Measuring Capital Mobility in Developing Economies

  • Stefano Manzocchi

Abstract

Part III of this volume — devoted to empirical studies of foreign capital in developing economies — starts with a chapter on capital mobility. The definition, and measurement, of the degree of integration in international capital markets is a widely debated issue with a number of related policy implications. Moreover, the evaluation of capital mobility in developing economies is a crucial prerequisite for the subsequent analyses of the determinants and consequences of capital flows in developing countries (DCs). If an individual economy is found to be insulated from world capital markets, it makes little sense to establish the role of foreign capital there. Conversely, if DCs are relatively well integrated in the international financial network, questions concerning the functions of capital flows become very relevant. The chapter is organised as follows: we first provide an introduction to the concept of capital mobility and highlight its major implications in terms of intertemporal allocation, financial diversification and economic policy (Section 5.1). We then review the debate on the measurement of capital mobility (Section 5.2). This is a ‘hot’ topic as many operational definitions and alternative empirical tests exist. A new methodology based on time-series econometrics is introduced in Section 5.3, and applied to the assessment of capital mobility in 33 DCs over 1960–88. In general, we cannot reject the hypothesis that several developing economies were integrated in international capital markets in this period. Some details on the statistical procedures adopted and the data sources can be found at the end of the chapter.

Keywords

Europe Propen Income Malaysia Argentina 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Notes and References

  1. 1.
    More on this topic in the book edited by Leiderman and Razin (1994).Google Scholar
  2. 2.
    Among other approaches, we signal those of gross capital flows (Golub, 1990), of international comovements in aggregate consumption (see Obstfeld, 1993, 1994) and of benchmark intertemporal optimising models (Ghosh, 1995; Dooley, Fernandez-Arias and Kletzer 1996).Google Scholar
  3. 3.
    See Appendix 1 (pp. 101-3) for the definitions of ‘integration’ and ‘cointegration’.Google Scholar
  4. 4.
    See Ghosh (1995) for a different perspective on this issue.Google Scholar
  5. 5.
    Their findings are reproduced here (Table 5.1, column (1)) but only the countries we analyse in Section 5.3 are listed.Google Scholar
  6. 6.
    See paragraph 2.2.2 in Campbell and Perron (1991, p. 153).Google Scholar
  7. 7.
    They argue that the cointegration between saving and investment is the statistical counterpart of the intertemporal solvency requirement, or no-Ponzi-game condition (this position is also expressed by Jansen, 1996). However, these authors neglect that not every theoretical model imposes the no-Ponzi-game condition, hence the statistical hypothesis of cointegration must be tested.Google Scholar
  8. 8.
    The definition of ‘integrated variable’ and other technical definitions are in Appendix 1 (pp. 101-3). The saving rate has been found I(1) in previous studies by Miller (1988) and Leachman (1991).Google Scholar
  9. 9.
    See Appendix 1 for the description of the system of hypotheses to be tested.Google Scholar
  10. 10.
    For the purpose of the empirical tests, the variable aidt is defined in Appendix 2 (p. 104).Google Scholar
  11. 11.
    A similar argument is advanced by Montiel (1993, pp. 34-5). In order to account for non-market financing, Montiel proposes also to include the variation of the official reserves of the central bank as an additional independent variable in the Feldstein—Horioka equation. However, this is likely to generate new difficulties as the change in official reserves already contributes to the definition of public, hence of aggregate, saving.Google Scholar
  12. 12.
    See Dolado, Jenkinson and Sosvilla-Rivero (1990); Campbell and Perron (1991).Google Scholar
  13. 13.
    Further details on this procedure can be found in Appendix 1.Google Scholar
  14. 14.
    As suggested by Campbell and Perron (1991, p. 156) the DF test can be preferred to the Phillips-Perron (1988) test whenever the DGP is characterised by negative serial correlation after one differentiation, which seems to be the case for many of the time series considered here.Google Scholar
  15. 15.
    This case occurs whenever st and it follow different deterministic trends: if this happens the Feldstein—Horioka equation (5.1) is misspecified as it should include a time trend.Google Scholar
  16. 16.
    As usual, the possibility of systematic statistical discrepancies cannot be ruled out.Google Scholar
  17. 17.
    The limitation on the number of countries studied in Chapter 6 is imposed by the availability of estimates on human capital stocks in 1960.Google Scholar
  18. 18.
    See Gundlach and Sinn (1992, p. 621); Mamingi (1993); Montiel (1993).Google Scholar

Copyright information

© Stefano Manzocchi 1999

Authors and Affiliations

  • Stefano Manzocchi
    • 1
  1. 1.University of PerugiaItaly

Personalised recommendations