Generalized Impulse Response Model And Its Applications In The Analysis Of Asymmetric Effects Of Monetary Policy

This paper attempts to construct a vector autoregressive model with more exogenous variables, use Monte Carlo methods to achieve the generalized impulse response function analysis, and the method was applied in the empirical analysis of Monetary Policy. At the same time, because the sample data belong to a small sample of data, process the data using bootstrap.Vector autoregression model is a econometric models, commonly used in economic or financial research. It was put forward by Christopher Sims (1980). This model uses the multi-equations simultaneous form, every one in the model equations, each endogenous variable using regression to lagged items of all the endogenous variables in the model, this is used to predict the relevant time series system and analysis of random disturbance term on the dynamics of the impact of variable systems. Impulse response function is used to measure endogenous variables of current and future value of the impact of changes in track, from the impulse of a unit of standard deviation of the random disturbance term, It is able to compare visually depicts the dynamic interaction between variables and its effects. For the traditional orthogonal impulse response function, following with the order of the variables into the VAR system is different, the decomposition of the impulse response function is also different, So Pesaran and Shin (1998) gives a new method-Generalized impulse response function, in order to overcome this shortcoming. the method does not require all of the impulse items are orthogonal, and is unaffected by the variable ordering of VAR model.Monte Carlo method is also called stochastic simulation or statistical test method. The basic idea of this method is:when you need to solve the problem is that the probability of certain events occur, or the expected value of a random variable, it could be through some "test" approach to obtain the appearance frequency of such events, or the average of this random variable, and use them as the solution of the problem. Bootstrap is a statistical methods, it is proposed by Efron (1979), in fact this is a sampling with replacement method. The advantage is that, even if a small sample data of relatively less informations can also be estimated and hypothesis testing, and it can obtain a inference analysis even when the distribution characteristics without strictly on the assumption.Generalized impulse response function of the basic objectives is the two conditional expectation. This paper describes in detail the steps of the method which use Monte Carlo mean method to simulate these two conditional expectation of generalized impulse response function, and compile the corresponding software program according to this analysis method, through this program, empirical analysis of the mutual effects between GDP and money supply in China.Through the empirical analysis to know, whether what level of money supply, between any money supply and economic at different times, the effect of the impulse respond are non-symmetrical, in economic development, according to this non-symmetrical relationship and economic development at different times, expansion and contraction, a conscious adjustment policies to guide chinese economy sustained and steady development.