中国货币政策效果检验与利率市场化建模研究

【作者】 高辉；

【导师】 赵进文；

【作者基本信息】 东北财经大学 ， 统计学， 2003， 硕士

【摘要】 本文首先利用协整理论，基于向量自回归(VAR)的Granger因果关系检验以及冲击响应函数方法，选取了货币政策工具变量与宏观经济的多个变量进行分析，对中国1993—2002年间货币政策执行效果进行了实证检验。单位根检验结果显示，在我们分析的样本期内，我国货币政策工具变量与宏观经济变量季度数据均是含有一个单位根的非平稳序列。Granger因果关系检验与协整检验的结果说明样本期内采用的货币政策工具：信贷、货币供应量MO、汇率对消费价格、消费、股价、GDP、储蓄存款的作用是显著的，而利率工具的作用是很小的。对于投资，样本期内采用的货币政策工具：信贷、货币供应量M1以及存贷款利率引导是显著的。对于净出口，信贷、货币供应量对其几乎没有引导作用，而利率(市场化利率、存款利率)、汇率的引导作用却是显著的。最后，对变量间的动态特性应用VEC模型进行了冲击响应分析。从最终作出的基于协整向量的误差修正模型的冲击响应曲线，可看到：所有货币政策工具变量对宏观经济变量的冲击在短期内均较微弱，从长期来看，冲击均有不同程度的增强。 因此，中国在1993—2002年间实施的货币政策对实体经济的效果是显著的，尤其是采用的信贷、货币供应量、汇率货币政策工具对引导价格、消费、储蓄、GDP、股价的作用是非常明显的，但同时我们也看到利率工具除了对投资、净出口有较强的作用外，对价格、消费、储蓄、GDP、股价作用很微弱。由于我国目前仍存在较大程度的利率管制，形成机理的非市场性决定了管制利率的调节容易滞后于宏观经济的变动，调节幅度不够灵活，货币政策的传导效率大打折扣。因此，推进利率市场化，实现货币调控方式由直接向间接的彻底转变，使利率真正起到宏观调控的作用。在利率市场化改革过程中，要真正发挥利率的宏观调控作用，构建有中国特色的利率市场化模型具有重要意义。 为了进一步研究利率市场化下的规律，我们从多个角度选取了可能影响市场化利率的多个变量，运用协整理论对利率市场化作建模分析。单位根检验结果中可以看出，在我们分析的样本期内，我国宏观经济数据均是含有一个单位根的非平稳序列。通过Granger因果关系检验，找到了影响市场化利率的内容提要Granger原因为:通胀率、GDP、价格因素(RP工)、存贷款、存贷款利率、货币供应量(MO，M1)。从长期来看，货币供应量MO比M1对市场化利率影响大，这与近年来央行多次调低利率对居民储蓄存款、货币供应量Ml、进出口调控效果并不显著的实际情况相符。以货币供应量M1为中介目标，长期来看对市场化利率的调控效果不佳。从最终建立的短期动态方程即误差修正模型，可看到:市场化利率受其自身滞后一季度到四季度的影响，但影响均不十分显著，模型系数均小于1。市场化利率受GDp滞后一季度到三季度影响，且影响逐步减弱。市场化利率受价格因素(RPI)当期及滞后二季度与三季度影响较显著。市场化利率受贷款当期与滞后一季度到三季度影响，且影响逐步减弱。市场化利率受货币供应量(M1)滞后一季度与滞后四季度影响较显著，且影响方向相反。误差校正项的系数不大，说明向均衡水平调整的速度较小。从最终建立的模型各项评价指标来看，该模型具有良好的统计性质，从拟合值及预测值结果看具有较好的拟合及预测精度。因此，该模型对市场化利率预测与控制具有较好的参考作用。 短期动态误差修正模型给我们的启示为:通过GDP、价格因素(RPI)、贷款、货币供应量(M1)就可以准确、及时预测出市场化利率，从而央行就可以及时调整存贷款利率，进一步引导投资、消费。在利率逐步走向市场化的今天，以利率为货币政策中介目标的同时，也要兼顾货币供应量的调控，从长期来说调控MO来引导市场化利率，从短期来说调控M1来调整市场化利率。以便央行能够准确，及时获取市场化利率信息，为进一步决策服务。更多还原

【Abstract】 In this paper, we firstly test the performance effect of Chinese monetary policy in Cointegration theory and Granger Causality Testing that based on VAR and Impulse-Response Function Analysis between 1993 and 2002. We select the tool variables of monetary policy and multi-variables of macro-economy to analyze. Unit-Root-Testing results show that Chinese quarter data of the tool variables of monetary policy and multi-variables of macro-economy are non-stationary and contain a Unit Root in our analytical term of sample. The results of Granger causality Testing and Cointegration Testing show that the monetary policy tools which were adopted such as credit, money supply MO, exchange rate have an special prominence effect on consumer price, consumption, share price, GDP, deposit in sample period. But the effect of interest rate tools is very small. Monetary policy tools, which were adopted such as credit, money supply Ml, interest rate of deposit and loan have a special prominence effect on investment. Credit, money supply has almost no effect on net export. But interest rate (interest rate of marketing, deposit), exchange rate have a special prominence effect on investment. Finally, we make impulse-response analysis to dynamic character of VEC model among variables. From impulse-response curves that were derived from VEC models based on Cointegration vectors finally, we can conclude as follows: All monetary policy tool variables have rather small impulse to macro-economy variables in the short run, and their impulse would boost up at different degree in the long run.Therefore, monetary policy that was carried into execution in China between 1993 and 2002 had marked effect on entity economy, especially monetary policy tools that were adopted such as credit, money supply, and exchange rate channel off markedly for price, consume, deposit, GDP, share price. But we noticed that interest rate tool had faint effect on price, deposit, GDP and share price except stronger effect on invest and net export. Non-market of form mechanism decided thatadjusting of regulative interest rate easily lag to changing of Macro-economy due to stronger interest rate regulation in China at present, extent of adjusting is not enough agility, transmitting efficiency of monetary policy discount highly. Therefore, we push interest rate marketing so that we implement conversion of macroeconomic management from direct mode to indirect mode in grain and bring interest rate into macroeconomic management action really. Models of interest rate marketing of characteristic of China that were built have important sense in order to bring into play macroeconomic management in deed in the process of reformation of interest rate marketing.In order to further research the law of interest rate marketing, we selected many variables which may influence interest rate of marketing from multi-angles. And we made Cointegration model-building analysis for interest rate marketing. We can conclude that macro-economical data of China are all non-stationary that contain a Unit Root in our analytical term of sample from Unit-Root-test results. We found Granger cause that influenced interest rate of marketing through Grange Causality Testing such as inflation rate, GDP, factor of price (RPI), deposit and loan, interest rate of deposit and loan, money supply (MO, Ml). Money supply MO has stronger effect on interest rate of marketing than Ml in the long run. This is consistent with the fact that Interest rate that was adjusted low by the central bank had no marked effect on deposit of resident, Money supply M1, imports and exports in these years. To use Money supply Ml as medi-target would have not good effect of macroeconomic management of interest rate of marketing in long run. We conclude from short-term dynamic equation (error correction model) that was built at last as follows: Interest rate of marketing is influenced by itself from lag one quarter to lag four quarter, but influence is not quite prominence, Coefficients of that model are all less than one. I更多还原