Der Dibakar Pal, thaks!

I'll check the links

Our problem is that in Brazil the Laws are made in a national context, which makes it difficult to work with public policy evaluation.

In this case, we need something that we can use as counterfactual, that is, if we had not received the treatment. At worst I thought of using synthetic control. In that case, I would like to know if anyone knows microdata that contain waste disposal questions.

I don't know if I understand the question right, but I have a simple suggestion.

I think the key of your question is to estimate the impact of certain treatment (policy intervention) to a certain dependent variable (let's say, amount of dump). The problem is that you do not have data for the control group, since the policy is done in national level.

You have data of '08-'09 and '15-'16, and treatment is done in '10. This means '15-'16 is effected by the policy, and '08-'09 is not

Although you only have 4 years of data, I think you have regional level data as well (since you mentioned 'microdata'). This will boost the number of observations.

Then, consider a simple regression with these observations to estimate the amount of dump in region j in year t (treatment = 1 for data in '15-'16, 0 otherwise):

dump(j,t) = b0 + b1*(t) + b2*(treatment) + bn*(other independent variables of region j)

if estimated b2 is significant, b2 would represent the effect of the treatment. Using b2, you can estimate the counterfactual data without policy intervention for data in '15-'16

Hello, Hyunhong Choi

Great suggestion. I thought about the possibility of using a synthetic control model. However, it would be necessary to find some countries with microdata available that have the desired information.

Thanks for the suggestions.