Econmetrics Help - Run a gregression of C on Y and N and estimate the regression coefficients.?

Question

In a cross-section study of the consumer spending behavior a sample of 10 observations on income (Y), consumption (C) and the number of children in the household (N) results in the following data (data are expressed in $1000).

C Y N

18 25 5

16 18 4

9 12 1

18 22 3

19 24 4

12 16 2

10 10 2

9 7 2

12 14 3

17 22 4

a- Run a regression of C on Y and N and estimate the regression coefficients.

b- Find the relevant statistics and analyze the significance of the coefficients.

c- Test for multi-colinearity between the independent variables.

d- Test for the heteroscedasticity in the regression error.

e- Write a 95% confidence interval for the estimated coefficient of Y and N.

f- Do a hypothesis test (at the .05 level) that the estimated MPC is less than 1.

g- Do a forecasting of consumption for a consumer whose annual income is $45000 and has three children.

It is a little different from the examples in class. Just wondering if anyone can help or point me in the right direction. Thanks a lot.

Answer 1

a -

For CN

Regression coeficent = 2.78571428571429

Intercept = 5.64285714285714

For CY

Regression coeficent = 0,617816091954

Intercept = 3,497126436782

b -

For CN

Standard Error =2.10229466468021

Significance F =0.00110967188312135

For CY

Standard Error =1.18160591092616

Significance F =0.0000102209086660734

c -

Variable... Tollerance...VIF

Y................ 0,333...3,002

N................ 0,333...3,002

e -

95% Confidence Interval

Variable..Lower Bound ........Upper Bound

C............ 0.855758622............ 5.788726899

Y............. 0.248681383............ 0.721127613

N............ -0.366275867............ 1.989196385

Answer 2

I don't get it, are you required to do it BY HAND? If not, try to do this with any statistical software package: once you know the right commands it'll take you a couple of minutes. But as a general comment, whatever coefficients you get, they cannot be considered as consistent, since your sample is not representative. If you need some theoretical background so that you understand what's going on, check out introductory Wooldridge, it gives (in my opinion) an excellent intuitive and well-presented exposition of the issues you have described.