SAS Takehome Final Exam Econometrics I

1. Using 2021 General Social Survey data, do the following:

a. Specify an econometrics model where family income in 2021 dollars

(????????????????????????????????????????????????) is your dependent variable of interest; you will need to add at least two additional independent variables to your model in addition to the following covariates: ????????????????????????????????????????????????, ????????????????????????????????????????????????, ????????????????????????, ???????????????????????????????? and ?????????????????????????????????????????????????. All variables must be indexed appropriately. Qualitative variables should be specified appropriately. All variables used should also be included in a summary statistics table.

b. Estimate a regression using OLS based on the econometrics model you specified above. Produce an econometrics model with the estimated

coefficients in place of the parameters (with standard errors below).

Interpret all coefficients, being sure to include statistical significance and

econometric significance in your interpretations. Be sure to recode and

transform variables as appropriate!

c. Test your model for functional form misspecification. What do you find?

d. Test whether ???????????????????????????????????????????????? and ???????????????????????????????????????????????? are jointly significant. What do you find?

e. Add an interaction term between ???????????????????????? and ????????????????????????????????. Interpret the coefficients for

the main effects and the interaction effects.

f. Test for heteroskedasticity. What do you find?

g. Estimate the regression from part (e), this time specifying heteroskedastic

robust standard errors. Interpret all coefficients, being sure to include

statistical significance.

h. Estimate the regression from part (g), this time transforming the dependent

variable. ????????????????????????????????: ???????????????????????? ????????????????????????????????????????? ???????????????? ???????????????????????????????????????????????????????????????????????????????? ???????? ???????????????????????? − ???????????????????????????????????????? ????????????????????????????????????????????????????????????????????????????????????????????????????????.

Interpret all coefficients, being sure to include statistical significance.

i. Is assumption MLR.6 closer to be being satisfied for the level-level or loglevel specification?

j. Create a table with four columns: each column should represent one of the

four sets of estimates you generated. Be sure to include standard errors and

asterisks indicating statistical significance, as well as rows at the bottom to

indicate if SE were robust, ????????, and ????????2. Your table should include all

appropriate notes which allow readers to understand what is included in the

table.

k. Are we estimating causal effects? Why/why not?

2. Using Wave 5 National Longitudinal Study of Adolescent to Adult Health data, do the

following:

a. We are interested in estimating the degree to which environment versus

genetics impacts current health. Given the variables made available to you:

i. identify proxies for “genetic health” or “biological predisposition”,

ii. identify environmental characteristics that potentially influence

current health (e.g. income),

iii. write out your econometric model, making sure to index all variables

appropriately, and create a summary statistics table that includes all

variables used, and

iv. estimate an OLS regression where you regress current health

(????????5????????????????1) onto the independent variables (RHS) that you identify

above. You should use a minimum of 6 RHS variables. Don’t forget –

you will need to create new variables, recode, or transform some of

the variables as appropriate! For example, if you choose to include

race/ethnicity, you should create a new variable (let’s call it

“race_eth”) coded 1 “white” 2 “Black” 3 “Hispanic” and 4 “other”; or if

you choose to include educational attainment, this should be coded in

a way that accounts for the diploma effect. Be sure to specify

qualitative variables appropriately in the estimation.

v. Report what you find by interpreting the coefficients, including the

statistical significance.

b. Perhaps we believe that there is a differential effect of some environmental

factor by sex at birth. How do you estimate this? Run an OLS regression that

expands on your previous specification and interpret the coefficients of

interest.

c. We may be concerned that we have omitted variables bias due to regional

variation in economic or environmental factors. We can attempt to deal with

this by adding region fixed effects. Expand the specification from part (b)

above to include region fixed effects. Interpret those coefficients.

d. Test for heteroskedasticity. What do you find?

e. Estimate the regression from part (c), this time specifying heteroskedastic

robust standard errors. Interpret all coefficients, being sure to include

statistical significance.

f. Create a table with four columns: each column should represent one of the

four sets of estimates you generated. Be sure to include standard errors and

asterisks indicating statistical significance, as well as rows at the bottom to

indicate presence of fixed effects, if SE were robust, ????????, and ????????2. Your table

should include all appropriate notes which allow readers to understand what

is included in the table.

g. Should we be concerned about endogeneity in our final specification?

Why/why not? Be specific.

pur-new-sol

Related Questions