. Empirical Exercise 7: Demand for cars.

Load the dataset PS_cars. dia. We will be conducting analysis on this dataset across several

problem sets to be able to see how different estimation methods compare to each other. This

exercise will also expose you to other realities of doing econometrics, namely, finding and

combining data and will test programming skills.

1. Answer the following questions and put the results in a simple table. Explain briefly

what Stata commands you used to calculate these numbers. Prefix "bysort" will be

helpful when answering these questions,

(a) What are the mean prices for cars in each year?

(b) What is the average price for cars that have AC as standard (given by "ac_standard" )

in each year?

(c) What is the cumulative market share of cars with above average price?

(d) What are the market shares of firms in each year? Firms are identified by different

values of the "firmid" variable.

2. Suppose that the prices are given in nominal dollar values and not real dollar values

(ie. not adjusted for inflation). You want to have all prices in the equivalent of 1990

dollars. You can do so in a few steps:

(a) Find the Consumer Price Index (CPI) for years 1970-1990 (you can use any Internet

resource);

(b) Weight the prices in years 1970-1990 by CPI to get their values in dollars of year

1990. To do so you will need to input the CPI data into Stata and combine it

with PS_cars.dta. Command "merge" is particularly helpful when combining the

dataseta.

Explain briefly what Stata commands you used.

3. You are interested in predicting a car's market share based on its attributes and price.

What regression would you run? Specify it, run it and report the results.