Explaining money demand in China during the transition from a centrally planned to a market-based monetary system

Abstract

Fundamental changes in institutions during the transition from a centrally planned to a market economy present a formidable challenge to monetary policy decision makers. For the case of China, we examine the institutional changes in the monetary system during the process of transition and develop money demand functions that reflect these institutional changes. We consider seasonal unit roots and estimate long-run, equilibrium money demand functions, explicitly taking into consideration the changes in the institutional characteristics of China's financial system. Using a newly compiled dataset that covers an unprecedentedly long period, 1984—2010, with quarterly frequency, we are able to draw conclusions on the transitions in households', firms' and aggregate money demand, on the role of the credit plan and interest rates, on the mechanisms of macroeconomic control during economic transition, and on theoretical questions in the development and money literature.

View correction statement:

Corrigendum

Notes

^1. Seminal works include Chow (1987) and Feltenstein and Farhadian (1987). Most works are based on a standard money-demand equation; they include Hafer and Kutan (1993), Huang (1994), Chen (1997), Chow and Shen (2005), Gerlach and Kong (2005), Mehrotra (2006), Laurens and Maino (2007) and Bahmani-Oskooee and Wang (2007). A few works have added institutional factors, including Qin (1994), Girardin (1996), Xu (1998), Hasan (1999) and Yu and Tsui (2000).

^2. In the literature, only Xu (1998) gives some thought to financial sector idiosyncrasies.

^3. On the concept of two circuits also see Dembinski (1988) and Holz (2000) and for the real economy Kornai (1979).

^4. The commodity inventory system specified that (working capital) credit be extended directly to the user in accordance with specific plans and for specific purposes on the basis of material inventories held by the economic unit. Enterprises were required to promptly pay back their loans when the commodities that were used to back the loan were transferred outside the enterprise (De Wulf and Goldsbrough 1986). The commodity inventory system represents the socialist economy's application of the ‘real bills principle’.

^5. The PBC today still sets all interest rates, subject to State Council approval, at nationwide uniform levels (albeit some, nowadays, with a bandwidth).

^6. A first bout of double-digit inflation in 1988–89 led to panic purchases of consumer goods and the withdrawal of household savings deposits. The PBC responded by inflation-indexing household deposits (but not enterprise deposits) with maturities of three years and longer for the period 10 September 1988–1 December 1991. A second period of inflation-indexing began on 11 July 1993, with the inflation subsidy dropping to zero in May 1997. See Jinrong shibao, 28 October 1997 and 6 December 1999, Finance Yearbook1997, p. 493, and Statistical Yearbook1998, p. 533.

^7.Price Yearbook1997, p. 482.

^8. According to Montes-Negret (1995, p. 31), the Industrial and Commercial Bank of China in the early 1990s could autonomously decide on less than 20% of its lending. Xiao (1997, p. 371) reports that approximately 40% of all state bank loans in 1991 were ‘policy loans’ (defined in the source as particular lending categories), while Hui (PCBC Shenzhen municipal branch and Hui 1994) claims that 20% of all Industrial and Commercial Bank of China loans were policy loans; for the Agricultural Bank of China the share was 30%, for the Bank of China 15% and for the Construction Bank of China 45%. Yet by 1993 the PBC explicitly encouraged banks to refuse loans to projects which were not economically viable.

^9. The national economic and social development plan of 1995 (NPC, 18 March 1995a) reported only one monetary target for 1995: this was the total increase in bank credit, by no more than 570bn yuan (with quarter-by-quarter limits for individual banks). Prudential ratios were not enforced for many years to come. Money supply data appeared for the first time in 1993, in the back pages of the PBC monthly magazine Zhongguo jinrong, together with all other quarterly financial statistics. The People's Bank of China Quarterly Statistical Bulletin, which started publication in 1996, carries retrospectively created data on M1 from 1990 onward.

^10. The South China Morning Post, 1 February 1996, reported Chinese growth targets for M₁ and M₂ in 1996 of 18% and 25%. The 1998 national economic and social development plan stated that credit limits on lending by the state commercial banks had ended. It did not report any monetary targets for 1998, except an increase in currency in circulation of 150bn yuan (NPC, 6 March 1998). The 1999 national economic and social development plan for the first time issued M₁ and M₂ growth targets, of approximately 14% and 14–15 % (NPC, 6 March 1999).

^11. For example, in early 2010 the head of the China Banking Regulatory Commission, Liu Mingkang, said that overall credit growth in 2010 would be restricted to 7.5 trillion yuan.

^12. The HEGY result suggests that the standard unit root tests wrongly reject the variable's stationarity because of the presence of a seasonal unit root. Indeed the HEGY test suggests that M₀ is stationary in the zero frequency but has a unit root in the semi-annual frequency, i.e. the dynamic of the variable differs between the first and the second half of the year. This may be due to the Chinese New Year holiday in the first quarter of the year, with a surge in household purchases and the use of newly printed cash for presents; the excess currency in circulation is then later withdrawn by the PBC.

^13. Usually the number of leads and lags, k, is two or three. Here, k = 2 to avoid too many coefficients to estimate. We have removed all insignificant lead and lag terms with a backward stepwise strategy.

^14. The average monthly interest rate is the arithmetic mean of the daily interest rates, and the average quarterly interest rate is the arithmetic mean of the monthly interest rates (see Appendix 1).

^15. We can equally well obtain quarterly yoy growth rates from the published monthly yoy growth rates. The difference between these two series is exceedingly small. Because we obtain real series of other variables by using 2002 as the base year, we take the same approach for the price indices. The choice of 2002 followed careful examination of the price indices in all years. 2002 was a year with average monthly CPI inflation of − 0.8% (and a standard deviation of 0.3 percentage points) and it followed two years of near-stable prices (see Appendix).