Annex 4 analytical methods

model description

general description

acronym: AGLINK-COSIMO

name: AGricultural LINKage - COmmodity SImulation Model

main purpose: A global agricultural economic model used to simulate the medium-term development of annual supply, demand and prices for the main agricultural commodities produced, consumed and traded worldwide. It has been extended to simulate the economic impacts of market uncertainties and climate extremes.

homepage: www.agri-outlook.org

Developer and its nature

ownership: Third-party ownership (commercial companies, Member States, other organisations, …)

ownership additional info: The OECD and the FAO are the sole owners of the model. The European Commission belongs to the users network and has a written agreement to use the model.

is the model code open-source?: NO

Model structure and approach with any key assumptions, limitations and simplifications

details on model structure and approach: The overall design of the AGLINK-COSIMO model focuses on the potential influence of agricultural and trade policies on agricultural commodity markets in the medium-term, typically 10 years ahead. Development on the basis of the (agricultural) economics literature, existing country-level models, and formal bilateral reviews has resulted in a modelling system that reflects the views of participating countries. To remain tractable, the model specification imposes some degree of uniformity across country modules. Taking this constraint into account, agricultural markets are modelled to best capture relevant settings and policies that are country- and commodity-specific. In undertaking projection work with the AGLINK-COSIMO model, individual country modules are calibrated on baseline projections that participating countries submit annually to the OECD and FAO in the form of structured questionnaires.

model inputs

Main inputs to the European Commission’s version of the model are:

The latest OECD-FAO Agricultural Outlook (issued every June), which is updated with
- the short-term outlook for EU commodity balances: crops (wheat, maize, coarse grains, sugar beet, oilseeds etc.), meat (dairy cattle, suckler cows, sheep, pigs, poultry etc.), dairy, sugar production and biofuel production.
- the latest macroeconomic and policy assumptions, and
- new model developments in terms of equations and data to better represent EU agricultural markets and policies.

Variables in the model can be endogenous (i.e., determined within the system) or exogenous (i.e., determined outside the system and simply inserted). Most behavioural equations are "double-log" which are popular in the estimation of supply and demand functions. In those functions, explained variables (on the left-hand side) and explanatory variables (right-hand side) are expressed in logarithmic terms; that is Y experiences diminishing marginal returns with respect to increases in X:

log(Y)= a + b*log(X) + log(r)

where a is the intercept, b is the Y-to-X elasticity (constant), and r is the residual (so-called ‘r-factor’). Numerous variations of this general form exist to represent real-world movements, such as technological change and cobweb-like market adjustments. Intercepts, which are time-invariant, and r-factors, which are time-variant, are interdependent and equation-specific calibration terms. These terms are endogenous during model calibration but remain exogenous in simulation mode (e.g., for scenario analysis). Year-specific shocks are implemented by changing the corresponding r-factors of endogenous variables or the actual values of exogenous variables. Oil prices and macroeconomic factors, such as GDP growth, inflation, exchange rates, energy prices, and population are exogenous.

model outputs

AGLINK-COSIMO generates projections on annual market balances for the next 10 years.

Key variables include production (e.g., crops, livestock), consumption (food, feed, biofuel, other industrial uses), trade (imports, exports), stocks (public, private), and prices (domestic producer, domestic consumer, domestic feed, global) of major agricultural commodities. The model covers over 100 commodities ranging from crops, such as wheat or maize, to processed goods and by-products, such as protein meals and distiller dried grains.

Intended field of application

policy role

The AGLINK-COSIMO model provides a yearly updated medium-term baseline to other market models used in the Commission, such as CAPRI, MAGNET and AGMEMOD. It is therefore used indirectly for ex-ante impact assessment (e.g., biofuels, climate negotiations, CAP reform, trade agreements). The partial stochastic analysis based on AGLINK-COSIMO is also used to analyze ex-ante the impacts of specific policy reforms.

An important activity of the European Commission is the annual production of medium-term (10 years) baseline projections for EU agricultural commodity markets (EC 2019), published annually by the Directorate General for Agriculture and Rural Development (DG AGRI) in the second half of the year. AGLINK-COSIMO, which is maintained at the JRC, is the key tool for building those baseline projections as well as for performing uncertainty and sensitivity analyses due to alternative macroeconomic environments and crop yields.

policy areas

Agriculture and rural development

Model transparency and quality assurance

Are uncertainties accounted for in your simulations?: YES - Uncertainties in input data and simulations are quantified every year with the partial stochastics module (Araujo-Enciso et al. 2017). Uncertainties in parameters are dealt with on a case-by-case basis using deterministic shocks.

Has the model undergone sensitivity analysis?: YES - Sensitivity analysis is performed every year with the partial stochastics module (Araujo-Enciso et al. 2017).

Has the model been published in peer review articles?: YES

Has the model formally undergone scientific review by a panel of international experts?: YES - Done formally by the OECD in 2009-10

Has model validation been done? Have model predictions been confronted with observed data (ex-post)?: YES - Informally, on a case-by-case basis. An ex-post exercise can be found in the OECD library (2013 version, Box 1.1).

To what extent do input data come from publicly available sources?: Based on both publicly available and restricted-access sources

Is the full model database as such available to external users?: NO - Property of the model consortium

Have model results been presented in publicly available reports?: YES

Have output datasets been made publicly available?: YES - Updated annually.

Is there any user friendly interface presenting model results that is accessible to the public?: NO

Has the model been documented in a publicly available dedicated report or a manual?: YES

Intellectual property rights

Licence type: Non-Free Software licence

application to the impact assessment

Please note that in the annex 4 of the impact assessment report, the general description of the model (available in MIDAS) has to be complemented with the specific information on how the model has been applied in the impact assessment.

See Better Regulation Toolbox, tool #11 Format of the impact assessment report).