Ch. D Fixed Effects, First Differences, Random Effects. Comparison of panel models (conclusion)

7 janeiro 2025, 10:00 Hugo Castro Silva

Lecture was moved up to allow for coordination of lab shifts before the final test.


Fixed effects vs. First differences
  • both are unbiased under similar assumptions
  • choice falls on presence of serial correlation
Unbalanced panels
  • potential Issues and sources
Random Effects
  • intuition
  • assumptions
  • use cases
  • estimation
  • comparison against FE and POLS
Comparison of POLS, RE, FD, FE
Applying panel methods to other data structures
  • revisiting the idea of fixed effects as group dummies