Chapter 4 - Uses of economic accounts for agriculture

Contents - Previous - Next

Structural analysis and economic indicators
Use of simulation and modelling


4.1 Agriculture is an essential component for sustainable economic growth. It is becoming increasingly clear that an appropriate long-term strategy for the agriculture sector is essential for both developing and developed countries. Decades ago the basic objective of agricultural policy was to meet the demand for food and fibre for growing populations. Today, the protection of natural resources and the preservation of the eco-system and additional objectives. Thus it has become essential to analyze the data carefully for policy purposes. The basic objectives for a policy-maker dealing with food and agricultural activities are:

(a) to monitor the current trends of agricultural growth within the process of overall economic development;

(b) to foresee the impact of various policy measures on agricultural development and the economic conditions of the population dependent on agriculture;

(c) to help policy-makers achieve optimum results in allocating increasingly scarce natural resources.

It is necessary to mention here that there are many different kinds of analytical work and the purpose of this Chapter is to illustrate only the possible role of SEAFA in this exercise. Separate work has also been done by FAO (see, for example, FAO, 1982 and 1986) in this direction. The objective of this chapter is twofold, namely, to list some of the common techniques or methods that must be taken into account by the compilers of the system to meet the requirements of policy-makers and to caution compilers about possible sources of errors that can invalidate the final conclusions.

4.2 One of the crucial problems for analytical studies is the chronic shortage of reliable data. Unless the available data are consistent with other related data and have similar coverage and concepts, meaningful analysis is not feasible. By definition, the systems of national or agricultural accounts represent consistent sets of statistical information. SEAFA has been developed partly on the model of a Social Accounting Matrix (SAM) structure by linking large amounts of other data to macro-economic flows in a flexible manner that suits the needs of different users in different situations. A SAM requires a large amount of data classified according to the economic issue under study. SEAFA proposes a limited set of internationally comparable economic accounts along with other kinds of data. (An overview of the areas that can be covered by SEAFA can be seen in Table 4.1.). Primary information related to different subjects may be suitably stored in disaggregated form and appropriate software developed to enable a variety of tabulations to be made, each of them tailored to different user requirements. It is possible that the data may not support a full balanced SAM, but the user should be able to draw better, more relevant conclusions. The three types of analysis mentioned above can be suitably grouped into structural analysis and the development of economic indicators; and the use of simulation methods and modelling. A brief discussion of these uses of SEAFA is given below.

Structural analysis and economic indicators

(a) The SEAFA's set of accounts
(b) Supporting statements
(c) The System as the basis for quantity and price index numbers

4.3 The term structural analysis is generally employed to cover all those areas of economic analysis are concerned with changes in the "structure" of an economy such as changes in the relative importance of various commodities, industries, economic regions, institutional sectors and economic classes. The agricultural data collected in various censuses and surveys and from administrative records may be of limited use for such analysis because of incomplete coverage and a lack of other relevant information. Today, when the economic situation is changing because of vast communication networks, any policy decision depends on the interplay of a large number of factors. For example, the attitude of farmers towards growing a certain crop may depend on the prices and output of similar crops, or of its substitutes, in other parts of the country or international market; on costs of production and on weather conditions. Even if the relevant data are available it is necessary to know their definitions and coverage. To achieve an overall favourable production pattern while trying to maximize farmers' incomes, the government may pursue certain pricing and subsidy policies. Accordingly, for efficient policy-making or plan formulation a sound information system is required consisting of aggregates and economic indicators based on up-to-date, reliable data and interconnected with each other. SEAFA provides the basic framework for establishing such a system.

(a) The SEAFA's set of accounts

4.4 The SEAFA includes five accounts using the accounting format recommended by the 1993 SNA. Each of these accounts provides an internationally comparable economic indicator to monitor specific phenomena:

Production Account Value added originating from an economic activity
Generation of Income Account The entrepreneur's share of the income payable out of the value added
Allocation of Primary Income Account The balance of primary income accruing to the entrepreneur
Goods and Service Account Classification of output (and imports) by type of end use
Capital Account Capital formation and its source of financing (domestic resources/external assistance)

4.5 As already indicated in Chapter 3 these five accounts may be compiled for different groups of establishments or institutional units. Accordingly, the indicators will monitor the current status of industries or sectors.

(b) Supporting statements

4.6 SEAFA recommends a set of statements to present detailed disaggregated information on the various flows included in the accounts. Such statement can be utilized for structural analysis and the construction of economic indicators as well as evaluation of past performance and monitoring of the current economic situation. Although the Handbook makes suggestions about the kinds of illustrative statements which may be compiled these statements can be varied according to the needs of different countries. For example, some countries may wish to disaggregate production by forms of organization rather than according to type of productive activity. Similarly, for some purposes it may be preferable to disaggregate gross domestic product by type of payment rather than by factor employed. To keep the system more comprehensive and flexible it is not recommended that a fixed set of statements be compiled. The statements may be varied according to the current needs of policy-makers using a primary database capable of producing different statements as required by various users. Some classification schemes that may be used follow:

(i) goods (e.g. cereal, pulses, oilseeds, sugar, fruits and vegetables, fibre crops, tree
crops, spices, medicinal and aromatic plants);
(ii) administrative regions, rural or urban areas, size of holdings;
(iii) agro-climatic or agro-ecological region;
(iv) natural resources (e.g. agricultural land by type of forest, forest land, animals, inland water coastal zone);
(v) uses (e.g. intermediate consumption, final consumption of households and
government, exports and capital formation);
(vi) international trade zone.

4.7 When creating databases using the above classification it must be possible to analyze different kinds of data drawn from different sources (i.e. not only output but also other data such as inputs, capital formation or utilization, labour force and producer or consumer prices of items of output and input). The data can be used to examine the development of a given economy and to get some idea of the main factors that determine such development. Further insight into the structural aspects of an economy can also be obtained from intra- and inter country comparisons. Apart from such general analysis, the following indicators are also relevant for policy-making:

(i) The ratio of the imputed value of output for own final use to total sales is of particular significance because it provides a measure of the degree of subsistence of the sector. The higher the ratio, the lower the share of output entering the market and the higher the degree of insulation of farmers from market forces and economic incentives.
(ii) The ratio of gross output to gross products (the difference between these two aggregates represents the goods and services purchased for intermediate consumption from other sectors) is an indicator of the organizational nature of the sector. As modernization of agricultural production takes place, the value of the above ratio will increase. However, the change may be comparatively slow as modernization may also result in an increasing trend in productivity.
(iii) The ratio of the prices of goods and services sold by the sector (producers) to the prices of goods and services purchased by the sector (producers) for final consumption, intermediate consumption and capital formation indicates changes in the purchasing power of agricultural producers; this is also known as the "terms of trade between agriculture and other sectors". Such changes are of particular significance in relation to incentives for agricultural production because, for instance, a shift of the terms of trade in favour of agriculture may encourage the use of new inputs and the adoption of more modern methods of production.
(iv) The share of agricultural commodities in total exports (or imports) shows its relative importance in foreign trade and the ratio of agricultural imports and/or net agricultural imports (i.e. imports - exports) to total exports shows the dependency of agriculture on external markets.

4.8 A set of similar indicators can be generated for, other economic activities included in SEAFA. Consider the case of the "Food and Nutrition" activity. In this case "food availability for human consumption" when expressed in terms of "calories (per caput per day), fats and proteins (grams per caput per day)" and compared over time shows the changing nutritional levels for a region or country. Similarly, a comparison of "average annual (percentage) rate of change" of per caput food availability for human consumption with the growth of the population gives a picture of the status of food availability which can be studied along with the food supply structure (domestic output, imports and exports) to study self-sufficiency in food production. A series of such statements can be prepared to meet various analytical needs. The role of SEAFA in this regard is to provide the basic data along with other databases such as the composition of food availability from different sources e.g. domestic output versus imports or crops versus animal products versus fisheries.

(c) The System as the basis for quantity and price index numbers

4.9 The important aggregates of the economic accounts, such as output, input, household and government final consumption expenditure, together with gross fixed capital formation, are generally compiled at constant as well as current prices. The compilation of constant price aggregates utilizes the entire data available taking account of differences in quality, location, etc. The data provide weights for the construction of index number (Laspeyres, Paasche, Chain base type). Implicit quantity and price index numbers can also be derived using current price data. Estimates at constant prices can also be presented in the form of index numbers using the same base year for the constant price series to show the relative volume movements. Similarly, the price series derived by dividing the estimate at current prices by that at constant prices can be converted into index numbers of implicit prices. These index numbers can be constructed for different aggregates (as well as their components) including gross domestic product.

Use of simulation and modelling

(a) Input-output
(b) Social Accounting Matrices (SAMs)
(c) Micro-macro linkages
(d) Simulation
(e) General equilibrium models

4.10 The role of economic accounts is not simply to provide economic indicators but to provide the empirical material for analysis and modelling for policy on developmental planning purposes. The structure of the economy can be seen more clearly from disaggregated information than from the aggregate accounts. Thus, to have a system of economic accounts with much analytical or modelling capability it is necessary to use appropriate classifications and to have sufficiently disaggregated data. A good primary database can serve a variety of users.

4.11 The first step in designing a system of economic accounts capable of supporting modelling and analytical needs is to identify minimum data requirements, including non monetized as well as monetized transactions. Lack of information (Richter, Josef, 1994) can lead to misuse of economic data. It is necessary to have adequate basic concepts and statistical units. Quite often theory rests on variables that cannot be directly observed and there may be a substantial discrepancy between the actual measure and the underlying theoretical concept. As a result of these difficulties proxy variables can be used whose behaviour may be entirely different. As an illustration take the case of establishing linkages between resource flows from non-agricultural to agricultural activities in developing countries. In developing countries agricultural activities are generally undertaken by unincorporated enterprises owned by households. Households, however, do not usually maintain accounts. A clear picture of the resources involved is not available. Borrowed funds may be used for personal rather than agricultural activities. For the analysis of production and its impact on the environment one type of statistical unit may be needed while for a study of measurement of income by sector a different type of unit may be needed. The aim of the investigation determines which kind of unit or classification scheme should be adopted, which degree of disaggregation has to be achieved, etc.

4.12 When a comprehensive set of accounts is prepared using a variety of sources, certain discrepancies may emerge between the two sides of an account. These discrepancies may arise because of incompatible assumptions used in deriving the various structural ratios used for compilation purposes. The size of the discrepancies reflects the degree of consistency of economic accounts. Any modelling exercise using data sets with in-built inconsistencies is difficult. Thus, before undertaking the analysis it is necessary to have a fresh look at the various assumptions underlying the estimates. If other methods are not feasible, it may sometimes be worthwhile to use statistical techniques, such as generalized least square or the RAS technique to adjust the data. When using these methods it becomes essential to check whether the adjusted estimates are plausible and fall within acceptable limits. There are two main advantages of these methods. First, they allow the maximum use of a wide range of available data to build consistent databases that serve a variety of purposes. Second, they provide systematic cross-checks which give users more confidence in the data.

4.13 The compilation of SEAFA is particularly important for developing countries because of the relatively large size of the agricultural sector and the dominant role it plays in achieving any growth target. The uses of SEAFA are discussed below under five broad headings.

(a) Input-output

4.14 Input-output is one of the most common methods regularly used for various policy and planning purposes. The data contained in the Economic Accounts for Agriculture may be used to prepare projections under alternative policy scenarios. Consider, for example, the case of a country whose population consumes food containing "x" calories with "a" amount of fat and "b" amount of protein. Now, suppose it is decided to raise the level of nutrition from "x" to "x1" calories which may contain "a1" and "b1" amounts of fat and protein respectively. The planner would be faced with the need to change the country's production to provide an optimum food basket, keeping in view the current status of domestic food production, the requirement of inputs, import and export of food, the cost of living of the population and also the likely financial burden on the country. The present SEAFA has been designed to answer such questions using input-output type analysis.

(b) Social Accounting Matrices (SAMs)

4.15 A SAM is a tool to analyze the structural features of an economy by establishing linkages between different flows in the economic accounts of institutional sectors SAMs are forms of accounts that focus on the role of people in the economy. Among other things, they require extra breakdowns of the household sector and a disaggregated presentation of labour markets (i.e. distinguishing various categories of employed persons). Key features of SAMs are integration and multiple classifications; in other words, conceptual and numerical linkages among all kinds of related monetary and non-monetary phenomena, which may be expressed in different measurement units. If, for instance, agricultural production is classified by agro-climatic zones, SEAFA can provide:

Cropping patterns in different zones,
A comparative picture of crop productivity for selecting suitable crops, Input requirements,
Thresholds for agro-environmental indicators,
Shares in national aggregates using flows that are consistent with other kinds of data in the system.

4.16 In a SAM it is desirable that the concepts, accounting structure and classifications should be tailored to the economy described, the specific purposes for which the SAM is constructed and the availability of data. A SAM provides a framework and consistent (base year) data for economy-wide models with detailed classifications of actors, such as industries, categories of employed persons and institutional sub-sectors, including various socioeconomic household groups. The abundance of data included in most SAMs may give the impression that they can only be constructed for countries with a wealth of statistical information. In practice, developing countries have taken the lead in compiling SAMs. Actually, it is in situations where the basic information and other statistical resources are very scarce that it is all the more important to make the best possible use of whatever data are available. Integrating the results from all kinds of costly censuses and surveys into a consistent framework may increase both their relevance and their reliability. Integration of basic data enhances the possibility of more policy issues being monitored and analyzed inter-relatedly. Moreover, the linkage of employment and income distribution data (or productivity and size distribution of holdings) to more macro-oriented objectives such as NDP growth, balance of payments equilibrium and stable price levels, comes within the reach of SAM.

4.17 The relationship between SAMs and models has several aspects. It has already been noted that for each model there is a corresponding SAM. The converse does not hold, however. For any given SAM there are a variety of possible models. The choice of SAM restricts the choice of models, but it does not determine it uniquely. The process of designing a model can usefully be divided into three stages. The first stage concerns the choice of what the model is to be about; what institutions are to be recognized; are asset holdings or flows of funds to be modelled; and what disaggregation of factors, activities and commodities is needed'? It is this part of the model design that determines the SAM and is uniquely determined by it. Within the framework developed at this stage, there is complete flexibility to choose the data sets and structural identities (within the overall framework of macroeconomic theory) to define the model. A distinctive feature of all SAM-based models is their reliance on complete balances, at a multi-sector level, of incomes and outlays of institutions and of supply and use of goods and services, usually including labour services. Another feature is that the structure of the model and the structure of the SAM are closely interlinked. To a lesser extent, this also applies to the parameter values. This means a departure from a model specification that hinges on associations between long runs of time-series. In turn, it implies that SAM-based models are less liable to the disadvantages of time series models, such as: (a) their use of proxy variables and independently estimated deflators for various transaction categories, which are not necessarily consistent; (b) their dependence on long time-series which often are not available at an intermediate level; and (c) their reliance on the constancy of structural relationships over a long period. Such relationships may change in response to various structural or external shocks (energy crises) and continual institutional reforms (lifting trade barriers, large-scale privatizations, etc.).

(c) Micro-macro linkages

4.18 Microanalysis deals with the behaviour of individual units without necessarily linking them to the overall economic system. The data for micro-analysis are generally obtained from sample surveys, administrative records and pilot studies which are heterogeneous and often give conflicting results. The major reasons for this are that (a) these data are seldom matched with respect to concepts, definitions and statistical units used for the collection of data, and (b) the coverage of the activity may be undefined and far from complete. In such cases, aggregation of the data raises problems. Basic rules are often not followed when presenting micro behaviour. In contrast, the macro data are generally comprehensive and consistent with one another when compiled within an overall economic accounting framework. However, these data can not easily be disaggregated to examine structural or behaviourial relationships. The increasing interest of policy-makers has led to micro databases becoming increasingly available. To understand the dynamic nature of the economic system in terms of the social, economic and demographic characteristics of the population it is essential to relate the micro data bases to the overall framework of economic accounts. This task requires adjustments on both sides. Some guiding principles follow:

(i) The domain for the collection of micro data should be the same as that for the macro economic accounts. The coverage of both should be examined and that of the macro data may be adjusted so that to match that of the micro data.
(ii) The collection of micro data should keep in view the need for uniformity with macro data.
(iii) The concepts and definitions used in the collection of micro data should be examined carefully and matched with those in the system of macro accounts. For example, values may have to be assigned to some activities or goods and services for which data are not available.
(iv) The macro data sets generally include some imputed values for non-market activities. The basic assumptions followed in these imputations should be examined and checked with corresponding micro data sets.

A system conceived in this way is well suited for the integration of micro and macro data sets for analytical purposes.

(d) Simulation

4.19 Simulation attempts to reproduce on a small scale the actual or essential features of an operating market, socio-economic system or other unit of inquiry. With increasing access to computers, simulation has become a powerful tool for policy analysis allowing the policy-maker to examine and investigate the impact of various policy measures on the economy. There are various types of simulation models but the following two are of particular interest:

(i) optimizing behaviour -- an optimizing model attempts to identify decisions that maximize profits, security or efficiency or that minimize costs, poverty, malnourishment or trade deficits; and
(ii) structural and predictive -- predictive models make it possible to forecast future outcomes while structural models emphasize causal relationships among the important variables of the system. These structural relationships are deduced using appropriate theory from disciplines such as economics and sociology. The structure of the model, once specified, determines what data and statistical analyses will be needed to estimate parameter values and test the model's validity.

4.20 With the increased use of simulation models microsimulation methods have been developed recently. The essence of microsimulation is to establish a link between the behaviour observed in various micro level data and macro aggregates in order to provide estimates of how various aggregates may change in response to policy innovation.

(e) General equilibrium models

4.21 The methodology underlying the formulation of General Equilibrium Models consists of three main elements; a preference function specifying the objectives of economic policy; a classification of variables useful for policy purposes; and a quantitative model reflecting the structure of the economy. The preference function is supposed to reflect the major policy objectives which the decision-maker is striving for as well as the relative importance of these objectives. The variables included in the model are divided into two classes: exogenous variables determined outside the system and endogenous variables whose value is determined by the model. The exogenous variables generally include policy instruments (i.e. exchange rate, interest rate, income tax rate, etc.) and data sets such as prices. The endogenous variables consist of target variables and other similar variables that reflect the impact of changes in target variables. The model is generally presented in the form of a set of simultaneous equations which represent different types of relationships, such as production functions, consumption functions, import functions and tax functions, and a few identities to close the system. The models are solved using various statistical methods appropriate to the situation. The equation system (the model) is designed to answer the following questions:

a) Given a set of known or projected values of exogenous variables for the coming year, what values of the endogenous variables are likely to emerge?
b) Given the desired values of the target variables and forecasts of the non controllable factors, what values should the policy instruments take in order to achieve the desired predetermined values of the target variables'?

4.22 One of the important questions of economic policy concerns the choice of the objectives (targets) and the relative importance attached to them. This is particularly true when some of the objectives may be in conflict; for example, when policies leading to a higher gross national product (GNP) growth rate result in lower consumption and a less equal income distribution in the short run. In such cases, economists can play an important role in clarifying alternative values of the target variables for comparison within a policy-maker's preference function.

4.23 The construction of general equilibrium models has been closely associated with the development of economic accounts. However, the representation of the agriculture sector in macro-models is not adequate. One reason for this unsatisfactory situation is that the sector is not as important in developed countries. In developing countries, however, lack of basic data and the difficulty of dealing with the problem of institutional constraints on agricultural development makes the task difficult. The increasing awareness that economists and policy makers have developed in recent years with regard to the positive contribution of agriculture to the overall process of economic growth has stimulated research work on the interrelationships between agriculture and other sectors and has led to the formulation of more comprehensive planning and policy-making models for economies that are predominantly agricultural. It is not possible to discuss these models in detail, but a large amount of literature is available on the subject. For example, FAO, 1982, contains four types of agricultural sector analyses (ASAs) and models: a) non-formal, general equilibrium consistency approaches; b) general, systems-simulation models; c) linear programming models; and, d) multi-level planning models. In addition, two Chapters are devoted to, respectively, issues in planning rural development and the operational usefulness of sector analysis and models to users. The final Chapter of the volume undertakes a critical evaluation of agricultural sector analysis.


1. FAO, 1982: Agricultural Sector Analysis and Models in Developing Countries, FAO Economic and Social Development Paper, No. 5.

2. FAO, 1984: Socio-economic Indicators relating to the Agricultural Sector and Rural Development, FAO Economic and Social Development Paper, No. 40.

3. FAO, 1986: Food and Agricultural Statistics in the context of a National Information System, FAO Statistical Development Series No. 1.

4. FAO, 1988: WCARRD - World Conference on Agrarian Reform and Rural Development -ten years of follow-up: Guidelines on Socio-Economic Indicators for Monitoring and Evaluating Agrarian Reform and Rural Development.

5. FAO, 1995: Report on Harmonization of Criteria and Indicators for Sustainable Forest
Management, FAO/ITTO Expert Consultation, Rome, Italy, 13-16 February.

6. OECD, 1994: OECD Agri-Environmental Indicators: The Measurement of a Proposed set of Indicators, Joint Working Party of the Committee for Agriculture and the Environmental Indicators, Meeting of Experts on Agri-Environmental Indicators, COM/AGR/CA/ENV/EPOC(94)93.

7. Pyatt, G., 1988: A SAM Approach to modelling, Journal of Policy Modelling, 10(3).

8. Pyatt, G., 1991: SAMs, the SNA and National Accounting capabilities, Review of Income and Wealth, Series 37, No. 2, June 1991.

9. Richter, J, 1994: Use and Misuse of National Accounts from a Modelling Perspective, The Review of Income and Wealth, Series No. 40, No.1, March 1994.

10. Ruggles, R & Ruggles, N.C., 1986: The integration of macro and micro data for the household sector, The Review of Income and Wealth, Series 32, No. 3, September 1986.

Table 4.1 An overview of SEAFA's linkage with the 1993 SNA

Contents - Previous - Next