BEEM011-计量经济代写-Assignment 1
时间:2022-12-15
12/14/22, 6:32 PM Assignment 1 December 2022 Student Template - Jupyter Notebook
localhost:8888/notebooks/BEEM011/assignment/Assignment 1 December 2022 Student Template.ipynb 1/5
BEEM011 Assignment December 2022
Instructions
Summative assessment
Direction: This Assignment consists of 3 questions. There are 100 marks in total.
For
each question, 5% of marks are reserved for evidence of best practice
in R coding, marks for sub-questions therefore add up to 95% of the
total marks for each
question.
Answer all the questions and upload your Jupyter Notebook and a PDF copy to the BART submission point.
WRITE YOUR R CODE AND ANSWERS FOR ALL OF THE QUESTIONS IN THIS JUPYTER NOTEBOOK. PLEASE INCLUDE YOUR CANDIDATE NUMBER IN A
COMMENT AT THE TOP OF YOUR CODE.
DO NOT PUT YOUR NAME ANYWHERE IN THE ASSIGNMENT.
Make
sure to include your Jupyter Notebook file (.ipynb) as an upload to
your submission. Please comment on each procedure to explain what you
are doing (or intend to
do). Submission is via BART. Submission of an incorrect filetype will result in the deduction of marks.
Read and answer each of the questions using your own code and words.
Collaboration
with others and plagiarism of other people's code is not permitted.
Presenting someone else's code as your own work is misrepresentation, an
academic
conduct offence.
Download the "investCEO.csv" dataset and save it in your BEEM011 folder
The Data
This
assignment takes inspiration from the work of Malmendier and Tate
(2005) who analyzed Corporate Investment and CEO overconfidence. The
objective was to
understand whether managerial overconfidence is a determinant of corporate investment distortions.
"Overconfident
managers overestimate the returns to their investment projects and view
external funds as unduly costly. Thus, they overinvest when they have
abundant
internal funds, but curtail investment when they require external financing." - Malmendier and Tate (2005, abstract)
Malmendier
and Tate tested their "overconfidence hypothesis", that the investment
of overconfident CEOs would be more responsive to cash flow than that of
non-
overconfident CEOs. To do this they constructed a panel data
set of the personal portfolio and corporate investment decisions of
Forbes 500 CEOs. They classified CEOs
as overconfident if they persistently fail to reduce their personal exposure to company-specific risk.
In
this assignment you will test the overconfidence hypothesis using a
different set of data. The variables included in the dataset are:
year
Year of the observation
firmid
A firm specific identifier
I
Investment (%), defined as firm capital expenditures as a percentage of capital at the beginning of the year
CF
Cash
flow, defined as earnings before extraordinary items plus depreciation
as a percentage of capital at the beginning of the year
MV
Market value of assets over book value of assets at the beginning of the year (%)
S
Percentage of total shares owned by CEO or CEO's family (%)
size
Logarithm of assets (£) at the beginning of the year
CG
Number of outside directors who are CEOs in other companies
TC
Press based overconfidence measure, the number of articles mentioning CEO as "confident" or "optimistic"
TD
A dummy variable equal to 1 if the majority of articles portrayed confidence and 0 otherwise
References
Malmendier, U., & Tate, G. (2005a). CEO overconfidence and corporate investment. The journal of Finance, 60(6), 2661-2700.
Malmendier,
U., & Tate, G. (2005b). Does overconfidence affect corporate
investment? CEO overconfidence measures revisited. European Financial
Management, 11(5),
649-659.
12/14/22, 6:32 PM Assignment 1 December 2022 Student Template - Jupyter Notebook
localhost:8888/notebooks/BEEM011/assignment/Assignment 1 December 2022 Student Template.ipynb 2/5
In [ ]:
Question 1 - Exploring the data
a i)
Load the data and store it in an object called "CEOdata" (2 marks)
In [ ]:
ii)
Explore the data -
What time period does it cover?
How many firms are included in the dataset?
How many observations are in the dataset?
Is the dataset balanced?
(3 marks)
In [ ]:
Provide your answer here
What time period does it cover?
How many firms are included in the dataset?
How many observations are in the dataset?
Is the dataset balanced?
iii)
Plot a histogram of Investment, , and describe its distribution (5 marks)
In [ ]:
Provide your answer here
b i)
Set the seed to your candidate number (1 mark)
Randomly
choose a one year time period from the dataset, save the randomly
selected time period as indext. Report the value of indext (1.5 marks)
Create a subset of the dataset for your chosen time period. Label your new dataset as mydata. (1.5 marks)
In [ ]:
b ii)
Use the subset, mydata, to answer the remaining parts of Q1
Plot
the distribution of Investment, , for CEOs with media portrayal
classified as overconfident, i.e. , and those who are not. Label your
plots carefully (4
marks)
= 1
In [ ]:
Compare and describe the two plots (4 marks)
# Set plot size to 4 x 3
options(repr.plot.width=6, repr.plot.height=4)
CEOdata <-
# Provide your code here
# Provide your code here
# Provide your code here
# Provide your code here
12/14/22, 6:32 PM Assignment 1 December 2022 Student Template - Jupyter Notebook
localhost:8888/notebooks/BEEM011/assignment/Assignment 1 December 2022 Student Template.ipynb 3/5
Write your answers here
c i)
Consider the following regression equation:
Describe the effect of TD in this model specification (3 marks)
= + + + + + + ( × ) +
β
0
β
1
β
2
β
3
β
4
β
5
β
6
Write your answer here
c ii)
Estimate the regression equation using your subset, mydata.
Interpret the results of your estimation (5 marks)
Based on the estimated parameters, write the two regression functions for the cases where TD is 1 and 0. (4 marks)
In [ ]:
Write your answer here
c iii)
Evaluate
whether the regression equation for CEOs portrayed as overconfident ( )
is statistically significantly different to the regression equation for
non-
overconfident CEOs. Discuss your method and conclusion. (4 marks)
= 1
In [ ]:
Write your answer here
c iv)
Write
down the predicted investment percentage for CEOs portrayed as
overconfident and other CEOs at the mean levels of cashflow, market
value, outside
directors and size (4 marks)
Comment on the differences between these predicted values. (2 marks).
In [ ]:
Write your answers here
Question 2
Continue to use mydata, your subset from CEOdata, in question 2.
i)
You decide to estimate a new specification of the model
Write down the two regression equations for CEOs - one for overconfident (TD = 1) and one for non-overconfident CEOs. (4 marks)
= + + + + + × + +
β
0
β
1
β
2
β
3
β
4
β
5
β
6
2
Write your answer here
# Provide your code here
# Provide your code here
# Case 1: Overconfident CEOs
# Case 2: Non-Overconfident CEOs
12/14/22, 6:32 PM Assignment 1 December 2022 Student Template - Jupyter Notebook
localhost:8888/notebooks/BEEM011/assignment/Assignment 1 December 2022 Student Template.ipynb 4/5
ii)
Estimate the model (2 marks)
In [ ]:
iii)
Calculate and discuss the marginal effect of CF at CF = 10, when TD = 0 and when TD = 1. (8 marks)
In [ ]:
Write your answer here
iv)
Test the joint significance of the interaction between CF and TD and the quadratic term . (3 marks)
2
In [ ]:
Write your answer here
v)
Is it meaningful to compare the standard error of the two regression specifications?
Explain your answer. (5 marks)
Write your answer here
Question 3
You should now work with the full panel data set.
i)
Explain why you would want to estimate a model with fixed effects.(4 marks)
Write your answer here
ii)
How
many fixed effects parameters will you estimate in a model with time
and entity fixed effects? Explain your answer (3 marks)
Write your answer here
iii)
Estimate a linear-linear model with time and entity fixed effects. The dependent variable should be Investment, . (5 marks)
In [ ]:
Interpret your results in relation to CEO overconfidence, TD, using the panel data method. (4 marks)
Do you find a statistically significant difference between overconfident and non-overconfident CEOs? (1 mark)
Write your answer here
# Provide your code here
# Provide your code here
# Provide your code here
# Load the plm library
# Estimate the regression
# Test the coefficients
12/14/22, 6:32 PM Assignment 1 December 2022 Student Template - Jupyter Notebook
localhost:8888/notebooks/BEEM011/assignment/Assignment 1 December 2022 Student Template.ipynb 5/5
iv)
Wider
research has found a relationship between the age of CEOs, investments
and cash flow sensitivity (Gupta, 2022; Belenzon et al, 2019).
Explain
whether it would be possible to re-estimate the model including a
variable capturing each CEOs age in each time period. (5 marks)
Explain
how you would test whether age affects the investment-cash flow
sensitivity of CEOs. Describe the model you would estimate and the
relevant test with the
null and alternative hypotheses and relevant degrees of freedom. (8 marks)
References
Gupta, G. (2022). CEO's age and investment‐cash flow sensitivity. Managerial and Decision Economics.
Barba
Navaretti, G., Castellani, D., & Pieri, F. (2022). CEO age,
shareholder monitoring, and the organic growth of European firms. Small
Business Economics,
59(1), 361-382.
Belenzon, S., Shamshur, A.,
& Zarutskie, R. (2019). CEO's age and the performance of closely
held firms. Strategic Management Journal, 40(6), 917-944.
Write your answer here