Determinants of foreign investment and divestment

This thesis investigates determinants of bilateral foreign direct investment (FDI) and foreign divestment (FDiV), based on different theoretical models of these capital flows and a variety of estimation methods. It consists of one chapter about data problems, followed by three stand-alone empirical chapters. Chapter 2 reports the main data issues throughout this research project, describing data collection, assessment, and details of the construction of the main data series. The first empirical study (Chapter 3) estimates the knowledge-capital model to find the dominant type of foreign direct investment – horizontal, vertical, or mixed – within Asia. The data include 31 Asian economies from 2001 to 2012. This study employs a comprehensive set of alternative estimation methods, including Tobit, Exponential Type II Tobit, Poisson Pseudo-Maximum-Likelihood (PPML), and Lognormal Hurdle (LH), to deal with excessive zeros in FDI between pairs of countries in Asia. After conducting a series of model selection and diagnostic tests, results show that PPML and LH are the most appropriate estimators. The findings signal that vertical FDI is the dominant type between Asian countries. A number of factors are shown to have a significant impact on the volume of intra-Asian FDI. These include distance, trade costs to both source and host country, GDP difference, and a common spoken language. A bilateral investment treaty has a positive impact on the probability of positive FDI, but not on its level. The second empirical study (Chapter 4) investigates determinants of bilateral foreign direct investment at a global level, on both margins: the extensive margin (whether to invest) and the intensive margin (how much to invest). I examine a global dataset comprised of 110 countries between 2004 and 2012. Estimation methods are random effects probit and correlated random effects probit for the extensive margin, and fixed effects and Chamberlain-Mundlak random effects for the intensive margin. Apart from conventional gravity variables, I find statistically significant impacts of the source country’s share of global technology capital and FDI share in production of the host country. Additionally, the two FDI margins are affected differently by corporate tax rates and common legal origin. Furthermore, breaking down the sample by country development levels reveals that FDI from less developed countries is not affected by many common variables. Hence, there is a need to develop more theories and empirical work to investigate FDI from less developed countries in particular. The final study (Chapter 5) examines determinants of foreign divestment on both the probability of divestment and the divestment amount, based on the foreign divestment theory developed by Boddewyn (1983a). The data include 137 host countries and 169 source countries, from 2004 to 2012. I apply parametric and flexible parametric estimators developed by Semykina and Wooldridge (2017) to correct for the sample selection problem in divestment data. Regression results show that the source country’s market size and GDP per capita, GDP growth of both source and host countries, bilateral investment treaties, and distance discourage divestment, whereas sharing a common currency and a common language surprisingly increases foreign divestment. In addition, political stability and common religion do not show any statistically significant impacts on either the probability of divestment or divestment amount