Assignment 2

  1. 1a

    Level-level model (wages vs education). Consider the R output (from the newdata object of HW1) of the regression of wage_per_hour against schooly.

    Report the estimation results as a formula and interpret the slope parameter.

  2. 1b

    The scatterplot in the assignment includes the estimated level-level model. Do you think the relationship looks linear?

  3. 1c

    For the estimated level-level model, what is the predicted hourly wage for an individual with 11 years of education? What about 17 years? What about no education?

  4. 2a

    Log-level model. The new variable ln_wage is defined in the assignment as the natural log of hourly wage, with a follow-up step that handles -Inf values.

    What does the R code do? Why is it necessary to include it before estimating the log-level model?

  5. 2b

    The scatterplot in the assignment shows log hourly wage against years of education. Does the relationship look linear?

  6. 2c

    The regression of ln_wage against schooly is reported in the assignment. Interpret the slope.

  7. 2d

    For the regression in (c), what is the predicted hourly wage for an individual with 11 years of education? What about 17? What about no education?

  8. 2e

    For the regression in (c), what is the marginal effect of a year of education on hourly wage at the sample average hourly wage?

    (This question is tricky. Hint: recall that the regression passes through the sample averages, and that log⁡(xˉ)≠log⁡(x)‾\log(\bar x) \neq \overline{\log(x)}.)

  9. 3a

    Sampling distribution of the OLS estimator. Show that the variance of the OLS intercept estimator is equal to

    Var⁡(β^0)=σu2[1N+Xˉ2∑i(Xi−Xˉ)2]\operatorname{Var}(\hat\beta_0) = \sigma_u^2\left[\frac{1}{N} + \frac{\bar X^2}{\sum_i (X_i - \bar X)^2}\right]

    Make sure to explain each step.

  10. 4a

    Modeling the money supply in R. Use money_v2026 (monthly US data, 1982 to 2024, N=516N = 516): MONEY (currency in circulation, billions of current dollars), INTEREST (federal funds effective rate, in %), CPI (0 to 100).

    Preliminaries. Import the data and summarize the variables in the dataset.

  11. 4b

    Create the variable MONEY_REAL, which measures the currency in circulation in units of real billions of dollars (hint: for each year divide MONEY by CPI to convert the nominal money supply to the real money supply), and summarize the new variable.

  12. 4c

    Plot the interest rate (x-axis) against the real money supply (y-axis).

  13. 4d

    Does the relationship look linear? Explain.

  14. 5a

    The log-log model. According to the theory of money supply the following model holds:

    ln⁡(Mi)=α+βln⁡(ri)+ui\ln(M_i) = \alpha + \beta \ln(r_i) + u_i

    where MM is currency in circulation adjusted to real dollars (billions) and rr is the interest rate.

    Create the variables for the money supply model: ln_M (the natural log of the variable from 4b) and ln_r (the natural log of the interest rate).

  15. 5b

    Estimate the money supply model parameters. Report the results as a formula.

  16. 5c

    What is the interpretation of the slope estimate?

  17. 5d

    What is the predicted amount of real money in circulation when the interest rate is 10%?

  18. 5e

    What is the predicted amount of real money in circulation when the interest rate is 1%?

  19. 5f

    What is the marginal effect of the interest rate on the money supply (at the point where both are at their sample average)? Interpret this result.

  20. 5g

    What is the rate of return at the point where both the interest rate and the real money supply are at their sample average? Interpret this result.

  21. 6a

    Classical assumptions of OLS. Plot the log real money supply against the log interest rate and add the estimated regression line. Does the relationship look linear? Explain.

  22. 6b

    Based on your plot, what (if anything) can you say about assumption A2? Explain.

  23. 6c

    Based on your plot, what (if anything) can you say about assumption A4? Explain.

Solutions are Alfie's submitted work, not an official key.

Related: Lec_12-Heteroskedasticity