A Study On The Dynamic Relationship Between The Stock Market Returns And The Inflation Rate Based On Bayesian Markov-switching Model

The fluctuation characteristics and the future trend of the infaltion, as a importantindex for making relevant macroscopical economic policy, have a direct effect on theeconomic growth. The stock market as a representative of the financial market, thevolatility will have a structural impact on social total money supply and aggregatedemand for money, and affecting the volatitly of inflation rate. Thus we study therelationship between the stock market and the rate of inflation, it is not only have apositive effect on helping our country to make the right macroeconomic policies, butalso have great significance to judge the market economic policy effect.The relationship between inflation and real stock return rate shows various forms.Gernerally, there are four possible cases: a positive correlation, negative correlation,uncertainty and irrelevant. Actually, the relationship between the stock market returnsand inflation rate is nonlinear and asymmetry in the real economic world. The Markovregime switching VAR model (MSVAR) has widely application in the research of thenonlinear dynamic relationship, because the Markov regime switching can depict wellfor the nonlinear switching of the time sequence. However, the MSVAR model hashigh dependence upon data, the number of the estimated parameter is large, and thegoodness of fit of this model is low. Combining Bayesian statistical inference methodsand MCMC sampling algorithm can solve these problems efficiently.Firstly, due to the problem of parameters uncertainty when we are investigatingthe relationship between the stock market and the rate of inflation, we modelingBayesian Markov regime switching VAR model.Then by analysing the statisticalstructure of the model, and then setting the prior probability distribution for the modelparameters, via Bayesian statistical inference method, we derive the posteriordistributions of the model parameters. Furthermore, we obtain the Bayesian estimationfor model parameters via employing the corresponding two Gibbs samplingalgorithm.Lastly, we employ the Bayesian Markov regime switching VAR model toexam the dynamic relationship among the stock returns, inflation volatility componentand inflation trend component. Research results show that: the Bayesian Markovregime switching VAR model characterizes the nonlinear dynamic fluctuations in therelationship between this three variables. In the period of the switching of regime,stock returns are positively correlated with persistent inflation and the temporary inflation rate, respectively, in the period of “market austerity regime”. A weak positivecorrelation with inflation rate is found in the short term of the “market expansionregime”. In the view of long term, the return rate are negatively correlated withinflation rate. This implies that the Fisher effect and Agency effects will be found indifferent market regime. The stock returns and volatility of inflation rate does notalways follow the Fisher effect, there may be multiple relationship.