On the measurement of efficiency and productivity under firm heterogeneity

The thesis addresses firm-heterogeneity in micro-level panel datasets and the necessity to take this heterogeneity into account to obtain unbiased measures of productivity (growth) and technical efficiency in a parametric frontier framework. Various existing and newly introduced econometric specifications of stochastic frontier models are discussed. The focus lies on these models’ assumptions about heterogeneity and how they attempt to distinguish heterogeneity from inefficiency. Two studies tackling empirical questions complete the thesis. In addition to elaborating on various forms of firm-heterogeneity that can complicate empirical work, these studies provide a detailed decomposition of sectoral labor productivity growth into between and within firm effects and a group- and chain-linked multilateral productivity index.