ECON3208/ECON3291(Arts)
ECONOMETRIC METHODS
PROJECT INFORMATION
Submission details: Students are required to submit a written copy AND an electronic copy of their
project. The electronic document must be a pdf file or a word file. Hard copies of the document
must be typewritten and stapled. Student name(s) and number(s) must be clearly indicated on the
cover page. You must use a cover sheet.
The hard copy of the major project should be deposited in the Economics Box 3 outside the northwest
doors to the ASB, by 5pm on Friday May 25th. The electronic copy of the major project must
be submitted on the course website by 11:59pm on Friday May 25th.
Submission via email or fax will not be accepted. Additional instructions on submission of the
electronic copy will be available on the website. You should keep a copy of all work submitted for
assessment as well as the returned marked assignments.
Late submissions: a penalty of 20% of the value of the project will be deducted for each day (24
hours) or part thereof past the deadline unless an extension is granted. This penalty will apply if
either the hard or the electronic copy of the project is late. Please note that NO extension to the
deadline would be granted.
All electronic copies of essays will be checked for plagiarism on the Turnitin software into which they
are uploaded. See notes on Plagiarism below and also note that the Turnitin software will
automatically check against all other assignments submitted.
Be reminded that the University regards plagiarism as a form of academic misconduct, and has very
strict rules regarding plagiarism.
For UNSW’s policies, penalties, and information to help you avoid plagiarism see:
http://www.lc.unsw.edu.au/plagiarism/index.html as well as the guidelines in the online ELISE
tutorial for all new UNSW students:
http://info.library.unsw.edu.au/skills/tutorials/InfoSkills/index.htm.
If you have any doubts or questions on what constitutes plagiarism and how it will be treated, please
ask me or your tutors.
Group work (Optional):
You may work in groups of at most THREE (absolutely no more!) people for the project. A cover
sheet for the project is posted on the course website. The group members are required to complete
a peer assessment sheet for the project. It is your responsibility to resolve any work load among the
group members. The peer assessment sheet will be used to assess each member’s contribution to
the project. In the event of a big discrepancy between group members’ contribution, the mark
allocated to each member will be an average weighted by their peers’ assessment.
The aim of your project and therefore its title is:
Measuring the monetary returns to education.
Measuring the effect of education on income is a matter of great importance for several decision
processes. The results of such analysis are relevant for government agencies responsible for
compulsory schooling laws, for school districts considering changes in school entrance policies and
also for parents deciding when to enrol their children to school. However, a problem is that
intellectual capabilities, which are usually not observed, not only influence education but also
directly affect income. Smarter students find school less difficult and may choose to obtain more
schooling to signal their high ability. So, even if extra years of education have no effect on income,
people with higher education will on average have higher incomes because of their higher abilities.
In this literature, there are two influential papers by Angrist and Krueger XXXXXXXXXXand Angrist and
Krueger XXXXXXXXXXwho consider the effect of education on income using data sets on individuals from
all states of the United States. Their great contribution was their use quarter of birth to form valid
instruments in order to account for possible endogeneity of the education spell. This choice of
instruments has generated a wealth of papers either questioning their validity or supporting their
results.
In this project, you are required to use the tools learned in Econometric methods and Introductory
Econometrics to make an assessment of the effect of education on earning in the US based on the
observed sample information. In other words, have your say about the findings in Angrist and
Kruger XXXXXXXXXXand join their critics and/or supporters.
This is also a great opportunity for you to get a sense of what it is like to have a real economic
question and a real data with all the challenges that economists and econometricians face when
deciding on a model specification. The aim is to find a “good” estimate(s) of the return to education
using the information available from the census data.
I. Some dot points to guide your analysis:
a) Read the reference papers and do a review of the literature. This will give you a background on
o Why we (governments and therefore researchers) care about the returns to
education and conceptual and empirical challenges to measure such an effect.
o What people tried in the past and why they are still looking for an answer?
o Whether the successes and failures are a matter of availability of data, or difficulty
to come up with a “satisfactory” model that is robust to misspecification
o How to come up with a model specification for the econometric analysis. You must
think about all the information you have in the data and how you can use it in the
analysis. Of course, not all variables are useful but you want to use as much
information as you can. You are free to choose what variables you put on the right
hand side. However, you MUST explain where you started and, at least in general
terms, how you arrived at your preferred model(s). For this, finding a
theoretical/Economic model to guide the econometric analysis is useful.
b) Get to know the data. Do the usual drills:
o summary statistics, plots of different variables of interest to check for patterns,
devide the data into groups and compare the differences in their summary statistics
(means, proportion...)
o know your variables: is the dependent variable special in any sense? What are the
potential regressors? Is there a theory (economic, behavioural or other) to back up
your choice of model specification?
c) Run simple regression to get a first sense of what is going on. Does your regression model satisfy
the Gauss-Markov assumption? Are your estimates BLUE, or CAN? Is your inference valid?
d) Why does Angrist and Krueger XXXXXXXXXXuse an instrumental variable to estimate the model?
Explain the reasons and their implications on the regression coefficients and especially your
estimate of the return to education. Do the instruments provided by these authors solve the
problems?
II. After you finish your econometric analysis and arrived to a satisfactory model with satisfactory
estimates of the returns to education, discuss the following comments from the literature
(Hoogerheide & Van Dijk XXXXXXXXXXabout the findings in Angrist and Kruger:
o “The Angrist-Krueger results on returns to education for the USA are almost completely
determined by data from a few Southern states;”
o “The quarter of birth does not only affect the number of completed years of schooling for
those who leave school as soon as the law allows for it, as these persons usually have
completed 9-13 years of education. Therefore, if one intends to increase the
understanding of the working of the quarter-of-birth instruments, it is a better idea to
focus on differences between states in school entry requirements and/or compulsory
schooling laws for children of age 5-7 than to concentrate on the differences in
compulsory schooling laws for students of age 16-18.”
III. The data: The file MHE7080.dta contains the complete original (real data) Angrist and Krueger
1970 and 1980 extracts, and all cohorts (men born XXXXXXXXXXin 1970, men born XXXXXXXXXXin 1980,
and men born XXXXXXXXXXin XXXXXXXXXXThis data set is obtained from “Mostly Harmless Econometrics
Angrist Data Archive”. The data was constructed by the authors in Angrist and Krueger using the
1970 and 1980 US census data. This provides a sample of 1,063,634 observations and 27
variables. The following is a description of the data from Angrist and Krueger (1991):
`` ... our extract combines data from three separate public-use files: the State, County group, and
Neighborhood files. Each file contains a self-weighting, mutually exclusive sample of 1 percent of the
population (as of April 1, 1970), yielding a total sample of 3 percent of the population. The data sets we use are
based on the question-naire that was administered to 15 percent of the population. The sample consists of
white and black men born between XXXXXXXXXXin the United States. Birth year was derived from reported age
and quarter of birth. In addition, we excluded any man whose age, sex, race, veteran status, weeks worked,
highest grade completed or salary was allocated by the Census Bureau. Finally, the sample is limited to men
with positive wage and salary earnings and positive weeks worked in 1969. Weekly earnings is computed by
dividing annual earnings by annual weeks worked. Annual earnings is reported in intervals of $100. This
variable was converted to a continuous variable by taking the average of the interval endpoints. Weeks worked
is reported as a categorical variable in six intervals, and was also converted to a continuous variable by taking
the mean of interval endpoints. Nine region dummies were coded directly from the Census Regions variable in
the Neighborhoods 1 percent sample, from state of residence in the State 1 percent sample, and from county
locations in the County Group file. If county groups straddled two states, the counties were allocated to the
region containing the greatest land-mass of the county group. The education variable is years of schooling
completed. The marital status variable equals one if the respondent is currently married with his spouse
present. The SMSA variable equals one if the respondent works in an SMSA XXXXXXXXXXCensus. The 1980 Census
micro data are documented in Census of Population and Housing, 1980: Public Use Microdata Samples
[Washington, DC: U. S. Department of Commerce, 1983]. Our extract is drawn from the 5 percent Public Use
Sample (the A Sample). This file contains a self-weighting sample of 5 percent of the population as of April 1,
1980. The extract we created consists of white and black men born in the United States between XXXXXXXXXX.
Birth year was derived from reported age and quarter of birth. We excluded respondents whose age, sex, race,
quarter of birth, weeks worked, years of schooling, or salary was allocated by the Census Bureau.”
IV. STATA support. I have provided some do files to help you read the data and help you start up
the analysis.
V. References
? Angrist, J.D., Krueger, A.B., 1991. Does Compulsory School Attendance Affect Schooling and
Earnings? The Quarterly Journal of Economics, Vol. 106, No. 4 pages XXXXXXXXXX.
? Angrist, J.D., Krueger, A.B., 1992. The effect of age at school entry on educational
attainment: an application of instrumental variables with moments from two samples.
Journal of the American Statistical Association 87, pages XXXXXXXXXX.
? Joshua D. Angrist and Alan B. Krueger XXXXXXXXXXSplit-Sample Instrumental Variables Estimates
of the Return to Schooling. Journal of Business & Economic Statistics, Vol. 13, No. 2, pages
225-235.
? Hoogerheide, L. and van Dijk, H. K., 2006. A Reconsideration of the Angrist-Krueger Analysis
on Returns to Education. Econometric Institute report EI XXXXXXXXXX
Writing up the empirical project
You should NOT think of the project as a tutorial assignment where you answer a set of detailed
questions. First, think about what are the possible determinants of women labour market
participation and how factors such as fertility, education and other socio-economic characteristics
can be a determinant factor. Think of any theoretical (population) model that may give you a
structure for the econometric analysis. A lot of work has been done in the literature to answer the
same question, you are encouraged to read and refer to this work.
Second, think about how you can apply what you have learned in your current and past econometric
classes (and other classes) to address this question with the data set.
You are not expected to use any technique or method that has not been covered in lectures and
tutorial questions. However, this is your chance to show that you understand this material. You are
free to use other relevant techniques but do not use methods that you do not understand or
cannot explain and interpret. Entering a STATA/SHAZAM command to get results from some
sophisticated econometric technique is not the purpose of this project, you should rather use what
you know, understand and can work with to answer and debate this question.
The paper should be around 8 to 10 pages of text and 3 to 5 pages of tables. You should use 12 pt
fonts and normal margins. In tables, it is acceptable to use smaller fonts but you should not use
anything smaller than 10 pts.
Broad guidelines
Read Chapter 19 in Wooldridge to get guidance on how to proceed when carrying out an empirical
analysis. As a guide, you could think of including the following section:
1. Introduction (1 to 1 ½ pages): Here introduce the topic and explain why it is important and
interesting to study the effect of schooling on fertility. Discuss previous results in the literature
(refer to one or two papers, and broadly describe their main results) and any policy issues and
implications.
2. Model (1 page): You can describe very briefly a theoretical model in order to motivate your
choice of variables and to help with the interpretation of the results. The model has to be
developed on the grounds of the properties of the dependent variable and the available
regressors. Then describe the econometric model(s) in some detail. This can also be done in the
results section.
3. Data (1 page): Describe the variables you choose to include in your model, the number of
observations, the main problems you had manipulating the data and the decisions you made.
Defend the decisions that you made. Basic descriptive statistics would be good here, and
probably some graphs too.
4. Results and interpretation (3 pages): This is likely to be the main section in your paper. Describe
your econometric models and your results. Interpret the results. Discuss likely problems with the
model and possible violations of your main assumptions. Try to say something about the effects
of possible biases on your findings. This should also include a careful account of relevant
hypothesis tests and their meaning.
5. Diagnostics and Sensitivity analysis (1 to 2 pages): This is also a crucial part of your paper. You
want to convince the reader that your results are real and not just an artifice of the particular
model you are using. Run different models, change some of the decisions you made in
manipulating the data and rerun the models. Describe your results and say if they remain true
with these changes or if they are sensitive to the exact model/data assumptions you are making.
Modelling of real life data is always difficult. It is therefore important that you are able to look
critically at your results and are able to identify drawbacks and shortcomings of your findings.
6. Conclusions (1/2 page): Restate your main results and say how you would expand the paper if
you had more time, more data, etc….
7. References (1/4 page): Reference your textbooks (if you used them) and any article you have read
and used as a basis for your discussions and decisions.