PANEL DATA ANALYSIS
7.1 WHAT IS PANEL DATA?
The discussion in the previous chapters focused on analytical methods for cross-section data and (to some extent) time series data. This chapter deals with the analysis of data from panel studies. Data from such studies consist of repeated measurements on cross sections over a period of time. In other words, in panel data there are repeated observations on the same subject over a time period as in longitudinal studies where, for instance, subjects are surveyed or followed over time. Here the term subjects will be used to refer to people, countries, companies, and so on.
As an example of panel data, consider the Cost for US Airlines example from Greene (2003). The data set has 90 observations for six firms from 1970 to 1984. This data set has been analyzed in Chapter 3, and it will be used to illustrate the different analytical techniques in this chapter. We have a cross section if we take data from a single year. The selected cross section simply gives a snapshot of the six airlines for the selected year. We have panel data if we use data for the six airlines from every year in the time period 1970–1984.
As another example of panel data, consider the case where a financial planning company collects data on profits generated by its financial advisors. For instance, the profit of the financial advisors may be monitored for several years. Various factors such as the regional location of the advisor, their age, and the wealth profile of ...