A decision support methodology for increasing public investment efficiency in Brazilian agrarian reform

Abstract

The Brazilian Agrarian Reform Program has subsidized the settlement of over 425,000 destitute families on previously unproductive land in what has become a very effective vehicle for social inclusion and productivity growth for those settlers who reach the final stage of the process and receive definitive title to the land. Unfortunately, there is a large difference in efficiency and productivity between more and less successful settlements -fewer than 10% of relocated families have received title and over 25% of them have abandoned the property to which they were assigned. This paper presents a decision support methodology for increasing the efficiency of public investments in agrarian reform that includes a data envelopment analysis model and a mechanism for building consensus among the various constituencies of the agrarian reform process, who not infrequently have conflicting objectives. The OR model described herein uses principal component analysis and data envelopment analysis to identify the most important success factors for relocated families leading to an increase in the chance of both autonomous integration with the market economy and definitive entitlement by these displaced families as well as an increase in the predictability of future settlement success. The model was implemented successfully in Rio Grande do Sul, the southernmost state of Brazil, and was partially used in a pilot project for the countrywide agrarian reform accelerated consolidation program.Emancipac¸a˜o de Assentamentos Resultantes da Reforma Agra´ria -PAC) of the Brazilian Department of Agrarian Development. The methodology applies data envelopment analysis (DEA) techniques to measure settlement efficiency and to modify each settlement's modus operandi in an effort to shorten the length of time to achieve self-sufficiency.The rest of the paper is organized as follows. Section 2 briefly describes the Brazilian agricultural sector and provides the context within which agrarian reform has been used as a policy instrument for wealth and income distribution as well as sustainable economic development. Section 3 describes the Brazilian agrarian reform program in detail and suggests some reasons for the proliferation of inefficient settlements. Section 4 offers a brief literature review of the DEA methodology including its use in rural applications and a description of different techniques for identifying the appropriate input and output variables. Section 5 describes the above mentioned decision-support methodology. Section 6 describes the application of this methodology in southern Brazil, including a description of the region, the characteristics of the settlements, and the various input and output variables used. Section 6 also presents results of this Rio Grande do Sul case study. The last section offers concluding remarks and comments on the methodology's possible role as an agrarian reform decision-support tool for the Brazilian government. L. Melgarejo et al. / Intl. PORGpond -qualitative variable describing the level of organi...

Augmenting efficiency of public investment in the Brazilian agrarian reform

“…Finally, principal component analysis and stepwise elimination were performed to reduce the parameter space to eight input and four output variables, which were then used for e ciency analysis. This nal dataset is fully tabulated in Melgarejo et al (2009), the reader is referred to this original reference for ne details on the variable selection process (the above description represents just a succinct overview).…”

“…In the current case of the Brazilian agrarian reform, the initial choice of variables was made by consensus of the interested parties (Melgarejo et al, 2009), but the agreed upon variable space turned out enormous:…”

“…by the above mentioned rule of thumb , for and one would need data on at least 5000 settlements only to begin considering DEA as the method of choice for this study. The subsequent reduction of the variable space to 8 input and 4 output variables resulted in considerable alleviation of the curse of dimensionality problem (Melgarejo et al, 2009), but the nal results remain somewhat ambiguous, considering the fact that ~ 50% of the DMUs turn out e cient, whereas only ~ 10% reach the stability objective. In the following section it is shown how the MDEA approach breaks down the tie between the numerous DMUs that comprise the standard DEA e ciency frontier.…”

“…In a previous work Melgarejo et al (2009) proposed application of Data Envelopment Analysis (DEA) (Farrell, 1957;Charnes et al, 1978;Banker et al, 1984;Charnes and Cooper, 1985;Färe et al, 1985;Färe et al, 1994;Charnes et al, 1994), aided with prior application of Principal Components Analysis (PCA) and stepwise elimination (Norman and Stoker, 1991) for selection of input and output variables, with the goal of identifying the faults in the settlement practice, and suggesting actions necessary for better use of public spending in the Brazilian agrarian reform. They used data on 90 settlements in the state Rio Grande do Sul (Melgarejo, 2000) with over 3600 families.…”

Single‐phase slack‐based centralized DEA for resource reallocation

“…Since Charnes, Cooper, and Rhodes proposed the first data envelopment analysis (DEA) model in 1978, several other DEA models have been developed and utilized (e.g., Melgarejo et al, 2009;Saranga and Phani, 2009;Lozano et al, 2011;Figueiredo and Barrientos, 2012;Mar-Molinero et al, 2014;Yu et al, 2013;Yadav et al, 2014). One of these models is the centralized DEA (CDEA) model proposed by and .…”