The saving and investment nexus for China: evidence from cointegration tests

Abstract

The saving and investment nexus as postulated by Feldstein and Horioka (FH) (1980) is revisited. The saving investment correlation for China is estimated over the periods 1952–1998 and 1952–1994, the latter culminating in a period of fixed exchange rate regime. Amongst the key results, it is found that saving and investment are correlated for China for both the period of the fixed exchange rate and the entire sample period. With high saving-investment correlation, the results suggest that the Chinese economy is in conformity with the FH hypothesis. This is a valid outcome, for in China capital mobility was fairly restricted over the 1952–1994 period as indicated by the relatively low foreign direct investment.

Acknowledgement

I appreciate and acknowledge comments and suggestions from George Saradakis, Michael McAleer, Farshid Vahid and Russell Smyth on earlier versions of this paper. I also thank Hashem Pesaran and Yongcheol Shin for sharing the GAUSS codes they used to produce the original set of critical values for the bounds testing approach to cointegration as reported in Pesaran and Pesaran (1997) and Pesaran et al. (2001).

Notes

Bayoumi (1990) has argued that the use of total investment may lead to spurious correlations with savings that reflect endogenous behaviour by private agents. In this light, it is useful to estimate the F–H equation using total business fixed investment. However, data on total business fixed investment is not available for China.

China was admitted to the WTO in December 2001. China began opening its economy, albeit gradually, since 1979. See also Branstetter and Feenstra (2002, p. 336).

Previous studies of the unit root hypothesis which have employed the Zivot and Andrews (1992) sequential trend break model have tended to use model C (e.g., Raj, 1992; Perron, 1994; Ben-David and Papell, 1995); however, there is no consensus on which model is superior.

Similar line of argument was taken by Narayan and Narayan (2005).

The Gregory and Hansen (1996) test only accounts for one structural break. However, this test is applied despite finding two breaks in the data series. This is done because there is no cointegration technique that accounts for two breaks. The break points given by the Zivot and Andrews (1992) and Lumsdaine and Papell (1997) unit root tests are somewhat different to those given by the Gregory and Hansen (1996) cointegration test though the majority of the breaks are either 1957, 1958 or 1959. The different break dates are due to the fact that the tests for unit roots search for breaks in a series while the cointegration test searches for a break in the residual of two series. For an analysis of structural breaks in China's GDP, see Narayan (2004c).

Foreign direct investment in China has increased rapidly in the 1990s, with real investment consistently been around 30% of real gross domestic product. ‘For the last 5 years – with the exception of 1999 – China has been the world's second largest recipient of foreign direct investment’ (Wang and Yao, 2002, p. 14).