Lesson 5.2: EXPAMOD

Icône de l'outil pédagogique Authors

Irina V. Bezlepkina, Ignacio Pérez-Domínguez, Thomas Heckelei, Alfons Oude Lansink, Erick Romstad


Icône de l'outil pédagogique 5.2 EXPAMOD – Extrapolation Model

Farms are the basic decision units in agriculture and, therefore, influence market outcomes, land use and the environment. Moreover, farmers perceive prices as given, since production in each production unit is small compared to total production in the sector. Farm level optimization models share a similar perspective. As long as the policies investigated are such that market prices remain stable, the error made by taking prices as exogenous is negligible. However, agricultural policies are horizontal and affect all farmers at the same time. Therefore, their aggregated response and their interaction may have profound impacts in agricultural markets, which in turn influence commodity prices. Agricultural sector models such as CAPRI (Common Agricultural Policy Regionalised Impact analysis) are able to capture price changes from policies.

The integrated project SEAMLESS aims at providing a consistent integration between a farm management model specified for a selection of farm types (FSSIM) and an EU wide aggregated agricultural model with an explicit market component (CAPRI). Given the ca. 250 NUTS2 regions and high diversity of farm types within the EU27, FSSIM is only capable of considering a subset of regions and farm types in detail. This is why a complementary procedure for expanding FSSIM results to all other regions is needed. The principle behind the presented methodology (EXPAMOD) is to make the regional supply modules of CAPRI behave like the aggregate of the FSSIM models of the same region (Figure 1).

 

Figure 1. Flow of prices (1, 2 and 4) and price supply elasticities (3) between models under a policy scenario.

 

In order to map the supply behaviour of farm management models to the market model, the EXPAMOD methodology comprises the following sequence of steps (see Figure 1):

(1) Price shocks are modelled in the existing FSSIM farm type models, with initial prices coming from CAPRI

(2) Estimation of supply responses as response functions of price variations, farm characteristics, regional soil and climate conditions; extrapolation of supply response to other farm types and NUTS2 regions; aggregation of supply response to level of CAPRI regions (administrative units) and CAPRI product categories;

(3) Calibration of regional supply modules in CAPRI to aggregated supply responses;

(4) A final run of FSSIM with market clearing producer prices from CAPRI results in the final consistent specification at the farm level.

 

Since the approach is still being tested, only some empirical implications of the theoretical approach are emphasized here. In order to arrive at the sensible number of FSSIM models to generate supply-price responses and enable the extrapolation from sample regions to all EU regions, a specific farm typology has been derived within the SEAMLESS project. According to this, 3 to10 most-representative farm types per sample region are selected, so that they adequately represent farm, soil and climate differences among regions. The current simulation design includes the generation of pseudo-observations by running the farm management models on different price sets for products (i.e. input data based on a sensitivity analysis of farm model behaviour, prices being changed one-at-a-time). The level of prices for the baseline scenario is kept at the 100% level to the initial prices levels obtained by CAPRI in a similar scenario. Prices of each product are set at 60%, 80%, 120% and 140% of the base level, providing a sufficient number of simulated observations. The choice of agro-environmental variables links closely to the determinants of farm typology (size, intensity, specialization) as well as agro-environmental typology (soil, climate) used in farm spatial allocation procedure. The list of selected variables in modelling the supply response function: economic size unit, machinery, labour, carbon content, root depth, minimum temperature, precipitation, radiation is supported by earlier studies of crop yield variability.

The response function estimated at one scale is not representative for the projection at another scale. Estimation of the response function is based directly on the simulated FSSIM data. Aggregation from farm types within one NUTS2 to the regional level, as needed by CAPRI, is performed by using the shares of land area of each farm type obtained during the farm spatial allocation exercise. The calibration of the response of the CAPRI supply module to the aggregated response of the FSSIM farm management models and the extrapolated regional models is done through regional price-supply elasticities.

Régi par la licence Creative Commons Attribution Non-commercial Share Alike 3.0 License

copyright Seamless 2009