YOUR ECONOMETRICS PAPER – DATA SOURCES FOR …



YOUR ECONOMETRICS PAPER

BASIC TIPS

There are a couple of websites that you can browse to give you some ideas for topics and data. Think about what you want to do with this paper. Econometrics is a great tool to market when looking for jobs. A well-written econometrics paper and your presentation can be a nice addition to your resume.

You are not expected to do original research here. REPLICATION of prior results is perfectly acceptable.

Read Studenmund's Chapter 11.

One of the most frustrating things in doing an econometrics paper is finding the data. Do not spend a lot of time on a topic before determining whether there is data available that will allow you to answer your question. It is a good idea to write down your ideal data set that would allow you to address your topic. If you find that the available data is not even close to what you had originally desired, you might want to change your topic.

Also, remember that knowing the location of your data – website, reference book, etc – is not the same as having your data available to use. It may take a LONG time to get the data in a format that EVIEWS can read. Do not leave this till the last minute.

For most data, I enter the data into Excel first. I save the Excel sheet in the oldest version, namely MS Excel Worksheet 2.1 . The reason is that format can be read by most programs whereas newer formats may or may not be read. Eviews easily reads an Excel sheet 2.1 version. You should use the first row to label your columns (variables). Be sure to follow the naming conventions in Eviews and do not use 'C' or 'RESIDUALS'. Eviews can only use numeric data.

A warning for time series analysis - In order to find interesting results, you need variation in your data. For example, it would be difficult to find the effect of defense spending on American manufacturing industries if defense spending does not change much from year to year. Education and health spending might be a different story, because you can take advantage of state and local spending in these areas as sources of variation. There are then two options – you could look at the variation across states in a single year, or construct a panel data set where you have information for a number of states for a number of years. If you find that you are looking at a topic that has little variation, consider doing a cross-country comparison.

Also, remember that it isn’t necessary to use the most recent data necessary. Econometrics is very useful in the field of economic history. The Census Data are available back to 1860 and there are many other historical data sources available.

GENERAL SOURCES

These are suggestions – they are not meant to limit your choice of a topic.

- RESOURCES FOR ECONOMISTS. This is a good overall site (and has a lot of good data).

- National Bureau of Economic Research. The NBER does a lot of really excellent empirical research. The papers on this site might be a bit technical, but browse through their titles for ideas.

- General site, the specific topics (health economics, sports economics) may be good sources for ideas. The DATA SOURCES may also lead you to a topic.

- Inter-University Consortium for Political and Social Research. A good source for all kinds of data. Just do a search for the subject you are interested in and see what comes up.

Wisconsin's topical index of data, most of the data is in the ICPSR (above) but the index is rather nice in choosing a topic.

- Census Data - this is a GREAT site. You can download individual level sample data from 1860-1990. 2000 data isn’t yet available in this format, but you can get aggregate census data from

Time series data library – single variable data sets, many rather esoteric (e.g., monthly public drunkenness in Minneapolis).

Social Science Data on the Net

EDUCATION

I am sure you are all familiar with the US News Rankings of Colleges. These data are available (in ICPSR) and can be used to answer a number of questions concerning education – questions about tuition, professor’s salaries, quality of students, etc.

Also, look at nces. - the National Centre for Education Statistics.

For a lot of education topics, you might want to pull in location data - i.e. if you are looking at local school funding issues, you would need other information about the area. A good source for location information is the City and County Data Books (available at the ICPSR site). These give very detailed information at the county level – things like average income, age distribution of population, housing values, etc.

HEALTH

There is a tremendous amount is econometric questions in the health field – questions concerning individual behaviour and health, retirement and health, health policies, insurance issues, international issues, labour issues (nursing shortages, gender of doctors) etc.

Possible Questions:

• Does smoking/drinking affect your salary? [1]

• Would a tax on junk food reduce consumption? I.e. how price sensitive is demand for French fries?

• Do states with higher unemployment levels have higher levels of suicide?

• What determines the level of health insurance held by an individual?

• Are individuals in cities more/less healthy than individuals in rural areas?

• Does trade extend life expectancy?

Sources

- National Health Statistics

- World Health Organisation A to Z list of health topics.

- Also NHS, this gives state level health statistics.

If you are doing a micro level study, look for surveys that ask health questions such as:

- General Social Survey - annual survey of Americans (is also available for other countries) that asks a nice variety of questions. Look under ‘subject’ for A to Z list of question topics. Has some great questions like – how happy are you?

- Health and Retirement study – survey on older Americans. A warning though, this dataset is a bit difficult to work with.

If you are interested in doctor’s characteristics and how they treat their patients, check out the Community Tracking Study Physician Survey which is at the ICPSR data page (see above for link).

For international data on broad health measures, check out the web pages of the United Nations the World Bank and the WHO. These have good data links.

URBAN ISSUES

As noted above, the County and City Data Books (ICPSR web page) are a great source of information at the county level. Housing values, commuting time, average income, industry information, etc.

There are a lot of interesting transportation issues – commute times, airline prices, train travel, metro buses, etc. Consider looking at automotive industry statistics, government spending information.

Check out the Bureau of Tranportation Statistics ().

LAW AND ECONOMICS

If you are interested in law, there are a number of interesting econometric topics. These can range from some analyses of the incarceration rates by states, crime statistics (again using the county and city data books), spending on public defenders, characteristics of inmates/victims, etc. You can also look at the effect that state law/regulation have on economic or demographic variables – i.e. if different states have different mandatory minimum sentencing for certain crimes, can you see some difference?

Check out the Bureau of Justice Statistics -

LABOR

There are tons of possible econometric papers dealing with labour topics – wages, gender issues, unions, retirement issues, unemployment, etc. And labour issues can be joined with many other topics to address the linkages. Do long spells of unemployment increase the likelihood of family violence, substance abuse, crime? Combine health and labour and consider topics surrounding the employment of doctors/nurses, midwives vs. Obs. Combine law and labour and look at the legal profession – are more lawyers a good thing?

Labor topics also fall neatly into macro studies. You might consider how government spending, war, monetary policy, etc. affect employment rates, or unemployment rates. You can be more specific (and potentially more interesting) by looking at specific employment rates – does a government policy have a different effect on the employment of men/women/young adults/highly educated, etc.

Data sources:

Bureau of Labor Statistics –

Census Bureau –

State labor information -

Consumer Expenditure Survey Homepage -

CROSS-COUNTRY COMPARISONS

- a United Nations site, provides links to the statistical agencies of countries. Warning – if you are gathering international data from a variety of sources, make sure that the definitions of variables are comparable. For example, the unemployment rate in the United States is defined differently from the unemployment rate in other countries. Try and find sources that have compiled cross-country statistics on your topic. They have likely already adjusted for definitional differences.

Check out the websites of the major international organizations – the OECD has a lot of data (, click on statistics).

The Luxembourg Income/Employment Study () is a great project that has collected survey data from a large number of countries. If you want to do a cross-country comparison of labour market situation at the individual level rather than the aggregate level, this might be a good place to look. The data isn’t accessible directly however, you have to submit your program to the site and they will send you back your results.

MACRO TOPICS

Macro (aggregate) data may be a little easier to find than micro data. Since the data is aggregated (i.e. unemployment rates versus whether a person was unemployed last year) there is little concern of confidentiality which hampers a lot of the gathering of micro-level data. The downside of using aggregate data is that the number of observations that you have are typically much lower and you may have less variation. Also, note that some data used in time series are collected annually, some quarterly, some daily. Make sure you are able to get consistent time periods.

Resources for Economists () is probably the best site to find aggregate data. has many of the US macro economic time series data, available in Excel files. Also the Census Bureau has good data on the 2000 census.

MODEL OF THE ECONOMY

- this is a complicated model of the US economy, and is mostly used for forecasting. I would suggest it ONLY if you really want to spend a LOT of time on a very interesting and worthwhile subject. YOU will have to do a lot of reading on your own since we do not have time to cover simultaneous equations in class.

FIRM LEVEL DATA

This type of data is notoriously more difficult to find, for confidentiality reasons. However, industry analyses make for interesting econometric papers. For example, one recent project looked at the effect of pharmaceutical advertisement regulation on sales. Or one could look at the impact of trade regulations or employment regulations on a certain industry. To do a project like this, I would suggest first looking at the annual reports of companies that you are interested in. It is easier to consider an industry with few major players – ie. Automotive, fast-food restaurants, etc. Also, remember that the fewer companies that you have the more years of data you will need.

FINANCIAL DATA

Financial data is useful, not just for finance related topics but as a way of measuring the performance of certain industries. Careful with this type of data – a lot of it costs money. Be sure that the data is downloadable and not just available.

fisher.osu.edu/fin/osudata.htm - provides a search engine for financial data sources

GROWTH OF ECONOMIES

See the web site section RESOURCES.

SPORTS

In prior years, there have been many sports papers. Although that is fine with me (because I believe you will learn more if you have innate interest in the topic), you MUST have references to the economics of sports literature.

It is very permissible to find an old paper and add to its data.

Performance data is useful to address a number of economic questions, especially in the field of industrial organization. Think about it – here you have these sports teams (firms) that are competing against other teams (firms). It is possible to clearly identify the output of individual athletes (workers) and how they work with others (individual stats and stats from certain combinations of players) and finally salary is assumed to be clearly related to performance, unlike in most other types of employment. The danger in writing a sports paper is forgetting to add any economic context whatsoever. This isn’t acceptable. Check out the Journal of Sports Economics to get an idea of the types of topics that you can ask.

pdf/sports_economics.pdf is an encyclopedic synopsis of sports economics, it might give you some topic ideas. There is also a sub section of that deals specifically with sports economics.

DATA: Be prepared to enter a lot of data in a spreadsheet by hand. Check league webpages. Data of past years may also be available in the reference area of the library.

- has data on baseball, legal situation, etc.

- info on stadiums and revenue

-----------------------

[1] Prof. Bart Hamilton (Business School) wrote an article on alcohol consumption and salary with some interesting results. "Alcohol and Earnings: Does Drinking Yield a Wage Premium?" (with V. Hamilton). Canadian Journal of Economics , 30 (1997) 135-151.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download