Why should we use panel data? From Baltagi’s book

1. Control for individual heterogeneity

2. Increase the variance of the dependent variables (less correlation, more variability)

3. Understand how populations change in composition and how treatment effects vary with time
4. Build a model of unit behavior (how it will act next)

But they also suffer from

1. Being expensive to measure many times and track individuals
2. Attrition