The ability to feed the world’s future population is deeply rooted in the development of food loss reduction policies and practices now. Increasing efficiencies in the global food system and advances in a collaborative research agenda on post-harvest loss (PHL) are vital factors. However, achieving less loss will only occur when new policies are enacted and changed practices drive results. To achieve these results one needs the ability to assess proposed infrastructure projects using specific criteria.

Around the world, policies for the development of new transport infrastructure have various economic variables – especially the total project cost and the reduction of transportation costs. In recent years, environmental and social issues have gained importance. It is now time to consider PHL metrics when developing new infrastructure projects. Food loss reduction from practical infrastructure investments will aid in the efficient delivery of much needed food that will in turn provide significant economic, environmental, and social outputs.

Soy Case in Brazil

Brazilian agro-industrial chains face high transport costs, increased dependence on road transportation, accessibility problems, and lack of storage capacity. These obstacles need to be overcome for the industry to become more competitive and operate efficiently.

For the soy value chain, we analyzed a series of proposed multimodal infrastructure projects being developed by both public and private enitites. The analysis incorporated economic, environmental and social impacts of each project. The study incorporated 3 variables to ultimately compare the overall sustainability of the main multimodal infrastructure projects for soybean transportation in Brazil - a total of 26 rail and waterway projects/stretches.

The study was divided in two parts. The first estimated the impacts of the new infrastructure projects based on a model for minimizing transportation costs in the Brazilian soy transportation chain (a network equilibrium model). The output of these modeling processes are:

• Economical variables: estimated cost of the investment project, the potential for transport costs reduction due to the implementation of the infrastructure project, the amount of cargo loaded by the infrastructure, and the utilization rate of projected capacity;

• Environmental variables: CO2 emission in soy transportation operations and an estimate of physical soy losses during soybean handling; and

• Social variables: infrastructure area of influence and the Municipal Human Development Index (HDI) changes due to transportation cost reduction.

In the second part, we used eight variables described above to analyze the selected projects' relative efficiency through a multicriteria method. We used Data Envelopment Analysis (DEA) to compare the investment plans. We developed an efficiency score as a metric of the projects' sustainability. The greater the efficiency, the more the project should be treated as a priority in Brazilian infrastructure policy.

It is also important to note that the cost of investment, CO2 emissions, and the of physical soy losses are variables with behavior inversely proportional to efficiency – the lower the level, the greater the efficiency score.

A summary of the individual results (part one) for each of the analyzed cases shows that the Ferrogrão project scenarios are the most beneficial in terms of potential cost reductions and CO2 emissions during soybean transport operations. This project is also characterized by a more significant decrease in road transportation dependence to move soybeans in Brazil.

The full use of modeled handling capacity occurred in these projects: Ferrovia Paraense, the Vilhena (RO) railroad, the expansion of the Malha Norte, the Marabá (PA) waterway, and Ferrogrão. Scenarios related to the North-South Railway (FNS) and the Central-West Integration Railway (FICO) showed the highest volumes of soybeans captured by transshipment terminals, with FNS project also seeing significant social benefits based upon the metrics adopted in the study.

The Malha Norte railroad expansion project and a stretch of the FNS in the southern part of the country are associated with the lowest level of physical losses during soybean transport.

To give more detail about the soybean physical loss during transport operations in Brazil, the model shows that, nowadays, we lose 0.3778% of our soybean production during road, rail, and waterway transport. By analyzing the implementation of each infrastructure project, Malha Norte railroad expansion (scenario "C-EMN-2") counts to a total loss of 442,144.90 tonnes (0.3858%) and the southern part of FNS ("C-FNS-SUL") to 442,671.78 tonnes (0.3863%), as presented by Figure 1. The Ferrogrão and Ferrovia Paraense projects are the worst cases of total soybean loss during transport operations – 0.4301% e 0.4197% of Brazilian production.

Addressing the second part of the study, Figure 2 presents the DEA model results based on eight variables. We highlighted only the well-ranked projects.