On the Impact of Policy Uncertainty on the Demand for Money in China: An Asymmetric Analysis

Abstract

One previous study that estimated the demand for money in China, included output volatility and money supply volatility as two measures of uncertainty. The study found their effects to be transitory in the short-run but not in the long-run. We suspect that the lack of long-run effects could be due to both uncertainty measures being less comprehensive. When we replaced the two measures with a relatively more comprehensive measure known as policy uncertainty, we too found only short-run effects. However, when we separated increased uncertainty from declines and engaged in asymmetric analysis by estimating a nonlinear money demand function, we found that in the long run while increased uncertainty has a significant effect on the demand for money, decreased uncertainty does not. Our findings indicated that in China, as uncertainty increases people prefer to hold more cash.

Keywords:

• Policy uncertainty
• China
• money demand
• asymmetry
• nonlinear ARDL

JEL CLASSIFICATIONS:

• E21
• F31

Acknowledgments

Valuable comments of an anonymous referee as well as those of the editor are greatly appreciated. Any remaining error, however, is our own.

Notes

1 For details see Bahmani-Oskooee et al. (2012).

2 See http://www.policyuncertainty.com/europe_monthly.html

3 Examples of factors contributing to volatility of uncertainty measure in Figure 1 are: role of fiscal policy in China’s money creation (Li et al., 2020), demand for bank lending and increasing defaulted bank loans (Shang et al., 2020), China-U.S. trade war (Liu, 2020), and foreign bank entry and financial liberalization (Wang & Giouvris, 2020).

4 This section closely follows Bahmani-Oskooee and Maki-Nayeri (2018b) who carried out a similar analysis for Australia.

5 Note that by deduction linear combination of lagged level variables is equal to lagged error term in (1). To see this solve (1) for the error term and lag both sides of the solution by one period.

6 This is known as normalization.

7 Indeed, the estimate of lagged error correction term must be negative if variables are to converge toward their long-run equilibrium values. Bahmani-Oskooee and Ghodsi (2018) have demonstrated that estimate of lagged error correction term in this ARDL model is exactly the same as the estimate of the coefficient attached to the lagged error-correction term in the Engle-Granger setting.

8 See Shin et al. (2014, p. 291).

9 This is called short-run cumulative or impact asymmetric effects.

10 For some other application of these new methods see Gogas and Pragidis (2015), Baghestani and Kherfi (2015), Al-Shayeb and Hatemi-J (2016), Lima et al. (2016), Nusair (2017), Aftab et al. (2017), Arize et al. (2017), and Gregoriou (2017), Istiak and Alam (2019), Olaniyi (2019), Bahmani-Oskooee and Kanitpong (2019), and Hajilee and Niroomand (2019).