I'm currently investigating the effect of pollution on companies preference towards debt finance. Due to the lack of accountancy variables for pollution i have differentiated polluting from non-polluting industry's.

I am planning on controlling as many variables as possible and then creating a year variable to get a sense of the polluting factor. Obviously this poses significant problems in regression.

Would it be more accurate to place my limited dependant variable as the dependant variable i.e. polluting or non polluting and use a logit or probit