Implementing the ERS Food Security Model Development

Co-Principal Investigators

OBJECTIVE:  The Food Security and Development Branch (FSDB) of the Economic Research Service (ERS) conducts an annual assessment of food security in 76 low and middle income countries.   The modeling framework used for the assessment has a strong focus on food availability, but is less robust in its ability to analyze the remaining three dimensions of food security--access, utilization and stability.   ERS has developed a new modeling framework which  strengthens the model’s capacity to analyze food access and improve its micro economic foundations.  ERS plans to use the new model for its 2016 international food security assessment.

The specific objectives of this cooperative research project are as follows:

  1. Jointly with ERS staff, develop, calibrate and run an R-based version of the newly developed model incorporating all new features currently implemented in its spreadsheet version.
  2. Together with ERS staff, create country specific models for countries included in ERS’s international food security assessment
  3. Assist ERS staff with model evaluation and validation

APPROACH:  Objective 1.  ERS is in the process of creating an R based model, consistent with an Excel spreadsheet prototype. There are several perceived advantages to the R-based model, including ease in updating data and populating country models, ease in documenting the model and its implementation options, consistency with USDA’s open data initiative and ease in disseminating the model to users outside ERS. A working R-based model will be the key output for objective 1.

Objective 2.  ERS conducts an annual international food security assessment which requires country-level model results, which can be aggregated to provide regional and international totals.  While country specific models all follow the structure of the prototype model, some unique country features, as well as unique parameters, will be part of each country model.  Many of these features will be options to be selected in the R-based model, although there may be some additional changes which will be identified as the country-specific models are developed.  Because ERS conducts its international food security assessment annually, new data are required for each year.  The R-based model will extract data from identified external sources, create data files, and populate country-specific models with new data.   Working country models will be the key output for objective 2.

Objective 3:  ERS needs to evaluate the new country models, and conduct both internal and external validations.   A comparison of the new model with results from the previous ERS model is a key agency requirement for moving forward with the new model.   ERS will work with the MSU cooperator to identify appropriate validations procedures, conduct them, and evaluate the results.  If required, changes will be made to country models.