Farming Behaviors. Farms do not occur in nature; they are a product of human and social behaviors. Therefore, accord-ing to a behavioral definition, it is sugge-sted here that a farm is a place where agricultural behaviors produce food and/or fiber from plants and/or animals on a sufficient scale for household consump-tion, sales and/or leisure.

Note that this definition does not depend upon the size or acreage of the place, upon the amount of agricultural goods that might be sold from such a place, whether it operates at a profit or a loss, or the way in which it is organized. Along the same lines, a farmer is one who engages in agricultural behaviors that attempt to produce food and/or fiber from plants and/or animals on a sufficient scale for household consumption, sales and/or leisure. Whether farms are small farms or large farms is a matter of degree and dimension; there are no absolute dividing lines between them. Some are simply smaller or larger than others in terms of a given characteristic or pattern of characteristics. A hundred acre farm, for example, may be a small farm for raising beef cattle, but it would be a large farm for growing strawberries.

The official definition of a farm in the United States is determined by the U.S. Department of Commerce. As might be expected, this definition has a commercial orientation and not the more general behavioral perspective on whether a place is a farm. Variations of this official definition have been around since 1850 when the first census of agriculture was taken (Harlan, Vacca, and Swaim 1993; Reilly and Hartwig 1995). Officially, the farm definition has gone through nine sets of criteria since 1850.

• 1850 and 1860

At least $100 worth of agricultural production for home use or sale and no acreage limitation.

• 1879, 1880, and 1890

Any agricultural operation of three or more acres; if less than three acres, at least $500 worth of agricultural products sold.

• 1900

An agricultural operation requiring the continuous services of at least one person.

• 1910 and 1920

Any agricultural operation with three or more acres; if less than three acres, $250 worth of agricultural goods produced for home use or sale. Or, an agricultural operation requiring the constant services of at least one person.

• 1925, 1930, 1935, and 1940

Any agricultural operation with three or more acres; if less than three acres, $250 worth of agricultural goods produced for home use or sale.

• 1945

Agricultural operations consisting of three acres or more of cropland or pastureland; or, $150 worth of agricultural products produced for home use or sale. If less than three acres, $250 worth of agricultural products produced for home use or sale.

• 1950 and 1954

If three acres or more, $150 worth of agricultural products produced for home use or sale. If less than three acres, $150 worth of agricultural products produced for sale.

• 1959, 1964, and 1969

If ten acres or more, at least $50 worth of agricultural products produced for sale; if less than ten acres, a minimum of $250 worth of agricultural products for sale.

• Since 1974

In the agricultural censuses of 1974, 1978, 1982, 1987, and 1992, a farm was defined as, ". . Any place from which $1,000 or more of agricultural products were pro-duced or sold, or normally would have been sold, during the census year" (U.S. Bureau of the Census 1994; vii).

Past criteria for official definitions of farms have included various combinations of the value of the farm goods produced, the home consumption of farm products, the sale of farm products, acreage and labor requirements. Today's narrower economic definition only uses the value of the annual commercial sales of farm products from a place. It does not use acreage, labor requirements or the amount of farm products consumed by the farm residents and not sold. Omitting home consumption of farm products from the official criteria for a farm serves to officially omit an unknown number of places from being counted as farms -- no doubt small farms and behaviorally important to those households and communities -- in the U.S. census and national agricultural database used by rural and agricultural scientists, public agencies and farm-related businesses.

Nonetheless, although the current, official, $1,000 definition of farms is not as broadly inclusive of the full range of farms as allowed by the behavioral definition outlined earlier, the $1,000 definition is much more inclusive of small farms than a $10,000 sales threshold would be.

Faced with a large budget reduction in late 1995, the Bureau of the Census proposed changing the current definition of a farm to a minimum of $10,000 of agricultural products sold or normally expected to be sold in the census year. This would have meant that only about half as many farms would be counted in the 1997 census, a number that could be included with the funding anticipated by the agency (Wimberley, 1996). However, there was much public resistance to raising the sales threshold of the farm definition for farming operations to be officially considered as farms. Many agency administrators as well as private citizens and organizations with agricultural interests opposed the pending change of farm definition. In other words, the answer to the question, "What is a farm?," mattered to many. Consequently, the budgeting issue was resolved by transferring the responsibility and budgeting for the forthcoming 1997 agricultural census from the Department of Commerce to the Department of Agriculture, a transfer that would be permanent. This marks a historic change in the census of agriculture. More importantly for the 1997 census, USDA announces that it will keep the $1,000 threshold in the farm definition for the immediate future.

Since the farm definition determines the very data farmers, scientists, public and private agricultural agencies and farm-related businesses use to inform their programs and policy or business decisions, it matters a great deal how farms are defined. There are public and environmental reasons, food-security reasons and quality-of-life reasons to keep the farm definition behavioral broad and inclusive. To illustrate, let us consider some of the effects -- an impact assessment -- of what would happen if the $1,000 definition had been pushed to a minimum of $10,000 in farm product sales. Such a change in the farm definition appears highly unlikely at this point since the authority for the 1997 census rests with USDA.

What if farms accounting for over 13 % of U.S. farmland, averaging 136 acres, having machinery and equipment values of $24 billion, and with farm real estate exceeding $132 billion were excluded from basic information on U.S. agriculture? What if more than one-fourth the farms showing profits were left out of the official U.S. count? And what if farms operated by 28% of the farmers whose principal occupation is farming were omitted from our nation's basic agricultural database? In other words, what if the farm census no longer covered half the U.S. farms and ranches? This would take a toll on each state's agricultural programs; many farm, food and fiber-related businesses; scientific research and applications; environmental and health information; small farms; and farms operated by minorities and women.

What droughts, floods, freezes, pests, prices, diseases, politics, legislation, and urbanization could not do to American farms, a definition could.

• Nationally

Nationally -- based on the most recent 1992 census count as a basis for future estimates (U.S. Bureau of the Census 1994; Reilly 1995), 0.9 million of the 1.9 million U.S. farms would no longer count as farms. That is, 47% of the farms would no longer officially exist. In essence, we would lose half the U.S. farms by a single swoop of the definition (Table 1).

• By state

In addition to the loss of information on half of our farms nationally, 29 states would lose at least half of their farms from the census data. These and many other states would be pressed to make up for the loss in agricultural data. Absolute declines of 50,000 or more farms would occur in four states: Texas would go from 180,644 to about 70,000 farms; Missouri from 98,082 to about 47,000; Tennessee from 75,076 to about 24,000; and Kentucky from 90,281 to about 40,000. Other states suffering large numerical decreases include Oklahoma from 66,937 to around 30,000 farms; California from 77,669 to about 41,000; Ohio from 70,711 to about 38,000; and North Carolina from 52,854 to about 25,000.

According to data from the Bureau of the Census (Reilly 1995), the state where farm numbers would fall proportionately the most is West Virginia where at 78% of the farms in 1992 grossed less than $10,000. Farms there would decline from 17,020 to less than 4,000 farms. Tennessee's loss would be approximately 68%; South Carolina's 67%; Alabama's and Mississippi's 64% each; Oregon's 62%; Texas 61%; and Virginia's and New Jersey's 60% each. Overall, southern states would appear most frequently among those with larger declines. Of the 16 southern states, six are among the top ten losers as are 12 of the top 25.

States losing relatively fewer farms by a change in census definition concentrate in the upper North Central region. These include North Dakota, South Dakota, Iowa, Nebraska, Illinois, Minnesota and Wisconsin. States that lost proportionately the fewest farms are likely to receive the most government payments to farms as reported in the 1992 census.

Overall, the number of states that would lose more than the national average of 47% of their farms by changing the farm definition from $1,000 to $10,000 far exceeds those losing less than 47%. The ratio is nearly two to one. Thirty-two states exceed the national percentage of loss; only 18 states lose less than the average loss. Although some states could lose more than others, none gain. With a more restrictive definition, everybody loses farms and information.

• Information for agricultural research, extension, businesses, farmers.

With such a change in definition, we would also lose much of the database used directly and indirectly for every U.S. county by federal, state and local officials; scientific researchers; agricultural extension services; agribusinesses; and, not least, farmers and ranchers them-selves. Furthermore, a change of definition in the census of agriculture would trigger the same change in the decennial census of population, housing, and socioeconomic information for farm and nonfarm, as well as rural and urban people and places.

• Environmental and health information

Furthermore, we would stand to lose baseline data on the environmental and health-related factors pertaining to various crop and animal products associated with soil, water and other conditions. Such information would be lost in both rural and urban areas where small farms exist. For example, farms grossing less than $10,000 annually amount to nearly 27 percent of the productive farms having some or most of their acreage in the Conservation Reserve Program. A $10,000 definition would move those that produce, but sell less in dollar values of farm goods, still further out of the U.S. agricultural database.

• Minority-operated farms

Nonwhite, minority farms would decrease by 65% from 43,487 to 28,235. Three-fourths of the farms operated by African-Americans would officially disappear as a result of the new definition, down by 76%, from 18,816 to only 4,567 farms. Black-operated farms have been disappearing at high rates for decades. However, the change of the farm definition would instantly push the numbers down to where they might otherwise be projected in another 20 to 30 years. Similarly, farms operated by women would drop from 145,156 to 50,292, a 65% decrease.

• Full and part-time farms

Nearly three of every ten people, 28% whose principal occupation is farming would no longer be included in the agricultural census if the $10,000 requirement were used in the farm definition. Nearly seven of every ten who are primarily part-time farmers would be dropped from the count and from the official database.

• Sales

The largest 2% of the U.S. farms, those with $500,000 or more in sales in 1992, produced nearly half of all the farm products sold. The 47% of the farms that grossed less than $10,000 in 1992 accounted for less than 2% of the sales of agricultural goods (U.S. Bureau of the Census 1994: 5). In the U.S., large farms produce more and small farms produce a smaller share of the agricultural goods marketed. This is not new. However, eliminating so many small farms from the definition and agricultural data would be new. Small farms and their characteristics are typically undercounted while large farms tend to be overcounted. In the 1987 census, for example, farms selling under $10,000 were undercounted by 22% while those selling more were overcounted by nearly 4% (U.S. Bureau of the Census 1990: xiii and 1). With a behavioral definition of farms that includes home consumption of farm products rather than sales alone, still more small farms would be found and counted in the nation's agricultural data.

• Small-farm commodity production

Large farms do not simply replace the roles played by small farms. Large farms do not exclusively produce all the food and fiber items that we consume or export. Certain commodi-ties are often produced on small farms. Smaller farms provide many specialty goods. These include nearly 30% of the pecan sales; 12% of the walnut acreage; 7% of the ginseng acreage; 19% of the blackberry acreage; 10% of the hay sold; 8% of the sheep, lambs, and wool that are sold; and 55% of the horses. However, farms grossing less than $10,000 contribute little to big-farm cash crops such as corn, wheat, soybeans, sorghum, Irish potatoes, and cotton or to poultry, dairy products, or pork. In today's agricultural structure, it appears inefficient for small farms to try to compete in the mass production of such crop and livestock commodities.

• The small-farm market for farm supplies and services

The 47% of farms in 1992 that grossed less than $10,000 account for about $5 billion in purchases of farm supplies and services including about $0.7 billion in local property taxes. Furthermore, the farms that would have been defined out of the database account for nearly $24 billion of the total, $93 billion value of U.S. farm machinery and equipment (U.S. Bureau of the Census 1994: 49-50 and 52). Small farms are big businesses. Regardless of the proportions that the smaller farms purchase and maintain, these billions represent major markets and revenues from purchases of farm inputs that are important to agricultural businesses as well as to the farmers.

• Small farm real estate and profit

The change in farm definition would eliminate from the census coverage about 27% of the farms showing profits where the value of product sales exceed production expenses. These farms represent over 13% of the U.S. farmland. They have an average size of 136 acres and a total real estate market value of over $132 billion. By any interpretation, a sizable amount of U.S. agriculture would be omitted from the census and information base with the higher sales requirement in the farm definition.

• A diverse farm structure and food security

All Farms are not alike. The overall structure of U.S. agriculture is multi-dimensional (Wimberley 1987). In addition to corporate-commercial and large-family-farm-area agristructure, one of the three major dimensions is small-farm agriculture. Small-scale structure is characterized by farms that are independently owned and operated by individuals or families who live on them. Their acreage tend to be lower. Their operators are often part-time farmers, and they frequently hire labor.

Small-scale agricultural structure is an added dimension of diversity for the strength of the nation's food security and sustainability. If there is strength in diversity, the diversity in the American food and fiber system is a basis for food and fiber security, agriculture's biological and economic robustness, and other agricultural strengths. A policy hypothesis is that the diversity of agricultural structure is important for a secure system of food and fiber and for a sustainable agriculture and environment. In a diverse structure of agriculture, it is unlikely that all farms will use the same practices, technology, genetic stocks, inputs, processors or marketing outlets.

• Small-Farm Structure and Quality of Life

It is not that small farms with low incomes are unimportant. Small farms can be quite important to families with low incomes. These farms supplement family incomes and food resources in many impoverished areas. Historically and even today, many low-income farm families depend upon what they grow to increase their incomes through cash sales. Low-income farm families can likewise consume some of the food they produce and thereby offset family budget expenses, improve their diets, or both.

Research on corporate-commercial, large-family-farm-area and small-farm agricultural structure have shown that the structure of agriculture is related to socioeconomic quality of life (Reif 1987; Lobao 1990). Essentially, high levels of corporate-commercial agricultural structure show little relationship to socioeconomic conditions. Large areas of family farm structure show positive benefits on socioeconomic well being. However, small farm structure is related to poorer socioeconomic conditions of poverty, low incomes, and unemployment. But does this suggest that poor places would be better off without small farms? Not likely. Rather, the presence of small-farm structure helps prevent the poorer quality of life in those places from becoming worse.

Losses that would be suffered by a narrow official definition of U.S. farms are far reaching. Keeping the farm definition behaviorally diverse in addition to or instead of a low sales threshold will keep the information needed to understand, serve and guide the diversity and strength of American agriculture and the biological, economic, social, and environmental security of our food and fiber system. In the future, the issue of what is officially a farm will emerge again. Those for whom the definition of farm matters -- farmers, other citizens, agricultural scientists, public agencies, and private agricultural business interests -- will have to determine what criteria are important for including places as farms in the national base of statistical information. When the time comes, perhaps a definition can be more clearly formulated to include behavioral criteria beyond annual farm product sales. If so, this will be to the advantage of keeping small farms in the data and information base on American agriculture.

As long as there are producers and consumers, it matters what a farm is -- in the behavioral sense of the term.

NOTES

1. This text accompanies a presentation to the USDA National Small Farm Conference in Nashville, Tennessee, September 12, 1996. The analysis is a contribution to Regional Research Project S-246, "The Transformation of Agriculture: Technology, Natural Resources, and Policy Implications," and was supported primarily by the Agricultural Research Service of North Carolina State University. The author is responsible for the ideas and interpretations presented here. For further information, contact the author at the Department of Sociology and Anthropology, North Carolina State University, Raleigh, North Carolina 27695-8107, telephone (919) 515-9026, or Fax (919) 515-2610.

2. Much of this analysis draws from an article, "Lose Half of American's Farms? A Definition Can," by the author (Wimberley 1996).

Harlan, William A., E. Ann Vacca, and Nancy L. Swain. 1993
"Farm Definition." Paper at the December meeting of the Census Advisory Committee on Agricultural Statistics. U.S. Bureau of the Census, Washington, D.C.

Lobao, Linda M.

1990 Locality and Inequality: Farm Structure and Industry Structure and Socioeconomic Conditions. Albany: State University of New York Press.

Reif, Linda

1987 "Farm Structure, Industry Structure, and Socioeconomic Conditions in the United States."

Rural Sociology 52 (Winter): 462-482.

Reilly, Joseph

1995 "Census of Agriculture: Results and Future Plans." Agriculture and Financial Statistics Division, Bureau of the Census, Washington, D.C. Paper presented at the meetings of the Rural Socioecological Society, Pentagon City, Virginia, August.

Reilly, Joseph and Douglas Hartwig.

1995 "Plans for the 1997 Census ofAgriculture." Paper presented at the May meeting of the Census Advisory Committee on Agricultural Statistics. U.S. Bureau of the Census, Washington, D.C.

U.S. Bureau of the Census 1990

1987 Census of Agriculture: Coverage Evaluation. Vol. 2, Part 2. Washington, D.C.: U.S. Government Printing Office.

U.S. Bureau of the Census 1994

1992 Census of Agriculture Volume I: United States Summary and State Data. Vol. 1, Part 51. Washington, D.C.: U.S. Government Printing Office.

Wimberley, Ronald C. 1987

"Dimensions of U.S. Agriculture: 1969-1982." Rural Sociology 52 (Winter): 445-461.

Wimberley, Ronald C.1996

"Lose Half of America's Farms? A Definition Can." The Rural Sociologist 16 (May): 37-41