65 TOTAL FACTOR PRODUCTIVITY GROWTH AND PUBLIC CAPITAL: THE CASE OF ITALY by Carmelo Petraglia 1. Introduction The aim of this paper is to contribute to the debate on the relationship be- tween public capital and productivity in the Italian regions, the main novelty being the decomposition of productivity growth into technical efficiency change and technological progress by means of Data Envelopment Analysis (DEA) and Malmquist productivity indexes. Applied work has commonly focused either on the so-called production function or on the traditional growth accounting approaches 1 . In the first case, an average production function which includes public capital is estimated and the relevance of public capital is measured in terms of output elasticity (Picci, 1999). The second approach involves two stages; in the first stage total factor productivity (TFP) growth is estimated starting from available data on inputs; in the second one, TFP growth obtained in the first stage is regressed on pub- lic capital (La Ferrara et al., 2000). These recent contributions have enhanced the literature on the topic, allowing for taking into account heterogeneity across Italian regions by means of panel data estimation techniques. However, their conclusions suffer from an important shortcoming: they do not take into account the “inefficiency issue”. This is due to the implicit assumption that the production process is fully efficient, which implies that the estimates of average production functions will be biased in the presence of inefficiency. Furthermore, if such an as- sumption does not hold, TFP growth will be identified with technological 1 A further approach followed in empirical work on Italy is the so-called growth theory ap- proach (Acconcia and Del Monte, 2000). Studi economici n. 78, 2002/3