1. Suppose we estimated the following linear probability model of labor force participation among women: inlf=...

  
 
 
 
 

Transcribed Image Text:

1. Suppose we estimated the following linear probability model of labor force participation among women: inlf= -5.5+0.50debt - 0.357kidbelow5 +0.030kidbelow5. debt (0.005) (1.54) (0.0014) (0.121) n = 753, R² = 0.264 Where inlf-1 if a woman answered yes to working outside the home, and zero otherwise. We assume labor force participation depends on the following variables: debt (in dollars) and kidbelow5 (a dummy 1 if the woman has a child below the age of 5). a.) Interpret the coefficient estimates on kidbelow5. What is the adjusted R2 for this model? Make sure to write out the formula you use for your calculation. (6 points) b.) Find the level of debt such that the predicted labor force participation of a woman with a child below 5 is equal to the labor force participation of a woman without a child below 5. What is the predicted labor force participation at this level of debt? Show your work. (7 points) c.) What happens to the regression output if we make the following change to the model above: We replace the dependent variable, inlf, with outlf, which is a dummy variable equal to 1 if the women is out of the labor force and zero otherwise. Make sure you specify what happens to the intercept, coefficient estimates, standard errors on the intercept and coefficient estimates, and R2. (7 points) 2. Consider a model of retirement contributions for individuals in Iowa. You observe each individual's contribution to their retirement, contrib, their earnings, earns, their age, age, and whether the individual lives in the south of Iowa (versus the North), south. The model is: contrib= 8 + 8,earns, + 8₂age, + 83south, + u a.) You suspect the above model suffers from heteroskedasticity. Describe how you would test for heteroskedasticity using the Bruesch-Pagan test. Be sure to state the null and alternative hypothesis, write down your test statistics, and indicate how many degrees of freedom your test statistic has. (7 points) b.) One possibility to test for heteroskedasticity is to combine features of the Breusch-Pagan and White test for heteroskedasticity by regressing u² on earns, age, south, contrib². Do you expect the R² from this regression be larger or smaller than the R² that you would obtained from part (a.)? Explain why. Would it make sense to include contrib as a regressor in the combined Breush-Pagan and White test? Explain why. (6 points) c.) Suppose that u = √earns e where: E(e) = E(elearns, age, south) = 0 and Var(e) = Var (elearns, age, south) = o2, which is a constant. Does the zero conditional mean assumption hold? What is Var(ulearns, age, south)? (7 points) 1. Suppose we estimated the following linear probability model of labor force participation among women: inlf= -5.5+0.50debt - 0.357kidbelow5 +0.030kidbelow5. debt (0.005) (1.54) (0.0014) (0.121) n = 753, R² = 0.264 Where inlf-1 if a woman answered yes to working outside the home, and zero otherwise. We assume labor force participation depends on the following variables: debt (in dollars) and kidbelow5 (a dummy 1 if the woman has a child below the age of 5). a.) Interpret the coefficient estimates on kidbelow5. What is the adjusted R2 for this model? Make sure to write out the formula you use for your calculation. (6 points) b.) Find the level of debt such that the predicted labor force participation of a woman with a child below 5 is equal to the labor force participation of a woman without a child below 5. What is the predicted labor force participation at this level of debt? Show your work. (7 points) c.) What happens to the regression output if we make the following change to the model above: We replace the dependent variable, inlf, with outlf, which is a dummy variable equal to 1 if the women is out of the labor force and zero otherwise. Make sure you specify what happens to the intercept, coefficient estimates, standard errors on the intercept and coefficient estimates, and R2. (7 points) 2. Consider a model of retirement contributions for individuals in Iowa. You observe each individual's contribution to their retirement, contrib, their earnings, earns, their age, age, and whether the individual lives in the south of Iowa (versus the North), south. The model is: contrib= 8 + 8,earns, + 8₂age, + 83south, + u a.) You suspect the above model suffers from heteroskedasticity. Describe how you would test for heteroskedasticity using the Bruesch-Pagan test. Be sure to state the null and alternative hypothesis, write down your test statistics, and indicate how many degrees of freedom your test statistic has. (7 points) b.) One possibility to test for heteroskedasticity is to combine features of the Breusch-Pagan and White test for heteroskedasticity by regressing u² on earns, age, south, contrib². Do you expect the R² from this regression be larger or smaller than the R² that you would obtained from part (a.)? Explain why. Would it make sense to include contrib as a regressor in the combined Breush-Pagan and White test? Explain why. (6 points) c.) Suppose that u = √earns e where: E(e) = E(elearns, age, south) = 0 and Var(e) = Var (elearns, age, south) = o2, which is a constant. Does the zero conditional mean assumption hold? What is Var(ulearns, age, south)? (7 points)

Answer rating: 100% (QA)

1a The coefficient estimate on kidbelow5 is 0357 The adjusted R2 for this model can be calculated as adjusted R2 1 1R2 n1np1 where n is the number of observations p is the number of predictors and R2 ... View the full answer