The second category in Table 1, retirement farms, is comprised of retired operators and their families. These households, accounting for 14 percent of all farms, had household income from off-farm sources averaging near the average level of US households and had nearly twice the wealth of US households. Hence, they could not be considered to be disadvantaged as a group although some households in the category undoubtedly were poor.

The third category is labeled residential/lifestyle households because, though living on farms, they do not call themselves farmers, but rather call themselves lawyers, factory workers, nurses, or another nonfarming occupation. They are numerous, comprising 41 percent of “farm” households, and had income and wealth well above those of the average US household in 1998.

Farms not in the first three categories but with farm sales under $100,000 are in the “farming occupation: lower sales” category in Table 1. They comprise a sizable 21 percent of farm households. Their total household income was low at $34,773 (they would have considerably more income if they relied solely on their off-farm income), but their wealth averaged more than double that of US households. Hence, they are not classified as economically disadvantaged.

The final three categories are classified as commercial farmers—farms sales are over $100,000. Their operators classified themselves as farmers and are not retired. Although they accounted for only 13 percent of all farm households, they accounted for 85 percent of farm crop and livestock sales. On average, households on each of the three categories of commercial farms had wealth more than double that of US households, hence by the standard defined above none of the three categories is economically disadvantaged.

The two largest size categories, those with sales of $250,000 or more, accounted for only 8 percent of all farms in 1998. However, those 8 percent accounted for nearly three-fourths of farm sales of crops and livestock. The majority of their income came from farming and their income and wealth were multiples of those for nonfarmers.

In summary, household income averaged over all farms in 1998 was 15 percent above US household income, and farm household wealth was 74 percent above US household wealth. Only one category of farms, limited resource farms, was classified as disadvantaged based on criteria used herein.

Because limited resource farms in Table 1 accounted for only 7 percent of all farm households and for a much smaller proportion of all farm sales, they cannot be helped cost-effectively by commodity programs that distribute benefits nearly proportional to crop production. Their need is for public assistance and human resource development programs of counseling, education, training, job search assistance, and relocation assistance.

Although farm household income or wealth was well above that of US households for each except the limited-resource farm category in Table 1, households on farms with crop and livestock sales of under $100,000 but not classified as retirement or limited-resource farms had income averaging only two-thirds that of other US households in 1998. Some of those farms benefit from commodity programs. Because noncommercial farms produce relatively little farm output and commodity program benefits are closely related to past output, commodity programs as currently formulated do not compensate these farms for technological change. However, direct payments under the 1996 farm bill could in principle be redirected to losers from technology change. A serious problem is that the losers are not easily identified.

Conclusions from Table 1 would not change much in the absence of federal commodity support programs. A special disaster appropriation raised direct payments to producers by $4.8 billion or by $2,373 per household in 1998. If that had not been available, income per farm household still would have averaged $57,361 or 11 percent above average US household income of $51,852 in 1998. Total government outlays for all farm support programs in 1998 totaled $10.1 billion or $5,015 per farm household. Elimination of all support programs would have brought income per farm household to $54,719 or 5.5 percent above average income of US households and still an all-time record—4 percent above the previous record set in 1997.

A most striking feature of Table 1 is the role of off-farm income in farm prosperity. Income of farm households from farm sources averaged only $7,106, but from off-farm sources alone averaged $52,628 or 1.5 percent above the average income of US households! Off-farm income has been rising and is expected to remain high; hence income of farm households is likely to remain above that of nonfarm households whether farm income is high or low. Mechanization technology has enabled farmers to farm while earning substantial off-farm income. As a result, over the years farmers have moved from incomes chronically lagging behind their urban cousins to a projected status of incomes chronically exceeding their urban cousins. For example farm income per capita was only one-third of nonfarmers’ income per capita in 1934 (US Department of Agriculture, 1960, pp.34-37).

The farming industry is headed for a paradox when over half of farm “operators” (41 percent in 1998) will not classify themselves as farmers.[4] Because of the inefficiency of small farms, hobby farming is expensive. It is becoming an avocation reserved for persons with substantial wealth and off-farm income against which to write off farm financial losses for tax purposes. Because farm household income will be above that of nonfarmers even when farm income is zero, it is time to reconsider what is a farm in order to make sound policy decisions. Commodity program advocates will argue that only farmers receiving the majority of their income from farming should be used to judge the economic position of “true” farmers, but Table 1 illustrates that these commercial farmers (group 5, 6, and 7) fared well in the depressed farm economy of 1998.

The New Economic Paradigm for Agriculture

The above data lend support to Tweeten and Zulauf’s contention that agriculture is under a new economic paradigm. That paradigm recognizes that farm commodities are not public goods that warrant public interventions, that markets work in agriculture, and that competent commercial farmers adjust quickly enough to technological change and other shocks to avoid chronic economic low returns on their resources. Why technology makes that possible is explained below.

Willard Cochrane’s well-known treadmill theory and Glenn Johnson’s fixed asset theory have traditionally provided the intellectual formulation explaining why farmers lose from technological change. These theories presume that aggregate farm demand and supply are inelastic. Empirical evidence indicates that aggregate supply and demand for farm output are indeed inelastic in the short run of up to 5 years but are elastic in a long run of 10 years or more (Tweeten, 1967; Gardiner and Dixit; Tweeten and Quance). Additional farm output coming from improved technology does not reduce farming receipts in the long run. Farm markets are sufficiently flexible in the long run to adjust to and reap gains from technological change as noted later.

It is also notable that technological change occurs gradually and is more foreseeable than annual or cyclical shocks to the farming economy. The shocks that constitute the principal economic problem facing commercial agriculture come mainly from weather, commodity and business cycles, and trade. Hence, unlike technological change that farmers anticipate and can plan for, these shocks cannot be predicted.

Farm commodity terms of trade or the parity ratio (ratio of the index of prices received by farmers for crops and livestock to the index of prices paid by farmers for inputs) in 1999 was only 36 percent of the 1910-14 average (US Department of Agriculture, June/July 2000, p. 30). This number has been interpreted by some to mean that farmers are only receiving 36 percent of a “fair” price. In fact, as explained below, real farm prices have risen!

Aggregate farm output and aggregate production input are measured over time by weighting physical quantities by constant dollar prices. Then these are added over all inputs to measure aggregate production input and are added over all crops and livestock to measure aggregate farm output. The multifactor productivity index (ratio of aggregate output of crops and livestock to aggregate farm production input) in 1999 was 3.94 times the 1910-14 level. Thus the factor terms of trade index, defined as the real price (purchasing power) of farm output per unit of farm production input, was 142 percent (36 x 3.94) of the 1910-14 level. In other words, real buying power of the average production input was 42 percent greater in 1999 than in the 1910-14 base period.

Alternatively, a productivity index of 3.94 in 1999 meant that farmers were “growing 4 blades of grass where one grew” in 1910-14 with the same input volume. Hence farmers needed only 25 percent as high a real price in 1999 to achieve the same real income per unit of production input as in 1910-14. In fact, however, real price in 1999 was 36 percent of that in 1910-14, hence real price in 1999 was 36/25 or 142 percent of that in 1910-14 after the parity ratio is adjusted for productivity growth. It follows that, comparing 1999 with 1910-14, the real parity ratio (factor terms of trade) was up by 42 percent rather than down 64 percent as implied by the conventional parity ratio (commodity terms of trade) of 36 percent.

Technology not only allowed real farm prices to improve, it also helped real farm income to improve. Real personal income per capita of the US population increased to 3.5 times its 1930 level by 1999, an average annual gain of 1.8 percent per year (Council of Economic Advisors, p. 341). Meanwhile, per capita income of farmers increased from 40 percent that of the average American to 113 percent of the average American (US Department of Agriculture, 1960, p. 38).[5] It follows that real income per person on farms was approximately 10 times (3.5 x 113/40) higher in 1999 than in 1930.

In an earlier study (Tweeten, 1994, p.7) I apportioned the 3.4 percent annual growth rate in income of farmers to its five principal sources for the 1930-1990 period as noted below:

(Proportion of farm household income growth by source)	(%)
Parity price ratio (commodity terms of trade)	-5
Multifactor productivity	19
Farm size	28
Government payments	2
Off-farm income	56
		100

Technology played a role in several of these sources of income growth. Partly because of productivity gains in excess of the increase in demand for farm output, a falling parity ratio would have reduced farm household income from 1930 to 1990, ceteris paribus. Multifactor productivity gains more than compensated. Other things equal, multifactor terms of trade accounted for approximately 14 percent (19 percent gain from productivity less a 5 percent loss from falling parity price ratio) of the increase in farm income per capita from 1930 to 1990. Government payments and farm commodity programs in general contributed little to growth in income of farm people.

Technology was central to the two major sources of farm income growth—farm size and off-farm income. Mechanization technology accounted for much of the 28 percent of the farm income gain attributed to greater farm size. Mechanization along with improved roads and vehicles also facilitated off-farm employment, which accounted for 56 percent of the income gain of farmers per capita.

The assumption that demand has to be elastic for technology to raise farm income applies only to factor-neutral technological change. In fact, much change in agriculture has come from labor-saving rather than output-increasing, factor-neutral technology.

Larger tractors, combines, milking parlors, and other technologies continue to improve farm productivity, allowing farms to expand in size and/or freeing farm operator and family labor to earn more off the farm. More recent technologies such as Roundup Ready soybeans and conservation tillage primarily save inputs rather than increase output. These technologies especially raise farm income because they do not reduce farm commodity prices as much as do output-increasing technologies such as irrigation, hybrid corn, or artificial insemination. By saving labor, Roundup Ready soybeans and conservation tillage leverage the time of the farm operator to cover more acreage. It follows that, in contrast to output-increasing technology, labor-saving technology retains more economic benefits on the farm but also makes for larger and fewer farms as noted in the next section.

Should Technological Change be Slowed?

The demand for farm output is inelastic in the short run (Tweeten, 1967), hence stifling innovation that moves the food and fiber supply curve to the right could temporarily raise farm prices and incomes. That strategy has several shortcomings, however.

The nation’s standard of living is established by technology. Nations succeed economically by saving, and then investing those savings in high payoff activities. This nation has a poor record of saving—personal saving rates are near zero. It has succeeded economically by providing a favorable climate for investment of savings from domestic corporations and foreigners in productive and profitable human, material, and technological capital. If productive technology essential for economic growth is encouraged for other industries, by what justification can it be withheld from agriculture? Withholding technology is difficult because the private sector now provides most agricultural research and development and cannot easily be shut down. Thus it is difficult as well as unwise to withhold technology from agriculture.

If technological progress could be materially slowed, farm people would be made worse off in time. With an elastic demand for farm output in the long run (Tweeten, 1967), the loss of foreign and domestic markets from retarded productivity growth could sacrifice competitiveness and could reduce income of farmers and nonfarmers.

In the new agriculture characterized by the strong tendencies towards economic equilibrium, income of farmers moves to equal income of nonfarmers (see Gardner; Tweeten and Zulauf). Slowing of technology and productivity gains in agriculture would slow the entire economy. Less income per capita for the economy would mean less income for farmers and nonfarmers alike.

Concluding Comments

The conclusion is that improved technology has on average benefited those who remained on the farm as well as well as those who left. A number of limited-resource farmers especially experience economic hardship, however. Their hardship probably has little to do with technological change. They are not helped much by commodity programs but can benefit from human resource development efforts.

This section made the case that the high living standards of farmers today could not have been achieved in the absence of major technological change—economic gains occurred because and not in spite of technological change. Nonetheless, it is well to look for policies that leave a larger share of benefits to farmers in general and especially for full-time operators in the “disappearing middle” size group (farms of 50-500 acres; smaller and larger farms increased their numbers between 1992 and 1997 according to the 1997 Census of Agriculture).

Public Policy Changes to Enhance Agriculture’s Ability

to Benefit from New Technologies.

Forsaking promising new technology does not appear to be a viable strategy for agriculture or the economy in general. Other strategies to help farm income in the face of continuing technological change appear to be more promising. Potential strategies include creating a higher and more elastic demand for farm output; emphasizing cost-saving, scale-neutral research; focusing on research and extension to benefit small farms; and targeting human resource development and commodity program payments to “at risk” farms—especially to mid-size and limited resource farms.

Creating a Higher and More Elastic Demand for Farm Output

The economic advantage to farmers of a higher and more elastic demand for farm output is illustrated in Figure 1. The original position with food demand D and supply S is an equilibrium at price P_o and quantity q_c.

Figure 1. Benefits of higher and more elastic demand with productivity advancing supply from S₀ to S₁.

An advancement in technology shifting the industry supply curve from S₀ to S₁ reduces industry price to P₁ and expands quality to q_ć. Economic gain to consumers (consumer surplus) is area a+b+c. With the advance in productivity, producers lose area a, but gain area e+f. If demand is inelastic, the loss of a is likely to exceed the gain from e+f, hence producer surplus falls with farm technological progress. The gain to the nation is the benefit to consumers (a+b+c) less the loss to producers (a-e-f) or b+c+e+f less cost of developing the technology.

Suppose that a highly elastic demand curve at X, originally slightly below P₀, is brought up to P₀. This could come about because of success in multilateral trade liberalization, because of technology creating new uses for farm output such as biomass for producing ethanol, or from cost-effective market promotion of farm products.

With farm productivity advance moving S₀ to S₁, given the new demand curve (following D down to intersect with S₀, then following X to the right), the new equilibrium is at price P₀ and output q_p. Quantity q_c continues to be utilized by the conventional market while q_p-q_c goes to the more elastic market for exports, ethanol, or new products. New technological progress does not increase consumer welfare but allows producers to reap the benefits of technological progress—producer surplus is raised by area b+c+d+e+f compared to the initial situation at supply S₀ and demand D.

The new elastic market demand X brings benefits to society and producers. The net gain to society from the new technology is b+c+e+f-C without the new market and is b+c+d+e+f-C΄ with the new market. C is the cost of developing the new technology and C΄ is that cost plus any cost of developing the new market. If the cost of market development is small, the gain to society from the new market is area d. All benefits accrue to producers, but in principle some gains could be taxed and thus shared more widely in the economy.

Emphasize Cost-Saving Scale-Neutral Research

The impact of output-increasing, scale-neutral technological change is contrasted with that of labor-saving, scale-biased technological change in Figure 2. In each case for the representative farm shown in the left side of panels A and B, the short-run average cost is SAC₀. For the industry as a whole, demand is D and initial supply is S₀. Initial price is p₀, firm production in q₀, and industry output is Q₀.

Figure 2. Farm firm and industry impact of scale-neutral and scale-based technology

Assume that an output-increasing, scale-neutral technology is introduced that reduces the short-run marginal cost to SMC₁ for the representative firm in panel A facing price p₀in Figure 2. Output of the firm expands to q₁. Firm pure profit is q₁(p₀-m). As other firms also expand output, the industry supply curve shifts from S₀ to S₁. The result is to reduce the price to the long-run equilibrium p₁. Firm size remains at q₀ but sufficient firms enter the industry to maintain industry output at Q₁.

Pure profit is squeezed from firms but society is better off by the national income (deadweight) gains shown as the crosshatched area in Figure 2. Laggards in adopting the new technology face losses of q₀(p₀-p₁), inducing them either to adopt the new technology or see their equity erode. Thus benefits of output-increasing, scale-neutral technology originally go to early adopters but eventually accrue to consumers through lower food and fiber prices as shown by panel A, Figure 2. Farm profit rates on resources are unchanged, given time for adjustment.

Examples of relatively scale-neutral technology include conservation tillage and genetically modified products such as bovine somatotropin (BST) and herbicide-resistant crops. These technologies work as well for crops and livestock on small farms as large farms and do not entail large, indivisible capital outlays. Still, it is difficult to find totally scale-neutral technology. Bioengineered technologies and conservation tillage are more prevalent on large than small farming operations.

The impact of labor-saving, scale-biased technology is illustrated in panel B of Figure 2. New technology shifts the representative firm’s short-run marginal cost curve to SMC₁. Profit of p₀-m on each unit when output is increased to q΄ brings total pure profit q΄( p₀-m). Because the new technology does not markedly increase yields and total area, and because firm numbers do not increase in the short run, the supply curve shifts little and the net social benefit to consumers and society is modest as illustrated by the crosshatched area in panel B of Figure 2.

Individual farms continue to make large profits, however. The profit arises from saving labor. Each farm operator will be able to farm more land with the new technology. As operators compete for land, pure profit is bid into the most fixed instrument, land, controlling access to that profit. Thus benefits of labor-saving technology accrue first to early innovators but eventually to landowners. Rents rise. Many farm families exit the sector as farms consolidate and expand size. Those who fail to adopt cannot cover all costs of production when land rent (or the opportunity cost of capital invested in land) is included. To survive, laggards must adopt the new technology, accept a low return on their resources, or leave the farm.

The situations differ between panels A and B, however. More farm operators are likely to be displaced by the labor-saving technology (panel B) than by scale-neutral technology (panel A), but laggards in adopting labor-saving technology reap gains from land price appreciation if they own land. Fewer operators are displaced by output-increasing technology, but laggards in adopting technology may need to find off-farm employment to support financially their farming operation as commodity prices fall.

Research for Small Farms

A third potential option to retain more research benefits for farmers would be to emphasize public research and extension (R and E) for small farms. R and E could focus on appropriate technology and social science research for noncommercial farms of less than $100,000 of sales. According to the Census of Agriculture (US Department of Agriculture, March 1999, p. 12), such units accounted for 82 percent of all farms in 1997. Thus social science research to improve marketing and management and in general improve human resources and technology on small farms would reach the majority of farmers. Small farm advocates have strongly urged such a policy (Berry).

Output-increasing technology for small farms would not have much impact on farm output. Noncommercial farms accounted for only 12.6 percent of farm crop and livestock receipts in 1997 (US Department of Agriculture, March 1999, p. 12). Hence a 10 percent increase in output on such farms would increase overall farm output only 1.3 percent. Farm commodity prices and receipts would decrease very little.

Focusing research on small farms has several serious shortcomings, however:

· It is very difficult to develop productive and profitable technology for small farms that does not provide even more benefits to large farms.

· Even if technology could be directed solely to small farms, the payoff to society could be low. Investments to improve farm technology have had a high payoff mainly by lowering farm production costs and food costs. If technology works mainly through lower cost farm output to benefit society, then the larger farms that produce most crop and livestock output must be included. It could cost just as much to develop technology for small farms as for all farms but the market would be only 12.6 percent as large. Cutting benefits to one-eighth the current level would make most investment in agricultural research and extension unprofitable to society.

· Farm consolidation would not be greatly slowed by redirecting public R and E to small farms. Spurred by new technology and ability to patent life forms, private firms now invest perhaps twice as much as the public sector in R and E. Furthermore, compared to the public sector, the private sector invests relatively more R and E in mechanization research, which is especially responsible for labor-saving, scale-biased technology that has brought fewer and larger farms.

· Finally, small farms are becoming gentrified. Except for a relatively few highly able, innovative operators on small farms, small operations entail such large losses per unit of output that only households with considerable wealth or off-farm income can afford to be a hobby farmer. As this gentrification process proceeds, the case for special treatment of small farms diminishes.

These shortcomings of targeting research do not mean that selective investment in physical and social science to benefit small farms cannot have a favorable payoff to society as well as to small farmers. Thus a wise strategy is for the public to continue its past policy of emphasizing scale-neutral investments in agricultural research and extension.

Targeting Farms at Risk

As noted earlier, census data show signs of the “disappearing middle” size of farm. Very small and very large farms are growing in numbers and in share of farm sales (US Department of Agriculture, December 1999). Mid-size farm numbers and share of sales are falling.

One means to save family farms is for the Farm Service Agency to target concessional credit solely to farms experiencing financial difficulty and facing foreclosure. A frequent criticism of this strategy used extensively in the early 1980s is that poor managers rather than good managers on farms were rewarded by government. The program was widely abused because bureaucrats were ill-equipped to sort out which farm operators were truly in need and could be made financially viable at a reasonable cost to taxpayers.

If the goal is to preserve limited-resource farm households, a mix of public assistance, credit, extension farm management and marketing, and other efforts could target such farms. The cost, however, of making such farms viable for taxpayers is likely to be very high.

Finally, the decoupled payments of the 1996 farm bill could be targeted more narrowly to mid-size farms. Payments of all types might be restricted to no more than (say) $50,000 per day-to-day farm operator. Only one operator per farm could receive payment and an operator farming several units could not receive over $50,000 in total payments. Some of the savings of government outlays compared to current payments could be used to fund human resource development and job finding assistance for those who wish to exit farming in mid-career. Some of the savings in farm commodity programs also could be invested in public research and extension—previously under-funded judging by the high rates of return on investment. This last option may offer the highest payoff to society of any of the alternatives considered in this section.

Conclusions

The foregoing analysis leads to the following conclusions:

· Based on data reported in this study, much income efficiency would be lost because of slowing the rate of technology advance to improve farm income. This nation’s economic progress is a function of investing savings in high payoff capital. The United States has a low (near zero for personal savings) savings rate. Its impressive economic success traces mainly to its innovativeness in allocating resources to high payoff investments in research and development that generate knowledge and improve technology in the farm and nonfarm sectors of the nation’s economy.

· The new economic paradigm recognizes strong equilibrium tendencies in the economy so that the income of farm people is determined not by farm technology but by income of nonfarmers. Gains in the nation’s income are principally the product of technology. Over the long term, farm household income is raised the most by allowing markets to work in an institutional environment encouraging public and private research and extension to foster technological progress. Thus a high economic payoff strategy is to target human resource development programs especially to the “disappearing middle” farm households and other at-risk households on farms while ending direct payments and other income transfers to commercial and hobby farms. Savings from sharply lower government transfers to farmers could be used to increase public funding for research and extension.

· In principle, decoupled direct payments originated under the 1996 farm bill could solely target at-risk farmers to keep them on the farm. Public credit programs could do the same. Targeting transfers solely to family farms at risk of financial failure or leaving farming requires triage—avoid subsidies to farms that are sure to fail anyway or that need no help while targeting farms where assistance will make a difference. However, the historic record is that politics and bureaucratic shortcomings conspire to render such fine-tuning elusive if not impossible as cost-effective means to preserve family farms.

· It is well to recall from Figures 1 and 2 the difficulty of circumventing the market indefinitely so that farmers retain the benefits from technology. Benefits of output-increasing technology are passed to consumers through lower market prices. Benefits of scale-neutral technology that does not increase output are bid into land prices and hence are passed to landowners and lost to renters, hired workers, and the new generation of landowners. Benefits of commodity programs also are eventually lost to farmers and instead accrue to landowners over time. Thus across-the-board income transfers to farmers are not a useful means to compensate farmers for technological change, which has not made farm people generally worse off based on data reported in this study.

· Commercial farmers are early adopters of technology and on average gain from technology. Hence, farm commodity programs targeting commercial farms are not cost-effective tools to compensate farmers for losses from technological progress. Nations have tried with only limited success to use commodity programs to circumvent forces of technology and markets pressuring for fewer, larger farms. The shortcoming of that strategy is apparent from Japan, for example, which lost farms at the rate of 2.2 percent annually (compared to 1.1 percent in the US) from 1980 to 1995 despite price supports approximately four times the US level (private communication from Mitsuhiro Nakagawa, Ibaraki University, Japan).

· To be sure, a number of farm people (including some who left farms in mid-career, some persons on limited-resource farms, some technology-adoption laggards, and some hired farm workers) are worse off because of farm technological change. However, it is impossible to identify precisely whether any past, current, or potential farm individual or household is worse off because of technology or because of other factors such as competition from imports, poor management, or bad luck. Thus, an optimal policy may be to use human resource development programs such as education, training, counseling, job relocation assistance, and public assistance to help people who are poor and disadvantaged for whatever reason and not just as compensation for technological change.

· Promising policies to help farmers retain benefits from technological progress include multilateral trade liberalization and other means to expand or make more elastic the demand for farm output. Scale-neutral public research and extension, especially basic research, can help American farmers to compete in a more open global market.

References

Alston, Julian and Philip Pardey. (1996) Making Science Pay: The Economics of Agriculture R and D Policy. AEI Press, Washington, DC

Ames, Bruce and Lois Gold. (1989) “Pesticides, Risk, and Apple Sauce.” Science. 244:755-757.

Berry, Wendell. (1977) The Unsettling of America: Culture and Agriculture. Sierra Club, San Francisco.

Braha, Habtu and Luther Tweeten. (1986) Evaluating Past and Prospective Future Payoffs from Public Investments to Increase Agricultural Productivity. Technical Bulletin T-163. Agricultural Experiment Station, Oklahoma State University, Stillwater.

Council of Economic Advisors. (2000) Economic Report of the President. US Government Printing Office, Washington, DC.

Gardner, Bruce. (2000) Explanations for the Growth of US Farm Income. Department of Agricultural and Resource Economics, University of Maryland, College Park.

Gardiner, Walter and Praveen Dixit. (1986) Price Elasticity of Export Demand. Staff Report AGES86408. Economic Research Service, US Department of Agriculture, Washington, DC.

Perry, Janet, Dean Schreiner, and Luther Tweeten. (1991) Analysis of Characteristics of Farmers Who Have Curtailed or Ceased Farming in Oklahoma. Agricultural Experiment Station Research Report P-919. Agricultural Experiment Station, Oklahoma State University, Stillwater.

Makki, Shiva, Cameron Thraen, and Luther Tweeten. (1999) “Returns to American Agriculture Research: Results from a Cointegration Model.” Journal of Policy Modeling 21(2):185-211.

Rausser, Gordon and William Foster. (1990) “Political Preference Functions and Public Policy Reform.” American Journal of Agricultural Economics 72:641-652.

Tweeten, Luther. (1967) “The Demand for US Farm Output.” Food Research Institute Studies 7:343-69.

Tweeten, Luther. (1989) Farm Policy Analysis. Westview Press, Boulder, CO.

Tweeten, Luther. (1994) “Is It Time to Phase Out Commodity Programs?” Countdown to 1995. Anderson Chair Publication ESO 2122. Department of Agricultural, Environmental, and Development Economics, Ohio State University, Columbus.

Tweeten, Luther and Leroy Quance. (August 1972) “Positivistic Estimates of Aggregate Supply Elasticities: Some New Approaches.” American Journal of Agricultural Economics 51:342-351.

Tweeten, Luther and Jay Coggins. (February 1992) “Political Preference Functions and Policy Reforms: Comment.” American Journal of Agricultural Economics 74:223-226.

Tweeten, Luther and Zulauf, Carl. (Fall/Winter 1997) “Public Policy for Agriculture after Commodity Programs.” Review of Agricultural Economics 19(2):263-280.

Tweeten, Luther and William Amponsah. (1998) “Sustainability: The Role of Markets Versus the Government.” Ch. 3 in Gerald D’Souza and Tesfa Gebremedhin, eds., Sustainability in Agricultural and Rural Development, Ashgate, Brookfield, VT.

US Department of Agriculture. (July 1960) The Farm Income Situation. FIS-179. Agricultural Marketing Service, USDA, Washington, DC.

US Department of Agriculture. (March 1999) 1997 Census of Agriculture: United States Summary and State Data. Volume 1, Part 51. AC97-A-51. National Agricultural Statistic Service, USDA, Washington, DC.

US Department Of Agriculture. (September 1999 and February 2000) Agricultural Income and Finance. AIS-72 and AIS-74. Economic Research Service, USDA,Washington, DC.

US Department of Agriculture. (December 1999) Summary Report 1997 National Resources Inventory. Natural Resources Conservation Service, USDA, Washington, DC.

US Department of Agriculture. (June/July 2000) Agricultural Outlook. AGO-272. Washington, DC: Economic Research Service, USDA.

* Professor Emeritus, Department of Agricultural, Environmental, and Development Economics, Ohio State University, Columbus. Comments of Cole Ehmke, Mariah Tanner, Allan Lines, and Carl Zulauf are much appreciated.

[1] Public and private sector research, extension and development outlays of $6.0 billion in 1993 (Alston and Pardey, p. 57) updated to 1999, plus education and infrastructure outlays bring total outlays to $11 billion.

[2] The above numbers vary from study to study depending on whether private investment in new technology is considered, the type of model, the study period, the control variables, and a host of other considerations.

[3] Total real return on farm equity capital in the seven years from 1993 to 1999 averaged 4.7 percent but averaged the lowest, 0.4 to 2.2 percent, in 1998, (US Department of Agriculture, September 1999, pp. 24, 45). Still the rate of return averaged a respectable 8.1 percent (excluding capital gain) on farms with sales of over $500,000 in 1998 and accounting for 57 percent of all farm sales. Rates of return were higher in earlier years (Tweeten, 1989, ch 4; 1994). For example, excluding capital gains, the rate of return on equity of farms with sales of $250,000 to $500,000 averaged 3.9 percent and of farms with sales of over $500,000 averaged 9.6 percent in 1997.

[4] The 1998 number is from the Economic Research Service’s ARMS survey. The 1997 Census of Agriculture (p. 10) reports that 50 percent of farm operators in 1997 reported an occupation other than farming

[5] Assumes that farm and nonfarm households had the same number of members in 1999.

Table 1. Income and net worth of US farm operator households and all US households in 1998.
Farm Category	Operator households	Total household income			Total net worth
(Sales)		From off-farm sources	From all sources	% of US average household	Average amount	% of US average household
	Number	*$ per household*	*$ per household*	%	*$ per household*	%
All operator households	2,022,413	52,628	59,153	115.2	492,195	174.2
Small family farms
1. Limited-resource ^a	150,268	13,153	9,924	19.1	78,718	27.9
2. Retirement (retired)	290,938	47,158	45,659	88.1	535,943	189.7
3. Residential/lifestyle (nonfarmer)	834,321	76,390	72,081	139.0	347,909	123.2
Farming occupation:
4. Lower sales (under $100,000)	422,205	37,186	34,773	67.1	576,402	204.0
5. Higher-sales ($100,000-250,000)	171,469	28,717	50,180	96.8	669,458	237.0
6. Large family farms ($250,000-499,000)	91,939	47,252	106,541	205.5	944,533	334.3
7. Very large family farms (Over $500,000)	61,273	33,240	209,105	403.2	1,508,151	533.9
^aHousehold income under $20,000, farm assets under $150,000, and gross sales under $100,000. Source: US Department of Agriculture, September 1999, pp.19, 24; Feb. 2000, p. 39.

Introduction

Have Farmers been Disadvantaged by Technological Change?

Identifying Losers from Technological Change Among Categories of Farms

Total household income