more tax transparent MNEs avoid taxes less aggressively. However, our assessment of different tax risk indicators reveals important variations between companies. 

We use novel leaked microdata to study offshore real estate in Dubai, a fast-growing tax haven. We find that the non-resident ownership share of residential real estate grew from nearly zero in 2005 to 30% ($68 billion) in 2019, a share much larger than other large global cities. We then go on to show that offshore real estate displays a strong gravity relationship and is more frequently owned at the upper end of the wealth distribution. Finally, we document that both tax evasion (with an 80% evasion rate) and sanctions driven capital flight are important motives for owning offshore real estate.