BOSTON COLLEGE
Department of Economics

EC 327 Financial Econometrics
Prof. Baum, Ms. Pumphrey, Ms. Phillips, Spring 2010

http://fmwww.bc.edu/ec-c/s2010/327/

Problem Set 1

Due at classtime, Tuesday 2 February 2010

Problem sets will not be accepted after their due dates. For the empirical problems in this exercise, you must use Stata, but will not make use of the RAW files listed in the text, as all of these data are available in Stata format over the Internet.

Write a 'do-file' (using the Stata do-file editor) and run it, producing a printout (log file) from Stata, annotated with your comments. The "log" icon in the Stata toolbar may be used to create a file containing both your commands and Stata's response. It may be easier to work with a text-format log; the command 'log using ps1.log' will create a text log. Use the 'Viewer' within Stata to examine and print the log.

  1. Load the prepared Stata dataset of US airline flights, 1997-2000, via
    use http://fmwww.bc.edu/ec-p/data/Wooldridge2K/airfare
    Use a new Do-File in Stata's Do-File Editor to perform all steps below. That is, edit the do-file to add each task and run it.
  2. How many observations are there in this dataset? How many variables? How many of these are string variables (rather than numeric variables)? [hint: help describe]
  3. How many observations contain flights originating in "BOSTON, MA"? How many terminating in "BOSTON, MA"? [Hint: help count]
  4. What is the shortest flight recorded? What cities does it involve? What is the longest flight recorded? What cities does it involve? Use scalars to extract these extrema; do not hard-code numbers in your Do-file. [Hint: return list]
  5. What does the variable -bmktshr- represent? What sanity check might you perform on that variable's values? What is the difference between that variable and the variable -concen-?
  6. Which of the origin-destination pairs are monopoly routes, and in which years?
  7. Is concentration correlated with length of the flight? With the log of length of the flight? [Hint: help correlate]
  8. What are the average fares charged for flights in 1997, 1998, 1999 and 2000?
  9. The US Bureau of Labor Statistics figures for the average consumer price index are
    year cpi
    1997 161.5
    1998 164.0
    1999 168.3
    2000 174.1
    Use these values to create a real (constant-dollar) fare variable, -rfare-. Did real fares rise or fall from 1997-1999? From 1997-2000?
  10. Create a real cost-per-mile measure as -cpm-. What are the quartiles (25, 50, 75th percentiles) of -cpm-? [Hint: help summarize]
  11. What is the correlation between cpm and market concentration? Does this agree with your priors about the effect of market concentration?
  12. In terms of (real) cost per mile, what was the cheapest flight, and in what year? The most expensive flight / year? [Hint: help sort]
  13. What flights / years attracted more than 7,500 passengers per day on average?
  14. Using the -lpassen- measure, what was the average rate of growth in air travel over the four years? Was this growth statistically significant? [Hint: help regress]
  15. Estimate a regression model explaining real cost per mile as a function of the average number of passengers, distance of the route and concentration in the market. Are these explanatory factors each significant? Do they have the expected signs? Does market concentration raise or lower average fares, cet.par.?
  16. In the prior model, is the effect of distance nonlinear? Include a squared distance term using Stata's factor-variable notation. Is it significant? How do you interpret (in words) the sign pattern on the distance and distance^2 variables? Does this change considerably improve the model's fit? What does that say about the prior model's estimates?
  17. Compute the average marginal effects of these three explanatory factors. [Hint: help margins]
  18. Include an interaction term (using factor-variable notation) between the average number of passengers and market concentration. Is it significant? Compute the average marginal effects for the model. How do they change from those which you calculated in the previous question?