changed. It is also interesting that the model accom-
plishes this while the real price of farm products falls
substantially.
 In the NIRAP model, U.S. agriculture does not inter-
 act with either the U.S. non-farm economy or with other
 economies in the world. However, it provides by 2030
 for crop and livestock exports more than 50 percent
 above current levels.
 Further insights into the next half-century are
 derivable from the projections produced by the relatively
 new USDA/IIASA/MSU model, hereafter referred to as
 the USDA world model.' In that model, U.S. agriculture
 does interact with the U.S. non-farm economy and with
 other countries via an international trade linkage com-
 ponent. This permits the model to project exports and
 non-farm uses of farm products rather than estimating
 the consequences of assumed exports and non-farm uses.
 The model also provides more flexibility than NIRAP
 in projecting the development and use of currently less
 productive, more fragile soils; and the intensity of crop-
 ping mixes. This flexibility permits the model to tell us
 more than NIRAP does about technological advances
 needed to permit the cropping of soils not now cropped
 and to produce more crops per year and more intensive
 mixtures of crops.
 When the USDA world model is operated using
 scenarios similar to those used in the NIRAP projections
 discussed above, the results with respect to land use are
 essentially the same for 2030. Both models show similar
 increases in yields per acre and in land utilization. In-
 creased intensity of cropping mixes is explicit in the
 USDA world model and implied by the NIRAP model.
 However, in the USDA world model, U.S. grain exports
 run into competition in both domestic and world
 markets. Together, exports and domestic use increase
 rapidly in both models. Though exports are assumed in
 the NIRAP model, the U.S. non-farm sector competes
 in the USDA world model, probably unrealistically,
 with other countries for the use of coarse grains as in-
 dustrial feedstocks for producing gasohol and organic in-
 dustrial products. Domestic livestock producers also
 compete with foreign grain and livestock producers to
 expand production and exports. This increases U.S.
 exports of livestock products in the world model and cur-
 tails U.S. grain exports to such an extent that coarse
 grains are no longer exported in 2030. Though it is
 doubtful that U.S. trading partners would permit such
 an expansion of livestock production and exports, the
 USDA world and NIRAP models agree on the need for
 capacity to produce much more to meet substantial in-


 These helpful projections were made by Douglas Maxwell
of the Economic Research Service (ERS) with the approval
of John Lee, ERS administrator.


creases in demand. Clearly a doubling of our capacity
to produce agricultural output in the next 50 years is a
reasonable target, even if we do not actually produce
such levels of output.
 The world model was also used to project the conse-
quences of a high level of world prosperity to see how
exports and pressure on U.S. agricultural resources
would be affected. In another attempt to see the conse-
quences of stressing U.S. agriculture, the model was run
using a scenario that reduced the yield-increasing effects
of technological and other research by half. The results
of these two analyses indicate that higher prosperity in
the world would not expand the use of land in the United
States in 2025 beyond the baseline projection of 436
million acres to meet export demands. The model's 2025
constraint of 450 million acres is not encountered. On
the other hand, halving anticipated yield increases
reduces acreage used because higher U.S. and world
prices decrease U.S. exports and domestic use. Produc-
tion abroad expands in response to the higher prices. (We
should note that the halving of expected yield increases
still generates yields in 2030 that are 32 to 40 percent
above present levels.)
 To project events, accomplishments and progress in
agricultural research and technology beyond a decade
is fraught with many uncertainties. More important still
is that the variables that determined the results are
handled in scenarios because they are difficult to model
and predict. In the above scenarios, levels of world pros-
perity and yield increases were treated as unpredictable.
Though yield increases may not be highly predictable,
they can be influenced by such controllable factors as
research appropriations; educational programs for en-
trepreneurs, scientists and educators; and institutions
and policies that make yield-increasing technology
profitable or unprofitable.
 It is our conviction that the United States can manage
the development of its agricultural technology, institu-
tions, human and natural resources, and capital
accumulations to make .it possible to attain the yields
used in the NIRAP I and world model scenario. Some
per acre yields for this scenario follow:


Commodity

Wheat, bu.
Soybeans, bu.
Grain Sorghum, bu.
Barley, bu.
Oats, bu.
Corn, bu.
Rough Rice, cwt.
Seed Cotton, lbs.


 NIRAP I
1980 2030


32.8
30.0
70.4
49.4
55.7
106.9
48.8
485.0


59.8
45.0
112.5
78.9
77.1
178.2
74.6
692.0


World Model
1980 2030


33.5
30.6
57.5
47.3
53.8
100.0
47.6
480.0


63.5
47.0
97.5
80.3
78.8
200.0
60.0
620.0


Neither the NIRAP nor the world model projections