Difference in Difference Panel Model
The objective of my research is to see how firms can overcome their liability of origin when they get internationalized by initating CSR activities (can be any initiative related to building schools or parks or start spending for environmental improvement).
For this purpose I am going to use Difference in Difference method (to examine how they were behaving before getting internationalized and what happen after internationalzation occurs).
My proposed model is following:
Note all firms are internationalized: period is 2008 to 2018. Total firms are 200.
Firm Performance = a + b1 CSR intensity + b2 Post + b3 CSR intensity * Post + b4 CSR intensity * Post * Firm ownership + b5 Sale + b6 Debt + u
Firm performance = EBIT
CSR intensity = Independent Variable
Post = Dummy variable taking value 1 when they start CSR activites and 0 otherwise
Firm Ownership = Dummy variable taking value 1 if a firm contains state-ownership and 0 otherwise
My questions are following:
1: Is the model correctly specified?
2: as there is not time fixed for change of firms' policy, because its a voluntary adoptation by the firms, so some firms might adopt this policy in year 2009 and other may adopt in year 2015. Will this effect be captured by "Post" dummy? Should I be careful about this issue?
3: As sample firms are diverse in nature, so heteroskedasticity might exist, using robust i.e. white SE DiD will catter this issue?
4: If there exist endogeniety as existing literature identified that firm performance also determine the sale and debt, so reverse causality may exit, How I should tackle this issue, if arise using DiD approach?
5: If some a firm is already using CSR activities, and do not change its intensity, shall I include this firm?
Your response will be highly appreciated.
For this purpose I am going to use Difference in Difference method (to examine how they were behaving before getting internationalized and what happen after internationalzation occurs).
My proposed model is following:
Note all firms are internationalized: period is 2008 to 2018. Total firms are 200.
Firm Performance = a + b1 CSR intensity + b2 Post + b3 CSR intensity * Post + b4 CSR intensity * Post * Firm ownership + b5 Sale + b6 Debt + u
Firm performance = EBIT
CSR intensity = Independent Variable
Post = Dummy variable taking value 1 when they start CSR activites and 0 otherwise
Firm Ownership = Dummy variable taking value 1 if a firm contains state-ownership and 0 otherwise
My questions are following:
1: Is the model correctly specified?
2: as there is not time fixed for change of firms' policy, because its a voluntary adoptation by the firms, so some firms might adopt this policy in year 2009 and other may adopt in year 2015. Will this effect be captured by "Post" dummy? Should I be careful about this issue?
3: As sample firms are diverse in nature, so heteroskedasticity might exist, using robust i.e. white SE DiD will catter this issue?
4: If there exist endogeniety as existing literature identified that firm performance also determine the sale and debt, so reverse causality may exit, How I should tackle this issue, if arise using DiD approach?
5: If some a firm is already using CSR activities, and do not change its intensity, shall I include this firm?
Your response will be highly appreciated.
