Understanding China's move into Africa: an empirical analysis

Abstract

An important new issue on the international scene is the upsurge in market and non-market South–South relations. The aim of this paper is to understand the dynamics that lie behind the recent Chinese move into Africa by empirically exploring the determinants of Sino-African relationships. In order to have a comprehensive picture, the analysis takes into consideration the main channels of commercial and political interactions: outward foreign direct investment (OFDI), trade and aid (international economic cooperation). The empirical analysis utilises a panel data set, from 1998 to 2005, for 43 African countries. The econometric estimates for three simultaneous equations are based on an instrumental variables method. Results show that the Chinese move into Africa is driven by strategic interaction among the three channels (FDI, trade and economic cooperation) as well as by pull factors, i.e. the characteristics of the receiving countries in terms of natural resource endowments and their market potential.

Keywords:

• China
• Africa
• FDI
• trade
• aid

JEL Classifications:

• F4
• F21
• F35
• P45

Acknowledgements

We are grateful to Tim O’Brian, Eunsuk Hong, Xiaohui Liu, Xiaming Liu, Yingqi Wei, Danilo Gambelli, and to an anonymous referee, whose suggestions have been fundamental at different stages of the research. We acknowledge with gratitude the help of Giovanna Hirsch, Liu Yuman and Valeria Pecchioni.

We benefited greatly from comments made by participants at the 18th Annual Conference of the Chinese Economic Association (CEA, UK) at the University of Nottingham and to the Associazione Italiana Sistemi Economici Comparati (AISSEC) Annual Conference at Parma University. We retain responsibility for the opinions expressed in the paper.

Notes

Notes

1. Data from the World Investment Report (UNCTAD 2006) record a total amount of US$31 billion inflows received (almost doubled compared with the US$17 billion data of the previous year, while the inward stock reached US$264 billion).

2. Data on Chinese approved overseas investments are provided by the MOFCOM and are classified as of two types: financial, which refers to investments in the financial sector; and non-financial, which includes manufacturing and resource extraction. Given the lack of data on financial investment, the data on non-financial investment is analysed in this study.

3. The China–Africa cooperation forum launched in 2000 consists of China and all its African partners (practically all, except for five countries that have not yet respected the only political agreement required by China, i.e. not recognizing Taiwan as an autonomous province). This political partnership has been reinforced by the numerous official visits made by the main Chinese authorities to all the African countries during this period, and by the statement of intent presented at the last forum held in Beijing in late 2006. In January 2006, ‘China's African policy’ was the first official ‘regional’ document ever produced by the Chinese Government for a developing region. China claims a special kind of cooperation with African countries; it presents itself as ‘the largest developing country in the world’ and describes Africa as the continent ‘which encompasses the largest number of developing countries’ proposing a sort of south–south cooperation that does not entail many political and economic strings (according to the ‘Five Principles of Peaceful Co-existence’) such as those required by western countries and international financial institutions.

4. In resource exploitation activities, China seems to be helped by the fact that it is less risk-averse compared with western countries. This implies a sort of first mover advantage in some SSA countries, consistent with the fact that major flows of FDI come from SOEs, which have access to low cost capital and can rely on longer term horizons. Thus, the volatile political conditions of many African countries can be considered an opportunity to gain from higher rates of return from risky investment rather than a constraint (Mwega 2006).

5. ‘China's African Safari’, Fortune, 20 February 2006.

6. This scenario reinforces the characteristic position of many African countries as exporters of almost exclusively natural resources. The impact of the growing demand from China (and in some cases India) is reflected in the prices of primary commodities. Where China (and India) are net importers there has been a considerable rise in prices compared with markets where the two countries are net exporters (e.g. aluminium for China) or ‘switching producers’ (Goldstein et al. 2006). Chinese ‘hunger’ for natural resources has generated such a strong impact in the global market that an inversion in the historical trend of terms of trade for developing countries has been observed (Kaplinsky 2005). As a consequence of this reversing trend, African countries whose exports are mainly primary commodities have shown larger rates of growth in their GDP (Broadman 2007).

7. No official figures on the aid provided by China are available. What is certain is the recent commitment by the Central Government to double the 2006 figures by 2009 and that the forms of assistance provided by China to African countries are the following (Reisen 2007; Davies 2007): grants (mainly in kind); zero interest loans; concessional loans.

8. With reference to this, it is relevant to state the distinction made by the World Bank (2004) with regard to the trade-investments nexus, identifying three stereotypes of investment that: (1) targets the Asian (internal) market; (2) targets the African market; and (3) targets the global markets.

9. For commercial relations, there is a simultaneity bias when aid is tied, since more tied aid will imply more imports from the donor. The risk is limited when working on aid commitment flows, and aid disbursements usually lag behind commitments, particularly for project loans or grants, which require new infrastructure (Hjertholm and Howard 2004).

10. For the full list of the countries see Table A in the appendix.

11. A few corrections have been made to the data set when obvious mistakes, arising from compilation or editing problems, were discovered during an examination of the distribution of the variables’ data series for specific countries.

12. The approach used to estimate Chinese OFDI stocks for each African country for the period 1998 to 2005 at year end is based on the FDI flows data and by the following formula: OFDI _t = OFDI _{t −1} + F_t  where OFDI _{t− 1} is the OFDI stock of the previous year already depreciated using US$ deflator and F_t is the flow of OFDI for the same year t − 1.

13. The official definition of CNC from the China Statistical Yearbook is reported in the Appendix.

14. This can be considered as a relevant variable in the case of Africa since, as outlined by Broadman (2007), 40% of the African population lives in countries with no access to the sea.

15. A conflict is defined as active when there are at least 25 battle-related deaths per calendar year in one conflict.

16. The net values mean that some data could be negative due to the repayment of the debt services.

17. u _it = a_i + ε_it ; v_it = b_i + ε_it ; z_it = c_i + ε_it are the individual and time variation effects. Typically country effects and ε_it are assumed to be independent. If the country effects are considered as constants, the econometric model is a fixed effect model (FE); if the country effects are considered as random variables, the model is a random effect model (RE). In both cases ε_it is a random variable.

18. In general, as part of the data analysis, the order of integration of variables is examined first in order to avoid spurious regression. In the case of panel data with limited years of observations, the problem is reduced.

19. We choose to estimate a single equation rather than a simultaneous equation system. For an equation system to be consistent and asymptotically more efficient than a single equation estimation, all equations in the system need to be correctly specified. If this is not fulfilled, none of the estimated parameters are consistent. Hence, we prefer the single equation procedure since it is more robust to specification errors.

20. For each equation, the Durbin–Wu–Hausman test regresses the potential endogenous variables on all exogenous variables to obtain the residual values for each equation. The residuals are then included as additional regressors in the original OLS regressions. The subsequent test to see if the coefficients on the residuals are different from zero gives a p-value of 0.0000, showing that the OLS is not a consistent estimator.