Theoretical Reserch And Application Of Quantile Spurious Regression And Error Correction Model

Quantile regression is a basic method to use the given variable X to estimate theconditional distribution of the dependent variable Y.It can measure not only theimpact of dependent variables to the center of a distribution, but also measure thistype of impact in extreme cases.It has some advantages that the classical least-squaresregression don’t have.Most of the time series in economic and financial field are not stationary. It mayfaced with the problem of spurious regression if directly make a least squaresestimation.By the use of error correction model,we can avoid spurious regression, atthe same time reveal long-term and short-term relationship between thevariables.Monte carlo experiment shows that quantile regression of non-stationarytime series is also likely to faced with the problem of spurious regression. And thereare few research on this question so far.This paper puts forward the quantile errorcorrection model.This model combines quantile regression method and traditionalerror correction model, and have advantages of the both.It can make a morecomprehensive study on the adjustment mechanism of the error correction model.Thespecific research idea is: first to use the least squares method to estimate thelong-term equilibrium relationship of the variables which are coingrated,second tomake a quantile regression on the error correction equations,and then discuss theadjustment mechanism at each quantile,especially the adjustment mechanism underthe condition of extreme distribution.Followly,the paper use the quantile error correction model to make a empiricalresearch on the function of price dicovery of CSI300stock index futures market andits information transmission with the spot market,and then compare it with the resultswhich rendered by traditional error correction model.Results show that theinformation efficiency of spot market gradually improved with the increase ofquantile.Spot market is better than futures market on information transmission.Themarket adjustment mechanism was overestimated by traditional error correctionmodel.