Applied Econometrics
Applied Econometrics
To complete the research project, you should follow the following steps (additional information
can be found in Chapter 11 of the Studenmund Econometrics text):
1. Planning for the project, based on Economic Theory
In general your paper will be testing an economic theory.
For example, you may want to test the law of demand or supply, or ascertain what determines the
level of imports, or savings, or consumption or investment or wages or the U.S. government
deficit or state tax revenues. Specify what variables are important in the model and the expected
signs on the coefficients of these variables. For example, theory would tell us that if a demand
function is to be estimated the relevant independent variables that influence quantity demanded
would be price, income, price of complements, and price of substitutes. The signs underneath
the variables indicate what type of relationship each has with quantity demanded. A minus sign
implies an inverse relationship while a plus indicates a positive relationship.
Qd = f( P, Y, Pc, Ps)
– + – + (the Betas in the regression equation are expected to have the
following signs)
Be sure to explain why a sign was picked. A minus under P indicates that as price increases
(decreases) consumers will buy less (more), this follows from the law of demand. Sometimes a
researcher will be uncertain as to the sign. For example a (?) may be placed under income (Y),
researchers may be uncertain as to whether the good in question is a normal (+) or inferior good
(-). If the results show a minus sign, then the good would be considered to be inferior, and if the
results show a plus sign, then the good would be considered normal.
The topic is due by October 17, 2013 (in class). Write one page outlining the theory to be
tested. Follow the information in #1 above.
2. Review of the relevant literature
Go to the library and use an on line service, such as Business Periodical Index or InfoTrack to
track down articles or books on the topic you have chosen. Google scholar is also a good place
to start. Based on the work of other researchers you may find that variables need to be added,
deleted or redefined, and it may also clue you in as to where the data can be found to estimate
your model. Tracing references from the bibliography of a few (a minimum 3, but 4-5 is
preferred) key articles concerning your topic will be most useful. You must cite these properly
using MLA or APA format.
Literature Review is due by November 7, 2013 (in class). (One to two pages)
3. Collect the data
Consider the sample size. All variables must have the same number of observations. A number
of computations might need to be made to get the variable you desire. For example you may
have total number in the labor force and total number of those working and what you seek is the
unemployment rate. It is your job to derive the unemployment rate from this data. Be sure to
define and explain how the dependent and independent variables are measured (for ex. in $, or
cents, or percentages or quantities, etc.) This specification is based on economic theory, an
independent variable is chosen because it partly explains the variation in the dependent variable.
“Regression gives evidence but does not prove economic causality.” For example, if you decide
to estimate a demand function you will need to choose a product (good or service) where
information is available for the variables chosen. Information is available in the Wall Street
Journal on the price and the quantity sold of cars. Don’t forget to check web sites; there is a link
to some economic data web sites on the Economics Department homepage.
Data selection is due by November 21, 2013 (in class). (One to two pages)
4. Estimate and evaluate the equation
Choose and estimate the appropriate model. Make sure you explain why the chosen model is
most appropriate. Once the model is estimated, check for errors. Usually additional model
developments or alternative testing will be required. If the regression results are unexpected, it
might be necessary to reexamine the model and underlying theory.
5. Specification tests
Have you tested for multicollinearity, heteroscedasticity, specification (the proper functional
form) and serial correlation? Have you corrected for the above problems if they exist? Be sure
to state in your paper what problems you looked for and corrected for if they existed. Please
note that if you run into serial correlation, you only know how to change the functional form to
correct for it. If you don’t want to do this, please talk to me about other options.
Document the results. A standard format for OLS is the following:
Yi = 32.9 – 0.70 Xi1 + .27 Xi2
(0.08) (.01)
t= -8.33 t=45.91
n=44 adjusted R2 = .984
The number in the parenthesis are the estimated standard errors ( 0.08, .01) of the respective
estimated coefficients (-.070 and .27). Explain all the regression output from the statistical
program. Explain the adjusted R-squared, F-value (probability associated with the F-value), the
meaning of each Beta coefficient, each t-value (probability associated with each t-value). Be
sure to write out the regression model results. Did you reject or fail to reject the null hypothesis
on the F-test? and each t-test? What does it mean using this data to reject or fail to reject the null
hypothesis? Did you get the signs expected on each B coefficients? For example, based on this
estimated regression equation, the meaning of the coefficient on Xi1 is: every time Xi1 increases
in value by 1 unit, Y decreases in value by .70 units holding Xi2 constant (be sure to state what
the units of value are in your regression).
The written documentation must contain enough information so that the reader can replicate the
study. If there is a series of estimated regression equations, then tables may provide the relevant
information for each equation. All data manipulations as well as data sources should be
documented fully. All papers should provide and explain descriptive statistics as well as
regression results. Are the results obtained consistent or do they run counter to the theory you
outlined under step one? If not, what is a possible explanation?
6. Policy recommendations
What policy prescriptions would you recommend? What would the possibilities be for future
research?
Final paper with the corrected model is due on December 5, 2013 in class. Length 10 pages,
not including the bibliography and regression output files. This is a research paper and footnotes
and bibliography are required. In the final paper provide an introduction to the research question
explaining why this is an important and relevant research question. Provide a literature review.
Provide a written estimation of the final model with all the explanation outlined under #5 above.
Provide answers to the questions posed under #6 above.
Send final regression results and paper to me via e-mail by the beginning of class on December
5th. I will respond to you and confirm that I have received your assignment. If I do not confirm,
I have not received it.
Final notes:
? You must include descriptive statistics in your paper. This includes the means and
standard deviations of each variable. Please ask me if you can’t figure out how to do this.
? You must cite all sources in MLA or APA style.
? You must describe in detail what is the source of your dataset. You must also cite this
on your works cited page.
? You must test for heteroskedasticity (this may simply be with a scatterplot) and serial
correlation.
? You must include the output containing all your results and tests. This can be saved and
attached to an electronic file or printed out.
? Even though you will have your results attached, you must clearly include them in your
paper. For example, if you run VIFs, you may want to create a table to show your results.
can be found in Chapter 11 of the Studenmund Econometrics text):
1. Planning for the project, based on Economic Theory
In general your paper will be testing an economic theory.
For example, you may want to test the law of demand or supply, or ascertain what determines the
level of imports, or savings, or consumption or investment or wages or the U.S. government
deficit or state tax revenues. Specify what variables are important in the model and the expected
signs on the coefficients of these variables. For example, theory would tell us that if a demand
function is to be estimated the relevant independent variables that influence quantity demanded
would be price, income, price of complements, and price of substitutes. The signs underneath
the variables indicate what type of relationship each has with quantity demanded. A minus sign
implies an inverse relationship while a plus indicates a positive relationship.
Qd = f( P, Y, Pc, Ps)
– + – + (the Betas in the regression equation are expected to have the
following signs)
Be sure to explain why a sign was picked. A minus under P indicates that as price increases
(decreases) consumers will buy less (more), this follows from the law of demand. Sometimes a
researcher will be uncertain as to the sign. For example a (?) may be placed under income (Y),
researchers may be uncertain as to whether the good in question is a normal (+) or inferior good
(-). If the results show a minus sign, then the good would be considered to be inferior, and if the
results show a plus sign, then the good would be considered normal.
The topic is due by October 17, 2013 (in class). Write one page outlining the theory to be
tested. Follow the information in #1 above.
2. Review of the relevant literature
Go to the library and use an on line service, such as Business Periodical Index or InfoTrack to
track down articles or books on the topic you have chosen. Google scholar is also a good place
to start. Based on the work of other researchers you may find that variables need to be added,
deleted or redefined, and it may also clue you in as to where the data can be found to estimate
your model. Tracing references from the bibliography of a few (a minimum 3, but 4-5 is
preferred) key articles concerning your topic will be most useful. You must cite these properly
using MLA or APA format.
Literature Review is due by November 7, 2013 (in class). (One to two pages)
3. Collect the data
Consider the sample size. All variables must have the same number of observations. A number
of computations might need to be made to get the variable you desire. For example you may
have total number in the labor force and total number of those working and what you seek is the
unemployment rate. It is your job to derive the unemployment rate from this data. Be sure to
define and explain how the dependent and independent variables are measured (for ex. in $, or
cents, or percentages or quantities, etc.) This specification is based on economic theory, an
independent variable is chosen because it partly explains the variation in the dependent variable.
“Regression gives evidence but does not prove economic causality.” For example, if you decide
to estimate a demand function you will need to choose a product (good or service) where
information is available for the variables chosen. Information is available in the Wall Street
Journal on the price and the quantity sold of cars. Don’t forget to check web sites; there is a link
to some economic data web sites on the Economics Department homepage.
Data selection is due by November 21, 2013 (in class). (One to two pages)
4. Estimate and evaluate the equation
Choose and estimate the appropriate model. Make sure you explain why the chosen model is
most appropriate. Once the model is estimated, check for errors. Usually additional model
developments or alternative testing will be required. If the regression results are unexpected, it
might be necessary to reexamine the model and underlying theory.
5. Specification tests
Have you tested for multicollinearity, heteroscedasticity, specification (the proper functional
form) and serial correlation? Have you corrected for the above problems if they exist? Be sure
to state in your paper what problems you looked for and corrected for if they existed. Please
note that if you run into serial correlation, you only know how to change the functional form to
correct for it. If you don’t want to do this, please talk to me about other options.
Document the results. A standard format for OLS is the following:
Yi = 32.9 – 0.70 Xi1 + .27 Xi2
(0.08) (.01)
t= -8.33 t=45.91
n=44 adjusted R2 = .984
The number in the parenthesis are the estimated standard errors ( 0.08, .01) of the respective
estimated coefficients (-.070 and .27). Explain all the regression output from the statistical
program. Explain the adjusted R-squared, F-value (probability associated with the F-value), the
meaning of each Beta coefficient, each t-value (probability associated with each t-value). Be
sure to write out the regression model results. Did you reject or fail to reject the null hypothesis
on the F-test? and each t-test? What does it mean using this data to reject or fail to reject the null
hypothesis? Did you get the signs expected on each B coefficients? For example, based on this
estimated regression equation, the meaning of the coefficient on Xi1 is: every time Xi1 increases
in value by 1 unit, Y decreases in value by .70 units holding Xi2 constant (be sure to state what
the units of value are in your regression).
The written documentation must contain enough information so that the reader can replicate the
study. If there is a series of estimated regression equations, then tables may provide the relevant
information for each equation. All data manipulations as well as data sources should be
documented fully. All papers should provide and explain descriptive statistics as well as
regression results. Are the results obtained consistent or do they run counter to the theory you
outlined under step one? If not, what is a possible explanation?
6. Policy recommendations
What policy prescriptions would you recommend? What would the possibilities be for future
research?
Final paper with the corrected model is due on December 5, 2013 in class. Length 10 pages,
not including the bibliography and regression output files. This is a research paper and footnotes
and bibliography are required. In the final paper provide an introduction to the research question
explaining why this is an important and relevant research question. Provide a literature review.
Provide a written estimation of the final model with all the explanation outlined under #5 above.
Provide answers to the questions posed under #6 above.
Send final regression results and paper to me via e-mail by the beginning of class on December
5th. I will respond to you and confirm that I have received your assignment. If I do not confirm,
I have not received it.
Final notes:
? You must include descriptive statistics in your paper. This includes the means and
standard deviations of each variable. Please ask me if you can’t figure out how to do this.
? You must cite all sources in MLA or APA style.
? You must describe in detail what is the source of your dataset. You must also cite this
on your works cited page.
? You must test for heteroskedasticity (this may simply be with a scatterplot) and serial
correlation.
? You must include the output containing all your results and tests. This can be saved and
attached to an electronic file or printed out.
? Even though you will have your results attached, you must clearly include them in your
paper. For example, if you run VIFs, you may want to create a table to show your results.
The post Applied Econometrics appeared first on Capital Essay Writing Blog.
CLICK HERE TO ORDER THIS PAPER………………………NO PLAGIARISM Get 100% Original papers from the writing experts
CLICK HERE TO GET A PROFESSIONAL WRITER TO WORK ON THIS PAPER AND OTHER SIMILAR PAPERS, GET A NON PLAGIARIZED PAPER FROM OUR EXPERTS
Comments
Post a Comment