高等宏观经济学KRlecture.docVIP

Introduction To KPR’s Matlab Program Chen Xiaoguang Sep 19, 2002 The six Programs writen by Rebelo in 1987, including dyn.m, cdpar.m, mdr.m, imp.m, plotimp.m and acv.m, enable us to do the artificial experiments in the neo-classical model economy specified by KPR ( 2001 ). The experiments include: 1. Impulse response; 2. Moment computation; 3. Stochastic simulation. This lecture includes two parts. Part A sets up the theoretical foundation for the matlab program. Part B demonstrates the function and operation of the program. The results of some experiments are also given in this part. Below is the detailed content of the lecture. Part A. The theoretical foundation for the matlab program Steady state growth in the basic neo-classical model Environmental and policy specification Technological and preferencial restrictions to ensure steady state Levels and ratios in steady state The transformed economy Near steady state dynamics Linear approximation Solution algorithm Solutions in alternative experiments Simulation, moment and impulse responses Part B. The introduction to the matblab program The flow chart of the program The function of each program Some results by running the program  . We partition the matrices as follows: Part B. The Introduction to the Matlab Program The Flow Chart of the Program The Function of Each Program There are six programs in the package. Each one undertakes the different work demonstrated in the flow chart above. Dyn.m set up the model and compute the dynamic decision rules for k and λ. ( cdpar.m is the analogy to dyn.m which sets up a two-sector model herein we do not consider.) Mdr.m uses the results given by dyn.m to compute the Markovian Decision Rules which relates the controlled states variables ( K, λ) , exogeneous variables ( A, G), control variables ( C, N ), flow variables ( Y, I, R, Y/N ) to the states variables ( K, A, G). Based on the Markovian Decision Rule computed by mdr.m, imp.m can do the impulse reponses experime