This paper concentrates on the description of the EAA (Economic Accounts for Agriculture) compilation within the Activity Based Approach. First, an overview is given on the background that led to the development of the concept of a Policy Information System for CEEC agriculture. The most important aspect in this is the process of transition in Central and Eastern Europe, changing a central planned economy to a more market orientated economy. This process affected the agricultural production structure and brought a decrease in production, a major sectoral implication. In addition, statistical institutions were affected. Statistics received a new function in society and the information flow changed, having many instead of a few economically units and often with no obligation to supply data. Furthermore the envisaged accession to the EU forced the institutions to adopt the "Acquis" in order to fulfil the necessary requirements for membership.

Statistical information on agricultural developments are not only important for policy and for the economy in Central and Eastern Europe. For the European Union comparable data are essential, especially as agricultural policy claims a high share of the EU-budget.

In a FAIR-project "Agricultural Implications of the CEEC Accession to the EU", financed by DG VI, it is the task of a research team in Bonn to establish a consistent sectoral database for CEEC-agriculture and to develop a concept for a policy information system. For this database the definitions of the EAA have to be applied and the compatibility ensured with the SPEL-System (Wolf, 1995), an information system e.g. used continuos at Eurostat.

Considering the production breaks and the institutional changes in Eastern Europe, it is quite difficult to develop an information system based on time series and constant data flows. To establish a database that has to meet the EAA requirements in the face of uncertainties the research team in Bonn had to collect and compile all relevant data. Deriving the EAA was one result.

This paper will focus on three major elements:

- the concept of a Policy Information System for CEEC agriculture

- the activity based approach, used for EAA compilation

- the technical realisation of the Sectoral Basic Data System

To explain briefly the important aspects of the Policy Information System for CEEC agriculture, this section will show the major aims of the Information System and the basic structure used to fulfil these aims. The realisation of this Information System can be done on a national level, focusing on one agricultural sector, or on a multinational level containing comparable information about the agricultural sectors of several countries.

Major aims for the information system are as follows:

1. A flexible data import is necessary to allow the inclusion of information from various sources and to accept data-flows coming from altered structures in one country to facilitate differences between different countries. As our research cannot undertake surveys, data will be collected from several existing statistics sources but compiled into a unique structure, a data bank offering time-saving routines and a structured documentation.
2. Collecting data from different statistics, consistency and plausibility checks are necessary at all steps of data compilation. Therefore the balancing of supply and demand and of input and output can be undertaken in the sectoral framework, also allowing the filling of statistical gaps.
3. Monitoring of the agricultural sector is possible, as most of the necessary data to do this are collected and checked. This includes the calculation of the EAA income indicators and the detailed analysis of crop and animal production activities.
4. The ex-post analysis of agricultural developments allows the examination of intra-sectoral interdependencies and sectoral interactions, assuming that the data is available in a comparable form over a time period.
5. The last updated data and the results of the ex-post analysis serve as foundation for Forecasts and Policy Simulations.

To fulfil these aims, the Information System is conceptualised in a modular structure (see Figure 1).

The concept of a Sectoral Basic Data System has been developed at the University Bonn. In the EAA-modelling group in Budapest a close co-operation has been established between the University Bonn, ASA-Bonn, the University Gödöllö and the AKII to establish the System. This includes a data bank and user-friendly interfaces, allowing CEEC partners to use the routines for data import, consistency checks, monitoring and ex-post analysis. As the final result of the Sectoral Basic Data System, the consistent data set will serve as the foundation for a Simulation System. Keeping the compatibility to the SPEL-System, it will be possible to apply several other modelling approaches developed at the University of Bonn (Henrichsmeyer et al 1996). After the technical and empirical realisation, the next objective will be to explore some approaches, if appropriate, for the special conditions in the CEEC.

From the structure in Figure 1 one can recognise that the forecast and simulation modules depend on the quantity and quality of data from national statistics, which are compiled in the framework. A continuos dialogue between the national statistics and the user of the data has to be ensured to discuss feedback and support each other’s work.

The most characteristic methodological feature of the information system is the Activity Based Approach, which differentiates the agricultural sector into a number of production-activities and use-activities in a consistent framework. The matrix of 51 production-activities and 14 use-activities (column structure) and 57 products and 43 input items (line structure) reflect the total sector, with the identification of intra- and intersectoral flows and interdependencies. The choice of outputs, inputs and use activities in the framework follows the guidelines of the Manual on Economic Accounts for Agriculture and Forestry. But, in addition to the intersectoral flows according to the EAA (net concept), the intrasectoral flows between production activities in the National Farm within the gross concept could be represented and an approximately realistic gross production value could be calculated.

The approach is divided in four segments (output-generation, output-use, input-generation and input-use - see Figure 2), which can be shown as physical and value component, linked by prices. In fulfilling the consistency requirements, each of these segments represents the aggregate gross value of production. In the input area the payments for factors (at factor-cost) are shown.

Source: Own description

This section concentrates on the compilation of the EAA according to the requirements of the Manual of Economic Accounts for Agriculture and Forestry. Subsequent sections consider the advantages of applying the Activity Based Approach for EAA-compilation in the transition countries. Afterwards the technical realisation is presented in a short overview.

The compilation of the final output according to the EAA definition takes place in the block output use. The use activities in the columns and the agricultural products in the lines (see Figure 2) enables the calculation of the gross and net production (final output). The recorded agricultural products in the line structure contain all elements of the list forming the branch "Products of Agriculture and Hunting". Production is taken into account irrespective of the institutional context in which their production occurs.

The description in Figure 3 illustrates that only four elements have to be taken into account to calculate the final output. For these reason the following four use-activities out of the framework structure are taken for the calculation of the final output.

• Sales to the market
• Processing by producers
• Own consumption; on farm
• Stock Changes; on farm (including own produced fixed capital goods)

The definition of these activities in the framework fit the definition of the manual for the EAA. The use activity "processing by producers" includes all quantities processed by agricultural producers (e.g. apple processed to make apple must and cider) which are produced for purposes other than own consumption. The "own consumption" activity includes the production consumed in the households of agricultural producers which they have produced themselves or obtained from intra-agricultural trading. Also, own products used as remuneration to staff employed in the holding are included. The intersectoral use component covers all sales made to branches outside of agriculture, including storage and intervention centres.

The valuation of the final production for each product is done with ex-farm prices and manufacturing prices. The sales are generally valued at the price actually obtained on the market (ex-farm price), by assuming that the obligation of the sellers ends at the farm gate. Changes in stocks are valued using manufacturing cost prices. For each product the observed prices are recorded in the database.

The calculation of the agricultural sector’s intermediate consumption takes place in the block "input generation" (see Figure 2). This block is defined with the use activities described above (columns) and a list of input items (lines) which correspond to the EAA definition. For the calculation of the EAA, on the input side the amount of these inputs purchased by agriculture from other sectors are recorded in the framework within the column "sales/purchases". Intermediate consumption goods which have been produced within the national farm are recorded as intermediate consumption only insofar as and to the extend that they have also been recorded as output. Intermediate consumption, calculated according to the EAA definition, is valued at the acquisition prices paid by farmers at the farm gate.

The gross value added at market prices can be derived as the difference between the value of output and intermediate consumption. This calculation is done by aggregating the measurements in the blocks of output use and input generation. The net value added at market prices can be calculated with the deduction of consumption of fixed capital. After the deduction of taxes linked to production and the addition of subsidies, the net value added at factor cost can be shown.

As a residual value, the Net Income from agricultural activity can be calculated after deducting paid rent, interest and compensation of employees from the net value added at factor cost. The values for the components subsidies, taxes, compensation of employees, rent and interest are shown in the line-structure of the framework in the area value added (see Figure 2).

At the beginning of this section it has to be pointed out that the application of the Activity Based Approach is in principle not necessary to compile the EAA. But the experience of establishing the sectoral basic data system for CEEC has shown that the application brings forth various advantages for EAA compilation and can be taken as a foundation for the forecast and simulation modules.

The treatment of production activities and the separation of intra-sector use activities are the main characteristics of the gross concept, compared to the net concept described in the previous section. The 51 production activities, including crops and animals, are represented as columns in the left part of the framework (see Figure 2). In these columns the outputs and inputs are covered for each activity. Furthermore the production-level (ha, head) for each activity is recorded. The second aspect of the enlargement to the gross concept is the differentiation of the use activities in the columns in the right part of the framework. This structure includes, besides the four components used in net concept (see Figure 3), intrasectoral activities like animal-feed used on the farm, seed used on the farm and use activities for certain animal categories (calves, heifers...). In the following part the advantages of this approach are summarised:

The consideration of output generation and output use enables the calculation of balances in the lines of the framework. The amount produced in the sector must be equal to the different use components (market, feed, seed, change in stocks, etc.) in the year depicted (calendar year). The compilation work has shown than in some areas there are deviations, especially for products like fruits, vegetables and wine. These inconsistencies can be caused by different information in the statistics about production and use. For the products named above, the high share of production within households is typical and could be a reason for the inconsistencies detected.

The consistencies of generation and use in monetary terms are checked by calculating a weighted producer price for the valuation of the outputs produced, which is compared with the prices for the certain use activities. By looking at the production-level for each activity a sectoral average yield can be calculated automatically. This average yield can be used for comparisons with previous years, other information sources and discussion with experts. These output-coefficients form a second step of plausibility check. If they seem to be unrealistically high or low, reasons have to be sought. The calculations of average carcass weights and milk yield have shown some inconsistencies in the different statistical sources.

In this approach inputs are also shown as total sectoral amount (input generation) and differentiated for each production activity (input use). This structure enables an balancing of use and generation, an approach analogous to that described for outputs.

The data required for input use per activity, for example the total amount of N-fertiliser or energy used for wheat production, are usually not available as original statistic information. For this reason input coefficients are used, taken from publications on technical information or from farm samples. In Hungary a sample of farms has been used, representing 20 % of the planted area of large-scale farms, to provide well-differentiated input data as coefficients per ha and (Kertes 1997). This source includes nearly 20 different input items per activity.

The main objective of the consistency check in the input area is to compare the sectoral input data, extracted out of national statistics, with the input use in the activities. For this comparison a sectoral "theoretical input use" (e.g. total fertiliser input for Hungarian agriculture) for each of the chosen input items has to be calculated first. The input items (per ha) have to be projected to a sectoral dimension per activity by multiplication. A "theoretical sectoral input" for each input item (e.g. total fertiliser input for Hungarian agriculture) can be calculated by the aggregation of the sectoral input item per activity over all activities.

The consistency checks have exposed significant deviations for some input items. For example the "theoretical" potassium input, calculated using coefficients, was 50 % higher than the sectoral amount shown in national statistics (Köckler and Quiring 1997). Therefore we can conclude that the coefficients chosen were not representative for the whole agricultural sector and have to be adapted. For input items like fodder and seed inconsistencies may be detected that have high influence on the EAA results.

The example of fodder needs to be discussed in more detail. The compilation of the EAA requires a data on the purchases from other sectors. This value has not to include the amount of fodder which is traded directly between farms, because this amount is not included in final production. The comparisons between the theoretical input use of fodder and the sectoral input (sum of purchased and intrabranch produced fodder) has indicated some double counting. The cause lies in the difficulty of finding sources of statistical data which make a distinction between fodder that does not leave the national farm and fodder that is purchased from other farms. If the cereals, used in fodder, are excluded from final output but included in the statistical data for the sectoral fodder input, this double counting will influence the value added significantly. This failing could be corrected by using a gross output concept and applying a fodder distribution step, which could be used to check if the assumed sectoral amount of fodder fits with the amount used for animal production.

The data processing work for EAA compilation in the CEEC has revealed various difficulties by separating the final production for the various products. For example it is quite difficult to separate between the share of seed which has been used directly in the national farm and the amount that has been sold to the other sectors but which returns later. Data sources do not readily provide this information, and shares have to be estimated.

It should be noted that if adjustments are necessary in the block output use, appropriate changes have to be made the block input generation. But one has also to consider that these adjustments in the use balance will influence the other input components, like energy, which are assumed to be used in the sector.

The transparency of the approach, with the depiction of sectoral gross values, physical measurements and coefficients per activity, enables the consideration of expert proposals in several stages of the work. Here are some examples:

• adjustments to allocate the original use components of the national statistics to the separation required in the EAA
• control of output coefficients
• evaluation of the adjusted input coefficients
• filling of statistical gaps
• adjustments to the original statistical data for rent, interest and compensation of employees

One of the most significant points of the future rearrangement of the EAA is the use of a local kind of activity unit (local KAU) as the basic unit. Here the consequences for the application of the Activity Based Approach are discussed briefly.

The local KAU as the basic unit for the agricultural industry will entail recording non-agricultural secondary activities where they cannot be distinguished from the main agricultural activity. To do this we need an enlargement of the framework to represent these forms of production. For example the activities Agro-tourism, sports and rural recreation and Agricultural services for third parties have to be added. Activities which represent a continuation of agricultural activity and which use agricultural products. (e.g. processing, grading and packaging of agricultural products) also have to be incorporated. With a more detailed differentiation of the product list and the collection of the relevant prices, this Activity Based Approach seems to be appropriate to these new requirements.

After the creation of the framework, the collection of data and the checking of consistency, analysis and forecast can start. Because these aspects are not the emphasis of this paper, only a few aspects are summarised.

• EAA income indicator
• distribution of the value added on animal and crop production
• gross value added of certain crop and production activities
• sectoral averages for the calculation of productivity per unit

The technical realisation of the Sectoral Basic Data System is shown here in a brief overview. The various structures and gaps in the CEEC statistics requires a flexible data collection and compilation mechanism. This has required a technical realisation for the Sectoral Basic Data System, that is quite different from comparable systems for the European Union. It has been developed with the software office professional from MICROSOFT.

The relational databank-programme ACCESS is used for all aspects of data-management and user-interfaces. Only the consistent framework module is realised in EXCEL spreadsheets. An important argument to in favour of using standard software was the need to develop a system that can be used by the research assistant in the project as well as by partners in the CEECs.

The following paragraphs describe the characteristics of certain elements of the Sectoral Basic Data System (Figure 4):

6.1 Data storage (ACCESS)

The data storage module saves all information in the structure of a relational data-base. The other modules use this data-base for the realisation of certain tasks. Each data element has a unique key, which enables precise identification and selection of data and offers high transparency for data preparation.

6.2 Collection module (ACCESS)

The collection module supports all aspects of data import. It contains user-friendly interfaces for manual data-editing as well as automatic import routines. The utilisation is characterised by the following aspects:

• assignment of a source-code to every single data point
• flexible import concerning units, degree of aggregation and comments
• import routines for different pre structured data files
• depiction of obvious inconsistencies and visualisation of related data during
  data collection work
• user-friendly surfaces (even with different languages for certain user-groups)

6.3 Compilation module (ACCESS)

The flexible import of the original data requires a transformation to a standard structure. In this working step different units like kilogram and metric tons are standardised and the degree of aggregation is adapted to the framework structure (e.g. calculating beans and lupines to the aggregate pulses).

6.4 Consistent framework module (EXCEL)

The consistent framework module presents the compiled data for one year and one sector (country) and is used for consistency checks. Identified inconsistencies require dialogues with the provider of the original data and modifications. Value Added for the whole sector and the production activities is calculated in this framework.

6.5 Exploitation module (ACCESS)

The exploitation module is a tool for flexible presentation of the stored data. The utilisation of pre structured surfaces offers selective data access and the module contains also pre structured print reports.

Eurostat (1989): Manual on Economic Accounts for Agriculture and Forestry, Luxembourg, Eurostat.

Henrichsmeyer, W., Cypris, Ch., Löhe, W., Meudt, M. (1996): Entwicklung des gesamtdeutschen Agrarsektormodells RAUMIS96 am Lehrstuhl für Agrarpolitik, Volkswirtschaftslehre und landwirtschaftliches Informationswesen der Universität Bonn. Agrarwirtschaft, Vol. 45 (1996), pp. 213 - 215.

Kertes, R. (1997): A mezögazdasagi nagyüzemek föbb agazatainak kölsegjövedelemhelyzete, Research and Information Institute for Agricultural Economics, Budapest.

Köckler, J., Quiring, A. (1997): Perspektiven der agrarsektoralen Entwicklungen in den MOE-Staaten; Notwendigkeit eines differenzierten Analyseansatzes. Paper presented at the Seminar on "Achtunddreißigste Jahrestagung der Gesellschaft für Wirtschafts- und Sozialwissenschaften des Landbaues e.V, Land- und Ernährungswirtschaft in einer erweiterten EU held in Freising-Weihenstephan vom 06. bis 08. Oktober 1997.

Wolf, W. (1995): SPEL System: Methodological Documentation (Rev. 1), Vol.1:Basics, BS, SFSS. Theme 5 Series E, Luxembourg: Eurostat.