You are on page 1of 1

14.

The following regression was run using quarterly data, amounting to


70 observations:

Bt  0.78  0.89 Pt  0.35St


(0.56) (0.78) (0.12)
R  0.76, DW  1.57White(5)  27.2
2

Where B is the demand for brokerage services, P is the price of the


services and S is the total number of brokers and all variables are in
logarithms (standard errors in parentheses). DW is the Durbin-
Watson statistic. White is White’s Test.
i) Comment on the specification of the above model
ii) Compute the t-ratios
iii) Does the above regression suffer from first order autocorrelation?
If so how might this have arisen?
iv) Does the model suffer from heteroscedasticity? Explain
v) Explain how you would test for normality assumption using
Jaque-Bera(JB) statistic
15. A student wanted to assess the effect of households’ income and
household size on food consumption expenditure. He obtained survey
data on weekly households’ food consumption expenditure, weekly
households’ income (both in USD) and household size based on a sample
of ten (10) households. Using STATA he estimated a multiple linear
regression with food consumption expenditure as a dependent variable,
and obtained the following results:

Source SS df MS Number of obs = 10


F( 2, 7) = 273.91
Model 8777.83767 2 4388.91884 Prob > F = 0.0000
Residual 112.162327 7 16.0231896 R-squared = 0.9874
Adj R-squared = 0.9838
Total 8890 9 987.777778 Root MSE = 4.0029

consumption Coef. Std. Err. t P>|t| [95% Conf. Interval]

income .4819976 .0231905 20.78 0.000 .4271607 .5368344


hhsize 3.887307 1.037112 3.75 0.007 1.434927 6.339686
_cons 15.45484 4.626001 3.34 0.012 4.516086 26.3936

Key: _cons = constant coefficient, hhsize = household size, coeff = coefficient


estimate, Std. Err. = Standard errors, df = degrees of freedom.
i. Provide a report on these results. Make sure that your report
includes interpretation and explanation on coefficient estimates and
their magnitude, t-statistic, p-values, standard errors and R-
squared.

You might also like