Research On Bayesian Estimation Of China's Sub-regional Residents' Consumption Model

Consumption as the starting point and the end of production,in the process of economic activity occupies a very important position.Consumption can stimulate production,create jobs,and promote sustainable economic development.Throughout the process of national economic development,residents' consumption plays a very important role in economic growth.In recent years,China's consumption growth rate showed a downward trend,simultaneously the residents' consumption rate is too low and the effective demand of the residents is insufficient,economic growth has also been affected.At the same time,the consumption patterns of residents in different regions are also different.Therefore,we need to study and analyze sub-regional consumption patterns of residents,to provide a basis for formulating reasonable and feasible consumption policy.After studying the relevant literature at home and abroad,it is decided to use the Bayesian estimation method to introduce the prior information of the parameters,so that the quality of the statistical inference is improved,and the parameter is set as a random variable,which is more in line with the economic facts.At the same time,the use of panel data,increased the model of observation data and degrees of freedom,including more information to make the model get more effective and reliable parameter estimates.On this basis,the Bayesian analysis of the static panel data model and the Bayesian analysis of the dynamic panel data model are carried out respectively,improve the validity and precision of parameter estimation by means of Bayesian estimation and panel data as much as possible.In the empirical study,China is divided into six regions: North China,Northeast China,East China,Central and South China,Southwest China and Northwest China,the per capita income and per capita consumption of each region are selected as the data,and the Bayesian static panel data residents' consumption model is constructed on the basis of absolute income hypothesis,the Bayesian dynamic panel data residents' consumption model is constructed on the basis of relative income hypothesis,to study the consumption patterns of sub-regional residents.It is found that the Bayesian static panel data residents' consumption model and the Bayesian dynamic panel data residents' consumption model are very effective,and the Bayesian static panel data model and the Bayesian dynamic panel data model can be effectively portrayed China's sub-regional residents' consumption patterns,and absolute income hypothesis and relative income hypothesis are more in line with China's national conditions;Bayesian dynamic panel data residents' consumption model can effectively describe the residents' consumption inertia,and compared with Bayesian static panel data residents' consumption model,it gets more accurate result,indicating that the use of Bayesian dynamic panel data model can better characterize the law of residents' consumption in China's sub-regions,and the relative income hypothesis is more ideal than the absolute income hypothesis;The marginal consumption tendency and the consumption inertia of the residents in the western region should be larger;The influence of the last period consumption on the consumption expenditure of the residents is greater than the impact of the current income on the consumption expenditure,the residents' consumption inertia is one of the main factors influencing the consumption expenditure.