1 Based on values as reported by country of origin. Therefore, excludes transportation and insurance costs, which have been added in table 2.

* Includes value of grain transhipped through Canada.

Sources: For the countries listed, data is from official country trade statistics. Values for countries listed under "other" includes both official trade statistics and estimates based on known quantities and unit values calculated from countries reporting both quantity and value of exports to the PRC.

THE EVOLUTION OF POLICY AND CAPABILITIES IN CHINA'S AGRICULTURAL TECHNOLOGY

BY THOMAS B. WIENS*

*Senior Research Analyst at Mathtech, Inc., Bethesda, Maryland. The author is indebted to Prof. Motonosuke Amano and Shigeru Ishikawa for providing invaluable access to materials in their possession, and to the Social Science Research Council, Joint Committee on Contemporary China, for partial support of this research.

The most important and often asked question about Chinese agriculture is whether it can generate a sufficiently rapid increase in total product to sustain at constant or improving consumption levels a population which will continue to increase for many years to come. At present it is only possible to answer with expressions of optimism or pessimism founded largely on faith or on the assumption that past trends will continue. A more convincing answer would relate past trends and future potential increases in "inputs"-land, labor, various forms of capital, and weather conditions-to realized or expected changes in "outputs"-production of agricultural products. But to begin to define the relationship between inputs and outputs, whether in qualitative or quantitative terms, one must have a prior understanding of the "technology" which defines the structure of the relationship, of the process generating changes in this technology, and of the effects these changes have had and will have on the stability of the input-output relationship.

Economists, including those studying China, have tended to confine their conception of technology to a "residual effect"-the "source" of changes in output which cannot be accounted for by changes in inputs within a static input-output relationship (or "production function")." This approach is of no value in predicting the future, when ceteris paribus is rarely an appropriate assumption, and can be terribly misleading in "explaining" the past. Econometric studies of production functions are unreliable unless technological change is explicitly accounted for, because changes in conventional input use are so

1 For an example of this approach, see Anthony M. Tang, "Policy and Performance in Agriculture,' in A. Eckstein et al., eds, Economic Trends in Communist China (Chicago: Aldine, 1968).

highly correlated with changes in technology that omission of the latter leads to biased estimates of the effects of inputs. For example, growing fertilizer use appears to have been closely correlated with historical increases in rice yields in Japan, until the rate of diffusion of improved seeds is introduced into the equation. Then, as Hayami and Yamada have shown, the estimated marginal effect of increased fertilizer use (ceteris paribus) becomes very small.2

The spread of improved seeds, however, is only one of the more quantifiable aspects of a complex process. This paper will avoid quantification, but will build a base on which quantification can proceed: To begin to predict future rates of technological advance and lag times before full diffusion, we need to know how technological change is generated and implemented in China; to begin to predict changes in the structural relationships among inputs and outputs, we need to understand the underlying strategy and objectives of Chinese agrotechnical policy and the extent of their harmony with the perceived self-interest of peasants; to begin to predict the quantitative importance of technological change to production levels, we need to assess the forces determining probabilities of success or failure and potential and realized economic impact of individual innovations. These topics are explored in this paper through a set of illustrative case studies of interrelated aspects of technology as they have developed over the past two decades and can be expected to unfold in the next few years. The aspects are loosely classed as improved seeds, fertilizer use, irrigation and water control, mechanization, and improved techniques. The case studies are not intended to cover all important developments over the period.

We begin with some working hypotheses, which may be appraised in the light of the case studies. First, to provide a crude paradigm of the process of technological change in China: it involves an interaction among scientists and technicians, politicians (i.e., higher level cadres), and peasants (including lower level cadres). Scientists and technicians are under pressure to make breakthroughs on problems deemed significant to economic production but a research achievement is in the first instance measured on scientific or technical, rather than economic, criteria. It undergoes peer review, which may pit younger researchers, untainted by preliberation or Western training, against an older (and sometimes wiser) group of colleagues, but this screening process is not as rigorous in China as elsewhere since the originator(s) may counterattack with the accusation that their disparagers have an insufficiently Marxist or Maoist spirit (i.e., are negativist), against which accusation there is no effective defense.

When a breakthrough is documented by evidence, albeit on narrow criteria of measurement, it may attract the attention of local politicians, in whose interest it is to obtain recognition by translating the breakthrough into increases in key production measures. This may be a low-risk proposition for the politician, who can blame misleading technical data if failure results. With political support, the innovation is subjected to large-scale experimental testing, but the tests may be hasty, unscientific, carelessly monitored, or again evaluated on exces

Y. Hayami and S. Yamada, "Agricultural Productivity at the Beginning of Industrialization," in K. Ohkawa et al., eds., Agriculture and Economic Growth: Japan's Experience (Princeton: Princeton University Press, 1970).

sively narrow criteria. If it passes these tests, the politicians are in a position to apply pressure for rapid and large-scale implementation in production, where for the first time any economic drawbacks may become fully apparent.

It is the peasants and grassroots cadres who are likely to notice and articulate any major economic weaknesses; if they are severe, and depending on the degree of voluntarism currently acceptable, they may abandon the innovation or resist its implementation. A high volume of promotional propaganda is frequently an indicator of such resistance. Peasants may not immediately perceive a net loss of social product, however, if the adoption process is subsidized in some way by the state, as must often be the case. Nor, for that matter, do the promoters ever publicly net out the cost of such subsidies in their published assessments of economic impact.

Failure eventually results in a period of reassessment, in which professionals who were previously reluctant to speak emerge to criticize the innovation, although less openly if the prestige of the political promoter(s) has been laid on the line. Rarely does such a discussion end in total discrediting of the innovation-usually the critique focuses on the limits or preconditions for its applicability, in part as a face saving device, and attention turns to means of removing these limits or fulfilling the preconditions. After a period of retrenchment, a modified or qualified version of the same innovation may be promoted once again with greater caution.

This paradigm suggests that the process of technological change in China, in contrast to other countries, embodies a less conservative or skeptical view of innovational potential, a higher risk of economic failure of implemented innovations, but also a very short lag between technical or scientific breakthroughs and large-scale diffusion. Elsewhere the process is slowed and made less risky by the need to demonstrate overwhelming economic superiority before a conservative peasantry can be persuaded to adopt.

Secondly, the underlying strategy and objectives of Chinese agrotechnical policy have been well-summarized by Ishikawa as: 3

(1) The utilization, as much as possible, of the local resources for agricultural inputs (designated as "traditional" inputs in contrast to "modern inputs");

(2) Economizing, as much as possible, of modern input (here defined as the inputs produced with the use of nonfarm resources); and

(3) Research and development of scientific methods for productivity increase on the basis of traditional inputs or the combination of traditional and modern inputs.

Another way of putting it is that the state has behaved as if it sought to maximize annual output (not output of each crop), subject to a virtually fixed constraint on cultivated acreage, tight constraints on the availability of liquid capital within the rural sector for purchase of outside inputs, but with no constraint on the supply of labor except perhaps in busy seasons. The obvious prescription for new technologies which increase the intensity of labor use (over the year as a whole) has been followed since the 1956 formulation and 1960 ratification of the national agricultural development program, the

3 Shigeru Ishikawa, "Agrarian Reform and Its Productivity Effect-Implication of the Chinese Patterns," in The Structure and Development in Asian Economies (Hitotsubashi University Institute of Economic Research, Paper No. 10, September 1968).