《金融计量学》习题及习题答案-上海财经大学.doc

诚实考试吾心不虚 ，公平竞争方显实力， 考试失败尚有机会 ，考试舞弊前功尽弃。 上海财经大学《 Financial Econometrics 》课程考试卷一 课程代码 课程序号 姓名 学号 班级 题号 一 二 三 四 五 六 七 八 九 十 总分 得分 Part 1 Term Explanation （20 marks） 1．White Noise 2．RandomWalk 3．Akaike Information Criterion 4．Jarque-Bera Statistic 5．Chow Test Important Point： 1．White Noise：White Noise is the special case of stationary stochastic process. We call a stochastic process purely random or white noise if it has zero mean, constant variance and is serially uncorrelated. 2．RandomWalk: Random walk means that the stochastic process is nonstationary and value of this period is highly related to the past values. For example, the stock price today may equal the yesterday’s price plus a random shock. Random walk without drift can be expressed as 3．Akaike Information Criterion: AIC provide a way to select the better regression model among several models by comparing their forecast performance. The lower the AIC, the better the forecast performance will be. AIC will also be used to determine the lag length in ARDL approach. 4．Jarque-Bera Statistic: The Jarque-Bera test is the test of normality. We first calculate the skewness and the kurtosis, and it is also based on the residual of the regression. The Jarque-Bera Statistic=, where S is the skewness and K is the kurtosis, n is sample size, and for normal distribution, S=0, K=3, if JB statistic is not significantly different from zero, p value is quite low, we reject the null hypothesis that the residual is normally distributed. 5．Chow Test: The test of structural change of the regression. The estimate of the parameter of the regression may not retain the same through the entire time period; we use the Chow test to test whether the relationship is stable and find the break point. It develop the F statistics=, the null hypothesis is the regression is stable. Part 2 Explain main purpose(s) of constructing following two models and making comments on the em