Measuring and analyzing performance in longitudinal data

The aim of this research is the modelling of longitudinal data, with specific reference to the analysis of business performance and factors influencing it. From the methodological point of view, we consider an innovative approach for the measurement of performance in panel data, based on the measure introduced by Kokic et al. (1997). This measure is derived from an M-quantile (MQ) regression (Breckling and Chambers, 1998) based on assumed production function. We extend this measure to panel data. With reference to the analysis of factors influencing performance, both static and dynamic error component regression models are considered. The empirical analysis is carried out using the Kauffman Firm Survey data.