n
A researcher analyses the effect of education on hourly wages with a random sample of 800 individuals from a particular country in 2018, by estimating the following equation: In(wage ) = Bo + Breduc; + us where In(wage,) and educ, are the natural logarithm of hourly wages and years of schooling of individual i respectively. (a) Suppose the data does not tell us the currency in which the hourly wages are denominated. Comment on whether this is a problem or not for interpreting $1. [Maximum 100 words] The researcher considers a second model with two new explanatory variables, exper; and mother_educ, which are the years of working experience and years of mother's education of individual i respectively: In(wage,) = Bo + Preduct + Byexpen + Bymother_educ + e; After estimating two alternative models, they obtain the following estimation outputs: In(wage) = 5.0 + 0.09educ TSS = 45000 (1) (1.0) (0.01) ESS = 13500 In(wage) = 5.0 + 0.05educ + 0.03exper + 0.06mother_educ TSS = 45000 (2) (1.0) (0.01) (0.01) (0.01) ESS = 27000 Standard errors are in parentheses. (b) At the 5 percent significance level, test whether the second model explains the sources of variations of In(wage) better than the first model. [Maximum 200 words] (c) Why do you think the researcher includes mother's education level as a regressor? Explain using appropriate economic and econometric terminology. [Maximum 150 words] (d) Comment on whether the observed decline in the estimated value of B, in estimation output (2) is sensible or not. [Maximum 150 words] (e) After the researcher looks at the related literature, they also want to include the gender of individuals as an additional explanatory variable. Write down and explain the new regression model. How you would interpret the coefficient on the gender variable? [Maximum 200 words]