英语论文网

the result, if the non-linear combinations of the independent variables have any power in explaining the dependent variable, means that the model is mis-specified.

Consider the model

? = E {y | χ } = βχ

The Ramsey test is used to test whether the (β1χ)2, (β2χ)3…,(βk-1χ)k has any power in explaining y. The Ramsey test is executed by calculate the following linear regression

? = βχ + β1?2 +…+ βk-1?k + ε

After examine the test, the means of the F-test is to determine whether β1 through βk-1 are zero. If the null hypothesis reveals that all regression coefficients are zero, means that the null hypothesis cannot be reject, the Ramsey test is unable to detect any misspecification. If the null hypothesis is rejected, means that the model is misspecification.

3.3.9 Jarque-Bera Test of Normality

Jarque-Bera test of normality is used to test the normally distributed. It is large-sample or an asymptotic test and based on the OLS. The test first calculates the skewness and kurtosis measures of the OLS residuals.

JB = n

Where the n = sample size, S = skewness coefficient, and K = kurtosis coefficient. The normally distributed variable, S is zero and K is three. Hence, the Jarque-Bera test of normality is a test of the joint hypothesis that S and K are zero and three, respectively. Therefore, the value of the Jaque-Bera statistic is expected to be zero.

For the null hypothesis the residual is normally distributed, asymptotically (i.e., in large samples) the Jarque-Bera statistic gives the chi-square distribution with two degree of freedom showed by Jarque and Bera (Gujarati 2003) For the alternative hypothesis the residual is not normally distributed. At 5 significant levels, computed p value is less than Jarque-Bera statistic, we can reject the null hypothesis that the residual is not normally distributed whereas computed p value is more than Jarque-Bera statistic, we do not reject the null hypothesis that the residual is normally distributed.

CHAPTER 4: RESEARCH RESULTS AND INTERPRETATION

4.1 Introduction

This chapter consists of the results and interpretation of the relationship between the unemployment rate and the economic condition in United State. First of all, this chapter will consist of the results from the unit root test and cointegration test. Next, the the estimated regression model results and interpretation will be shown by using ordinary least square (OLS) method on the second part of this chapter. Besides that, multicollineartity, autocorrelation, heteroscedasticity and misspecification errors will also be tested.

4.2 Stationary Test

Table 1 below showed that the stationarity results of Augmented Dickey-Fuller (ADF) unit root test. In examining the stationary, Augmented Dickey-Fuller unit root test has been used in executing the stationary test for all variables to test the level and first different. The results of Augmented Dickey-Fuller unit root test are presented in Table 1 which includes test constant and linear trend.

H0: p = 0 (the variable is non stationary or a unit root process)

H1: p ≠ 0 (the variable is stationary or non unit root process)

Table 1: Results of Augmented Dickey-Fuller unit root test (level)

Level

_______________