a) Design a simple econometric project to identify the factors that affect the demand for a good or service of your preference. Estimate the
significance of these factors using multiple regression analysis (at least two explanatory variables) and using at least 20 data points. Use a full range of econometric techniques in your project including estimating your model using absolute values and natural logs of these values.

Interpret the coefficients, their individual statistical significance, joint significance, and model goodness-of-fit in order to evaluate the usefulness of the econometric techniques in identifying the factors that affect demand.

b) Estimate the demand equation in (a) with only one explanatory variable and compare your results with those in (a) identifying the consequences of this type of specification error.

Learning outcomes assessed: a, b, c

Question

a) Give two examples of economic variables that can introduce a multicollinearity problem into your OLS estimation and two other examples for heteroscedasticity.

b) Discuss the problems and practical issues both multicollinearity and heteroscedasticity may create on the variables you have chosen in (a).

c) Briefly explain the possible solutions you could apply to resolve the issues identified in (b).

Question

The following equation is estimated to assess the effects of gender and education on wages (standard errors in brackets):

ln(wage) = 0.389 - 0.227 female + 0.082 educ - 0.0056 female . educ
 
                          (0.119)  (0.168)             (0.008)          (0.0131)
 
n = 526, R2  = 0.441

Where wage is the wage in dollars per hour, female is a dummy variable indicating female gender (1) or otherwise (0), and educ  is the years of education.

 a) Interpret the equation and the goodness-of-fit.

 b) Test the individual significance of the coefficients for female, educ, and the interaction effects at the 1%, 5%, and 10% levels.

 c) Is this estimation likely to be affected by autocorrelation? Explain why.