P1.T2.321. Univariate linear regression

Nicole Seaman

Director of CFA & FRM Operations
Staff member
Subscriber
Questions:

321.1. Using Stock & Watson's CPS08 dataset, we regressed wage and age data for 7,711 workers. The following is the resulting regression model of average hourly earnings (AHE) against worker's age:

P1.T2.321.Univariate_linear_regression_001.png


Are the slope and intercept coefficients significant with 95% confidence?

a. Neither is significant with 95% confidence
b. The intercept (1.0823) is significant, but the slope is insignificant
c. The slope (0.6050) is significant, but the intercept is insignificant
d. Both are significant with 95% confidence

321.2. Samantha Xiao is trying to get some insight into the relationship between the stock returns of Facebook (FB) and the NASDAQ-100 (QQQ). Using historical returns, she estimates the following:
  • The mean return for FB is 9.0% with volatility of 30.0%
  • The mean return for QQQ is 6.0% with volatility of 20.0%
  • The correlation coefficient between FB and QQQ is 0.40
If she uses the same data to regress FB returns (dependent variable) against QQQ returns (independent variable), which of the following models will she obtain? (note: variation on GARP's 2013 Practice Exam Part 1, Question 8)

a. return(FB) = 1.00% - 0.10*return(QQQ) + e
b. return(FB) = 3.20% + 0.30*return(QQQ) + e
c. return(FB) = 5.40% + 0.60*return(QQQ) + e
d. return(FB) = 7.80% + 0.90*return(QQQ) + e

321.3. Using Stock & Watson's dataset (Growth), we performed a univariate regression of average annual percentage GDP growth (Growth, the dependent regressand) against average number of years of schooling of the country's adult residents (YearsSchool, the independent regressor) for 65 countries. The regression model is given by the following:

P1.T2.321.Univariate_linear_regression_002.png


Each of the following is true about this regression model EXCEPT which is false?

a. The coefficient of determination, R^2, is approximately 0.1094
b. The 95.0% confidence interval for the "true" slope coefficient is approximately 0.158 to 0.336
c. SER = 1.80 = SQRT(SSR/df) = SQRT(205.1/63)
d. The correlation coefficient, r, is approximately 0.331

Answers:
 
Top