Skip to main content
Log in

Cross-border real estate investment: a different animal? Comparative evidence from bilateral flow data

  • Original Paper
  • Published:
Review of World Economics Aims and scope Submit manuscript

Abstract

Despite its growing importance and potential financial stability implications, cross border CRE investment unlike other types of capital flows has seen little research, a gap this paper tries to fill. Leveraging on a unique private dataset of bilateral CRE transactions covering 894 country pairs available from 2007q1 to 2018q4, matched with recently available bilateral Balance of Payment flows dataset, it provides a comparative analysis of the covariates of bilateral CRE flows versus traditional flows. First, higher relative GDP growth in the host economy is a stronger covariate of CRE flows than other flows. In contrast, CRE flows are not associated with interest rates and the differential between source and host country, unlike portfolio flows, nor with host country institutional variables, unlike banking flows. Second, CRE flows appear to comove as much with the global financial cycle as portfolio and banking flows. Third, among gravity factors, contiguity and distance are core covariates of CRE flows, unlike colonial relationship or common language. The paper finally finds that tightening of macroprudential policy and liberalisation of capital controls on outflows in the source countries are associated with higher outward CRE flows, while host country policies are found insignificant.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1

Source: RCA, author’s elaboration

Fig. 2

Source: RCA, author’s elaboration

Fig. 3

Source: RCA, author’s elaboration

Fig. 4

Source: RCA, author’s elaboration

Similar content being viewed by others

Availability of data and materials

Underlying data on bilateral commercial real estate investment may not be shared as protected under NDA agreement with RCA, a private data provider. Other variables can be shared upon request.

Code availability

Stata code can be shared upon request.

Notes

  1. Lizieri and Mekic (2018) estimate that around 65% of major global office transactions from 2007 to 2014 occurred in just 20 cities.

  2. Definitions of the origins of the buyer and seller are different between RCA (explained in Sect. 2.1) and the BoP, the latter of which is solely based on the concept of residency. In addition, there are no possibilities of negative flows in the RCA dataset as it records transactions and not divestments.

  3. See Barcelona et al. (2021) for a discussion of real estate flows entering as statistical discrepancy in the Balance of Payments of the US and errors and omissions in China.

  4. They also find that a lack of transparency within the legal framework, administrative burdens of doing real estate business, socio-cultural challenges and political instabilities deter international real estate investors.

  5. As mentioned in Sects. 2.1 and 2.3, RCA doesn’t record divestments, all bilateral flows in the CRE datasets are positive as a result. The log transformation of the dependent variable aligns and make comparable our CRE data and the BoP data in this respect, the latter of which may take negative values in case of outflows (negative inflows). Our results should thus be understood as the covariates of investment/ inflows.

  6. Due to its multiplicative form, the Poisson specification adequately deals with zero flows and provides a better alternative than arbitrarily adding a small positive number to be able to log the data. According to Santos Silva and Tenreyro (2006), the use of PPML also helps address two usual caveats of traditional gravity models: First, the log transformation can lead to the under-prediction of large flows and the under-prediction of the total flows. Because the Poisson regression model is estimated by maximum likelihood, the estimates are adapted to actual data. Second, when there are many cases with small observed and expected flows, the homoscedasticity condition can be violated, while the Poisson regression estimates operate consistently in the presence of heteroscedasticity and are reasonably efficient. The same equation is thus run, replacing the log dependent variable with the level of flows and replacing missing values in the RCA dataset by “zero flows” as discussed in Sect. 2.1.

  7. Examples of such macroprudential policies covering commercial real estate transactions, including some actions taken during the COVID shock, are presented in the supplementary material (Annex C).

  8. There is important potential for endogeneity in those tests whereby policy actions may be taken in conjunction or because of important flows. I see the following analysis as a first exploratory step towards understanding the impact of these policies on CRE flows.

References

Download references

Funding

Not applicable.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Etienne Lepers.

Ethics declarations

Conflicts of interest

No conflict of interest to be declared.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

I warmly thank Real Capital Analytics for kindly providing access to the bilateral cross-border CRE investment data for this academic project and for help in navigating the data. I also thank Laura Alfaro and Kathryn Niles Russ (editors) and an anonymous reviewer for constructive feedback. Views expressed here are mine and not those of the OECD or any of its member countries.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 31 kb)

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lepers, E. Cross-border real estate investment: a different animal? Comparative evidence from bilateral flow data. Rev World Econ (2023). https://doi.org/10.1007/s10290-023-00505-5

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s10290-023-00505-5

Keywords

JEL Classification

Navigation