On the Composite Indicators for Food Security: Decisions Matter!

AbstractDuring the past decades, there has been much debate on food security. A variety of indicators have been proposed in order to establish which countries are in need of improved food security status. The heterogeneity of existing indicators and the lack of consensus on how to compare and rank countries have motivated international organizations to build composite indexes to synthesize the information. The process of building composite indexes involves multiple choices that influence the outcome. This analysis aims at understanding how relevant and discretional may be the analyst's choice of algorithms to compute composite indexes for food security. To this extent, we have computed several composite indexes for food security by using data provided by the Food and Agriculture Organization, which includes a large set of proxies for food security, as emerged from the Committee on World Food Security Round Table. We compare different methods to impute, homogenize, weight, and aggregate data, in order to compute composite indexes and show how relevant are the choices to be made. We show that normalization and weighting are not very crucial decisions, whereas special attention has to be paid in choosing the data imputation and aggregation methods. By commenting on the implications that different measurement choices may have in terms of global index, we show that the index construction decisions matter. Keywords: Composite indexDevelopmentFood policyMeasurementMillennium goal AcknowledgmentsThe author thanks the Statistic Division of the Food and Agriculture Organization of the United Nations for having provided data on food security indicators. The author is very grateful to Carlo Cafiero, Adam Prakash, two anonymous referees, and the editor for their valuable comments. All remaining errors are the author's. Notes1 More information on the Medium Term Plan is available at: http: //www. fao. org/docrep/meeting/027/mf490e. pdf. 2 More information on the new metrics that measure food security is available at: http: //www. fao. org/fileadmin/templates/ess/globalₛtrategy/GSHighLevelMeeting/GS2012EM4Scorecards₂0121204. pdf. 3 The views expressed in this article are those of the author and do not necessarily reflect the views or policies of FAO. 4 More information on the Committee on World Food Security's final report can be found at: http: //ftp. fao. org/docrep/fao/meeting/018/k7197e. pdf. 5 A three-indicator index would be as follows: (ax1) ρ (bx2) ρ (1 − a − b) x3) ρ−1/ρ, where a and b ranges from 0 to 1, and −1 ≤ ρ ≤ ∞. 6 The database was released on October 9, 2012, and revised on March 15, 2013. The author is grateful to the ESS FAO Division for having provided the data set. 7 The latter is adopted to compute the Global Food Security Index 2013. The formula is as follows: . More information on the GFSI can be found at: http: //foodsecurityindex. eiu. com/. 8 Geometric aggregation is intended to capture the hierarchical structure of the phenomenon. As Barrett (Citation21) pointed out, "Availability, access, and utilization … are inherently hierarchical, with availability necessary but not sufficient to ensure access, which is, in turn, necessary but not sufficient for effective utilization. " The structure calls for further research on how to aggregate the subindex. An alternative, yet not empirically investigated, is to use a quasi-linear aggregation method or to use a Stone-Geary-type function. This is a promising research area. 9 It is worth noting that most of the proposed indexes (e. g. , the Global Food Security Index proposed by the Intelligence Unit and the Global Hunger Index proposed by the IFPRI) are slight variants of our baseline scenario.