US Food and Fiber—Abundance or Austerity?

The United States

The U.S. economy began strengthening in 1983, but for industrial countries as a whole, growth is and will be modest in the immediate future, with relatively high levels of unemployment. Near-term recovery prospects are tenuous in many of the developing countries because of the combined effects of declining export earnings, severe problems in servicing external debts incurred under inflationary expectations of the 1970s, and stringent deflationary policies that have reduced imports and slowed economic growth.

The three decades of global economic development provided major impetus to the growth of U.S agriculture. Output expanded by nearly 70 percent and the United States became a dominant supplier of coarse grains (mostly corn and sorghum), cereals, and oilseed in world trade and an increasingly important balance wheel of the world food system.

In the fifteen-year period from 1950 to 1965, the nominal value of U.S. agricultural exports grew at about 4 percent per year (figure 1). Throughout much of the 1950s, the value of U.S. agricultural imports exceeded that of exports. But by the middle of the 1960s, foreign economic growth was stimulating a sustained growth of U.S. exports and the balance of agricultural trade became positive, in the range of $1.5 to $1.9 billion annually. Then a probably unique combination of events generated dramatic increases in U.S. exports—a global shortfall in agricultural production, large-scale entry of the USSR in world agricultural markets, rapid economic growth in many countries, declining value of the dollar in international exchange, and expectations of commodity shortages and inflation. In 1973 alone, the value of exports surged some 88 percent. In the middle 1970s, with U.S. farmers planting "fencerow to fencerow," exports ranged between $22 and $23 billion per year, with an annual trade surplus of $10 to $12 billion. In the late 1970s, the value of exports jumped once again to a new plateau, capped in 1981 by $43.3 billion and a surplus balance of trade totaling nearly $27 billion. Between 1970 and 1980, the nominal value of exports grew at an astonishing 19 percent compound annual rate. In real terms, the total increase was close to 100 percent.

All major groups of commodities shared in the expansion of exports. By the late 1970s and early 1980s, exports of U.S. wheat equaled nearly two-thirds of expanded annual production. At the same time, more than one-quarter of coarse grain production, 60 to 70 percent of rice production, and 50 to 60 percent of soybean and cotton production were being exported. The U.S. share of world trade had grown to 40 percent for wheat, 55 to 60 percent for coarse grains, 60 to 66 percent for soybeans, 30 to 40 percent for cotton, and about 20 percent for rice.

Geographically, U.S. exports to all regions increased. Exports to Asia, dominated by Japan, grew at an annual compound rate of 19 percent between 1970 and 1980 and by the end of the decade accounted for more than one-third of all U.S. agricultural exports by value. Even with the import protection policies of the European Economic Community (EEC), exports to Western Europe grew at 17 percent per year and absorbed about 30 percent of total exports by the end of the decade. The largest increases were to Eastern Europe and the USSR, reflecting both slow growth in agricultural output in that region and central government decisions to improve the food supply; to North Africa and the Middle East; to Latin America, mainly Mexico; and to other middle-income developing countries where rates of economic growth were high. And in the late 1970s the People's Republic of China also emerged as a major U.S. market ($2.2 billion in 1980), primarily for cotton and grains. Although both the value and volume of aggregate U.S. exports have come off their 1981 peaks, they remain at levels far above those of preceding decades (see figure 1).

Less well recognized is the fact that the United States now is the second largest importer of agricultural commodities, mainly from the developing countries of Latin America, Asia, and Africa. The nominal value of these imports, two-thirds of which compete directly with domestically produced products, quadrupled to $17 billion between 1950 and 1980—a 28 percent increase in real terms. Although the domestic market still is the primary outlet for U.S. agricultural products, the sector clearly has become increasingly trade oriented. The growth in U.S. trade has been highly and positively correlated with world economic development. For the remainder of the century and beyond, the pace and pattern of global population, economic, and agricultural development will be crucial to U.S. agriculture.

World population

Changes in regional and world population constitute the fundamental and most important force likely to increase the demand for U.S. commodities. Rapid population growth means larger potential markets for exports, but it also may diminish growth in living standards and dampen demand.

With only one exception, every region in the world where 1970s population growth rates exceeded 1 percent is experiencing declining rates today and expects much slower growth by the end of the century (table 1). The exception is Sub-Saharan Africa, where growth rates are expected to increase through the 1980s before beginning to slow in the late 1990s.

Declining growth rates notwithstanding, the global prospect is for swelling economic pressure from population. As noted at the outset, the increase expected over the last two decades of the century is enormous—an average of 87 million persons per year. Furthermore, the distribution of the growing population is critical. Some 93 percent of the growth during the coming two decades will be in the world's six least developed regions, with more than 25 percent in South Asia alone and 20 percent in Sub-Saharan Africa. By 2000, only 19 percent of the world's people will live in the six most developed regions, down from 23 percent in 1980.

Economic growth

Important as are the implications of population growth and its distribution, most of the developing countries have the potential for rapid, sustained economic growth. Many did grow rapidly during the 1970s, a pattern interrupted by the world-wide recession, which has had an inordinate impact on developing countries.

We have assumed that during the next two decades real economic growth will average well below the rates of the 1970s. The recession of the early 1980s and the impact of policies implemented by both developed and developing countries to constrain inflation will result in comparatively slow economic growth through the rest of the 1980s and into the 1990s (table 2). We assume growth rates will pick up during the 1990s and, by the end of the century, possibly exceed the average rate of the 1970s. We project annual petroleum price increases averaging between 1 and 2 percent. In general, we see the most difficult problem facing the developing nations as balancing their needs for economic growth and for more jobs for a rapidly growing labor force against the need to reduce large debts and high inflation rates.

In each region, economic growth and population growth are closely intertwined. We project that, for example, by 1990, Sub-Saharan Africa and the EEC will have similar rates of real economic growth, with the rate in Sub-Saharan Africa slightly higher (table 3). However, the population growth rate there is so high and the EEC rate so low that the per-capita-income growth rates differ greatly. The EEC economy will grow a moderate 2.2 percent per capita per year, while we see economic growth in Sub-Saharan Africa declining by 0.5 percent annually through 1990.

General policy assumptions

"Each is prone to see agriculture according to his lot; only a few see it as a whole, and fewer see it as an integral part of an interdependent economy." —Theodore W. Schultz, Agriculture in an Unstable Economy

Food and agricultural policies

Projected global demand

The effects of economic changes on food and fiber consumption are extremely complex. Generally, food consumption increases with income to a point where consumer needs become satiated. As incomes increase, consumers first attempt to meet their caloric needs, then to increase the variety and quality of their diets, usually by substituting animal protein for carbohydrates—more meat, eggs, and dairy products and fewer grain-based foods. Table 4 shows the general dimensions of the change in world diets we expect to result from projected patterns of economic growth and food availability. For the last two decades of the century, we project average annual increases in per capita meat consumption smaller than during the 1970s, along with a continued—although slower—decline in per capita milk consumption. The combination of substituting animal products for food grains and increasing feed use of cereals almost is offsetting, and results in per capita cereal consumption only slightly above the level of 1978-80. Oilseed consumption likely will continue to increase, but not as rapidly as during the 1970s. The increases in aggregate demand are, nevertheless, very large once population growth is taken into account. We project world demand for meat to increase by 89 million tons, or 64 percent; for milk, by 168 million tons, or 36 percent; for grains, by 722 million tons, or 46 percent; for oilseed, by 95 million tons, or 62 percent (figure 2). The largest percentage increases tend to be in the developing world. The larger demand for livestock products will contribute significantly to the growth of aggregate demand for cereals and oilseed, over and above the increased demand for grains and oils for direct human consumption tied to the population increase. We project global consumption of natural fillers to grow 37 percent by 2000, or about 1.5 percent annually. Consumption of cotton, as a result of somewhat less intense competition from synthetic fibers, is projected to increase 1.85 percent annually to 2000.

Global production

Detailed examination of land and water resources and of the possibilities of raising yields by increased applications of inputs and technology suggests that all regions will meet increased demand largely out of domestic production. While the rates of growth of demand are particularly high in some of the developing regions, these generally also are the regions where production technology has been lagging and the potential for productivity growth is, therefore, relatively large. Realizing this potential will require massive investments in land improvement, irrigation, research, and extension; adequate inputs of fertilizer, pesticides, and improved seeds; improved marketing and storage facilities; and adequate economic incentives.

The projections presented here represent our judgment of what is likely to happen under present or modified policies. They take into account considerations of comparative advantage and consumer welfare that favor increased trade, but also foreign exchange constraints and self-sufficiency and farm income objectives that may work in the opposite direction.

Cereals

Accounting for about three-fourths of the world's crop area and production, cereals clearly are the most important commodity group. The 1.8 percent projected annual rate of increase in production is significantly less than the 2.6 percent annual growth experienced in the 1970s (table 5). Most of the increase in production will come from improved technology rather than additional land, as has been the case for the last several decades; indeed, we expect area expansion to contribute only 15 percent of the production growth, compared to 25 percent in the 1970s. In absolute magnitude, harvested grain area, which increased by 46 million hectares in the 1970s, is projected to increase by another 40 million hectares in 1980-2000. [One hectare (ha) equals 2.47 acres.]

Oilseed

The 1970s brought a sharp increase in oilseed production as nations attempted to improve food quality and the efficiency of animal production. In 1969-71, nearly 113 million ha were used to produce oilseed—about 17 percent of the area devoted to cereals. By the end of the decade, the share increased to 21 percent. Nine-tenths of the increase was in soybeans, mainly in North and Latin America.

The annual rate of growth of world oilseed production projected for 1980-2000 (2.1 percent) is only half that experienced in the 1970s (table 6). For North America, the projections call for 1.8 percent annual growth, only one-third of that in the 1970s; for Latin America, 1.9 percent, down from 11.3 percent in the 1970s.

As with cereals, most of the increase will come from increased yields. This contrasts sharply with the 1970s, when area expansion contributed more than half of the increase in production.

World trade projections

World trade in commodities depends not only on consumption needs not satisfied by domestic production, but also on economic and trade policies, and on the financial and foreign exchange position of each nation.

In the short run, trade policies frequently are determined less by comparative advantage in production than by balance of payment constraints, self-sufficiency considerations, or farm income objectives. Many nations tightly control prices of agricultural products, holding them at target levels by import quotas, variable fees and levies, or directly administered prices.

Although such policies are vitally important to world trade and changes are difficult to anticipate, the prospect by and large is for policy structures similar to those currently in place, except for slight modifications in some regions where forces for change are building. These are primarily the EEC, USSR, China, and Sub-Saharan Africa. We project that the world will depend more on trade by 2000, and assume implicitly that current protectionist pressures will abate as economic growth resumes.

National financial conditions are major considerations in world trade projections. Two diverse impacts arise from the heavy debt burdens and weak financial situations that now exist in many developing nations, primarily in Latin America, but also in Africa and Asia, and in Eastern Europe. The first is that the stronger dollar has made both oil and U.S. commodities more expensive to purchase and thereby tends to reduce many importers' ability to import these products while enhancing U.S. competitors' positions. The second is the pressure such conditions place on developing nations to increase exports, even when world markets are depressed and it is uneconomical for them to do so, thus further increasing competition with U.S. exports.

We expect the current difficult financial conditions for the Third World to continue through the mid-1980s at least (and longer for the most debt-ridden countries), but to improve steadily through the 1990s. The increases in per capita food consumption projected for the end of the century—modest as they are—depend heavily on world trade. Cereal consumption is expected to exceed production in ten of the twelve regions by 2000, and meat consumption in five of twelve.

Projected trade volumes in major commodities derived from the preceding consumption and production projections are presented in figure 3.

• For meat, the aggregate volume of regional net imports is projected to triple, from 2.6 million metric tons in 1978-80 to 7.4 million in 2000.
• Trade in dairy products, as measured by aggregate regional net imports, is projected to increase by 60 percent, from 15.7 million metric tons (milk equivalent) in 1978-80 to 26 million in 2000.
• Trade in cereals is projected to almost double, with aggregate regional net imports rising from 131 million metric tons in 1978-80 to 242 million by the end of the century. This, however, represents a much slower rate of growth than that experienced in the 1970s.
• For oilseed, the aggregate volume of net regional imports is projected to increase by 52 percent, from 51 million metric tons in 1978-80 to 78 million in 2000. This, again, represents a much slower rate of growth than prevailed in the 1970s.
• World trade in natural fibers, as measured by regional net imports, is projected to grow by 25 percent, from 4.8 million metric tons in 1978-80 to 6.0 million in 2000, also a slower rate of growth than in the 1970s.

Some major uncertainties

All projections, including ours, are at the mercy of uncertainties at every step in the analysis, with changes in public policies undoubtedly the most important sources of uncertainty. For example, an EEC policy in which food production, consumption, and trade were geared to world prices could swing the region from being a net exporter of cereal grains to being, as we project, a net importer of those products by 2000. The difference could be as much as 20 to 30 million metric tons. Similarly, a reorganization of Soviet agriculture could have major effects on the global food system. And other uncertainties should be noted, as follows.

Population growth

For many years following World War II, future population growth tended to be underestimated, mainly because forecasters did not fully foresee the effects of progress in medicine and sanitation in reducing mortality in the developing world. More recently, the tendency has been to overestimate growth because of a lag in recognizing the decline in fertility rates in the wake of urbanization and industrialization; more widespread education and family planning programs; and the gradual enhancement of the social status of women. Even small downward revisions of population growth rates can have significant effects on food projections.

Income growth

Assumptions about future economic growth are critical. Accordingly, we estimated changes in the demand for meat, milk, cereals, and oil-seed, and in the corresponding trade levels to 2000 resulting from "high" and "low" growth rates. Even though the alternative growth rates differ, in general, by only one-half of a percentage point from those used in the baseline projections, the resulting world demand estimates for meat for the year 2000 vary from minus 7.6 percent to plus 8.6 percent around the baseline estimates. For the other, less income sensitive commodities, the differences are smaller. The effects on trade are even more significant: the high-income growth variant raises the aggregate regional net meat imports by 42 percent; the low variant reduces it by 29 percent.

Income elasticities

As in the case of population, analysts often have to work with outdated income-consumption relationships and thus are slow to incorporate changes that already have taken place, let alone changes that might occur in the future. Moreover, past income growth may have delayed future effects because it may take years—even a generation—for national diets to adjust to higher levels of affluence.

Income elasticities also are affected by changes in income distribution, which are not considered here. Poor people spend a higher proportion of any additional income on staple foods than their more affluent compatriots. Consequently, the average income elasticity of demand for foodgrains would be significantly higher in South Asia, say, should the income distribution become more equal.

Competitive factors affecting trade patterns

Trade patterns are highly sensitive to changes in competitive conditions. The United States has a substantial comparative advantage in grain and soybean production, but other countries, including Canada, Australia, Argentina, and Brazil, have become major competitors. The United States is not now a significant exporter of livestock products and is not likely to become one, but competitive conditions could become more favorable. A great deal of uncertainty exists about the extent to which importing countries will satisfy the increased demand for livestock products by importing meat and dairy products or by developing their own livestock industries, based on imported feed-stuffs.

Beyond 2000

World population probably will continue to increase through the first two decades of the twenty-first century to perhaps 7.8 billion, despite the projected decline in the growth rate—1.23 percent annually from 2000 to 2020 compared with 1.67 Percent from 1980 to 2000. The significance of a nearly 3.5 billion increase in the next forty years is difficult to assess from 1984's vantage point, but it clearly implies mounting pressure on resources and technologies to meet human needs for food and fiber.

The projections in figure 4 are highly aggregated, but they do indicate the general order of magnitude that might prevail for principal variables by 2020.

Production of cereals required to meet projected 2020 consumption levels would be 90 percent above the 1979-81 world average production; that of oilseed and meat more than double (123 and 140 percent, respectively); and that of cotton is nearly two-thirds above the 1979-81 average. Although an increase in the cultivated land base (106 million ha) would be necessary from the current stock of unused arable land—estimated to range between 1.5 and 2.0 billion ha—most of the increased production probably would be derived from increased crop yields, because it is more economic. The projections suggest increases of 70 to 75 percent in global yields of cereals and oilseed would be needed by 2020 relative to 1979-81. Such an increase would equal an average annual yield increase of 1.3 to 1.4 percent over the forty years. Increases of this magnitude on a sustained basis would be possible only with major new technological advances, expanded irrigation, double cropping, and other types of intensified production. An even greater reliance on world trade would be needed.

Projected demand for U.S. food and fiber

Since the late 1960s, the U.S. policy for grain, oilseed, and cotton has been geared to the world market: as domestic demand leveled off, production responded primarily to trends and fluctuations in export demand (table 7). The sharp rise in export volume in the 1970s was obtained by increasing acreage and yields, without lasting effects on real prices. By comparison, growth in both domestic and export demand for U.S. products is likely to be more moderate in the next two decades, particularly in the 1980s.

Projected total demand for cereals is 31 percent greater in 2000 than the 1979-81 average; that for oilseeds about 42 percent greater; and for meat, milk, and cotton about 20 percent greater than in 1979-81. The implied annual growth rates, in the range of 0.9 to 1.6 percent, all are well below those of the 1970s. Under alternative economic growth rates, cereal demand might range between 379 and 405 million metric tons (a 24 to 35 percent increase relative to 1979-81), and demand for oilseed might range between 41 and 44 percent above 1979-81.

If realized, these projections will reinforce the trend of increased dependence on foreign markets. By 2000, close to 60 percent of the total consumption of U.S. cotton and soybeans and 40 percent of consumption of U.S. cereals will be in foreign markets, while projected per capita domestic consumption by 2000 will be only 3 to 6 percent above 1979-81 levels.

The long-term productive capacity of U.S. agriculture

The future supply of natural resources and other production inputs.

The adequacy of the natural resource base to sustain continuing expansion of agricultural production came under intense scrutiny during the 1970s. And well it might: take the case of the most basic natural resource-the land itself. High rates of growth in exports resulted in the return to production of some 60 million acres of cropland held in reserve under government programs in the late 1960s. With the incentive of high commodity prices in the mid-1970s, expectations of continued expansion of export demand, and the ready availability of low-cost capital, planted cropland expanded to a record high 390 million acres in 1981. The emergence of large stocks of grain in 1981 and 1982 and the withdrawal of nearly 83 million acres from production in 1983 muted—but did not eliminate—public concern about the adequacy of the land base.

"Man's relationship to the natural environment, and nature's influence upon the course and quality of human life, are among the oldest topics of speculation of which we are aware. Myth, folktale, and fable; custom, institutions, and law; philosophy, science, and technology—all, as far back as records extend, attest to an abiding interest in these concerns." —Harold J. Barnett and Chandler Morse, Scarcity and Growth: The Economics of Natural Resource Availability.

At the same time, mounting evidence of soil erosion, water pollution, and overdrafts in water use in some regions, combined with concern about the environmental impacts of agricultural chemicals raised serious questions about the long-run consequences of continued high rates of growth of "high-tech" agriculture. Some statistical evidence of a slow-down in agricultural productivity growth rates added to the worry.

Two major reports in the early 1980s, Global 2000 and the National Agricultural Lands Study, further fueled public concern and debate. Global 2000, incorporating and extending high 1970s rates of growth in export demand for U.S. farm products, foresaw a world with rising real costs of production, higher real prices of food and fiber, and increasingly serious environmental problems caused by soil erosion and agricultural chemicals. The National Agricultural Lands Study, also assuming high rates of growth in export demand, pointed to losses of prime agricultural cropland to urban and industrial uses and an eventual national "cropland crisis." In the late 1970s, government policy decisions to slow public investments in developing water supplies, coupled with threats of further pollution and increasingly intense competition among urban, industrial, and agricultural uses of water in the western United States, added further to what some foresaw as impending shortages and serious degradation of the nation's natural resource base by the end of the twentieth century, if not before.

These are serious fears and charges, and they deserve to be addressed seriously. At the outset, however, we should make clear that we do not accept a basic premise of this line of argument—that the world's resources are akin to a pie of a certain and fixed size that is doomed to be consumed at a more or less predictable rate.

The specter of impending crises in the availability of land and water resources arises in part from preoccupation with their physical limits. But the demand for and use of resource services from that physical stock are determined by human choice and powerfully influenced by social and economic criteria. Scarce resources are socially valuable resources. Economic scarcity of a resource induces an increased supply of technology to substitute for the scarcer, higher-priced resource service. And as a resource becomes scarcer and more socially valuable, users conserve that resource by substituting other resources and by adopting resource-saving technologies and management practices. This principle of substitution is dramatically evident in the performance of U.S. agriculture in recent decades.

Several complex, sometimes countervailing, forces will shape the competition for natural resources into the twenty-first century. In an increasingly interdependent society, competition for natural resources doubtless will heighten. On the margin, the value of water and generally the value of land in nonagricultural uses exceeds its value in agriculture. Thus, where markets are operating unfettered and efficiently, agriculture in many locations in the twenty-first century will be in a weak competitive position for use of those resources, much as it is now. Somewhat related is the likelihood of continued erosion of the political power of agricultural interests at the national level and in many states. By the next century, agricultural policy makers will find it more and more difficult to obtain or even maintain "special-interest" policies for water, other resources, or, for that matter, agricultural commodities themselves.

The growth in nonagricultural demand for resources will be highly uneven, given projected demographic and economic growth patterns to the twenty-first century. Broadly, agriculture close to urban centers, and in southwestern states in general, will experience the greatest—in some areas irresistible—competition from non agricultural forces.

Agricultural land

Economic pressure on the agricultural land base from nonagricultural demand probably will be lower in the next three decades than in the past three. Most important, U.S. population growth rates are diminishing. The dramatic migration from metro to nonmetro areas, prominent in the 1970s, may slow. The rate of household formation is likely to decline, beginning in the 1990s, because of the age composition of the population. Housing starts, retarded in the early 1980s because of high construction costs, high real interest rates, and recession, may increase in the late 1980s, but ultimately can be expected to slow because of declining rates of household formation. Construction rates for new airports, water and highway transport systems, dams, and reservoirs—all significant past claimants on cropland—already have slowed. There are, of course, important regional and local exceptions to such generalizations. In the Sunbelt states, for example, competition for land and water will intensify, posing critical choices among agricultural, urban, industrial, and other enterprises in the use of resources well before and into the twenty-first century.

Seen in the light of the principle of resource substitution and prospects for lesser growth in competition for land for nonagricultural uses, an impending "cropland crisis" seems less likely than was popularly depicted in the 1970s.

The national cropland base is estimated to be about 540 million acres—413 million acres of current cropland and 127 million acres of pasture land, rangeland, forestland, and other land of "high" or "medium" potential for growing crops (figure 5). Although the annual net conversion of 875,000 acres of cropland in 1967-75 was highly dramatized, the cumulative conversions during those nine years constituted only slightly more than one-tenth of 1 percent of the cropland base. But preoccupation with a single, national level statistic can be misleading. Soil characteristics differ, and land in one area may not be a perfect substitute for land in another area in either a physical or economic sense. The cropland base is a valuable national asset that warrants prudent husbandry.

Of the 127 million acres considered to have "high" and "medium" potential for conversion to cropland, about 50 million are in pasture and 70 million in rangeland. This land would be an obvious first choice for conversion, but it clearly would have negative effects on livestock production. Another 30 million acres of the potentially cultivable land are privately held forests. Putting that land under cultivation generally would cost more than converting pasture or rangeland, and it would marginally depress the production of forest products.

Growth in demand for food and fiber is likely to be expressed not only in demand for additional cropland, but also in more intensive uses of cropland already in production. Increased intensity could be achieved in several ways, among them greater use of inputs such as fertilizer, pesticides, labor, and water, shifts from lower to higher value crops, higher plant population per acre, and reductions in crop failure and summer fallow.

Double-cropping is another way to produce more from the same amount of land. According to Robert Boxley, 14 million acres were doubled-cropped in 1981—three and a half times the 1969 level 1. Boxley points to several factors conducive to expanding double-cropping—faster-maturing plant varieties with shorter growing seasons, improved machinery and equipment, and minimum or no-till technology. He suggests that, by 1992, double-cropped acreage might be 22 million acres, with much of the increase occurring in the Delta and southeastern states.

Expanding the land base and using it more intensely means capital investment, improved management, and increased use of other inputs. But it also means several types of social costs accruing over long periods but not incorporated immediately or fully into markets for resources or products—environmental degradation and loss of wildlife habitats, for example. Of the various environmental threats, soil erosion probably is the most significant, through its effects on water quality and potential productivity losses on agricultural cropland. These effects, however, tend to be highly localized. Regardless, the potential to exacerbate such environmental and productivity costs increases as the cropland base expands. The supply of cropland probably will not physically or economically limit U.S. food production by the beginning of the twenty-first century, but actions to preserve it and planning to regulate its use in environmentally acceptable ways at the local level are neither irrelevant nor unnecessary. Indeed, these issues and choices are likely to grow ever larger as the next century approaches.

Water resources

Most people are unaware of agriculture's tremendous thirst. In fact, however, agriculture accounts for 80 to 85 percent of the total amount of water consumed each year in the United States. Some other basic points:

• Irrigated land in farms doubled between 1950 and 1978 to a total of nearly 51 million acres.
• Some 93 million acre-feet of water—that is, 22 inches of water per irrigated acre—were used on this land, about 54 percent of it from surface watercourse and the rest from groundwater aquifers.
• The western United States, where more than one-half of the value of crops derives from irrigation, contains 86 percent of the nation's irrigated acreage.
• Irrigated agriculture accounts for roughly one-quarter of the nation's crops and nearly one-seventh of the nation's cropland.

Initially, when the West was settled, water was heavily subsidized by governments and treated almost as a free good by western agriculturalists. The original users not only were allowed to use water without charge, but also were granted water rights as long as what they used was put to "beneficial" use. Also, until recently, low energy prices coupled with federal or state subsidies have helped to keep the costs of moving water low, and this has been important to developing groundwater supplies.

According to Kenneth Frederick, however, "The days of cheap water are ending in the West. In many areas, irrigators now depend on essentially nonrenewable water supplies. Current irrigation levels with average precipitation result in the mining of over 22 million acre-feet of water from western aquifers."

In the United States as a whole, the average rate of groundwater withdrawals has risen 3.8 percent per year since 1950, almost twice the rate of increase in the use of surface water; much of the increase was in the water-short areas of the West. Nationally, nearly one-quarter of the groundwater withdrawn is not replenished: the water table is declining an average of up to 6.6 feet per year under 15 million acres of land irrigated by groundwater. From the Rio Grande to Nebraska, in Arizona and California, falling groundwater levels and higher energy costs are boosting substantially the costs of groundwater irrigation. Nor is surface water likely to pick up the slack, since the demands on the rivers and streams of the nation's principal irrigated areas already commonly exceed available supplies. The availability and price of water and energy, rather than land, may be the critical natural resource variables for agriculture in the West in the years ahead. Public policy for water in the West is moving from developing additional supply to managing the increasingly more valuable current supply.

What seems likely to ensue over the next several decades is a series of marginal agricultural adjustments to higher priced water—more efficient water application, reduced rates of application, and shifts from lower to higher valued crops. The potential to conserve water in these ways is high. However, the investment capital needed to achieve more efficient irrigation systems is substantial, running into several billions of dollars.

The physical requirements for water to meet projected urban and other nonagricultural uses in the West to the year 2000 are small relative to the total quantities now used in agriculture. Nevertheless, the water issue will force many difficult, controversial choices. One of the major challenges is to develop institutions to reduce distortions arising from policies predicated on an abundant, low-priced natural resource.

Beyond water's physical and economic dimensions lie major issues of water quality. Groundwater contamination from agricultural (as well as nonagricultural) sources is serious in many parts of the country, and western irrigation practices have increased groundwater salinity. As Frederick notes, "Perhaps one-quarter of the lands currently under irrigation in the West are heavily dependent on nonrenewable water supplies, and the productivity of several million additional acres is threatened by rising salt levels." And pesticides and fertilizers may seep into groundwater.

Other water quality problems—dissolved oxygen, suspended solids carrying bacteria, nutrients, pesticides, excessive phosphorus and nitrogen nutrients—can be laid at agriculture's feet in part, and occasionally in major part. The 1972 Federal Water Pollution Control Act set forth ambitious goals to improve water quality. Progress has been made—scattered evidence suggests that some of the worst pollution problems may be abating—but much remains to be accomplished. Maintaining or improving water and environmental quality may require modifying agricultural production practices and ultimately may mean higher direct costs of agricultural production.

Other production factors

Capital

As noted, the growth of U.S. agricultural productivity and total output in recent decades springs from the substitution of capital-embodied technologies for land and labor. Several factors were at work.

• Public and private investments in research and development made available a vast array of new or improved technologies, including labor-saving machinery and equipment, improved stocks of seeds, more effective inorganic fertilizers, a plethora of pesticides manufactured for specific control purposes, and scientifically blended, nutritionally balanced animal feeds.
• Institutions such as the agricultural extension services and extensive public and private communication and information systems helped to transfer newly developed technical information from scientist to farmer and assisted its adaptation to local use. Education and improved management skills and techniques speeded the transfer and adoption process.
• Fertilizer use increased from an average of 22 million tons in 1951-55 to more than 50 million tons currently, and pesticide application jumped from 194 million pounds in 1964 to nearly 480 million pounds in 1982.
• Direct energy use in agricultural production now approximates 2,000 trillion Btus annually.

Capital requirements in agriculture for investment and operating purposes will grow substantially in the decades ahead and competition for capital for domestic and international development and for the financing of public debt may maintain real rates at levels well above those of recent decades. Nevertheless, capital availability and costs are not likely to seriously impair growth in U.S. agriculture in the longer-run context of 2000 and beyond. Corporate business structures will become increasingly prevalent among the larger, more highly capitalized and biggest producing farms. And it seems plausible that a growing number of these corporations will be securing investment capital directly from equity markets by the end of the century.

Economic incentives

In addition to technology and institutions to effect its transfer, powerful economic incentives reinforced the "technological revolution." Low real prices of petroleum and other energy supplies induced the substitution of mechanical power and fertilizer for higher-priced labor and land. An expanding spectrum of more efficient pesticides offered opportunities for higher crop yields and lower per unit production costs. Between 1950 and 1978, the average value of land increased 650 percent and farm wage rates some 380 percent. But the price of gasoline increased only 150 percent and that of fertilizer about 84 percent—powerful incentives for resource substitution. And the competitive economic structure of agriculture provided an economic imperative for farmers to make widespread use of the technologies. The results are apparent in table 9. Total inputs in 1980-83 were all but identical to the average of 1950-59, but output averaged nearly 70 percent more. Total factor productivity—the ratio of total output to total input—was nearly two-thirds higher in 1980-83 than in 1950-59.

Labor

These technical changes greatly enhanced resource use efficiency in farming, expanded production and kept food and fiber prices low for domestic and foreign consumers, and freed resources from agriculture for producing goods and services elsewhere in the economy. Massive structural adjustments ensued. The number of farms fell from 5.6 million in 1950 to an estimated 2.4 million in 1982. Production became even more concentrated, with 28 percent of the farms now accounting for 88 percent of product sales. Farm employment fell from 7.2 million in 1950 to 3.4 million in 1982. In the two decades between 1940 and 1960 a net of 21.5 million persons moved to the cities or to urban nonfarm residences—an average of more than 1 million persons every year.

The plunge in farm employment came as farm and nonfarm wage rates approached equilibrium and labor-saving production technologies came on-stream. With labor now representing only 12 percent of all inputs used in farming and farm employment constituting only 3 percent of total U.S. employment, future absolute declines in employment and labor inputs obviously will slow compared to historic rates. However, assuming national economic growth and continuing advances in labor-saving technologies, much of the unskilled labor now in farming may well disappear by 2000. On the other hand, demand will grow for labor with highly technical skills to manage increasingly complex technology and production systems. Overall, however, farming will be a shrinking part of the national labor market, and the same is likely to be true in allied input and product marketing systems in automation and other labor-saving technologies advance as expected.

Input prices

With respect to future prices of off-farm production inputs, that of energy is perhaps most uncertain. Although agriculture adjusted fairly readily to the sharp increases in energy (and energy feedstock) prices in the 1970s, it did so under conditions of generally rising farm product prices. The real price of fossil fuel energy is not expected to rise much for the remainder of this decade and possibly to 2000, but the events of the last decade are fresh enough to remind us that expectations are not necessarily realized.

Energy uncertainty highlights the importance of U.S. public policies that will encourage energy conservation, development of indigenous energy supplies—including nonpetroleum sources—and development of energy-conserving technologies and production systems. Energy is an important component of inorganic fertilizers, but fertilizer prices have increased less rapidly than the price of energy in recent years. Assuming only gradual increases in real energy prices, fertilizer prices seem likely to increase at rates close to general price levels in the decade or two ahead. Further, some project much slower rates of growth in fertilizer use, in part because of higher fertilizer—farm price ratios and in part because of possible environmental regulations.

As for other types of major manufactured inputs, such as machinery, equipment, feed, and pesticides, long-term supplies seem to pose no major constraint to 2000. Their prices, too, are expected to follow general price trends in the economy.

Governmental regulation

Environmental policies to restrict the use of fertilizer and pesticides may become important enough to significantly change the mix of production inputs by 2000.

About 1,000 new chemicals are introduced each year in the United States, with some 55,000 to 60,000 marketed primarily from domestic manufacturers. Only a small proportion of the total is used directly in agriculture, but it is a proportion that in absolute numbers has been growing exponentially in the last twenty years. Comparatively little is known about the potential toxicity of many of these chemicals, how they are used, whether and how they enter the food chain and other ecosystems, and their ultimate effects on human health and on other species. Controls on pesticides in agriculture have become more stringent, and progress has been made in less toxic but effective pesticides and in integrated pest-management systems that reduce application rates. Nonetheless, pesticides remain a solid part of the production of major field crops.

The issues surrounding agriculture and the quality of the natural environment are neither transitory or ephemeral. Nor are solutions simple or absolute. It is impossible to reduce the environmental risks of high-tech agriculture to zero: tradeoffs between food production, food quality, and quality of the environment are inevitable. By the twenty-first century, the choices will be more complex, more difficult, and more important to both agriculture and the remainder of society.

Projected production and resource use, 2000 and 2020

Production increases to meet projected demand for U.S. food and fiber might be met through various combinations of natural resources, other production inputs, and production technologies. Three such possibilities are presented in figure 6.

No breakthroughs

In general, the scenarios in figure 6 (column A) might be regarded as a "static technology" option. With no increase in per acre yields of major crops, an additional 95 million harvested acres would be needed to meet projected requirements in 2000; more than 200 million harvested acres would be required by 2020. Clearly, demand for cropland of this magnitude would place enormous pressure on the cropland base. The result would not only be dramatically higher land prices, but also higher costs of crop and livestock production, higher food and fiber prices to consumers, and economic pressure on the forestry sector, with possible higher prices for forest products. Soil erosion and water pollution surely would worsen as production expanded to more and more ecologically fragile cropland.

However, such a land-using course of development represents an unlikely outer boundary. Increased demand for land would stimulate substitution of other inputs and management practices. And if prices of farm products rose as a result of higher production costs, the projected increase in consumption of food and fiber would slow. Thus, even without new technologies to enhance crop yields, the "equilibrium" demand for land would fall far short of the forgoing projected land requirements, albeit at levels well above the present.

Optimism

A much different scenario would ensue from the "most probable' yield increases projected by scientists participating in the 1982 RCA Symposium (see figure 6, column C). Were those yields attained—generally on the order of 1.5 to 2 percent per year gains relative to the 1979-81 average—harvested cropland requirements would decrease by as much as 55 million acres by 2000 and, by 2020, still would be nearly 30 million acres below the 1979-81 average.

The RCA Symposium projections of yields are based on technology that already exists but has not been adopted—the best efforts on today's test plots could be the farmer's average yields in the year 2000. Similar optimism prevails for livestock productivity: probable production gains per breeding animal of 25 percent for beef and pork and 30 percent gain in milk output per cow are projected for the turn of the century. Productivity gains projected to 2020 are based on recently conceived technologies that could be made practical through further research and development during the next forty years. Implicit in these high-tech projections are assumptions of favorable economic circumstances to encourage farmer adoption of technology, improved managerial capabilities, regional and interregional shifts in resource use, increases in double-cropping, and increased or more effective use of inputs such as fertilizers, pesticides, and livestock feed additives. But major gains also would come from such emerging technologies as nitrogen fixation, insect-and disease-resistant and drought-resist-ant crop varieties, and a variety of genetic-based livestock technologies.

A middle course

An intermediate scenario is presented in column B in figure 6. The yield projections are based on long-term trends adjusted to conform with what we believe to be plausible growth rates considering available technologies, probable product-factor price relationships, and the expansion of crop production on lands likely to be less productive than current cropland. Under these conditions, an increase of about 25 million acres of harvested cropland would be needed by 2000-17 million acres to sustain increased production of cereals, 13 million acres for oilseed, and slightly less than 1 million acres for cotton. Reductions of nearly 6 million acres in harvested cropland for barley, oats, and other cereals would partially offset increases in cereal and oilseed acreage. By 2020 nearly 50 million additional harvested acres relative to 1979-81 would be required.

Several potentially important implications can be drawn from these projections. Under the high-tech projections (see figure 6, scenario C), less pressure would be put on the cropland base than has existed in recent years. Indeed, the possibility of major crop surpluses would have to be monitored carefully. However, with intensive use of land in production, increased use of fertilizers and pesticides, and a relative shift in land use to soybeans and corn, problems of soil erosion, water quality, and environmental pollution could be exacerbated. Under the land-using scenario (figure 6, column A), the tendency would be to plant "fencerow to fencerow," using relatively less of the high-tech inputs. Soil erosion, water quality, and environmental degradation problems also are part of this picture, but they come from a different source—the expansion of production on land that is ecologically more fragile. Thus, in either case economic and ecologically plausible resource conservation technologies, management practices, and public policies are likely to play critical roles.

In the intermediate projection (figure 6, column B) problems of several types might emerge, with erosion heading the list. Again, this is because the projected net expansion of 25 million acres of harvested cropland would require substantial additional capital investment and in some areas of the Midwest, Delta, and the Southeast might push production onto erosive, fragile land. Also, the conversion of land now in pasture could lead to economic adjustments in the livestock-dairy sector in several regions.

Thus, while attaining projected output levels seems reasonable with the economic supply of natural resources, technologies, and nonfarm inputs we expect for the next two decades, it will not be achieved without substantial adjustments throughout the sector, including adjustments of human capital. Nor will such adjustments be distributed evenly among regions, farms, and farm families.

Science and education

Public and private investments in science and education have been mainsprings in the American economy and in improved standards of living, and in no sector is this more evident than in agriculture. Annual internal rates of return to public investment in agricultural research range from 30 to 35 percent, sometimes higher. The benefits are many. Abundant, relatively low-cost food and fiber are available for domestic and foreign consumers. Human resources have been released to produce goods and services elsewhere in the economy. Export markets have been developed and foreign exchange earned for consumption of imported goods and services.

But along with the benefits have come social costs and unanticipated side effects associated with technical change—human and institutional adjustments not readily reduced to dollars, and environmental externalities and risks to human health. And the net benefits of science-and education- based technical change have been unevenly distributed within the population.

Some analysts note evidence of an apparent slowing in long-term total factor productivity growth rates since the mid-1960s as gains from the technologies of an earlier era wane with widespread adoption. Some are concerned that the near-static real total expenditures for agricultural research in the U.S. Department of Agriculture and the state agricultural experiment stations for almost twenty years may further slow productivity gains in the decade ahead. Others allege declining emphasis and reduced real funding levels for basic research in recent years that will magnify the possibility of a future slowdown in productivity growth unless such trends are reversed.

Our projections suggest that long-term planning for agriculture should be predicated on a 70 to 100 percent increase in effective global demand by 2020 for U.S. food and fiber at real prices close to those of recent years. Moreover, we do not allow for the great uncertainty inherent in developing global agricultural systems, or for food assistance to persons in the United States and abroad who are unable to secure adequate nutrition in the marketplace. In short, we believe that meeting the projected demands for 2020 without additional productivity-enhancing technologies and improved production systems would place inordinate pressure on the nation's natural resource base, induce serious environmental consequences, and substantially increase real prices to consumers of food and fiber.

Given these uncertainties and possible undesirable outcomes, it seems prudent public policy to base research, development, and education policies for the long-term future on continued advances in productivity at rates approximating those of recent decades. This means planning: major lags remain between initiation of research, subsequent development, and application. Today's strategies for investment in science, research, and education must be based not on circumstances of the moment but on perceptions of the needs in a distant and uncertain future.

The planning of future agricultural research, however, must be based on more than technology per se or a simple multiplication of product output, for innovation and productivity gains come not only from discoveries in the physical and biological sciences. Emphasis should be given to the development of socially appropriate technologies that take into account not only long-term demand for food and fiber but also national goals concerning environmental equality, natural resource conservation, and human health and nutrition. Equally important is innovation enhancing the effectiveness of institutions that govern the use of technology, human development, capital accumulation, and social science research that improves understanding of human and institutional behavior.

What kinds of things do we have in mind when we call for more emphasis on research and development? Johnson and Wittwer identify the following as among the most promising.

• Mechanization and automation, particularly that employing improved electrical sensors and controllers to monitor biological and environmental stress on plant cultural and harvest operations; improved machine designs for, conservation tillage, and reduced fuel and labor consumption
• Thematic mapping and multispectral imagery, including remote sensing techniques, for monitoring resource use, crop conditions, water and energy availability, and weather forecasting
• Improved, more economical, livestock feeding using nonprotein nitrogen combined with forages, byproducts, and waste
• Improved animal health through better management systems, vaccines and other products of microbial synthesis, hormonal regulation, and genetic selections and engineering
• Biological nitrogen fixation as an alternative to chemical fixation to increase crop productivity; more effective use of fertilizers through better understanding of nitrification and denitrification processes
• Genetic improvements and genetic engineering to achieve more dependable and higher yields, disease resistance, greater uniformity, climatic and poor-soil adaptability, and improved nutritional value. Increased preservation of genetic diversity for future research
• Basic research in plant growth regulators to increase yields and quality, hasten maturity, extend storage life, and aid mechanical harvesting
• New approaches to pest control through basic research in plant protection and integrated pest management to reduce the number of pests resistant to chemical pesticides; to establish resistance in economically important crops; and to insure health and safety of agricultural workers, nearby communities, and consumers
• Improved reproductive efficiency of livestock through basic research on genetic improvements, semen preservation , pregnancy detection, multiple births, super ovulation, and nonsurgical embryo transfer and implantation
• Food science research to enhance food safety, product storage, handling, marketing, and nutrition, and to develop fabricated foods
• Research to enhance photosynthesis and reduce photorespiration, including identification and possible control of the mechanisms that regulate respiration; identification of growth regulators; and improvements in plant architecture, anatomy, cropping systems, planting designs, and cultural practices
• Assessment of biological effects of increased atmospheric carbon dioxide, ozone, and other air pollutants and trace elements on productivity
• Basic social science research regarding risk bearing; conservation and investments in natural resources; organization and control of institutions and Institutional decision making and administration; and measurement of human values.

Johnson and Wittwer estimate that a 45 percent increase in agricultural output incorporating about a 30 percent increase in crop yields would be attainable by 2020 by maintaining public funding of agricultural research at the real level of 1983 and as a constant proportion of the value of agricultural output throughout the forty years. To obtain a 75 to 80 percent increase in output and a 55 to 60 percent increase in crop yields in 2020, they estimate that public funding of research would need to be increased at a compound annual rate of 10 percent between 1983 and 1994 and as a constant proportion of the value of agricultural output over the forty years.

Summing up