Agricultural Investment Data Analysis (AIDA)

Overview

The Agricultural Investment Data Analysis (AIDA) tool is embedded in IFPRI’s Rural Investment and Policy Analysis (RIAPA) modeling system. The integrated AIDA-RIAPA framework is used to analyze the economywide impacts of investments in the agri-food system (AFS) and help inform the design and prioritization of policies and investments given their impacts on development outcomes, such as economic growth, job creation, poverty reduction, or dietary improvements. Whereas AIDA is used to translate new investments in, say, irrigation, extension, or rural roads into changes in productivity, a recursive-dynamic Computable General Equilibrium (CGE) model, which sits at the core of the RIAPA system, measures the direct and indirect (spillover) effects of those investments on the rest of the economy. For example, investments in extension not only increase farm productivity of those receiving extension visits, but also create new jobs and income opportunities in downstream nonagricultural sectors, such as food processing, trading, and transport. This has welfare implications for households and individuals, which are analyzed using a series of microsimulation add-on modules also embedded in RIAPA.

How does AIDA measure direct investment impacts?

While productivity changes are usually introduced as exogenous shocks in CGE models, unpacking the sources of those productivity increases is important for policymaking. AIDA tracks the levels and composition of public expenditures in the AFS and translates these into productivity changes using information on investment costs (e.g., spending on irrigation), coverage rates (e.g., the share of farmers using irrigation), and impact coefficients (e.g., the impact of irrigation on farm-level productivity). AIDA compiles information for a wide range of investment options, including investments targeting the farm (e.g., irrigation, input subsidies, or extension services), and investments targeting the broader AFS (e.g., processing facilities, rural roads, or education). The tool therefore endogenizes the direct productivity shock, allowing analysts to experiment with different investment portfolios and examining how changes in the mix of policies and investments affect development objectives.

What data is required to set up an AIDA model?

The AIDA module is calibrated with information from a wide range of sources. Government reports provide high-level information on agricultural investment plans or flagship policies. Government budgets published by Ministries of Finance, or public expenditure analyses carried out by the FAO MAFAP program or as part of the World Bank Public Expenditure Reviews (PERs), provide detailed expenditure by policy or investment type. Ex post evaluation studies of large government and donor projects (e.g., conducted by governments or institutions such as IFAD) provide information on investment costs, coverage, and unit costs. Lastly, analyses of household survey data (e.g., randomized trials or econometric analysis of LSMS panel data), or biophysical models (e.g., crop models) provide estimates of crop responses to investments on the farm or in the broader AFS.

Figure 1 shows how these different data sources and analyses feed into the AIDA module to estimate changes in productivity. These productivity changes, alongside the investment requirements, feed into the RIAPA CGE and microsimulation models to produce estimates of economic and welfare outcomes.

Figure 1. The AIDA-RIAPA framework

How can analysts or policymakers use AIDA?

AIDA comes in two versions: an online version that is easy to use and allows users to run combinations of pre-coded scenarios; and an offline (desktop) version for more advanced users that allows for more flexibility in defining scenarios and model assumptions.

Online AIDA provides flexibility for the user to design a package of investments that differ by type (e.g., irrigation or extension), location (i.e., subnational regions), and level of spending. Users can also adjust certain model parameters, including the initial impact of the investment on productivity or the unit cost of expanding coverage. AIDA takes this information and, using pre-coded and pre-solved RIAPA results, reports the economywide impacts. Currently, online AIDA is available in four countries in the MENA region: Egypt, Jordan, Tunisia, and Yemen.

Offline AIDA is available (or under development) in all countries with RIAPA models. There are two steps to a typical AIDA-RIAPA analysis. First, the model is used to estimate the marginal returns to the next dollar spent on each policy or investment option in terms of growth, jobs, poverty, or diets. Outcomes vary across policies and investments because of their differential impacts on productivity growth, as estimated by AIDA, and because the links between investments and outcomes vary by sector being targeted. For example, productivity gains in livestock sectors are generally more effective at diversifying diets than the same level of productivity gains in cereals, since the latter already dominates diets of the poor. This step produces what is a called a RIAPA Policy Stack, a ranked list of policies and investments based on their cost-effectiveness in achieving a given set of outcomes.

Second, the estimated marginal returns are compared to the current pattern of public spending to determine if there are potential gains from reallocating spending towards higher-return policy and investment options. This step is known as Budget Rightsizing and produces useful information for governments as they consider how scarce public resources should be reallocated. A recent example of a typical AIDA-RIAPA analysis in Rwanda can be found here.

For more information, please contact IFPRI-RIAPA@cgiar.org.