An Empirical Analysis On Premium Income Forecast Of China Based On The Theory Of Structural Breaks

Premium income is the fee that insurance companies charge from policy-holders,whereas the insurance company also need to perform the corresponding obligations stipulated in the insurance contract.Since the policy of reform and opening up,China's financial market is developing at a very high speed.As one of the three pillar industries,insurance industry go through an obvious growth,yet inferior by comparison when contrasting with the banking and securities industry.On the one hand,premium income is an important index for a country or a region to measure the insurance industry development level.On the other hand,premium income visually reflects an insurance company's business scale and its ability to directly operate the fund.To a great extent,premium income affects an insurance company's future decisions.Therefore,predicting the possible number of future premium income and the accuracy of the prediction are particularly crucial for an insurance company's decision,let alone the whole insurance industry.It would also influence the government's guidance to insurance industry in the future.This paper first defines and describes the premium income in our country,and then introduces the structure break theory,which includes structure break estimation model,assumption,limiting distribution and three methods of structure break inspection,as well as stationary process that contains the structure break trend.On this basis,1980-2012 China's premium income data is chose as the research object,respectively establishes regular time series ARIMA model and stationary process that contains the structure break trend.Through the test and comparative analysis,finally selects ARIMA(2,1,2)as the regular time series model.Although the model fitting effect is not ideal,its prediction error rate is too small,which made the predicted results acceptable.Then utilizes the Exp-LM test to determine the structure mutation point location and model regression to confirm the structure mutation point type.From which,it is found that the premium income time sequence in our country mutates at point 1998,and this is a slope mutations which along with a trend.Further regresses residuals and makes AOADF inspection to ascertain the residual and the stability of the model,this paper finally establishes the complete model of premium income includes the structure break and trend stationary.Then use the model to forecast the premium income in our country,model fitting effect and prediction results are better.By comparison with the conventional time series ARIMA model and trend stationary model which contains structure mutation,it was discovered that the trend stationary model which contains the structure mutation is superior to the conventional time series model.