- Current research grants
- Finished research grants
- Current Dissertation Projects
- Finalized Dissertations
Development of a regionalized EU-Wide operational model to assess the impact of Common Agricultural Policy on farming sustainabilityProject for IPTS/DG-JRC, Sevilla, Call for tender No. J05/30/2004
- applies CAPRI
- and develops a farm type module based on FADN
- training activities for IPTS staff
- reference run and countervailing scenarios to test new system
Development of an operational modeling tool to assess the impact of agricultural policies on regions
- It is Pan-European (currently covering EU 25, inclusion of Romania and Bulgaria is proposed in here) and regionalized at NUTS II level.
- It combines regional programming models with a global multi-commodity model for main agricultural products, thus allowing for endogenous prices for agricultural outputs.
- Special modules or parts of the regional programming models deal with pricing of young animals, own produced feed and organic fertilizer.
- The model is linked to environmental indicators (N,P,K balances, output of CH4, NH3 and NOx, calculation of Global Warming Potentials in line with IPCC guideline, water scarcity indicator; a module to break down the NUTS II results to small homogenous mapping units, allowing a link to bio-physical models and landscape indicators is under development).
- The model is realized in GAMS.
- It is typically used for medium-term analysis (8-10 years from the current base period).
Development of a farm typology and integration of farm system models for a regionalized agricultural sector model
- typical rounding errors in FADN records. Many farmers will for example round up reported hectares. It is hence rather probable that a reported 5 may be somewhere between 4.5 and 5.5. That means changes ranging from 100 to 100.5 and 1 to 1.5 are considered equally probable, whereas changes beyond this range should be weighted based on the relative difference.
- the importance of activities and products for the classification of a specific farm type. Whenever possible, the cereal production and hectares of a specialized cereals farm should stay stable, whereas more pronounced changes should be allowed for other small elements in the production pattern (e.g. small animals herds) as long as their total contribution to farm income remain minor. Equally, any residual farm type categories should be allowed to change more easily to achieve consistency.
- representativity of aggregation weights. Farm types aggregated from a small number of actual FADN records provide more noisy data information than one based on a large number of farms.
- Aggregation weights per farm type multiplied with activity level (hectares, herd sizes) are equal to given NUTS II activity levels.
- Aggregation weights per farm type multiplied with physical production are equal to given NUTS II production.
- Physical production per farm is equal to output coefficients times activity levels.
- Farm classification (by size class and specialization) is kept fixed for each farm, and defined as a function of activity levels according to the standard gross margin concept of FADN.
Reference run and validation of the operational modeling tool
- A complete update of the data base (market balances, import and export flows, tariffs [globally for individual countries], FEOGA budget, CAP policy [EU level], economic accounts [Member State level], land use, herd sizes, yields [at regional level])
- Trend analysis for yields [regional level], efficiency in fertilizer use [Member State level], as well as market balances and prices [Member State level]. Technically, in a first step independent trends for all variables to project are calculated. From these trends, supports are derived which are linear combination of the trend value and the last known values, where R▓ is used as the weight for the trend value and (1-R▓) as weight for the last known values. These supports drive the final step, where the minimal deviation to these supports are determined with market and land balances, feed requirements etc. as constraints.
- Literature review on results from agricultural projections [e.g. FAPRI, DG-AGRI, FAO, IFPRI], which are used for comparisons. For everything not EU-27, no own trend analysis is undertaken, and shifts in the market model for these countries/country blocks are completely based on external studies.
- Development of scenarios to assess the impact of policy on farming systems sustainability
- The efficiency of resource use in terms of trends of productivity (i.e. total factor productivity), net value added, cash flow, profit, and agricultural income per unit of resource used such as land, labor, or capital).
- The regional on-farm employment consequences of different scenarios will be assessed using a module estimating labour demand depending on the production program which was developed within the CAPRI network. Off-farm employment consequences will be discussed based on the importance of agriculture as an employer in the regional economy.
- The environmental impact will be asserted by considering the positive and negative externalities associated to each farming system model. The evaluation will be based on environmental indicators implemented in the model.
The partial equilibrium agricultural sector model CAPRI matches the major requirements mentioned in the call for tender and therefore applied for the project:
The main work in the context of the project is the documentation of the enlarged system and training for IPTS staff.
The structure of the individual aggregate supply models in CAPRI will continue to be identical across farm type models (see: project for DG-ENV), differences will be expressed in farm type specific parameter sets. That requires that the model equations and variables are general enough to build a framework which captures the main interactions between activities on farm and between the farm, markets and the environment. Nevertheless, scarce resources plus the pan-European approach with endogenous prices define limits for the complexity of the single farm programming models, and will certainly lead to distinct differences to typical single farm models.
The most obvious difference between the proposed farm type model and typical, well developed single farm type model will be the use of a non-linear cost function approach with adjustment costs whose parameters will be derived both from econometric estimations, literature reviews and engineering knowledge. The cost function approach serves multiple purposes:
(1) it allows for perfect calibration of the model to given ex post data,
(2) it allows for technically elegant and easy way to represent complex interactions between activities on farm where a primal representation by technological constraints would be too expensive or even impossible given missing data, and
(3) it avoids the typical overspecialized corner solution in LP models with a limited number of constraints, and
(4), perhaps most important, links model behavior to observed behavior of farmers.
The second major question is that of the farm typology. It must first be realized that technical constraints set limits to the overall number of independent farm type models in the system. The current layout with regional models (about 200 for EU25 plus Norway) will run for about 1.5 hours on Intel Workstations, a version of 5 typical farms per regions would certainly double to triple solution time. As the majority of runs are development runs where new features/assumptions are tested, especially in the context of the reference run, solution time and result quality are highly interlinked. Further on, the number of farm type models does not only influence the time of model runs, but the complexity of model analysis as well.
We assume currently that probably between 5 to 10 typical models per NUTS II region are technically feasible, so that a total of 1000-2000 models will be later on linked into the complete system. We propose a two-dimensional typology by specialization and size, as many other elements related to technical, economic, environmental and social dimensions are very closely linked to farm specialization and size. Specialization will be based on the FADN 3-digit typology which is linked to the concept of the Standard Farm Income, where size could be measured either by farm endowment (hectares, number of animals) or by economic size classes.
A third major challenge is the question of consistency to regional statistics. We propose in here to ensure perfect aggregation from farm types to NUTS II level. The reasons are as follows. (1) given the sensitivity of policy instruments as subsidized exports, intervention sales, or import quantities and tariffs to slight changes in the degree of EU self sufficiency, even differences well below 10 % may change important results. (2) Perfect aggregation to regional CAPRI statistics allows aggregating the farm types in two directions consistently to Member State by EU level: by farm type or regions. (3) It ensures that the farm types cover the whole agricultural usable area, an important aspect for environmental analysis.
Methodologically, posterior density estimators will be used to ensure minimal differences in production pattern and aggregation weights necessary to achieve consistency. The a-priori densities attached to uncertain variables and certain correction factors will reflect
The resulting problem is hence a non-linear constrained optimization problem with the following constraints:
Aggregation weights, yields, activity levels and production quantities per farm are estimated values during the process of achieving consistency to regional statistics. The change in these estimated values compared to the original data found in FADN are determined maximizing the posterior probability density given the assumed a priori density functions for these estimates subject to the restriction mentioned above. During solution of the model version developed in the context of that proposal, the aggregation weights will remain constant, whereas the SEAMLESS IP will add a module to change aggregation weights endogenously in model application.
Given the short time frame, major changes regarding the current proceeding used in CAPRI are not feasible. It is hence proposed to fine tune the existing approach. The basic steps for build a reference run are the following:
Due to the tight time schedule, up to five alternative policy scenarios will be developed and assessed against the reference situation developed in task 3. Policy relevant scenarios will be chosen such that the full range of possible assessments with the CAPRI system and its here in updated farm system module is illustrated.
The detailed scenario specification will be discussed at the kick off meeting. An aspect of particular interest is the flexibility of Member States in regard to the decoupling of direct payments as mentioned above. It is proposed that scenarios such as alternative decoupling schemes, harmonization of decoupling, single payment scheme versus hectare premium should be on the agenda.
Alternative policy scenarios will be evaluated with respect to the three dimensions of sustainability i.e. economic feasibility, environmental safety, and social equity. For each farming system model the following dimensions will be considered in particular:
Current project team in Bonn
Last updated: Thursday, January 05, 2017
- Brexit: an economy-wide impact assessment on trade, immigration, and foreign direct investment, Jafari, Y., Britz, W. (2020): Empirica 47(1): 17-52.Efficiency differentials in resource-use among smallholder cassava farmers in southwestern Cameroon, Molua, E. L., Tabe-Ojong, M. P., Meliko, M. O., Nkenglefac, M. F., Akamin, A. (2019): Development in Practice: 11 pages.New insights on efficiency and productivity analysis: Evidence from vegetable-poultry integration in rural Tanzania, Habiyaremye, N., Tabe-Ojong, M. P., Ochieng, J., Chagomoka, T. (2019): Scientific African 6(e00190): 11 pages.Explaining farm structural change in the European agriculture: a novel analytical framework, Neuenfeldt, S., Gocht, A., Heckelei, T., Ciaian, P. (2019): European Review of Agricultural Economics 46(5): 713-768.Machine learning in agricultural and applied economics Storm, H., Baylis, K., Heckelei, T. (2019): European Review of Agricultural Economics, jbz033: 44 pagesInterdependence between cash crop and staple food international prices across periods of varying financial market stress Amrouk, E. M., Grosche, S., Heckelei, T. (2019): Applied Economics: 16 pages.Household Welfare Implications of Better Fertilizer Access and Lower Use Inefficiency: Long-Term Scenarios for Ethiopia Engida, E., Srivastava, A. K., Kuhn, A., Gaiser, T. (2019): Sustainability 11 14(3952): 24 pages.International Interdependence between Cash Crop and Staple Food Futures Price Indices: A Dynamic Assessment Amrouk, E. M., Heckelei, T., Grosche, S. (2019): Journal of Agricultural and Applied Economics: 1-22Economic Valuation of an Urban Lake Recreational Park: Case of Taman Tasik Cempaka in Bandar Baru Bangi, Malaysia Othman, J., Jafari, Y. (2019): Sustainability 11 (11)(3023): 15 pages.