Eric Doviak
Doviak.net 
Economics and Public Policy Analysis
 
Development 730 - Message Board
 
paper submission
(Mon. 14 May 2007 at 02.50)
 
Development Students,
 
Click HERE for instructions on submitting your papers. Many of you have asked me to look at drafts of your papers and asked other questions. I'll try to respond to all of your requests, but please understand that my time is very limited and I already have a backlog of requests.
 
At this point, I can no longer read and provide comments on drafts. I simply do not have the time. I can however answer questions via email. Just please be patient. I cannot provide immediate answers.
 
All the best,
- Eric Doviak
...
 
Ridge Regression
(Sun. 22 April 2007 at 10.45)
 
Development Students,
 
In class yesterday, I used Ridge Regression to estimate the effect of changes in state tax rates on income inequality in the United States. Here's a copy of the R script that I used in class. The script computes the standard OLS estimates, the Hoerl-Kennard Ridge estimates and Obenchain's two-parameter Ridge estimates. The script stops shrinking when Obenchain's CLIK (a minus two log likelihood ratio) is minimized.
 
In order to run the script, you need to install and load Bob Obenchain's RXshrink package. To install and load it in R, go to the "Packages" drop-down menu and select "Install package(s)..." Choose a nearby CRAN mirror and then select "RXshrink." Once you've installed the package, you have to load, so go back to the "Packages" drop-down menu and select "Load package..." and then select "RXshrink" from the dialog box.
 
 Now that information is useless if you don't know how to use R or if you don't know what Ridge Regression is. So let's take a step backward.
 
For those of you who are unfamiliar with R, I created a new page on this site: Getting Started with R.
 
For those of you who are unfamiliar with Ridge Regression, take a look at my Notes on Ridge Regression and take a look at Obenchain's Introduction.
 
You should also look at the articles by Hoerl and Kennard (Technometrics, 1970a, 1970b and 1975) the article by Hines, et al. (Inverse Problems in Engineering, 2003) and the article by Obenchain (Stata Technical Bulletin, 1995). Due to copyright restrictions, I cannot post PDFs of the articles to my website, but you can obtain them from CUNY's horrendous BlackBoard system. The files are in the "Course Documents" page. See instructions below.
 
 Happy Shrinking!,
- Eric Doviak
...
 
today's spreadsheet
(14 April 2007 at 21.00)
 
In class today, I created a spreadsheet with the US inequality data and the NBER tax data. For your convenience, I have uploaded the sheet that I created in class -- click HERE. It's an example and is NOT intended to replace the other data that I gave you. One sheet contains the data in levels, another contains the data with 2-year lagged differences and the last sheet contains some of the tax data.
 
As I've mentioned in class, the paper is now due at the end of the semester.
 
As you recall, you have the option of writing a literature review or doing an empirical paper. I'm strongly encouraging everyone to do an empirical paper because I think you need the practice. The empirical paper should have a short literature review so that you can place your results in the context of other work.
 
If you decide to write a literature review and don't do any empirical work, then I expect an analysis of the literature. In other words, don't just repeat the authors findings. Critique their arguments. Which arguments do you find convincing? Which don't you find convincing? Most importantly -- WHY?
 
Hope this helps,
- Eric Doviak
...
 
tax and international data
(24 Mar. 2007 at 23.10)
 
Because so many of you have asked ...
  • Click HERE to get the international income inequality dataset by Deininger and Squire.
  • Click HERE to get state tax rates from the NBER's TAXSIM 
  • Click HERE for my preference:  Table 4 -- Sum of federal and state average marginal tax rates 
  • Click HERE for historical US state populations by race (with breakdowns by percentage slave and free)
  • Click HERE for the spreadsheet of all 50 states and DC
OR feel free to use another dataset. For example, incorporate historical information, trade data or industry composition and see what you find. As I mentioned in class, I would be very interested in a paper that looks at how slavery and/or the Great Black Migration affected income inequality in US states.
 
Hope this helps,
- Eric Doviak
...
 
clarification on dataset
(06 Mar. 2007 at 11.40)
 
Students,
 
Just a clarification on the dataset. You do NOT have to use the dataset that I gave you. I just gave it to you to help you get started.
 
In retrospect, it might be easier to download a complete dataset from one of the sites on my Economics Links page.
 
Alternatively, you can assemble your dataset from multiple sources. If you choose that option, then you can use the US state data and the UN country data as a place to start. Or you can choose not to. It's up to you. Once again, my Economics Links page will take you to some good sites where you can obtain additional data.
 
Once you've assembled the dataset, the regressions will be easy.
 
Hope this helps,
- Eric Doviak
...
  
paper submission dates and data
(04 Mar. 2007 at 20.10)
 
Students,
 
A couple of you asked me for another EViews tutorial on Sat. 10 March. I'll be happy to provide one.
 
Many of you have asked about the deadline for the mid-term paper. To give you maximum flexibility and to provide you with the assistance that you need, I have decided to offer two submission dates.
 
The first will be on Sat. 24 March. The second will be on Sat. 14 April.
 
If you submit a full draft on Sat. 24 March, I will read it and return it to you with critical comments when we meet on Sat. 31 March (the last class before spring break). Then if you would like to revise your paper -- and get a better grade -- you may submit the revised version on Sat. 14 April.
 
I will not accept any papers (or revisions) after Sat. 14 April.
 
In class yesterday, I showed you a dataset that contains observations on income and educational inequality in US states. I have incorporated that dataset into the original spreadsheet that I gave you -- click HERE. This dataset REPLACES the original. Please delete the original!
 
During the revision, I noticed an error in the program that I wrote to calculate the mean and variance of net income in the US state data, so if you want good regression results, then delete the original!
 
I have also typed an addendum to the paper topic document which describes the construction of the human capital variable -- click HERE. The description is on page 4 of the document.
  
Finally, some of you have noted that I only gave you a handful of variables and asked me how you should work with such a dataset. The simple answer is that YOU have to expand it. You're graduate students. You have to learn how to create a dataset. You have to learn how to perform econometric analysis. You have to learn how to present the results of your research.
 
I know this is a tough assignment, but you'll thank me for it one day.
  
To get started, think about this: "What do you think I should have given you?" ... Once you've thought about that question, you'll know what you need to find.
 
Now where do you find it? Start by clicking on some of the links in my Economics Links page. Most of the linked pages have data that you can download for free.
 
Also ... If you would like to incorporate dividend and interest income into the US state data that I gave you, then let me know and I'll give you the files that I used to generate the spreadsheet (the original Census Bureau downloads and the R script file I used to generate the data). It would probably be a simple and easy way for you to learn R.
  
You can do it!, 
- Eric Doviak
...
 
mid-term paper
(13 Feb. 2007 at 13.00)
 
Development Students,
 
I have finally finished writing your mid-term paper. Sorry for the delay.
 
The paper topic and suggested reading list is in this PDF file -- click HERE. Because an empirical analysis is another option, I am also providing you with a spreadsheet that contains observations on income inequality, average income and the growth rate of average income -- click HERE.
 
If you have any questions, then let's discuss them on Saturday.
 
Good Luck!
- Eric Doviak
...
 
Welcome Message for Spring 2007 Students
 
Welcome to the course website. Please check this site regularly.  
 
I've posted a copy of the syllabus -- click HERE. 
  
Almost all of the readings for this course come from academic journals, for your convenience, I have posted them to BlackBoard. See the instuructions on logging into BlackBoard below.
 
The readings are best viewed using Adobe Acrobat 6.0 or 7.0. If you're still using an older version of Acrobat, then some of the files may not open. To download the latest version of Acrobat Reader -- click HERE. 
 
When you print the readings, please DOUBLE  SIDE. Save the trees. Save the Earth. 
 
Also ... SAVE your EYESIGHT !!!  When you print the readings, look for the "Page Scaling" option and  select "Fit to paper."
 
For example, pages from the Journal of Political Economy are 5 inches wide by 8 inches high. When you select the "Fit to paper" option, the page image will expand to fill the paper you're printing on and the text will be 130 percent larger and much easier to read.
 
BlackBoard
 
Almost all of the readings come from academic journals, so it would be convenient to obtain those articles from a single source (instead of searching the library's databases for each article). That convenient source is BlackBoard. (The articles are copyrighted, so I cannot legally post them to my website).  If you don't have a Portal Login yet, then it's high time you got one! Click the "Register Now!" link and the system will verify your status by using your last name, Social Security Number, date of birth, etc.  You're going to learn a lot this semester and I'm looking forward to working with you. 
   
All the best, 
Eric Doviak