Applying sample selection methods for panel data to analyse determinants of foreign direct divestment

ABSTRACT

This is the first empirical study to examine determinants of both the probability and the volume of foreign direct divestment (FDD) between countries, applying sample selection methods for panel data. The data cover 137 host countries and 169 source countries, from 2004 to 2012. Empirical evidence shows that market size and GDP growth of both host and source countries, as well as a bilateral investment treaty, discourage divestment. By contrast, sharing a common currency has a positive impact on divestment. Distance negatively affects the divestment amount but does not show any statistically significant impact on the probability of divestment. Neither do political stability, colonial relationship, common language, or common religion play an important role in FDD. This study opens up an area for future research in terms of both theory and empirics since FDD is still a neglected area in international economics.

JEL CLASSIFICATION:

• Foreign direct divestment (FDD)
• foreign direct investment (FDI)
• sample selection
• parametric
• flexible parametric

JEL CLASSIFICATION:

• F23
• F21
• C33

Acknowledgment

The author thanks, without implicating, four anonymous referees of the journal for very helpful comments.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the author upon request.

Notes

¹ There are also studies employing only FDI stocks in their analyses such as Stack, Ravishankar, and Pentecost (2015), Nguyen et al. (2020) and Nguyen and Cieślik (2021).

² Political stability is highly correlated with other indicators (voice and accountability, government effectiveness, regulatory quality, rule of law, control of corruption) in the World Governance Indicators (>0.85).

³ Semykina and Wooldridge (2018) suggest that a cubic specification can also be considered.

⁴ At a country pair level, total annual FDI should be reasonably higher than 1,000 USD.

⁵ Note that these illogical observations appear mainly in the original data from both UNCTAD and OECD and are not attributable to the perpetual inventory method.

⁶ Bilateral FDI data published by UNCTAD are up to 2012 only, which explains the period of analysis in this study.

⁷ Regression results by country income groups are similar to the whole sample results and available on request.