Panel Data FE vs RE Visualizer

📊 Panel Data: Pooled OLS vs Fixed Effects

// lihat bagaimana pooled OLS bias kalau ada unit heterogeneity

True β (effect X→Y) = 1.5

Unit Heterogeneity = 5

Correlation u_i, X = 0.7

POOLED OLS β̂

FIXED EFFECTS β̂

RANDOM EFFECTS β̂

Coba: Set Heterogeneity = 5 dan Correlation = 0.7 (typical scenario di IDX firms). Pooled OLS akan biased. FE consistent recover true β. RE OK kalau correlation=0 saja.

Aturan: Kalau correlation u_i dengan X ≠ 0 → Hausman reject → pakai FE.