Technical and Scale Efficiency in the Italian Citrus Farming: Comparison between SFA and DEA Approaches

Abstract

1. IntroductionSince the Farrell's (1957) seminal paper several procedures have been proposed in order to calculate efficiency and productivity. However two main approaches have been mainly proposed in literature: the Stochastic Frontier Analysis (SFA) and the Data Envelopment Analysis (DEA). Both approaches have their advantages and disadvantages and the suitability of method to the data depends on the industry to be examined (Ruggiero, 2007). Literature has shown several studies on comparing the two approaches (Gong and Sickles, 1992; Hjalmarsson et al. 1996; Sickles, 2005) and more papers have focused attention on agriculture (Kalaitzandonakes and Dunn 1995; Sharma et al. 1997; Wadud and White, 2000; Theodoridis and Psychoudakis, 2008; Minh and Long 2009). They have mostly investigated on differences between technical efficiency scores and their distribution on the observed sample even if discordant results have been found (Sharma et al., 1997; Ruggiero, 2007; Minh and Long, 2009).On the contrary, poor relevance has been given on comparison in scale efficiency scores1. Scale efficiency is a measure inherently related to the returns to scale of a technology at any specific point of the production process (Forsund and Hjalmarsson 1979). It measures how close an observed plant is to the optimal scale, i.e. it describes the maximally attainable output for that input mix (Frisch, 1965). A great number of papers have estimated scale efficiency in agriculture, generally using a DEA model (Bravo-Ureta et al., 2007). However choice of the method is a crucial issue because differences in scale efficiencies interpretation and scale properties might derive from inherent differences between parametric and non parametric models. It is expected to obtain some differences according to the methodology applied for estimating scale efficiencies in terms of scores - e.g., caused by divergences in properties of the technology itself or in evaluation of distance to the frontier - and their distribution on the sample (Banker et al., 1986; Forsund. 1992; Orea, 2002). Therefore, the issue is determining the true efficiency of a firm (Andor and Hesse, 2011). Using empirical data, it is impossible to evaluate the performance of the methods or to demonstrate absolute advantages of a method over its competitors. However, comparison between the two methods allows us to put on evidence if an estimated differences in efficiency measures exists and eventually to estimate the influence factors that lead to these differences.The aim of this paper is to contribute in the existing literature providing a comparison between SFA and DEA approaches for estimating efficiency in agriculture with particular attention on scale efficiency. Specifically, we estimated efficiencies in the Italian citrus farming. This sector has been historically characterized by presence of small scale farms and structural disadvantages. If significant technical and/or scale inefficiency were found, this would indicate that structural problems prevent farm expansion and the rational use of technical inputs. It is the first attempt of comparing SFA and DEA scores in the Italian agriculture.2. The Italian citrus farmingCitrus fruit growing is one of the largest categories in the Italian vegetable and fruit sector (Giuca, 2008). The value of production amounts to about 1.5 billion of Euros that corresponds to about 3% of the total gross domestic product from Italian agriculture (Ismea, 2011). In terms of value, oranges contributes to more than 50% to Italian production and about a 40% is equally distributed by tangerines and lemons.Citrus farming is performed by more than 80,000 farms, mostly located in the southern regions of Italy. Specifically, more than 70% of the farms operate in Sicily and Calabry, whereas the rest of the farms are sited in Apulia, Campany, Sardinia, and Basilicata. During the period 1985-2014, the number of farms and the land area covered by citrus fruits have decreased by about 35% and 45%, respectively (Istat, 2015). …