Determinants of Net Capital Flows in Developing Countries
A fundamental question concerning capital movements in developing economies is tackled in this chapter. After suggesting a (qualified) positive answer to the question: ‘Were developing countries (DCs) open to international capital movements during 1960–88?’ in Chapter 5, we now ask: ‘What determined the distribution of net capital flows across DCs in the same period?’ While Chapter 5 addresses an issue which is preliminary with respect to the central topics of this volume, this chapter goes right to core of the growth-theoretical approach to capital flows — indeed, it would not be arbitrary to view it as the empirical counterpart of Chapter 3.
KeywordsHuman Capital Capita Income Capital Flow Regional Dummy Capital Movement
Unable to display preview. Download preview PDF.
Notes and References
- 1.Uzawa (1965) argued that ‘human capital’ is only partially approximated by education and it includes other items such as health and some kinds of infrastructures and public goods.Google Scholar
- 2.See Gundlach (1994), OECD (1998); the topic of the returns to investment in education is extensively treated in Psacharopoulos (1994), among others.Google Scholar
- 3.A drawback of Romer’s (1990) regressions is that he uses data on the literacy ratio — admittedly of poor quality — as a proxy for the level of human capital.Google Scholar
- 4.Klenow (1998) provides an empirical comparison among alternative endogenous growth models based on US industry data.Google Scholar
- 5.Alternatively we also tried total GDP as a scale indicator; the two variables are strongly correlated.Google Scholar
- 6.Levine and Zervos (1993) argue that this should always be the correct economic interpretation of cross-section regressions.Google Scholar
- 7.Similar results are found using GDP as a scale variable.Google Scholar
- 8.See Edwards (1993, 1998a) for a thorough evaluation of the degree of trade openness and of its contribution to economic growth in DCs.Google Scholar
- 9.All other developing economies in the sample belong to Latin America.Google Scholar
- 10. The initial value of per capita income is already included in the regression. The interaction term (the saving rate times initial per capita income) is not statistically significant.Google Scholar
- 11.We neglect the contribution of the elderly population that is likely to be quantitatively less relevant, and more controversial, in DCs (see Bloom and Williamson, 1997, p. 16).Google Scholar