ECON 104 Lecture Notes - Lecture 34: Dollar Sign
Document Summary
3 different lines with 3 different intercepts, constrained by same slopes. If cali observations are mainly above texas observations -> higher intercept for cali than texas. Without panel structure, each entity has individual observation. In this structure, every state has multiple observations, we can identify what the intercept must be separately. If we try to do single observation per state, we can match each state"s observation perfectly. Instead of putting fixed effects, put dummy variable for each of these states. Dummy variable for cali, =1 if it belongs to cali, 0 if not. Dummy variable for texas = 1 if it belongs to texas, 0 if not. Leave out massachusetts to avoid the dummy variable trap. With 48 states, add 47 dummies - bothersome and messy. For fixed effects, use entity subscript in intercept term. Must first tell software that you are working with a penl data.
