How to control for industry effects in regression?

Question

Right now I'm working on an analysis of influence of cultural aspects on investment mode preference. However I have to control for many other factors, for example industry, since some industries, for example, are more direct investment intensive, whereas others are more appropriate for licensing. But I have no clue how to manage this issue. Which kind of information do I need in order to be able to control for industry influence and calculate variables expressing the influence of each industry on dependent variable, which is a ratio of royalties to FDI income. I guess I need to include industry dummies, however, I do not really understand how to estimate them and based on which kind of information.

Answer 1

Which kind of information you need is more of a theoretical question than a statistical one. However, if you only aim at controlling for industry specificities, you do not need much information, apart from the knowledge of which industry each of your observations belongs to. Creating a categorical variable that holds industry affiliation is equivalent to having one intercept per industry (every modern statistical package, such as R, will take care of selecting one of the industry as the baseline, helping you avoid perfect multicollinearity, known as the "dummy variable trap"). This assumes that between industry, the relationship between the dependent and independent variables is the same (the slope of each group is the same), but with a different mean for each group (the intercept is varying by groups).

Answer 2

Each regression includes a variable equal to the median Tobin’s Q for the firm's industry in the year of the turnover, net of that year’s median (across all industries). 16 Using the industrylevel Q should provide a better measure of growth opportunities the firm faces at a point in time