Error component model

Summary

• A common model for the error term in a LRM with panel data is the error component model . There are three versions of the error component model:
• The one-way error component model with individual specific effects

\[ε_{i,t}=α_i+μ_{i,t}\]

• The one-way error component model with time specific effects

\[ε_{i,t}=γ_t+μ_{i,t}\]

• The two-way error component model

\[ε_{i,t}=α_i+γ_t+μ_{i,t}\]

• \(α_i\) is called the individual specific effect and \(γ_t\) is called the time specific effect .
• The error terms \(μ_{i,t}\) are typically assumed to be homoscedastic and not autocorrelated.