Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
IB93F0 Research Methodology – Individual Assignment 1
Students must do all questions of the three parts using the MATLAB. Any other computer software
package (e.g., STATA, MS Excel, etc.) will NOT be accepted. Students must submit the answer, the
MATLAB script (‘.m file’), and the used dataset so the marker can run submitted codes to check with
written answers.
IMPORTANT: Students must not share their MATALB codes. Plagiarism will be seriously dealt with
when the answer/script are marked, and will give ZERO mark.
Marking Criteria: Marking will be based on the overall quality and accuracy of answers and
MATAB codes: 70+ (distinction), 60+ (merit), 50+ (pass), and 50- (fail)
Word Limit: 1,000 words
IB93F0 Research Methodology – Individual Assignment 1
2
Part 1. Asset Pricing Methodology
Objective: Compute the average return and Fama-French alphas of book-to-market (BTM)
portfolios.
Data to Download:
(1) Stock prices, returns, and the number of shares outstanding from the CRSP database
(2) Book value of equity from the CRSP/Compustat Merged (CCM) database
(3) Fama-French three factors from the CRSP database
MATLAB Task:
1. Construct BTM quintile portfolios.
Choose any 20 stocks on your discretion for the five year’s period: July 2011 to June
2016
Construct five portfolios (portfolios 1 to 5) by sorting stocks at the end of June of
each year based on the BTM ratio.
2. Calculate the monthly returns for BTM quintile portfolios and a long-short portfolio.
Report the average return and its t-statistics
Report the FF3 alpha and its t-statistics
Expected Output:
1. Present the table of portfolio returns and alphas in the form as below:
PF 1
(Low)
PF 2 PF 3 PF 4 PF 5
(High)
PF 5 –
PF1
Avg Return
t-statistics
FF3 Alpha
t-statistics
2. Explain whether your results in Problem 1 are consistent with the existence of value
premium.
IB93F0 Research Methodology – Individual Assignment 1
3
Part 2. Event Study Methodology
Objective: Compute the cumulative average abnormal returns (CAAR) around the S&P 500
index addition/exclusion.
Data to Download:
(1) Stock returns from the CRSP database
(2) Fama-French three factors from the CRSP database
MATLAB Task:
1. Compute the CAAR of the added firm for the 5 days before- and 5 days after the
addition.
Consider the following firms among others that were newly added to the market index
on 20/12/2010:
Company Name Ticker PERMNO Addition Date
Netflix NFLX 89393 20/12/2010
Cablevision CVC 68857 20/12/2010
F5 Networks FFIV 86964 20/12/2010
2. Compute the CAAR of the excluded firm for the 5 days before- and 5 days after the
exclusion.
Consider the following firms among others that were excluded from the market index
on 20/12/2010:
Company Name Ticker PERMNO Exclusion Date
New York Times NYT 47466 20/12/2010
Eastman Kodak EK 11754 20/12/2010
Office Depot ODP 75573 20/12/2010
Expected Output:
1. Present the table of the CAAR for (−5, 5) in the form as below:
Index Addition Index Exclusion
CAAR for (−5, 5) t-Statistics CAAR for (−5, 5) t-Statistics
2. Plot the graph of the CAAR for (−5, 5) for the index addition/exclusion: x-axis for days
from -5 to +5, and y-axis for CAARs.
3. Interpret your results in Problems 1 and 2 to see whether economic values are created by the
market index addition/exclusion.
IB93F0 Research Methodology – Individual Assignment 1
4
Part 3. Panel Data Methodology
Objective: Estimate the effect of AFDC benefit levels on birth rates.
Data:
The provided data set contains information on the level of AFDC benefit in each of the United
States from 1970 to 1996 (excluding 1995), as well as information on birth rates at the same
level from 1973 to 1992. AFDC is the acronym for the Aid to Families with Dependent
Children program, which was the cash transfer program for single mothers (and, in some states,
two parent families with both parents out of work) from the 1930s to the mid-1990s. A large
literature examines the effect of AFDC benefit levels on various outcomes such as birth rates,
out-of-wedlock birth rates, marriage rates, divorce rates and so on.
Variable Definitions
FIPS = Federal Information Processing System = variable indicating the state. These
values are called “FIPS codes”.
YEAR = The last two digits of the calendar year corresponding to the observation.
LAFDC = The natural log of real monthly AFDC benefits for a family of four with
no other income deflated using the CPIU, base period 1982-1984.
LWB2024 = The natural log of the birth rate for white women ages 20-24 in a given
state and year. Births occurring in the first nine months of a given calendar year are
attributed to the previous calendar year.
MATLAB Tasks and Expected Output
1. Estimate a simple OLS linear regression of the log of the birth rate on the log of real
AFDC benefit levels ignoring the panel nature of the data. Interpret the resulting
estimates. For a state with the mean benefit level in 1980, how much would a $100 per
month increase in benefit levels change the birth rate? Is this effect plausible in light of
the simple economics of the problem?
2. Describe and interpret the assumptions required for the estimates from the previous
question to have a causal interpretation.
3. Estimate the fixed effects model using dummy variables for states. Interpret your
results and compare them to the results from Problem 1.
4. Estimate the fixed effects model using dummy variables for states and years. Interpret
your results and compare them to the results from Problem 3.