Eviews6a Karts

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Introduction to Quantitative Methods

Dr Aris Kartsaklas, Email: [email protected] Eviews Class 1

The two-variable regression model


Using the data contained in the les cons_inco.xls and wages_autosales.xls answer all the following questions.

1
1.1

Consumption and Income


Regression
Run a regression of consumption on a constant and on disposable income using the sample 1959:1 to 1995:2. Before estimation we need to specify the sample we want to use. To do so enter the following command: smpl 1959:Q1 1995:Q2. Once the sample is specied, we can estimate the equation using OLS. To do so enter the following command: equation eq01.ls gc=c(1)+c(2)*gyd. Eviews will create the object eq01. Double clicking on eq01 provides the results

1.2

Tests
Derive the t-statistics for the null that the coe cients are equal to 0. To do so use the estimated coe cients and their standard errors. Is the coe cient attached to the constant signicant at the 5% level? Is the coe cient attached to disposable income signicant at the 5% level? Test the null hypothesis that the coe cient attached to the constant is equal to 1.

1.3

R-squared and Analysis of Variance


Draw a graph of the actual-tted-residual decomposition. To do so, click "resids" on the estimation output window:
5000 4000 3000 2000 80 40 0 -40 -80 -120 1960 1965 1970 Residual 1975 1980 Actual 1985 1990 Fitted 1995 1000 0

Show that the square of the S.E. of regression equals the sum of squared residuals (or Error Sum of Squares, ESS) divided by N-k, where N is the sample size and k the number of regressors. Compute the total variation of Y (Total Sum of Squares, TSS). Hint: you can compute it directly, or derive it from the variance of the dependent variable. Compute the explained variation of Y (Regression Sum of Squares, RSS). Hint: you can compute it directly, or derive it from ESS and TSS. Show that the R-squared is equal to the ratio between RSS and TSS.

Wages and Retail Auto Sales

The le wages_autosales.xls contains quarterly data on wages and retail auto sales. 1. Export the data in eviews and run a regression of sales on wages and a constant using the sample 1959:1 to 1995:2. 2. Provide the estimation results in a table such as the one shown in the previous exercise on consumption and income. 3. Comment the results. In particular: Derive the t-statistics for the null that the coe cients are equal to 0. To do so use the estimated coe cients and their standard errors. Is the coe cient attached to the constant signicant at the 5% level? Is the coe cient attached to wages signicant at the 5% level? Test the null hypothesis that the coe cient attached to the constant is equal to 5. Draw a graph of the actual-tted-residual decomposition. Compute the total variation of Y (Total Sum of Squares, TSS). Compute the explained variation of Y (Regression Sum of Squares, RSS). Show that R-squared is equal to 1 minus the ratio between ESS and TSS.

You might also like