Money and price relationship in China

Abstract

This paper investigates the effectiveness monetary policy by Granger causality tests in the two regimes of inflation and deflation, respectively. The surplus lag rolling estimation is applied to deal with the problem of the frequent structural changes in the Chinese monetary system. We found that the monetary policies have become less effective in stabilizing the price level in the deflation era that started from 1998. There is also empirical evidence to suggest that money was endogenous in China during the inflation period. This implies that the People's Bank of China had difficulty exercising the power of money supply to reduce inflation if the endogeneity was the result of the market behaviour. However, if the endogeneity was due to the government inflation-targeting rule, then there is no evidence to suggest that this rule has been effective for M0, M1 and M2 instruments, except for the M0 instrument during the inflation period of April 1990 to March 1995. Although it was found that money ceased to be endogenous in the deflation periods, it does not support the proposal of utilizing the money supply as a policy instrument, as we found that money is impotent in influencing price in the deflation regime. Our findings provide some empirical evidence to support the Chinese government adopting alternative policy instruments such as an active fiscal policy in the era of deflation.

Keywords:

• Money
• price
• Granger causality
• China
• JEL classifications: E52
• E31
• C32

Acknowledgements

The paper benefited from the comments by Professor Haiyan Song, the editor of JCEBS, two anonymous referees, Chor-Yiu Sin, and the participants from the 14th Chinese Economic Association (UK) Annual Conference, Middlesex University, London, UK, held from 14–15 April 2003. This research was supported in part by a Competitive Earmarked Research Grant (No. LU3110/03H) from the RGC of Hong Kong SAR Government, and a grant from Lingnan University, Hong Kong. However, we are responsible for any remaining errors.

Notes

 Or in log-levels in our case.

 Whilst the concept of endogeneity is relatively easy to define, the definition of an exogenous variable is subtle and sometimes subject to controversial complication (see, for example, Engle et al., 1983). Therefore, we avoid the term ‘exogeneity’ in our paper.

 Similar interpretation applies to the horizontal axis in all of the subsequent figures.

 The full results may be obtained from the authors of this paper on request.

 In a dynamic VAR model, an increase of money supply at time period t will affect the economic variables in subsequent periods.

 We chose the starting date of estimation as January 1999 because the orders of all three VAR models (Lmj, Lp, Ly) (j = 0, 1, 2) have been stabilized to one lag. This simplifies the measure of M1 to only the coefficient on the lag one of Lm1, i.e.

(cf. equation (4)).

 See note 6.