1 Introduction
A major blind spot of existing economic statistics is the lack of data on cross-border real estate.
While progress has been made in the estimation and analysis of offshore financial wealth, little is known about real assets owned by households outside of their country of residency. How large is offshore real estate? Which households are more likely to own properties abroad? Does this form of wealth typically evade taxation? These questions are increasingly relevant for policymaking, because real assets are not covered by the multilateral automatic exchange of information between tax authorities that entered into force in 2017. This coverage gap creates incentives for tax evaders to rebalance portfolios away from financial assets toward real estate. There are also long-standing concerns that offshore real estate may sometimes be used to launder money and to evade international sanctions.
In this paper, we attempt to shed light on this aspect of financial globalization by analyzing property ownership in one of the world’s largest offshore financial centers, Dubai. We do so by exploiting micro-data capturing the ownership of about 800,000 properties in this territory, one of the seven emirates forming the United Arab Emirates. These data were provided by confidential sources to the Center for Advanced Defense Studies (C4ADS), a US nonprofit organization dedicated to analysis and reporting of conflict and security issues worldwide. The dataset contains property-level characteristics, including in most cases the owner’s nationality as of 2020. To our knowledge, this is the first time that such comprehensive and granular data on the ownership of cross-border real estate is analyzed in any country. Our main contribution is to use this data to estimate the value of real estate owned in Dubai by country and to analyze country-level patterns. We do not study individual cases: all the statistics presented in this paper aggregate many observations. Our analysis delivers four main findings.
First, Dubai offshore real estate is large. We estimate the total market value of properties in Dubai at USD 533 billion in 2020, of which about 27 percent is foreign-owned. Because we cannot identify the nationality of about 7 percent of the owners, foreign ownership is likely to be even higher. By our estimate, offshore real estate in Dubai adds up to at least USD 146 billion. To put this number in perspective, pioneering work by Bomare (2019) suggests that offshore real estate held in London through shell companies amounted to about USD 66 billion in March 2019. Foreign-owned real estate in Dubai thus appears to be about twice as large as in London, despite the fact that Dubai (with a population of 3.5 million) is only a third the size of London (population of 9.0 million).
Second, geographical proximity and historic ties are important determinants of foreign investments in Dubai. The bulk of foreign-owned properties in Dubai belong to owners from the Middle East, South Asia, Europe, and Central Asia. The largest foreign owners (both by the aggregate value of properties owned and by the number of owners) are Indian nationals: about 35,000 Indians own Dubai properties, worth almost USD 30 billion (20 percent of total offshore Dubai real estate). The United Kingdom comes next (23,000 unique owners, with properties worth USD 15 billion, 10 percent of the total). The remaining top countries by aggregate values include countries in the wider Middle Eastern and Central Asia region (e.g., Pakistan, Saudi Arabia, Iran, Jordan, and Russia) and large economies (e.g., Canada, United States, and China). About 8 percent of offshore Dubai real estate belongs to owners from the European Union. These patterns remain when we focus on the most affluent neighborhoods, where the share of real estate owned by foreigners is particularly large, around half. The main difference is that while India remains the largest owner, its share of foreign-owned real estate falls, while the share of Russia is multiplied by two, to reach 6 percent in the most expensive districts.
To better quantify these patterns, we estimate gravity-like models of foreign investment in Dubai properties. As in the literature on cross-border financial investments (e.g., Portes and Rey (2005) and Lane and Milesi-Ferretti (2008)), we find a statistically significant and economically large effect of distance and of the size of investing countries on the ownership of Dubai real estate. There is also a significant positive correlation between real estate held in Dubai and financial assets held in Switzerland, the largest offshore financial wealth center, even after controlling for gravity-type variables. The most notable exceptions are Iran, Iraq, and India, which hold large amounts of Dubai real estate but comparatively little in Swiss banks. Investments in Dubai properties thus share commonalities with other offshore investments, although Dubai appears to be particularly attractive to neighboring countries.
Our third finding is that a number of conflict-ridden countries (e.g., Afghanistan, Syria, Yemen) and countries under autocratic rule (e.g., Eritrea, Azerbaijan, and Kyrgyzstan) have large holdings in Dubai real estate relative to the size of their economy, equivalent to 5%–10% of their GDP. For example, Syrian nationals own the equivalent of 7.4% of Syria’s GDP in Dubai properties. Our data also reveal significant ownership by tax havens. Saint Kitts and Nevis, a major provider of citizenship by investment, has by far the largest ratio of real estate in Dubai to GDP. Tax havens like the British Virgin Islands, the Cayman Islands, the Bahamas, and the Seychelles also own significant amounts, reflecting ownership through shell companies. Allocating these properties to their ultimate beneficial owners would increase the size of Dubai properties owned by non-haven countries.
Fourth, the probability to own offshore real estate rises with wealth, including within the very top of the wealth distribution. To establish this result, we analyze the anonymized records of Norwegian owners of Dubai properties matched to tax records in Norway. Because Norway has a wealth tax, the authorities in this country collect individual-level information on wealth, allowing us to rank owners of Dubai properties in the wealth distribution. The probability to own Dubai real estate rises with wealth all the way up to the top 0.01% of the wealth distribution, similar to the pattern found by Alstadsæter, Johannesen, and Zucman (2019) and Guyton, Langetieg, Reck, Risch, and Zucman (2021) for the ownership of offshore financial assets. Moreover, about 70% of the properties owned in Dubai by Norwegian residents were unreported to the tax authority and thus potentially evaded taxation.
Our results have implications for the analysis of financial globalization and for international tax cooperation policies. Because offshore real estate typically goes unrecorded in official international investment statistics, the net foreign asset position of low-income economies with sizable holdings in Dubai is significantly larger than officially recorded. This finding has implications for the sustainability of the external debt of these countries and for macroeconomic modeling, as the net foreign asset position is a key state variable in standard open-economy macroeconomic models.3 Second, our results suggest that the lack of cross-border exchange of information on real estate ownership is an issue for tax enforcement, depriving governments of income tax revenues (on rental income and capital gains) and wealth tax revenues (for the countries that tax wealth). Expanding the forms of international cooperation that currently exist for financial assets to real assets could help address this issue.
The rest of this paper proceeds as follows. Section 2 relates our work to the literature. Section 3 provides background and summary statistics for Dubai real estate. We analyze the distribution of its ownership across countries in Section 4 and across wealth groups in Section 5. Section 6 concludes.