Answer Happy

QUESTION THREE You are investigating the factors determining earnings. Models (1) and (2) are estimated using information on a cross-sectional sample of data on wages for 525 UK workers recorded in 2010. Model (1): Earn = 323.70 +5.15 × Age₁ - 169.78 × Female₁, R2 = 0.13, SER = 274.75 Model (2): In Earn, = 5.44 + 0.015 × Age₁ - 0.421 × Female₁, R2 = 0.17,SER = 0.75 where Earn are weekly earnings in £, Age is measured in years, and Female is a dummy variable, which takes the value of one if the individual is a female, and is zero otherwise. (a) Clearly explain what the estimated values of the slope coefficients in the two models imply about the relationship between each explanatory/independent variable and earnings (you do not need to discuss the intercept/constant coefficient in each model). (15 MARKS) (b) You are unsure which of the two functional forms above is more appropriate, so you use Ramsey's RESET. While performing the RESET tests, two powers of the fitted values (i.e. Ŷ² and ³) are included in the RESET regression. The values of the RESET test F-statistics are 4.05 for model (1) and 2.52 for model (2). Using a 5% significance level, what are the outcomes of the two RESET tests and what do these results imply about the suitability/unsuitability of the functional forms used in models (1) and (2)? Explain how you reach your conclusions. (15 MARKS) Your peer points out to you that age-earning profiles typically take on an inverted U-shape. To test this idea, you add the square of age to your log-linear regression (t-ratios given in parenthesis). In Earn, = 3.04 + 0.147 × Age₁ - 0.421 × Female; - 0.0016 Age?, R2 = 0.28 (3.18) (5.35) (-3.25) (-15.05) (c) Are there strong reasons to assume that this specification is superior to the previous one? (8 MARKS) (d) What are the possible consequences if you omit Age² variable. (12 MARKS)