Scribd
Upload
2 views

Assignment 3( QM)

The document outlines an assignment involving regression analysis using specified datasets and models, with various tasks related to estimating regression coefficients and testing hypotheses. It includes questions about the significance of variables, optimal class sizes, and the implications of findings in the context of SAT scores, stock returns, and campaign expenditures. Additionally, it addresses the conditions for good instrumental variables and the derivation of estimates in a simple regression model.

Uploaded by

wanvyingc
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

Assignment 3( QM)

The document outlines an assignment involving regression analysis using specified datasets and models, with various tasks related to estimating regression coefficients and testing hypotheses. It includes questions about the significance of variables, optimal class sizes, and the implications of findings in the context of SAT scores, stock returns, and campaign expenditures. Additionally, it addresses the conditions for good instrumental variables and the derivation of estimates in a simple regression model.

Uploaded by

wanvyingc
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Assignment 3

Deadline: Feb 5
Submit your work in class.

1. Using the data in GPA2.dta, the following regression model is estimated:

𝑠𝑠𝑠𝑠𝑠𝑠 = 𝛽𝛽0 + 𝛽𝛽1 ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝛽𝛽2 ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝛽𝛽3 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 + 𝛽𝛽4 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 + 𝛽𝛽5 ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝑢𝑢

The variable 𝑠𝑠𝑠𝑠𝑠𝑠 is the combined SAT score, ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 is size of the student’s high school graduating
class, in hundreds, hsizesq is the square of hsize, 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 is a gender dummy variable equal to one
for females and zero otherwise, 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 is a race dummy variable equal to one for blacks and zero
otherwise, and hsrank is the student’s ranking in graduation class

The estimation report is as follows.


. reg sat hsize hsizesq female black hsrank

Source SS df MS Number of obs = 4137


F( 5, 4131) = 163.00
Model 13244288.1 5 2648857.62 Prob > F = 0.0000
Residual 67129558.2 4131 16250.1956 R-squared = 0.1648
Adj R-squared = 0.1638
Total 80373846.3 4136 19432.7481 Root MSE = 127.48

sat Coef. Std. Err. t P>|t| [95% Conf. Interval]

hsize 24.62948 3.672897 6.71 0.000 17.42863 31.83034


hsizesq -.382122 .5118367 -0.75 0.455 -1.385598 .6213535
female -52.30055 4.020457 -13.01 0.000 -60.18281 -44.41829
black -136.7292 8.684767 -15.74 0.000 -153.756 -119.7023
hsrank -.7944603 .0395439 -20.09 0.000 -.8719877 -.7169329
_cons 1038.577 6.032024 172.18 0.000 1026.751 1050.403

(a) What is the possible reason ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 included in the model?

(b)Is there strong evidence that ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 should be included in the model? From this estimated
regression model, what is the optimal class size?

(c) Holding ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 fixed, what is the estimated difference in SAT score between nonblack females
and nonblack males? How statistically significant is this estimated difference?

(d) Test the null hypothesis 𝐻𝐻0 : 𝛽𝛽1 = 𝛽𝛽2 = 0 against the alternative hypothesis 𝐻𝐻1 : 𝐻𝐻0 is wrong at
the 5% significance level. The estimated restricted regression model is as follows. What is your
conclusion?

(e) Is the Assumption 5 satisfied? How do you know?


2. Regression analysis can be used to test whether the market efficiency uses information in valuing
stocks. For concreteness, let return be the total return from holding a firm’s stock over the four-
year period from the end of 1990 to the end of 1994. The efficient markets hypothesis says that
these returns should not be systematically related to the information known in 1990. If firm
characteristics known at the beginning of the period help to predict stock returns, then we could
use this information in choosing stocks.

For 1990, let dkr be a firm’s debt to capital ratio, let eps denote the earning per share, let
netinc denote net income, and let salary denote total compensation for CEO.

(a) Estimate the following regression, using the data in RETURN.dta,


𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 𝛽𝛽0 + 𝛽𝛽1 𝑑𝑑𝑑𝑑𝑑𝑑 + 𝛽𝛽2 𝑒𝑒𝑒𝑒𝑒𝑒 + 𝛽𝛽3 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 + 𝛽𝛽4 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 + 𝑢𝑢.

(b) Test whether the explanatory variables are jointly significant at the 5% level. Is any explanatory
variable individually significant?

(c) Now, reestimate the model using the log form for netinc and salary:
𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 𝛽𝛽0 + 𝛽𝛽1 𝑑𝑑𝑑𝑑𝑑𝑑 + 𝛽𝛽2 𝑒𝑒𝑒𝑒𝑒𝑒 + 𝛽𝛽3 log (𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛) + 𝛽𝛽4 log (𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠) + 𝑢𝑢.
Do any conclusions from part (a) change?

(d) Should we try to use log(dkr) and log (eps) in the model to see if these improve the fit?

(e) Test the null hypothesis 𝛽𝛽1 = 1.5𝛽𝛽2 against 𝐻𝐻1 : 𝐻𝐻0 𝑖𝑖𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤.

(f) Test the null hypothesis 𝛽𝛽2 = 𝛽𝛽3 = 𝛽𝛽4 = 0 against 𝐻𝐻1 : 𝐻𝐻0 𝑖𝑖𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 .

(g) Overall, is the evidence of predictability of stock returns strong or weak?

3. The following model can be used to study whether campaign expenditure affect election
outcomes:
𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 = 𝛽𝛽0 + 𝛽𝛽1 ln(𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒) + 𝛽𝛽2 ln(𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒) + 𝛽𝛽3 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 + 𝑢𝑢

where 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 is the percentage of the vote received by Candidate A, 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 and 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 are
campaign expenditure by Candidates A and B, and 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 is a measure of party strength for
Candidate A (the percentage of the most recent presidential vote that went to A’s party).

(a) What is the interpretation of 𝛽𝛽3?

(b) Is homoskadasticity assumption met? (Use the 4-step procedure that we discussed in class.)

(c) Use the LM statistic to test for heteroskedasticity and report the p-value.

(d) Do we omit nonlinear (X) terms in the original model? Use the Ramsey RESET to detect
nonlinearities. [Hint: to calculate predicted y, use the command > predict yhat]
4. Consider a simple regression 𝑦𝑦𝑖𝑖 = 𝛽𝛽0 + 𝛽𝛽1 𝑥𝑥𝑖𝑖 + 𝑢𝑢𝑖𝑖 , 𝑖𝑖 = 1,2, . . 𝑛𝑛, where 𝐶𝐶𝐶𝐶𝐶𝐶(𝑥𝑥, 𝑢𝑢) ≠ 0.

(a) State the conditions needed for a good instrumental variable 𝑧𝑧.

(b) Derive the 𝛽𝛽̂1,𝐼𝐼𝐼𝐼 .

(c) Show that 𝛽𝛽̂1,𝐼𝐼𝐼𝐼 is consistent.

(d) Show that 𝑉𝑉𝑉𝑉𝑉𝑉(𝛽𝛽̂1,𝐼𝐼𝐼𝐼 ) is larger than the standard 𝑉𝑉𝑉𝑉𝑉𝑉(𝛽𝛽̂1,𝑂𝑂𝑂𝑂𝑂𝑂 )

You might also like