英语论文网

lihood (ML) estimators. There are the best linear unbiased estimators (BLUE). Besides, the least-square estimators are best unbiased estimators (BUE); it means that they have minimum variance in the entire class of unbiased estimators.

3.3.5 Multicollinearity

Multicollinearity shows the two or more independent variables in a multiple regression model are highly linearly related. The multicollinearity test is perfect if the correlation between two independent variables is equal to 1 or -1. Multicollinearity will occur when there is a strong linear relationship among two or more independent variables.

The equation below is refer the variables is perfectly multicollinear if there exist one or more exact linear relationships among some of the variables.

Estimates for the parameters of the multiple regression equation is

The ordinary least squares estimates include inverting the matrix

XTX

where,

It indicate that if the linear relationship (perfect multicollinearity) is exactly with the independent variables, the rank of X is less than k+1 and the matrix XTX will not invertible.

One of the detection of multicollinearity is used detection-tolerance or the variance inflation factor (VIF) for multicollinearity

where R2j is the coefficient of determination of a regression of explanatory j on all the other explanators. Tolerances of less than 0.20 or 0.10 or a VIF of 5 or 10 and above reveal a multicollinearity problem.

3.3.6 Breusch-Godfrey Serial Correlation LM Test

Breusch-Godfrey Serial Correlation LM test is a test of autocorrelation that is basically allows for nonstochastic regressors such as the lagged values of the regressand; higher-order autoregressive schemes such as AR (1), AR (2), etc and higher-order moving averages of white noise error terms such as t.

Two variable regression models to illustrate the test, regressors can be added to the model and also lagged values of the regressand can be added to the model.

Yt =β1 +β2Xt +ut

The error term ut assume that the pth-order autoregressive, AR (p),

Ut = ptut-1 + ptut-2 + …+pput-p + t.

where t.is a white noise error term.

The null hypothesis H0 can be show as

Ho: p1 = p2 = … = pp = 0 (no autocorrelation)

At 5% significant level, if the computed p value of Chi-square is less than Chi-square tests, do not reject the null hypothesis, meaning that there is no autocorrelation problem. If computed p value of Chi-square is more than Chi-square tests, reject the null hypothesis, meaning that there is autocorrelation problem.

3.3.7 Autoregressive Conditional Heteroscedasticity Test

In econometrics, Autoregressive Conditional Heteroskedasticity (ARCH) model assume that the variance of the current error term is related to the previos one. Autoregressive Conditional Heteroskedasticity model is used to model the time series with time-varying volatility such as stock price.

3.3.8 Specification error

Ramsey Regression Equation Specification Error Test (Ramsey RESET test) is used to examine the specification error. The specification test for the linear regression model. More specifically, it is used to test the specification error in the equation. As