Lag Effects Of Money Supply On Prices In China

The relationship between money supply and prices is one of key issues in macroeconomics.Although much work has been done to investigate the lag effects of money supply on prices by using different samples and methods,and have obtained much findings,there are also controversial views.The lag effect of money supply on prices refers to the persistent influence of money supply(in practice,means the change in the growth rate of money supply)on prices in the future.Duration and intensity are two important aspects of the lag effect.Different from some studies using correlation,this paper will use the predictability of money supply to future change in prices to measure the intensity of the lag effect.This paper aims to investigate the lag effect of money supply on prices in China.Vector Auto Regressive Models of money supply,prices and output are constructed based on the Quantity Theory of Money.Multiple Horizon Granger Causality test(MHGC)and Rolling Windows Multiple Horizon Granger Causality test(R-MGHC)are proposed to investigate the lag effect,using the monthly data from Jan 1996 to Dec 2016 as the sample.The overall characteristics and dynamic characteristics of the lag effect are investigated in this paper.The basic idea of Multiple Horizon Granger Causality test(MHGC)is to estimate the model VAR(p_t,m_t,y_t)of money supply,prices,output and its marginal process model VAR(p_t,y_t)separately.Using the covariance matrices of the forecast errors for forwardly predict p(t+h)using the two models separately,Intensity of Granger causality is measured,mhgc(h).The period parameter h describes the time duration of the lag effect,and the Granger causality describes the intensity of the lag effect.Bootstrap is applied to construct numerical confidence interval for hypothesis testing and statistical inference.The basic idea of Rolling windows-Multiple Horizon Granger Causality test(R-MHGC)is to construct a series of time windows using the rolling extension method and conduct multiple Granger causality tests on samples of each time window to estimate mhgc(h,i).The Granger causal intensity series,mhgc(h,i)(i=1,2,…,w)describes the dynamic change of the lag effect of money supply on prices with h lags.Investigate the overall characteristics of the lag effect using Multiple Horizon Granger Causality test(MHGC)and the whole sample.Empirical results show that the overall characteristics are as follows:(1)M2 has a lag effect of 1 to 7 periods(months)on CPI,the one-month lag effect has the strongest intensity.(2)M2 has a lag effect of 1 to 8 periods(months)on PPI,the three-month lag effect has the strongest intensity.(3)The lag effects of money supply on prices of production areas(PPI)and consumption areas(CPI)are significantly different.The lag effect on CPI has a shorter duration and a weaker intensity.Investigate the dynamic characteristics of the lag effect by using the Rolling Windows-Multiple Horizon Granger Causality test(R-MHGC).According to the change point of the trend of"The Granger causal intensity series”and considering the major changes in international and domestic economic development,the sample period is divided into three phases:Dec 2001 to Sep 2008,Dec 2008 to Mar 2014,Jun 2014 to Dec 2016.The backgrounds of two changing points are:In Sep 2008,the international financial crisis turned into a global financial tsunami;in May 2014,China officially made the judgement of entering the new normal era.Empirical results show that,the lag effect presents two distinct dynamic characteristics:First,there are significant differences in the three phases.In the first phase,the effect is weak.In the second phase,the effect quickly strengthens and maintains strong effect.In the third phase,the effect rapidly weakens and maintains weak.There are some possible reasons for the changes in third phase,such as China's economy has entered a new normal,growth rate and growth model have entered the transition phase;supply-side structural reforms have improved the relationship between supply and demand in the market;overflow of money supply to the real estate market;and the basically completion of interest rate liberalization.Second,dynamics of lag effect of money supply on prices of production(PPI)and consumption(CPI)shows significant differences:(1)The intensities of lag effects on PPI and CPI are different in each phase,and the relative strength of intensities also switch in the three phases.In the first phase,the intensity of effect on PPI is relatively weak.In the second stage,the intensity of effect on PPI strength quickly and be much stronger than the intensity of effect on CPI;In the third stage,both intensity of the effect weakened,the intensity of effect on PPI is just slightly stronger than that on CPI.(2)The lags with strongest intensity are differences.In the first stage,the strongest lag effect on PPI and CPI are lag 1 and 2(months);in the second stage,the strongest lag effect on PPI is in lag 3 moth,while on CPI is still in lag 1 month;not only the intensity becomes stronger in the phase,but also the duration of effect becomes longer.In the third phase,the stronger lag effect on PPI switche back to lag 1 month;not only the intensity becomes weaker in the phase,but also the duration of effect becomes shorter.This paper may contribute the literatures as follows:Firstly,this paper explores the dynamics of relations between money supply and prices rather than an overall static one;Secondly,the methods Multiple Horizon Granger Causality test(MHGC)and Rolling windows-Multiple Horizon Granger Causality test(R-MHGC)one VAR models proposed realize the identification of time-lag durations,and dynamics;Thirdly,providing new findings of overall and dynamic characteristics of the lag effect of money supply on prices,especially the finding of dynamics of the lag effect.These findings are also practical meaningful for economics agents.