The PANEL Procedure


The following notation represents the usual panel structure, with the specification of ${u_{it}}$ dependent on the particular model:

\[  y_{it}=\sum _{k=1}^{K}{x_{itk}{\beta }_{k}} + u_{it} \hspace{0.2 in} i=1,{\ldots }\mi {N} ; t=1, {\ldots }\mi {T} _{i}  \]

The total number of observations ${\mi {M} = {\sum }^{\mi {N} }_{i=1}\mi {T} _{i}}$. For the balanced data case, ${\mi {T} _{i}=\mi {T} }$ for all ${i}$. The ${\mi {M} {\times } \mi {M} }$ covariance matrix of ${u_{it}}$ is denoted by ${\mb {V} }$. Let ${\mb {X} }$ and ${\mb {y} }$ be the independent and dependent variables arranged by cross section and by time within each cross section. Let ${\mb {X} _{s}}$ be the ${X}$ matrix without the intercept. All other notation is specific to each section.