Panel Data QR Models
Panel data consist of observations made on individuals over time. We shall consider models in which these observations are discrete. If observations are independent both over individuals and over time, a panel data QR model does not pose any new problem. It is just another QR model, possibly with more observations than usual. Special problems arise only when observations are assumed to be correlated over time. Serial correlation, or temporal correlation, is certainly more common than correlation among observations of different individuals or cross-section units. In this section we shall consider various panel data QR models with various ways of specifying serial correlation and shall study the problem of estimation in these models.