Case Study:
Case study concerning interpretation of:
- Logit regression
- Explanatory variables
- Dummy variables
- Statistical tests
Economists are often puzzled by the way domestic producers react to changes in the price of imported products. Sometimes when changes in tariffs and exchange rates result in a rise in the price at which imported products are sold in Australian markets domestic suppliers raise their price and lose some of ther competitive advantage. Now, we would expect prices to change if labour costs, and especially wages, changes and if the prices of raw materials (whether imported or not) change. So any study of price changes following a rise in the price of imported products would have to allow for this. Also it is usually argued that if the industry has only one or a few sellers in it then the domestic suppliers and the importer’s agents are likely to collude and have various agreements to share the market and, as a result, they don’t really compete with each other and instead their prices rise at the same time and by about the same amount.
Information has been collected from the managers of 33 Australian firms on whether or not they raised their prices in the period following a large devaluation of the Australian dollar, and thus a rise in the price of imports, in 1997. Since the researchers are primarily interested in whether, after the change in the price of imports, domestic producers reacted by raising their price or not, they used logit regression to estimate the relationship between their dependent variable and various explanatory variables. The dependent variable was defined to equal 1 if the firm’s manager responded “yes” to the question: “did the price of your product rise following the devaluation of the dollar in 1997?” and equal to 0 if the answer was “no”. All of the data is derived from survey questionnaires administered to managers of the 33 firms. The questionnaire asked whether their price rose in the period following the devaluation and also contained a number of other questions designed to elicit information on the nature of the industry and also information on whether or nor there were changes in labour costs, and/or in the prices of the raw materials they purchased over the period.
The list of explanatory variables used in the logit regression can be found down the left hand side of the table below. Note that some of the explanatory variables are themselves dummy variables; this is indicated by the phrase (yes/no) after the name of the variable. The coefficients (and estimated standard errors of the coefficients – these are given in brackets).
Increase in the prices of raw materials produced in Australia 0.41 (0.10)
during 1997 in dollars
Increase in hourly wages of workers during 1997 in dollars 0.62 (0.16)
In the production of our product we use a lot of imported 0.51 (0.09)
raw materials (yes/no)
There are very few sellers (besides the importers) in the 1.75 (0.06)
market for the product we produce (yes/no)
Interpret these results. (Be sure to set out clearly any statistical tests you perform and to explain why you are doing them).
Provide complete and step by step solution for the question and show calculations and use formulas.