The need for indicators and accounts that provide a frame of reference for assessing the economic performance, viability and equity in agriculture are even greater today than they were when originally designed. However, the list of policy issues to be addressed has become much broader and more complicated and the economic structure of the sector has changed dramatically, making equity questions much different from those that occurred when this part of the data system was designed.

These changes result in a problem of conceptual obsolescence for the economic indicators and sector accounts. These data do not measure some of the right things and do not provide information for addressing some of the important economic policy questions. Instead of clearly indicating who gains and loses as farm programs, production and prices change, the system provides incomplete data at best and misleading information at worst.

The farm sector was the major concern when the economic indicators and sector accounts were designed. Now the issues relate to a broader food and fiber sector including agriculture service, farm input, food processing and food distribution industries. The greatly increased importance of these other industries in providing food for the consumer make their economic performance an issue today. The equity questions today include the market power and level of returns to each of these industries in relation to food prices. The equity questions also include the relative level of returns to each of these industries in relation to the return received by the farmer from the consumer's food dollar.

The equity issues within the farm sector have also shifted as the structure of the sector has changed. The widely diverse types of farm operations require new distributional data to measure the returns to large and small, corporate and sole proprietorship, and single establishment and multi-establishment types of businesses as different target groups.

Concept Problems, Data Gaps and Recommendations. Since most of the indicators and accounts were designed for the farm sector, the most significant concept problems are related to this sector. Concepts and standard classifications used for farm sector data perpetuate incorrect descriptions of the economic structure of farming. The resulting information from the indicators and accounts data thus fails to clearly show the distributive impacts of program and income changes.

Some of the more important information problems caused by incorrect structural concepts used in the indicator and sector account data are:

1. All income from farming goes to individual landowners or to farm operator families. Instead, the situation is that some unknown portion of income from farming goes to large corporate firms who own farmland, operate a farm, or contract with farmers for the production of commodities.

2. All income from farming goes to the farm population, which consists of households that operate farms. The situation is that farm population is a residential definition rather than an occupational or business definition and the two are increasingly divergent. Many people living on farms do not operate farms while an increasing number of farm operators do not live on farms. This fact, plus the importance of other sources of income to farm operators, makes a comparison between per capita incomes of the farm and nonfarm population a very poor indicator of the need for changes in farm price and income programs.

3. Average net farm income per farm is a frequently used indicator of the adequacy of returns to farm operators and to the resources used in farming over time. However, the changes that have taken place in the structure of the sector make this a very weak indicator to use. One reason is the unknown portion of income from farming that is going to the nonfarm corporate sector. A second reason is the greatly increased concentration of farm production on large operations. Thus, the size of the numerator of this indicator is primarily influenced by the management of the large farm operations while the denominator includes the relatively large number of very small operations.

4. Each farm business has one and only one operator household related to it. This concept has long been outdated because of formation of partnerships and family or other closely held corporations. Continued use of this concept in derivation of farm income per operator figures provides an inaccurate picture of the relationship between business and household incomes:

5. Farm operations consist of an inseparable business establishment and household unit. While this is true for many farm operations it is not true for nearly all of them, as it was when

the indicator and sector account data were designed. One important change has been the relatively large amounts of wage, salary and other income sources earned by farm operator households. The majority of farm operator households do not have to rely only on income from farming for their well-being. A second change making this concept obsolete is that farm production has become an integrated or diversified activity for a significant number of nonfarm business corporations. Concern for the welfare of a farm operator household as prices or incomes change is not transferable to this type of economic structure.

6. Food and fiber production is the only economic activity that takes place on farms which are characteristically independent, small, single-establishment and family-owned businesses. Again, this is true for many farm operations but the extent and change of nonfarm corporate involvement in farming is not measured well by the indicator and sector account data. Nor is the importance of secondary business activities on farms adequately measured by the data system.

One final major concept problem for the farm sector is encountered when trying to determine what price levels provide a fair return to the farmer. One indicator used for this purpose is net income for the farm sector relative to the historical trend. Two other indicators used, for which weaknesses have already been discussed briefly, are the historical trend in net farm income per farm and the comparison of per capita income between the farm population and the nonfarm population.

Another indicator that has long been used for this purpose is the parity price ratio. This is based on prices paid and received by farmers in a base period assumed to be a time when farmers received a fair return from their business operation. The parity ratio, based on a 1910-14 base period, is badly out of date because of the drastic changes in production practices and even in what commodities are produced. An indication of this occurred in the 1973-75 period when farm prices near 100 percent of parity resulted in profits large enough to bring extremely large increases in land prices and rents to the land resource. Changes in the economic structure of the farm and farm input sectors make the 1910-14 period obsolete as base for determination of a fair price or return to farming activity.

In recent years a new concept based on cost of production has been partly substituted for parity prices. This concept also has a number of problems

associated with it. Statistical measurement of costs across a large number of producers with different production methods is difficult. Wide variance in efficiency of farming operations makes average cost. figures a rough guide at best. Then questions exist about what to include in the measure for return to land since profits get reinvested in land causing a ratcheting effect on farmland prices.

These very significant concept problems that have been discussed require major investment in renewal of the economic indicator and sector account part of the agricultural data system.

One major change that should be made in the farm sector accounts is to separate the business and household sectors. Separate income and wealth accounts are needed that represent the business activity only. This procedure would require the income estimates to be compiled and published as income from farming, rather than as the current concept of farm operator's income; the latter implies who receives the income.

As a second necessary step, income from farming should be disaggregated to the several major groups who receive it. A useful set of categories or groups would be farm operators receiving self-employment income from farming, nonfarm business firms that operate farms as a minor activity, nonfarm business firms that contract for farm production, and nonoperating landowners. The group receiving selfemployment income from farming could be further disaggregated into two groups based on whether farming is or is not their primary occupation. Average income per farm could be calculated for volume of sales categories in each of these two subgroups as useful indicators over time of the farm price and income impacts on these more homogeneous business operations.

Self-employment income from farming cannot be properly transformed into a per farm operator basis (implying an income level per household) without more adequate information on the number of households receiving self-employment income from farming. Income for the household sector should be separately accounted for through existing, planned, or modified household surveys. The subgroup of households of primary interest are those receiving self-employment income from farming. Although several household surveys exist or are planned for income data, they may not include a large enough sample for estimating the income of this target group. Samples should be augmented to provide the data needed on this minor portion of the population.

A second major change that should be made in farm sector data is to update the standard classifications

used for generating statistics about relevant subgroups. The need for statistics distributed by target subgroups is not being met because the classification system has not been renewed.

One part of the need is to develop new categories for classifying data by economic structure. Legal form of organization is one classification used currently but is not sufficient by itself. One proposal developed in recent years is to implement use of a system that labels all farm operations as primary, part-time, or business associated according to the major occupation of the self-employed operator or the major income source of the corporation. Other ideas need to be developed. These could include concentration and specialization ratios and other concepts of economic structure and control. Careful consideration of whether establishments with a farming activity should all be called farms or classified by major economic activity, as is now done for other economic sectors, is an important part of classification system renewal.

A second need is to evaluate the current use of demographic characteristics of the household for classifying production data for the business sector. One problem this creates is to perpetuate the concept of a one-to-one linkage between a farm establishment and a household. Characteristics such as age, race, and ethnic origin of a self-employed farm operator are relevant for classifying the household income data but are questionable for classifying production or business data.

A related topic that has received significant attention in recent years is what statistical definition to use for a farm. This also has been an issue because of the inability of the agricultural data system to provide information needed on relevant target groups. This resulted in the frequent use of an aggregate indicator, net farm income per farm, SO proposals were made to make this a more meaningful economic indicator. This would be accomplished by excluding more of the very small operations from the statistical count of farms used in the denominator for income per farm calculations. The justification was that these small units were more like home garden or household subsistence operations and contributed very little to total sales of farm products. Thus, it was misleading to use them in the count of farms for deriving an indicator that was used in assessing the adequacy of major farm price and income programs.

The proposal adopted in August of 1975 by the Departments of Agriculture and Commerce was to count a unit as a farm only if it sold $1,000 or more of product. Some type of minimum size criterion has been in use since 1850 and been changed eight times

since. The criteria used since the last change in 1959 were (a) a place with 10 or more acres and $50 of product sales or (b) a place under 10 acres which had at least $250 of product sales.

The new definition of a farm was adopted for the final tabulations of data from the 1974 Census of Agriculture. Data collected from the small units were also published to enable data users to adjust historical data series. But the issue is not closed because of opposition to the change from rural development and rural fundamentalist groups. The issue was considered by the 94th and 95th Congress because of the opposing arguments.

Recommendations discussed earlier would improve the information on who receives farm income and would eliminate calculation of net income per farm across the diverse types of operations. However, the statistical definition of a farm adopted in 1975 should be fully implemented unless an alternative definition is required by law. This change in minimum sales value cutoff is justified by price inflation and improved efficiency of statistical programs.

The income accounts are the most important of the farm sector accounts. Some concept problems related to these accounts were discussed above and some data gaps related to these accounts were discussed in the supply, use and price data section. However, several other concept and data gap problems also need correction.

There have been a number of studies that recommend that income accounts for the farm sector be more nearly like the national income accounts in concept and definition. This would require a clean separation of capital items from the current account and accounting more directly for value added in the sector. These important improvements should be made so there will be more comparable measures of gross output and economic activity for the sector.

Several data gaps for the farm income accounts are identified in the Gross National Product Data Improvement Project report. One of the most important recommendations from that report is to collect current quarterly data on farm production expenses for major cost items. Others are to collect annual survey data on cash rents and in-kind share rents separated for dwellings and land, and data on current costs for farm dwellings.

To fill these economic data gaps a more comprehensive establishment data program needs to be developed. The Census of Agriculture has traditionally been the only periodic survey providing economic information for farm establishments. But even this survey is largely devoted to detailed land use and pro

duction statistics. An Annual Economic Survey program has been conducted by ESCS for the past four years but needs considerably more development to provide the regular flow of establishment data needed.

One of the gaps in the economic data for establishments is the lack of data on the division of expenses and income between farmers and nonfarm firms that contract with them for commodity production. These data will be needed from the improved establishment data program to estimate who receives the income from farming as discussed earlier in this section.

Two other sector accounts that provide important information on the farm sector are the balance sheet and the productivity accounts. A priority need for improving the balance sheet is to separate the business and household sectors so that capital accounts include only the business sector.

Recommendations for improvement in the productivity accounts will be forthcoming from a current evaluation of these accounts. This study is being conducted by the Economic Statistics Committee of the American Agricultural Economics Association and ESCS.

Continued concern about economic equity and viability for the farming sector requires improved economic indicators for the sector. A comprehensive study should be made of the indicators used for this purpose (parity, cost of production, income and price indexes) and other indicators that might be developed. More adequate measures to determine the place of farmers on the economic ladder are needed. Concepts of parity of income and measures of return on equity capital should be evaluated.

A final major data gap in economic indicators and sector accounts are the other industries in the food and fiber sectors. A more complete program of data on costs, incomes, capital investments and economic performance of these industries should be developed. A program started by ESCS to measure operating margins of food wholesale and retail firms should be continued.

Rural Community and Demographic Data

The third major part of the agricultural data system is data on the people and communities in nonmetropolitan areas. This includes people living on farms but also about four times as many who live in these rural areas but not on farms.

This part of the data system has been less developed than supply, use and price data or the sector accounts. The available demographic data has

come from the decennial censuses of population and housing and more recently from the annual housing survey and an expanded program of current population surveys. New demographic surveys including the authorized mid-decade statistical effort for population and housing data will close more of the data gaps on rural people. Planning for existing and new demographic surveys should include identification of data needs for the rural population and related target subgroups.

Rural community data is even more difficult to find and is a major problem for programmatic decisions of the Federal, State and local governments. Data on availability and needs for sewer, water, health and transportation services are greatly needed to target loan and grant programs and evaluate their effectiveness.

Other rural people and community data needs include educational services, employment opportunities and important causes of community change. A significant amount of concept and development work needs to be done for development of an adequate data system on rural people and communities.

Organizational Issues

Responsibility for the agricultural data system is pretty well concentrated in ESCS, the statistical and analytical agency for USDA, and in the Agriculture Division of the Census Bureau. However, like other parts of the Federal statistical system, there is enough decentralization to require periodic examination of roles and responsibilities.

A task force should be formed to conduct a comprehensive evaluation and make recommendations and plans for improving the agricultural data system. The task force is needed to examine the many concept and data gap problems identified in this chapter and develop a long-range plan for correcting these problems. An Interagency Committee on Agricultural Statistics should be formed to guide the work of the task force and react to ideas and recommendations developed by the task force. The interagency committee is needed to ensure inclusion of all relevant topics and concerns in the task force study. The committee can also help bring a breadth of perspectives and increased objectivity to the deliberations.

One of the organizational issues that should be examined by the task force is the conduct of statistical work by program agencies in USDA. The main areas for examination are the foreign commodity statistics program of the Foreign Agricultural Service and the market and price statistics programs of the Agricultural Marketing Service. Consideration

should be given to whether the quality, timeliness, efficiency, or effectiveness of these programs would be significantly improved by transferring them to the statistical center for USDA. Needs for more joint planning or resource competition between these programs and the statistical programs conducted by ESCS should be considered. Any possible lack of objectivity in these statistical programs or influence from policymakers should be identified.

The second major organizational issue that should be examined by the task force is the role and relationship of the Census Bureau and ESCS. Many times the questions start with why there is a need for both in the agricultural data system. But the more correct questions are how they are each serving the broad economic and social data needs of the agricultural data system and what opportunities there are to develop more complimentary and cooperative roles for meeting the data needs.

Comparison of the data needs with the strengths and capabilities of each agency justifies a continued role for both. Census has the most capability to provide the data needed on rural people and communities, on the other industries in the food and fiber sector, and improved economic structure and control data on the farm sector. ESCS has the most capability to furnish very current and timely data on supply, use, and price of farm commodities. Who should be responsible for other data needs and whether the overlapping programs for the farm sector can be more compatible are questions that need to be resolved. One final issue is the cooperation between the agencies in development and maintenance of sampling frames for the farm sector and other food and fiber industries.

Methods and Procedures Issues

This topic covers a variety of miscellaneous issues in agricultural statistics that need attention. The most recent major effort by the statistics unit of ESCS for improving the quality of their statistics is development and maintenance of a master list frame of farm operators and other establishments from which data are needed. More controlled and more efficient sampling can be done by ESCS when this list is fully operational. As resources are made available, this list frame should be used to conduct more probability surveys. low-response mail surveys still used for some programs have an unknown statistical error and an unmeasured response bias.

Although the intent of the Census of Agriculture is to survey all farm operations, it is difficult in practice to identify all operations, especially very small ones. The degree of incompleteness in the Census of Agriculture has been measured by comparing aggregate results with data from SRS sample survey programs. Although this is helpful, a far better procedure would be to use a land area sample frame to make the census more complete by collecting data from a sample of farms that have not completed a mailed census questionnaire. This procedure should probably be used to achieve completeness at the State level although county-level measurement might be considered in States with large proportions of small farms.

New survey methods for agricultural statistics should continue to be researched and implemented where possible. Many of the recent research efforts have gone into use of satellites and remote sensing. The major payoff from this new approach is probably in improved foreign crop production data. Use of producer and consumer panels needs to receive more research attention.

Agriculture Handbook No. 365, Major Statistical Series of the U.S. Department of Agriculture: How They Are Constructed and Used, is a fairly comprehensive documentation of most of the existing agricultural statistics. The 11 volumes of this handbook should be updated and include more detail on all agricultural and rural statistics programs. Other efforts underway in USDA to document and provide general access to computer readable data bases should continue. A central point for locating and learning about agricultural and rural statistics has been under development by ESCS. This should prove to be a valuable service to other users of agricultural statistics.

A few other issues affect agriculture statistics just as they do other statistical programs. Statutory protection of data collected for statistical and research purposes is needed. The Freedom of Information Act as well as increased auditing and regulatory activity in government make it necessary to have this authority to hold data confidential. The timeliness of release for some agriculture statistics can be improved. Some data definitions, table formats, and presentation can be improved to make statistics more comparable and useful.