Direct, indirectly

FDI

The trouble with treating FDI inflows as a national report card is that it’s often overly difficult to know what is, and what is not, FDI. 

In September 2024 the Grand Chamber of the Court of Justice of the European Union ordered Apple to repay 13 billion euros in back taxes to Ireland. The ruling turned on a question about location: where, exactly, the intellectual property behind iPhones, Macs and software services was generating its value. Apple’s position was that the rights sat with engineers in Cupertino. The European Commission ruled that they were lodged with two Irish subsidiaries which employed nobody to develop them. The case was about tax, but the same accounting question drives how a large share of foreign direct investment is recorded in national statistics, and which countries top the headline league tables.

Rebeca Grynspan, secretary-general of UN Trade and Development (UNCTAD), presented the World Investment Report in Geneva in June 2025 and reported global FDI up four per cent, at 1.5 trillion US dollars. Productive FDI, she went on to say, had fallen 11 per cent. Two months earlier the Organisation for Economic Co-operation and Development’s FDI in Figures had recorded global flows up one per cent, to 1.49 trillion dollars, and down nine per cent once a handful of European outliers were excluded. The European outliers in question are Luxembourg, the Netherlands and similar jurisdictions through which multinationals route intra-group capital for tax and treasury reasons.

Phantom FDI

Jannick Damgaard of Denmark’s central bank, Thomas Elkjaer of the International Monetary Fund’s statistics department and Niels Johannesen of the University of Copenhagen worked on quantifying the gap between official and real positions. Their updated paper in the Journal of International Money and Finance in 2024 put phantom FDI (capital lodged in empty corporate shells with no link to any local economy) at about 40 per cent of the global total. Roughly four trillion US dollars of inward FDI was parked in Luxembourg as of 2017. Ten jurisdictions, including Bermuda, the British Virgin Islands, Hong Kong and Mauritius, hold over 85 per cent of the worldwide phantom stock.

Patrick Honohan, a former governor of the Central Bank of Ireland, has documented the most famous example. In July 2016, the Irish Central Statistics Office (CSO) published a revised set of national accounts showing the economy had grown 26.3 per cent the previous year. Senior CSO statistician Michael Connolly had transmitted the encrypted data to John McCarthy, the Department of Finance’s chief economist, only hours before publication, after intense discussions with the multinationals concerned about how to record their activity. Paul Krugman tweeted that afternoon, coining the phrase “leprechaun economics”. The next day Honohan wrote that the distortions had grown so large as to “make a mockery of conventional uses of Irish GDP”. The 26.3 per cent number had no counterpart in Irish activity: Apple had moved a tranche of intellectual property to Dublin for tax purposes, and the asset shift had entered the national accounts.

Ireland’s response was eventually to invent a parallel measure called Modified GNI, stripping out the worst of the distortions. Other governments have taken the opposite approach. Péter Szijjártó, then Hungary’s foreign minister, posted on Facebook in January 2025 that Chinese companies had been the largest source of FDI to his country in 2020, 2023 and 2024. Evidence, he said, of the success of the ‘Eastern Opening’ strategy. The figures he cited counted announced project values, not money actually deployed. A joint update by the Rhodium Group and the Mercator Institute for China Studies, published in May 2025, found that Chinese investors ranked third in Hungary’s net inbound FDI for 2024, behind European and South Korean sources. Much of the announced Chinese capital sits in planning or construction phases and is not yet captured by Magyar Nemzeti Bank data. 

China has the opposite problem at the receiving end. A World Bank working paper put round-tripped capital at 30 to 50 per cent of FDI flowing into China between 1994 and 2008. Chinese firms moved money out via Hong Kong or the British Virgin Islands and brought it back as foreign investment, qualifying for tax breaks meant for non-residents. The 2008 corporate income tax reform and subsequent 2014 measures cut the practice. The Reserve Bank of India’s 2015-16 census on foreign liabilities recorded Mauritius as the source of 20.8 per cent of cumulative inward FDI into India, a legacy of a 1983 bilateral tax treaty.

FDI by purpose

Maria Borga, a former senior statistician at the OECD and now at the IMF, edited the fifth edition of the OECD’s Benchmark Definition of Foreign Direct Investment, which the world’s statistical agencies are meant to use as their compilation manual. The new edition was published in March 2025. It asks compilers to identify pass-through funds separately and to report positions by ultimate investor economy and ultimate host economy, not just by the immediate counterparty. It also asks them to classify FDI by purpose, distinguishing greenfield investment, M&A and financial restructuring. Andrew Jowett of the United Kingdom’s Office for National Statistics has chaired the OECD working group steering the rewrite since June 2022. Full implementation is meant to happen by 2029.

In July 2024 the Central Statistics Office released Ireland’s 2023 national accounts. They showed GDP falling 5.5 per cent on the year while Modified GNI grew by five per cent. Jack Chambers, Ireland’s finance minister, welcomed the GNI figure as confirmation of strong growth in the domestic economy. The 2024 accounts, released a year later, showed the divergence widening: GDP at 562.8 billion euros against Modified GNI of 321.1 billion euros. The difference consists almost entirely of redomiciled corporate profits and the depreciation of foreign-owned intellectual property held in Irish subsidiaries. The Irish Department of Finance now uses GNI as the denominator in its debt-to-income ratio for budget purposes.


Photo: Dreamstime.

Privacy Preference Center

Strictly Necessary

Cookies that are necessary for the site to function properly.

gdpr, wordpress_[hash], wordpress_logged_in_[hash], wp-settings-{time}-[UID], PHPSESSID, wordpress_sec_[hash], wordpress_test_cookie, wp-settings-1125, wp-settings-time-1125, cookie_notice_accepted

Comment Cookies

Cookies that are saved when commenting.

comment, comment_author_{HASH}, comment_author_email_{HASH}, comment_author_url_{HASH}

Analyze website

Cookies used to analyze website.

__hssc, __hssrc, __hstc, hubspotutk

Targeting/Advertising

Cookies for provide site rankings, and the data collected by them is also used for audience segmentation and targeted advertising.

__qca

Google Universal Analytics

This cookie name is asssociated with Google Universal Analytics.

_ga, _gid

Functionality

This cookies contain an updated page counter.

__atuvc, __atuvs