Unit Roots Test And Monte Carlo Simulation On The Interval Structural Breaks

The violent impact(such as financial crisis or macroscopic economic systemvariety etc)may cause data generate process (DGP) occurrence structural breaks,but traditional unit root test doesn't have the in?uence which considers a struc-tural breaks. So the structural breaks unit root tests of economic time seriesshould have the value and research of domestic and international scholar. Thispaper analyzes this new structural breaks form, make use of theories to proveand Monte Carlo Simulation research DF statistics of structural breaks unit root.Detection at the DF statistics also convergent. At the same time used of MonteCarlo Simulation experiment to study DF unit root test power and distributionof DFρstatistics then put forward contain a interval structural breaks variableof DF unit root test model, used Monte Carlo Simulation experiment methodsto prove the superiority of the new model.The textual and main contents can begeneralize as follows:1.Chapter 1 introduce research background and result of structural breaksand foundation knowledge of structural breaks unit root test.2. Chapter 2 research interval structural breaks unit root which put forwardby local professor Wang Shao Ping. For constant occur the interval structuralbreaks unit root process this new form that is carried on DF unit root test andtheoretically proved to statistical:Got theorem 2.2.1. This conclusion for DF unit root test provided a theoriesbasis, also studied a work to lay foundation for after time's substantial evidence.3. Chapter 3 used Monte Carlo Simulation to carry on 3 sets of experiment.The first test studied to when structural breaks occur in the di?erent position toDF unit root test output of in?uence. When the position of structural breaksafter the sample data of a half segment, DF unit root test has invalid. Compareddi?erent structural breaks position histograms of DF_ρdistribution, analyzed thereason of test invalid, for resolve a problem to put forward new test model tomake preparation further. The second test studied a di?erent structural breaksextent to output in?uence of the DF unit root test. Find that structural breakextent is more big the power of DF test to be more low. Compared di?erentstructural breaks extent histograms of DF_ρdistribution, analyzed the reason oftest invalid. Test three, aim at to test one problem.For improve the power of DF unit root test, put forward to contain a interval structural breaks variableof DF unit root test model. Use Monte Carlo Simulation to prove the powerof this new model has be improved. Compared di?erent model histograms ofDF_ρdistribution,say further the usefulness of this break form that the contain ainterval structural breaks variable of DF unit root test model aims at this textresearch.4. Finally made summary to the textual work and generalized the problemof interval structural breaks will need to be research. Point out that there arestill have a more controversy to the structural breaks unit root test, so can notbecome one complete tightly science system of from theories to applied.