—- aqua-nu. - «1‘ ‘7"‘""‘~""""‘§l _ THE IMPACT OF DWARF WHEATS 0N RESOURCE PRODUCTIVITY LN WEST PAKISTAN'S PUNJAB Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY JERRY BRUCE ECKERT 1970 THE‘S’S This is to certify that the thesis entitled The Impact of Dwarf Wheata on Resource Productivity in was: Pakistan's Punjab presented by Jerry Bruce Eckert has been accepted towards fulfillment of the requirements for Ph.D. Agricultural Economics degree in kgifiisflfl :5 414%!qu Major professor Date June 12, 1970 0-169 LIBRARY Michigan State University A 0'00. 2007 JUN 2 8 2.007 ABSTRACT THE IMPACT OF DWARF WHEATS ON RESOURCE PRODUCTIVITY IN WEST PAKISTAN'S PUNJAB BY Jerry Bruce Eckert Dwarf wheat varieties from Mexico were introduced into West Pakistan on a commercial scale in the winter of 1965-66. Since that time the varieties have Spread over approximately 70 per cent of West Pakistan's irrigated wheat land. Dwarf wheats are highly responsive to fertilizer, resist lodging and produce yields that are several-fold larger than those of indigenous wheats. Rapid change in several aSpects of wheat production followed the adoption of the Mexican imports. The rapidity of change placed new and severe stresses on agricultural price policy. Policy makers, however, lacked a firm understanding of the changes being wrought within the production environment of individual farmers and were unable to make the necessary policy adjustments with precision. This thesis attempts to quantify the differences in mar- ginal productivity of farm resources when used to grow dwarf wheats as opposed to native wheats.: Detailed sampling of Jerry Bruce Eckert Punjabi wheat farmers during the 1968—69 season provided the raw data; Cobb-Douglas type production functions were esti- ”—-——-— “at“ /-—-‘-‘ Mm-. .i - ..-. , “mg”. mated using identical constructions for both wheat varieties. The resulting production elasticities are tested for signifi- cant differences and where possible applied to several current policy questions. Independent variables used included: bullock labor in land preparation, seeding rate, number of irrigations, pounds of nitrogen, pounds of phOSphate and pounds of farmyard manure. In addition, zero-one variables were used to identify the effect on yields of different dates and methods of sowing and the possible bias resulting when farmers responded to someone they thought might use their answers against them. The sample permitted useful conclusions only in the cases of water and nitrogen inputs.‘ Dwarf wheats appeared to be highly responsive to water inputs while no response could be measured for native wheats at an acceptable level of signifi- cance. The implications of this differential are traced for adoption patterns, pricing of water supplies and on incentives to overinvest in tubewelI installation. Dwarf wheats were found to be approximately twice as responsive to nitrogen inputs as were native varieties?’ The implications of this conclusion were developed for fertilizer pricing and subsidizing and the wheat support price level. A suggested framework is presented to adjust these two price variables to attain a predetermined quantity of wheat produc- tion.£ THE IMPACT OF DWARF WHEATS ON RESOURCE PRODUCTIVITY IN WEST PAKISTAN'S PUNJAB BY Jerry Bruce Eckert A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1970 ACKNOWLEDGMENTS Any well done dissertation contains much of the sub— stance of its author. And that substance is the developmental result of an infinite variety of experiences. Consequently acknowledgment sections must be either woefully inadequate for their task or tediously long. While realizing that many de- serving people will be overlooked, this section is of the shorter variety. The series of teachers, who were at the same time friends and associates, to whom I owe a great intellectual debt, begins with Dr. Morris Kelso of the University of Arizona and continues with Dr. Jimmye Hillman of that institution; Dr. Raymond Seltzer, now with Agri-Research, Inc.; Dr. Bruce Johnston and Dr. Roger Gray of Stanford University and Prof. Dr. Heinrich Niehaus of the University of Bonn. At Michigan State University, Drs. Dale Hathaway, Glenn Johnson, Lester Manderscheid and Lawrence Witt must be cited. The inspiration and personal attention of each of these men is deeply appre- ciated. Each had a substantial impact in the moulding of Jerry Eckert. Particular thanks are due to Dr. Witt who first guided me in a difficult decision in 1962 and has con- tinued to do so, often beyond the normal call of duty, to this day. The opportunity to come to Pakistan and all that re— sulted from it I owe to him. ii Michigan State University and its Agricultural Experiment Station enabled me to pursue my studies while employed as a Research Assistant. And had the Ford Foundation in Pakistan not considered the results of this study of interest to their agricultural programs, this research would not have been completed. Mr. Robert Havener's encouragement and constructive suggestions in this regard have substantially enhanced the end product. To my parents I owe a great deal, including most of my education and my total commitment to an international involve- ment. To my wife, Sue, thanks for the constant encouragement and for relieving me of many time consuming chores in the interest of academic concentration. Much of the sweat and worry that went into this document were hers. Lastly my apologies to Erin and Scotty who went without a father at home so frequently and for so long in order that this thesis might be written. May it not be necessary again! iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS. . . . . . . . . . . . . ii LIST OF TABLES o o o o o o o o o o o o o Vii LIST OF FIGURES. . . . . . . . . . . . . ix LIST OF APPENDICES. . . . . . . . . . . . X Chapter I O PROLOGUE O O O O O O O O O O O O O 3 Setting 0 O 0 O O O O O O O O O 3 Problem . . . . . . . . . . . 4 General Characteristics of West Pakistan's Agriculture . . . . . . . . . . 5 II. THE ACCELERATED WHEAT IMPROVEMENT PROGRAM IN WEST PAKISTAN O O O O O O I O O O O 12 Historical Background. . . . . . . . 12 conception O O O O O O O O I O O 15 Objectives . . . . . . . . . . . 18 Attainments . . . . . . . . . . . 20 Seed Imports . . . . . . . . . . 21 Adeption Patterns . . . . . . 22 Strengthening Wheat Research. . . . . 25 Wheat Production. . . . . 26 A Catalyst for Changing Attitudes. . . . 34 III. THE SURVEY: DESIGN AND IMPLEMENTATION. . . 37 Dual Purposes Sought . . . . . . . . 37 Problems of Survey Research in the Farm Sector of West Pakistan . . . . . . 37 Respondent Apprehension . . . . . . 38 Analyst's Understanding . . . . . . 39 Language 0 O O O O O O O O O O 4 1 Recall and Measurement. . . . . . . 41 Vagaries of Nature . . . . . . . . 43 The Right Respondent . . . . . . . 45 iv Chapter IV. VI. VII. Page Sample Design . . . . . . . . . . 46 District Selection . . . . . . . . 46 Sample Frame. . . . . . . . . . 50 Village Selection . . . . . . . . 50 Farmer Selection. . . . . . . . . 51 Variety Selection . . . . . . . . 52 Production Function Variables . . . . . 52 Water . . . . . . . . . . . . 53 Land Preparation. . . . . . . . . 54 Seeding Rate . . . . . . . . . 56 Farm Yard Manure (FYM). . . . . . . 56 Yield . . . . . . . . . . 57 Interviewer (zero- one). . . . . . . 57 Date of Planting (zero- one) . . . . . 59 Method of Sowing (zero-one) . . . . . 60 Cost Variables . . . . . . . . . . 61 Implementation of the Survey . . . . . 63 WHEAT PRODUCTION TECHNOLOGY IN THE PUNJAB - PART I . . . . . . . . . . . . . 66 Structure of the Final Sample . . . . . 66 Input Levels by Variety . . . . . . . 69 Input Levels by Farm Size . . . . . . 71 Input Levels by Interviewer. . . . . . 74 WHEAT PRODUCTION TECHNOLOGY IN THE PUNJAB - PART II . . . . . . . . . . . . . 77 Zero- -one Variables. . . . . . . . . 81 Dl(Interviewer) . . . . . 82 D2 and D (Dates of Planting) . . . 83 D2(Lide Sowigg) . . . . . . . 84 Marginal Physical Products (Actual Input Levels). . . . . ., . . . . 85 MPP of Irrigation Water . . . . . . 87 Estimates of Bullock Labor . . . . . 93 Estimates for Seed Rate . . . . . . 94 Estimates for PhOSphate and Farm Yard Manure. . . . 95 Marginal Physical Products (Adjusted Input Levels). . . . . . . . . . 96 COSTS OF PRODUCTION . . . . . . . . . 99 SUMMARY AND IMPLICATIONS . . . . . . . 106 Need for Objective Crop Estimates. . . . 106 Water Pricing and Water Management . . . 108 V Chapter Page Water Pricing . . . . . . . . . . . 108 Water Management Management. . . . . . . 113 Nitrogen Fertilization . . . . . . . . . 114 Summary . . . . . . . . . . . . . . 118 BIBLIOGRAPHY . . . . . . . . . . . . . . 121 APPENDICES . . . . . . . . . . . . . . . 126 vi LIST OF TABLES Acreage Devoted to Major Crops in West Pakistan and Yields Obtained . . . . . . Effects of Seed Import Program on Multi- plication of Dwarf Wheats . . . . . . . Recent Changes in Wheat Cultivation in West Pakistan . . . . . . . . . . Use of Chemical Fertilizer in West Pakistan. . Rankings by Selected Progressivity Criteria for Wheat Growing Districts in Bureau of Statistics Sample. . . . . . . . . . Characteristics of Various Types of Fertilizer Available in West Pakistan on September 15, 1968 O O O O O O O O O O O I O 0 Size Distribution of Usable Questionnaires . . Reported Input Levels for Dwarf and Native Wheats in Sahiwal District. . . . . . . Reported Input Levels for Dwarf and Desi Wheats by Farm Size in Sahiwal District . . Input Levels Reported to Interviewer for Dwarf and Desi Wheats in Sahiwal District. . Results of the Ordinary Least Squares Regres- sion Estimation of Production Functions for Mexican and Native Wheats . . . . . . . Chan es in the Coefficients of Determination (R ) Assuming Zero—One Variables Were Not Included. . . . . . . . . . . . . Estimated Yields of Wheat Under the Effects of Different Methods and Dates of Planting. . . . . . . . . . . . . vii Page 10 23 27 33 48 55 68 70 72 75 79 81 83 Table Page 5.4 Estimation of the Marginal Physical Products (MPP) from Inputs to Dwarf and Native Wheats O O O O O O O O O O O 0 O 86 5.5 Approximate Probability Levels for a Type I Error Associated with Rejection of Selected Hypothesis . . . . . . . . . 87 5.6 Total and Marginal Product Curves for Water on Mexican Wheat in West Pakistan's Punjab . 90 5.7 Total and Marginal Product Curves for Fer- tilizer Applied to Mexican and Native Wheats C O I O O I O O O I O O O 92 5.8 Comparison of Native Wheat with Dwarf Wheat Under Two Levels of Input Usage . . . . . 97 7.1 Marginal Value Product (MVP) of Water Applied to Dwarf Wheat in West Pakistan's Punjab . . 109 7.2 Marginal Value Products for Nitrogen Applied to Mexican and Native Wheats . . . . . . 115 7.3 Expected Fertilization Rates Under Different Prices of Wheat and Nitrogen . . . . . . 119 viii LIST OF FIGURES Figure Page 2.1 Trends in Production, Acreage and Yields of Wheat for West Pakistan and those Districts Formerly in the Punjab State . . . . . . 14 2.2 Acreage, Production and Yield in West Pakis- tan Before and After Dwarf Wheats . . . . 28 6.1 Distribution of Farmers by Cost of Production per Maund . . . . . . . . . . . . 101 6.2 Cumulative Per Cent of Farmers with Production Costs per Maund Between Zero and Indicated value. 0 I I O O O O I I O O O O 102 ix LIST OF APPENDICES Appendix A. List of Indigenous Terms and Measures . . . B. Questionnaire . . . . . . . . . . . C. Geometric Means of Input Levels Applied by Sample Farmers to Mexican and Dwarf Wheats. Page 127 129 143 CHAPTER I PROLOGUE Setting World awareness of the threat of growing food shortages increased rapidly during the first half of the sixties. This sense of urgency, heightened by crop failures on the Indo— Pakistan subcontinent and in China in 1965 and 1966, culmin- ated in dire predictions of massive famines in the near future.11 Immediate, coordinated action was felt to be needed on several fronts. As stated by one authority, "the scale, severity, and duration of the world food problem are so great that a massive, long-range, innovative effort unprecedented in human history will be required to master it."2 Developing countries, where this problem is most acute, lie generally in the tropical and subtropical regions of the earth. Conversely, yield increasing develOpments in food production have occurred largely as the result of concerted, scientific research and innovation. Many of these varieties and techniques are only marginally applicable to a tropical 1William and Paul Paddock, Famine, 1975! (Boston: Little, Brown and Company, 1967). 2President's Science Advisory Committee, The World Food Problem (The White House, May, 1967), I, p. 11. or subtropical peasant farming milieu.l Productive agricul- tural research on those crops and environments where it was most urgently needed was thus sorely lacking in a time of crisisfli In response, the Ford and Rockefeller Foundations col— laborated to establish two international research centers located in developing countries and with the primary purpose of increasing yields of the world's most important food grains.v At Los Banos, The Philippines, the International Rice Research 'Institute (IRRI) began work in 1962. And in Mexico the long standing Rockefeller research programs in corn and wheat were expanded and reorganized into the International Center of Maize and Wheat Improvement (CIMMYT).2 The success of both institutes has been unparalleled in its rapidity and significance. Both in wheat and rice a new concept in plant architecture, dwarfness, was evolved which has revolutionized the yielding potential of each crop." Essen- tially, dwarfness enables the plant to absorb much higher levels of nutrients without lodging from excessive vegetative growth.J Short stature is combined with an erect leaf charac- teristic which minimizes shading of lower leaves and maximizes leaf area exposed to sunlight thus enhancing photosynthesis.x4 1Ibid., p. 20. For a discussion of the limited trans- ferability of genetic technology and the need for indigenous research programs, see John W. Mellor, The Economics of Agri- cultural Development (Cornell University Press, 1966), p. 268 ff. 2Letters of the official title--Centro Internacional de Mejoramiento de Maiz y Trigo. Yields can be as much as doubled or tripled even under native farming conditions. Coupled with rapid growth in available supplies of fertilizer, water and locally adaptive research and extension programs, these varieties have led West Pakistan to self-sufficiency in wheat and rice in only three years since their widespread introduction was begun.‘“ Dwarf wheats are so far superior to indigenous varieties under West Pakistan's irrigated conditions that farmers appear likely to shift over almost completely in the irrigated areas. Since West Pakistan was only 10 per cent deficient in food- grains in the worst crop year of the last decade,1 the pros- pect of doubling wheat yields throughout implies a whole new set of problems facing policymakers and planners:/ After the import substitution phase, the overriding importance attached to the goal of expanding cereal production dissipates. The dwarf varieties become just another production alternative, even though one whose enhanced productivity must require extensive adjustments in the existing allocation of resources. Price policies must therefore be modified to reflect the new set of priorities after self-sufficiency is attained as well as the new competitive relationships among productive alternatives. 1Estimates for fiscal year 1965-66 show total disappear— ance for consumption, seed and losses (changes in stocks are not available) of 6.989 million long tons while gross production was 6.416 million long tons. Charles Elkinton and Aziz Sayeed, Agriculture in Pakistan (Prepared by US-AID, Karachi, 1967), p. I89. Problem These are primarily economic questions which necessitate an economic assessment of the shift from native to dwarf wheats. This study attempts just such an assessment. It asks the ques— tion: Are there measurable differences on farmers' fields in the marginal products yielded to a given input by the two types of wheats, dwarf and tall statured? To answer this ques- tion, 115 farmers were interviewed who were growing both wheats.J Levels of inputs and the resulting yields were quantified for specific fields.' Chapter III discusses the survey in detail. From these data, I have sought in Chapters IV and V to develop a pair of comparable Cobb-Douglas production functions, one for each wheat type. The coefficients derived are used to generate marginal physical products which are then compared to determine the impact on the farm of the transition from native to dwarf varieties. The remaining portions of this thesis apply the technical conclusions to some of the agricultural policy issues facing Pakistan at the onset of the Fourth Five Year Plan (1970). The results should have a dual use. First, as already mentioned, the possible transition to foodgrain surpluses must bring some basic and far-reaching changes in agricultural policy, particularly those policies concerned with product and input prices and resource allocations. Hopefully, these data and conclusions will be a useful and used foundation for those decisions. Second, development strategists have come to realize that due to the uncertainty facing the peasant farmer, a new innovation or technology, particularly one that requires cash inputs, must often promise a high rate of annual return before it will be widely adopted.i Yet such returns are also of such magnitude that they can be expected to create short term dis- tortions in resource allocation and product mix and an imbalance between supply and demand. This has happened in wheat in West Pakistan. Crops such as the basic foodgrains with low income elasticities and a very high preponderance in local diets often experience the constraint of a slowly expanding demand shortly after a production breakthrough. Examining the transi- tion to dwarf wheats in West Pakistan should provide both the theorists and practitioners of agricultural development an insight into the types of problems that may be expected follow- ing the introduction of technology that is sufficiently remun- erative to be rapidly adopted.£ General Characteristics of West Pakistan's Agriculture West Pakistan is an area of diversity and contrasts. Ranging from the Himalayan "roof of the world" to below sea level in one of the world's most uninhabitable deserts, the Rann of Kutch, the province includes a very wide range of ecological conditions, crOpping patterns and crops.1 1Only a brief review can be given here. A more complete treatment may be found in: Government of Pakistan, Report of In all, the province covers 198 million acres but only 40 million acres are considered "cultivated" in the official data. Seventy-six million are in grazing land and other uncultivated used while 82 million, largely in the loosely administered Tribal Areas, are listed as "unreported." Most of the cultivation occurs in the alluvial plain through which flow the Indus and its tributaries. This area includes much of the former Punjab and portions of the former Sind which lie along the Indus course. Over the last century the five rivers that give the Punjab its name (literally "five waters")2 have been developed into the world's largest canal irrigation system. As distrib- utary canals were extended to new areas, progressive farmers from the more populous areas were invited to settle on the land, thus forming the canal colonies. Normally half a square of land (12 1/2 acres) was allocated although one square or more could be obtained for special purposes.3 Construction of new barrages and colonization continues today gradually to increase the cultivated acreage. the Food and Agriculture Commission (Karachi, 1960); Charles M. Elkinton and Aziz Sayeed, Agriculture in Pakistan (Prepared by USAID, Karachi, 1967); Mohammad Hasan Khan, The Role of Agri- culture in Economic Development: A Case Study of Pakistan IWageningen, Holland: Centre for Agricultural Publications and Documentation, 1966); Department of the Army, Area Handbook for Pakistan, DA Pam. No. 550-48 (Washington, D.C., October, 1965). 2The Ravi, Chenab, Sutlej, Jhelum and Beas; all tribu- taries of the Indus. Four flow through West Pakistan at pre- sent; the Beas joins the Sutlej while still in India. 3 For a list of indigenous terms, see Appendix A. Canals are generally of two types: inundation canals, which depend on seasonal overflow from the rivers; and peren- nial canals which are fed by deflecting the normal flows of the rivers with elaborate headworks or barrages. The non- perennial canals receive water only six months of the year during the summer period of runoff from the Himalayan snow— pack. Many of the earlier canals were also designed for the less intensive cropping pattern and sparser settlement density of the period and consequently provide adequate water for only a portion of the area they command.l Hence, the canal system does not offer a complete water supply, particularly during the winter months. Supplemental water has historically been obtained from countless thousands of persian wells.2 In a new development during the present decade, nearly 78,000 tubewells3 have been installed by private parties and over 8,000 more on government account. Normally of one cusec capacity, they provide assured water for at least 50 acres of most crops and have been a major factor in expanding cropped acreage as well as intensi- fying the cropping pattern. Finally, the world's largest 1The "command area" of a canal is all of that area which lies downhill from the canal and could theoretically be irri- gated from it. 2An animal powered wheel lifting water 10 to 25 feet from an open well by means of a continuous chain of buckets. 3A small diameter (2"-6") borehole lined with sieves and casing which may be 2-300 feet in depth and is powered by an electric or diesel engine. earth-fill impoundment, Mangla Dam, has been completed and work is well along on Tarbela Dam. Together they should store some 13.1 million acre feet of active water reserve, greatly enhanc- ing the adequacy of West Pakistan's irrigation system, partic- ularly during the winter season.1 One hundred years of water development in the province has created an irrigation system that is the major necessary precursor for an agricultural revolution, and it is largely in these irrigated areas that the present rapid growth is con- centrated. Population is also concentrated on these productive areas and farms are accordingly small and fragmented. Median farm size in West Pakistan is approximately 5.2 acres. Seventy-four per cent of the 4.86 million farmers cultivate an areas of less than 12.5 acres. In fact more than 50 per cent of the entire population of the province depends primarily on farms of less than 12.5 acres for a living.2 These small farms remain basically traditional in their resource endowment. Twelve and one half acres is approximately the limit which can be handled effectively by one pair of bul- locks plus family labor. It is generally too small to make a a tubewell or tractor remunerative. The few simple tools used‘ 1International Bank for Reconstruction and Development, Program for the Development of Irrigation and Agriculture in West Pakistan, Vol. V (London, May, 1966). 2Calculated from: Government of Pakistan, Census of Agriculture, 1960 (Karachi, 1964). are fashioned by village craftsmen, paid for in produce and are perhaps best viewed, as suggested by Mellor, as the em- bodiment of labor rather than a capital investment.1 Only three productive factors are generally used which require cash expense; fertilizer, water and, more recently, improved seed. Where water is available, crops may be grown during both winter and summer seasons. The 1960 Census of Agriculture estimated a cropping intensity of 121 per cent among wheat growers. Since then,with the increase in supplemental water from tubewells and the incentive prices established for wheat, this figure has been rising.2 Table 1.1 shows the area devoted to major crOps and the yields obtained in the years just before the introduction of dwarf wheats. Wheat dominates both production and consumption in West Pakistan. Sixty—five per cent of all farms grew wheat in 1960 as compared with only 41, 24 and 18 per cent for fodder, cotton and rice respectively, the next most prevalent crops.3 On the 1John W. Mellor, "Toward a Theory of Agricultural Devel— 0pment," in Agricultural Development and Economic Growth, edited by Herman M. Southworth and Bruce F. Johnston (Ithaca, N. Y.: Cornell University Press, 1967), p. 37. 2For all crOpped land in West Pakistan the multiple crop- ping index has risen from 109 per cent at the time of Partition to 113 per cent in 1965. Asian Development Bank, Asian Agri- cultural Survey (Seattle: University of Washington Press, 196FT. 3Government of Pakistan, 1960 Pakistan Census of Agricul— ture, Vol. II (Karachi: October, 1963). 10 TABLE l.l.-—Acreage Devoted to Major CrOps in West Pakistan and Yield Obtained. (Six year average, 1960/61 - 1965/66) Area Per cent of Yield Crop (000 acres) Cropped Area (long tons/acre) Wheat 12,435 32.78 .33 Rice 3,137 8.27 .37 Gram 2,850 7.51 .22 Bajra (millets) 2,026 5.34 .18 Jowar (sorghum) 1,285 3.39 .19 Maize 1,211 3.19 .41 Barley 447 1.18 .25 Food CrOps 23,391 61.66 Cotton 3,523 9.29 .lO(1int) Sugarcane 1,214 3.09 1.15(gur) Rape and Mustard (oilseeds) 1,173 3.09 .18 Tobacco 116 0.31 .66 Total 29,417 77.55 a - Total cropped area taken as 37.93 million acres based on the average of 1962/63-1964/65. - The balance of the cropped area is very largely winter and summer forages for which annual data are not avail- able. Source: - Central Statistical Office, Pakistan, 1966 and 1967 issues (Karachi, Statistical Pocket Book of 1967 and 1968). 11 other hand, wheat supplies at least 60 per cent of the caloric content of local diets.l It is obvious that the successful introduction of a high-yielding wheat would be expected to affect a vast majority of Pakistanis, either as producers or consumers or both. 1Computed from data in: G. C. Hufbauer, "Cereal Consump- tion, Production, and Prices in West Pakistan," The Pakistan Development Review (Summer, 1968); National Science Council of Pakistan, Protein Problem of Pakistan (Islamabad, July, 1968); and Directorate of Nutrition Survey and Research, Preliminary Report, West Pakistan Nutrition Survey 1964-66 (Islamabad, 1968). CHAPTER II THE ACCELERATED WHEAT IMPROVEMENT PROGRAM IN WEST PAKISTANl Historical Background Wheat has been the subject of research in Pakistan for over three quarters of a century.2 As early as 1893 foreign varieties were being tested for their local adaptability. Be— ginning with the pure lines isolated from local wheats by the Punjab Board of Agriculture in 1906, a succession of "improved" varieties Spread in waves over the land. While each contained some advantage over its predecessor, they were all traditional long stemmed varieties which could not utilize heavy doses of nutrients. Even today it is a mark of pride with some farmers when these varieties are grown well enough to reach five feet in height. For the purpose of this paper they will all be classified as native, or desi, as opposed to dwarf varieties. 1A detailed account is available in a forthcoming Special bulletin by Ignacio Narvaez Morales to be published under this same title by CIMMYT. 2See for example: Mian Anwar Hussain, Agricultural Re- search at the Ayub Agricultural Research Institute, Lyallpur (Lahore: Government of West Pakistan, 1964); and Ch. M. A. Aziz, Fifty Years of Agricultural Education and Research at the Punjab AgriCultural CoIIege and’Research Institute, Lyallpur (Lahore: Department of Agriculture, Government of West Pakistan, 1960). 12 13 Despite three—quarters of a century of these efforts, the crucial variable, yields, remained essentially static. Whatever increased potential the new varieties might have offered was effectively nullified by increasing salinity and the extension of wheat cultivation to progressively less pro- ductive lands. Figure 2.1 shows that the gradual growth in production is almost exclusively due to expanded acreage. Data from two available series are plotted for the Pakistan Punjab from 1931-32 to 1957-58 and for the Province as a whole from Partition until 1966-67. In addition to the cor- relation of trends in production and acreage, Figure 2.1 shows the dominance of the former Punjab in the wheat acreage and production of the province as a whole, as well as the somewhat higher yields due to the concentration of irrigation. Neither acreage nor production in either series grew at more than 1.1 per cent per annum. Since Partition in 1947, wheat production has grown particularly Slowly, expanding at 0.5 per cent annually. Clearly this was inadequate when pop- ulation in the province has been estimated to increase at annual rates of as much as 3.5 per cent.1 The inevitable food grain deficits began to appear in the mid-fifties. Although agriculture received somewhat more attention in the Second Five-Year Plan (1960-1965), than it had in the First Plan, the emphasis remained on industrial acceleration and lLee L. Bean, et a1., Population Projections for Pakis- tanLgl96O-2000, Monographs in Economics of Development, No. 17 (Karachi: Pakistan Institute of Development Economics, 1968), p. 14. I.[ ixii iii .oumum common on» :M hauosuom muofluumfio mmone pcm cmumfixmm umoz new noon: mo mpauwa can mamouo< .:0quo=poum CM monouella.m musmqm coma omma ovma mmma . . . o la M mmmuw>¢ wcfl>oz Mom» m H a \ x W oa . mpawaw Hmouoc l\\, S . . m -m NH . .UM\pE . m naoaw :OHuUSpoum : v In m \/\l\/\\ .. m l. 4 >\/\I\. , n l . w wmmmuo< 1 : /\ a S I 3 K . z muowuumflo SH \ BASE I 2 $333 umms mouu¢ mo "WCOB macadawz 15 agriculture's Share of the overall development program (13.3%) was smaller than four other major sectors.l Due largely to surface and ground water development, agricultural growth did rise to a rate of 3.5 per cent annually during the plan period compared with 1.3 per cent per annum for the previous decade.2 This was, however, only slightly more than sufficient to off- set population increases. Food grain imports of approximately 1,000,000 tons continued to be required each year for West Pakistan alone. Conception Beginning in 1961, the Rockefeller Foundation's wheat research program in Mexico had provided a limited amount of assistance to West Pakistan's wheat program. They had sup- plied samples of some of the best Mexican varieties for test— ing under Pakistan's conditions and four Pakistanis had received eight months training in applied techniques of wheat research. Then in 1964, Mr. Haldore Hanson, Ford Foundation Rep- resentative in Pakistan visited the research program in Mexico and returned quite impressed with the potential it held for Pakistan. The combination of his enthusiasm and the earlier contacts with the Mexican program led Malik Khuda Bakhsh, then 1Planning Commission, The Second Five—Year Plan (1960-65) (Karachi: Government of Pakistan, 1966), p. 12. 2Planning Commission, The Third Five—Year Plan (1965-70) (Karachi: Government of Pakistan, 1965), p. 2. 16 Secretary of Agriculture, Government of West Pakistan, to explore the possibilities of an expanded collaboration with Ford and Mexico. Mexico had shown a greater increase in its average wheat yields than any other country during the pre- ceding decade. Drawing on the Mexican experience seemed per— haps the only way of achieving Pakistan's goal of doubling wheat yields during the Third and Fourth Five-Year Plans, 1965—75. Dr. Ignacio Narvaez of the Mexican National Agricultural Research Institute visited Pakistan twice during the winter season of 1964-65 to prepare a comprehensive evaluation of the country's wheat research program.1 In the experimental Mexican wheats then under adaptive testing programs in Pakistan, he saw the potential solution to Pakistan's chronic deficits. Malik Khuda Bakhsh then requested financial assistance from the Ford Foundation to support the cooperation between Mexico and West Pakistan.2 Under the accelerated program de— signed by Narvaez, the Government of West Pakistan agreed to: lIgnacio Narvaez, "Accelerated Wheat Improvement Program for Pakistan", Memo to Mr. Haldore Hanson, Ford Foundation Representative in Pakistan (Karachi, November 18, 1964), and Ignacio Narvaez, "Accelerated Wheat Improvement Program in West Pakistan," Memo to the Secretary of Agriculgure, Government of West Pakistan (Lahore, March 30, 1965). 2Malik Khuda Bakhsh, "Accelerated Wheat Improvement in West Pakistan," letter D. O. No. V-ER(71)DDPAI/64, to Mr. Haldore Hanson, (Lahore, May 20, 1965). 17 Provide all recurring local currency costs of the research in Pakistan. This was essentially a com- mitment to double annual expenditures on wheat research from rupees 150,000 to 300,000. Take necessary administrative measures to acceler- ate the work of the Agricultural Research Service with respect to wheat, to facilitate communication between its scattered units and to maintain high morale. For its part CIMMYT agreed to administer the Mexican participation in the program and to: 1. Release a senior Mexican scientist to serve as resi- dent consultant to the project in Pakistan. Recruit other short-term consultants as needed and agreed upon. Permit export of the most advanced seed stocks emerging from CIMMTY'S breeding program. Provide training facilities for Pakistani scientists at Mexican research institutions. Ford Foundation agreed to cover foreign exchange costs of this project including: 1. Resident technical consultant. Short-term consultants. Project equipment. Travel and study awards. Shipment costs of experimental seed. 18 The Foundation's initial grant for $300,000 was made in 1965 to CIMMTY which then provided the scientific and techni- cal administration for the program and dispatched Dr. Narvaez as the first wheat advisor.l Objectives AS its basic goal, the Accelerated Wheat Improvement Pro- gram sought to accelerate agricultural research so as to gen- erate a rapid rise in commercial wheat production. Wheat research in Pakistan was finally to be put to the test of yield increases under farm conditions. The primary focus of effort was on a breeding program to develop new high yielding varieties based on a blend of the genetic advances made in Mexico with the desirable characteristics of the indigenous wheats. The desired outcome was a dwarf, stiff—strawed, widely adaptable variety that would be highly responsive to fertilizer and yet contain inherent resistance to the diseases and climatic vagar- ies of Pakistan. An eating quality acceptable to Pakistanis was also sought. Innovative within the Pakistan context was the incorporation of broad agronomic and economic research to develop the accompanying package of cultural practices which would be most suitable to various parts of the country.2 Team research was stressed both within and between agri— cultural research institutes. Scientists from related lNarvaez, op. cit., March 30, 1965, pp. 3-4. 21bid., pp. 1-2. 19 disciplines were to be brought into the design, conduct and evaluation of experiments wherever appropriate. The cereal botanists from each participating institute were to meet together with the resident advisor and other to plan annual breeding and testing programs of maximum complementarity. It was realized that eventually wheat research in Pak- istan would have to mature into a self-sustaining program, run by Pakistanis without foreign advisors or assistance. Consequently, the major portion of Ford Foundation assistance in all three wheat grants to date has been for manpower devel- opment overseas. In his progress report of 1965, Borlaug told Pakistan and Ford, "I regard this continuous staff improvement of your research personnel at all levels—-the research assis- tants, senior scientists and the administrators who head the stations--to be the heart of your drive for accelerated wheat research."l Accordingly plans were made to send, over a five year period, 30 research assistants to Mexico for one year of applied training under scientists from CIMMYT and elsewhere. Five senior scientists were to receive Ph.D. degrees in aspects of wheat breeding, and each supervising cereal botanist was to make a one or two months study tour of Mexico.2 Subsequently 1Norman E. Borlaug, "Progress Report on the Accelerated Wheat Improvement Program in West Pakistan," Karachi, November, 1965, p. 14. (mimeographed) 2S. A. Qureshi and Ignacio Narvaez, "Annual Technical Report, Accelerated Wheat Improvement Program West Pakistan, 1965—66," Lahore, August, 1966, pp. v—vi. (mimeographed) 20 this list was expanded further to include advanced graduate training for related disciplines which would contribute to a balanced overall research program.1 Implicit in the overall goal was a commitment to enhance the adoption and Spread of the results of this research. Among the devices intended to accomplish this within the research program were a series of Micro-Plot Trials and Semi—Commercial Trials of promising advanced lines. These trials were conducted on the fields of cooperating farmers throughout the province. In addition to generating research data, they served to multiply seed and demonstrate its dramatic superiority. WideSpread insertion of selected varieties into farmers' hands was achieved by allowing cooperators to keep the seed from the Semi- Commercial Trials after they had seen its yield when grown next to their own desi wheat. Attainments The first results of preliminary yield trials in Pakistan identified three Mexican varieties as sufficiently superior under local conditions to warrant immediate multiplication and widespread release. Consequently, 200 kilograms composed of Penjamo-62, Lerma Rojo-64 and Sonora-64 were imported in 1964. They were multiplied on the fields of the three main research institutes during the 1964-65 winter. These three varieties 1S. A. Qureshi and Ignacio Narvaez, Annual Technical Report Accelerated Wheat Improvement Program West Pakistan, 1966-67, Lahore, July, 1967, p. 81. (mimeographed) 21 were first grown in large acale by farmers the following year and covered the great bulk of the total dwarf wheat acreage planted in 1965/66 and 1966/67. At the same time, seed from the CIMMYT experimental line II-8156-IR-2M-4R was being selected and stabilized at the re- search institutes within the country. A handful of this seed had been brought back from Mexico in 1962 by one of the four Pakistani research assistants who had been studying in CIMMYT. The strain, a cross between Penjamo-62 and Gabo, an Australian wheat, proved exceptionally adapted, was rust resistant and had the desired plant type. However, it was still segregating for color. Reselection in Pakistan produced two stable lines, a white grain named Mexipak-65 and a red grain called Indus-66. To the farmers these two sisters became White Mexipak and Red Mexipak. Both topped all yield trials under a wide range of conditions during the 1965/66 season. In addition to their yield performance, Mexipak White had the grain color pre- ferred by Pakistanis for their chapatis (flat bread) while Penjamo, Sonora and Lerma Rojo were red wheats and made an inferior chapati. With this it was realized that Mexipak would replace the earlier dwarf varieties as rapidly as seed could be made available and intensive multiplication was begun the following year. Seed Imports A major strategem of the wheat program as it developed was to rely on imports of seed wheat wherever possible to 22 accelerate the multiplication process. The first such ship- ment was a purchase of 350 tons of Penjamo and Lerma Rojo in 1965 for which the government paid $100,000 from their foreign exchange resources. The following year they sought to purchase 100-300 tons of Mexipak seed with US-AID financing; however, only 50 tons could be purchased. This multiplied the available supplies of the variety by 3.5—fold. In 1967, West Pakistan negotiated the largest Single purchase of seed in agriculture's history: 41,720 tons of Mexipak-65 and Indus—66 were brought in. The $5,000,000 cost of this purchase also is the largest single financial commitment of any type by the Government of West Pakistan to the accelerated Wheat Improvement Program. Table 2.1 summarizes the effects of imported seed on the multiplication process. It can be seen that seed imports in— creased the availability of Penjamo and Lerma Rojo by 20-fold and white and red Mexipak by 28—fold. When one considers that an approximate 40—fold multiplication is possible in one season under good conditions in a large scale multiplication effort, it is obvious that these Shipments saved at least one—half year each in the multiplication program. Adgption Patterns Differences in the adoption pattern of the two generations of dwarf wheats are also partly shown in Table 2.1. Only 250,000 acres were planted to Penjamo and Lerma Rojo when it first began to have an impact on consumption. This is Shown by 23 .pmmm “Ow pm>mm mmz mono umm£3 mum3p m50fl>mum may no Ham mmEdmmm "Hmowumnuomam I o .SOAuMUAHQMuHsE Edfifixms new .om\.mna mm um pmpmmm I a .cowuoSpoum unfluumeeoo How mumu acaommm pmpcmsfioomm I m wmw.a oom.H mmeu o.m omh.He ooo.m mmh com.» I I oooo.mh~ mmnwwma no.e m.H «mean m.m om om omm mam I I oom.vH smIeemH nmomm nmma I I I. I I NH oa mmEfiu om ommamqooa maamqm mmImmmH Aooov Aooov uouomm .mcouv “macaw Aooov Aooov uouomu Amcouv Annouv pmucmHm mom\¢om@ coHumo muuomEa can: pmucmHm mom\tom® cowumo muuomEfl can: mmmmuom mvmmumm Iwamwuaaz Ummm co pmmm mommuom mmmmuom Iwamwuasz pmmm :0 Ummm Hmsuoc manwmmom Hmouoc manwmmom Saamon Saamoa Iumuomne Immuomce motmapCH mam mmemmaxmz OEmncmm can onom mEumq .mummnz mumzo mo cowumoflamwuasz co Emumoum unomEH pmmm mo muommmmII.a.m mamme 24 the fact that less acreage was planted than was possible under average seeding rates, implying that some of the preceding crOp had been consumed. Mexipak-65 and Indus-66, on the other hand, did not appreciably enter consumption until they had spread to well over 1.6 million acres. Two reasons may be hypothesized. First, Penjamo and Lerma Rojo were the first Mexican dwarf wheats introduced, and they encountered all of the hesitation and Skepticism usually expected when farmers discount promised returns for the uncer- tainty involved in a new technology. Their success rapidly convinced farmers of the profitability of dwarf wheats and led them to a much greater effort to multiply the Mexipak and Indus varieties when they became available. With their reticence overcome, each farmer sought (and often paid very high prices for) a few seers of seed which he treasured and grew through one or two seasons in order to be able to plant his entire acreage. Second, Mexipak and Indus are visibly better yielders than Penjamo and Lerma Rojo; farmers could see the difference especially with adequately controlled water and high fertility. This fact probably also contributed to the more exhaustive mul- tiplication efforts which kept farmers from eating much of this grain until after the harvest of 1968. The seeds value as an advanced technology was Simply too great to allow its use as a consumption item. 25 Strengthening Wheat Research One of the intentions that lay behind the Ford Foun- dation grants to the program was to develop a revitalized wheat research program within Pakistan's existing institutions that could sustain productive research after Foundation support to the accelerated program terminated. They wanted Pakistan to expand and strengthen its wheat research and to reorganize it in several ways so that it might become more productive. Nar— vaez had recommended from the outset that it would be necessary for the Government of West Pakistan to double their eXpendi- tures on wheat research in order to accomplish this.1 The basic additional commitment was forthcoming in the form of a scheme for a "Wheat Improvement and Production Project (Microplot Trials of Wheat Varieties in West Pakistan)" which was sanctioned in 1965. The scheme sought to Spend RS. two million over a five year period for the network of micro- plot trials mentioned earlier. Dual purposes were pursued—- to evaluate new varieties under farm conditions and to further the adoption and Spread of the best of these varieties. It is interesting that this basic commitment was to a program which was largely extension. Evidently the government and those who advised them had, by this time come to View the immediate prob- lem as one of selecting the most suitable of the available technologies and then rapidly infusing it into West Pakistan's agriculture. Actual expenditures under the "Micro-Plot Trials lNarvaez, "Accelerated Wheat Improvement Program," (March 30, 1965), p. 12. 26 Scheme" have run somewhat below projected levels with the result that after four years of the scheme's five year life only RS. 1,122,000 have been Spent. Subsequently, a second scheme was approved for a summer nursery in the mountain areas. By conducting breeding programs there during the summer a second generation of these winter wheats could be obtained each year. Development and multiplication time for advanced lines would accordingly be halved. The scheme envisioned a one—time expenditure of RS. 733,000 Spread over a five year period with an annual recurring expenditure of Rs. 50,000 after completion of the project. Difficulties in site selection and acquiring land restricted progress under this scheme and little or nothing had actually been spent by the end of 1969. However, it remains an active scheme, and it now appears that it will be fully completed over the next few years. Wheat Production There has been dramatic progress in wheat production since the accelerated wheat improvement program began. Physi- cal attainments are Shown in Figure 2.2 and Table 2.2. Expansion of acreage must be carefully interpreted. A strongly rising trend has been evident since the early 1950's when foodgrain deficits became critical. For the period 1947-48 to 1966-67, this trend may be expressed in the follow- ing equation where A is thousands of acres planted and T is the number of years beyond 1946-47. 27 .H.m magma Eoum mmmmno< ummnK mumzo .xaeaa .sm mean .muoemqv cmumaxma umms mo munmNmo one =.meImeaH new» are now mono ummn3 mo mmmemvmm pcoomm: Scum meImemH How mama .xmema .snmscme .mnormqv fl .02 .H .Ho> .moflumflumum mocmEHomumm mo cflumaasm maumpumso .cmumflxmm ummz mo ucmEcum>ow pamEunmmmo acmEQOHm>mo cam msficcmam .moflumflpmmm mo :mmnsm Scum mmlhmma nmsoucu ammo q.sm mme.m me.aa mam.e Hmm.mH me\memH o.eH mem.m NG.HH sam.e mme.sa mexseaa m.H emm ma.m Gem.q mom.ma ee\eema H.o NH mm.m smm.m mme.ma Ge\momH I I am.m ”mo.s eHm.NH me\aema IHexoemH Awe Hence Locos mmmmuoa A.oa\.mezv re ooov Aoooc eofluma mane mo Dame Hmm ummnz mumBQ pamflw cofluozponm mmmmuoé mm mumza ummnz Hence .cmumwxmm ummz ca coflum>fluasu mums: CH mmmcmnu pcmommII.m.N mqmme 28 Millions of Tons or Acres 14 12 10 6 ,_ 4 i;//“*" Production Md./Ac. . 12 2 4- j 11 J, 10 Yield .. 9 u/K f 0 8 I 7 Figure 2.2-—Acreage, Production and Yield in West Pakistan Before and After Dwarf Wheats. 29 A = 9555 + 175 T (R2 = .86) This trend is superimposed on Figure 2.2. The acreages for 1965—66 and 1966-67 followed the pattern quite closely. The forecast acreage for 1967-68 would have been 13.230 million acres (95% C.I. = 12.232 - 14.228). Obvi- ously, acreage was significantly higher than the trend alone would have suggested. Three factors probably contributed. First on April 1, 1967, the support price for wheat was raised from 14.50 to 17.00 rupees per maund.l Since the 1966-67 harvest had already begun, this announcement had no effect on acreage for that year. Furthermore, wheat prices remained above the support level for most of the harvest period and the Food Department procured only 21,700 tons.2 Consequently, when the 1967-68 crop was planted there was no reason to doubt that the higher support price would effectively guarantee profitable cultivation of wheat. Part of the extra acreage may have been in response to this more certain expected return. ELecond, the 1967 planting was the first time dwarf wheats were available to a significant portion of the province's farmers. lAnonymous, "Review of Pricing Police Including Subsidies arui Support Prices," Agricultural Wing, Ministry of Agriculture auud Works (Islamabad: Government of Pakistan, 1969), p. 2. hnimeographed) 2Government of West Pakistan, Quarterly Bulletin of Per- iermance Statistics, Vol. I, No. 1 (Lahore: Planning and Devel— opment Board, 1969) , Table 34. 30 The increased profits promised by dwarf wheat probably led many of the more progressive to eXpand their wheat acreage at the expense of other winter crops or fallow. To most of the small farmers dwarf wheats were an experiment, to verify under their own conditions the reports they had heard of the vari- eties' excellence. And as with any risk situation there must have been some discounting of expected returns. Therefore, while they may have grown some dwarf wheat, they did not re- duce g§§i_acreage by an equivalent amount and thus increased total wheat planted. This probably happened again in 1968-69 as perhaps one million additional farmers tried dwarfs for the first time. Data from the survey reported in depth below showed that for every added acre of dwarf wheat only .672 acres were with- drawn from de§i_wheats with the result that total acreage increased by some 19 per cent. The third Probable contributor to the acreage jump in 1967—68 was the favorable weather during the season. Barani wheat requires the residual moisture from precipitation for adequate germination, emergence and seedling development. During planting (October-December) a series of intermittent rains occurred over much of West Pakistan's wheat area.. Each one enabled farmers to plow and plant additional acreage. Even in the irrigated areas the phenomenon was observed. Here farmers usually plant wheat and fodder on those acres which they expect normal canal flows and tubewell water to supply. 0 Their remaining acres are left fallow or planted to oilseeds 31 and gram, both of which require very little water. Yet as each additional rain raised soil moisture, many of these acres were planted with wheat instead. These later plantings were then grown under near barani conditions. Acreage rose even further in 1968-69 in Spite of several changes from the situation a year earlier. Only limited rains fell at planting, and they were particularly scarce in the irrigated areas. Also, under pressure from the large crop harvested the year before, the price of wheat actually re- ceived by farmers fell as much as 2-2.50 rupees below the support price. Hence most farmers could not realistically count on the profit levels promised by the 17 rupee support price. On the other hand, a prolonged cool period with frequent rains occurred in the spring of the 1967-68 crop which en— abled maximum development of the wheat plant. With the longer growing season, additional carbohydrates were formed which became more grain at harvest. Consequently, the yield poten— tial of dwarf wheat was amply demonstrated to farmers through- out the province. In fact the demonstration may have been misinterpreted overoptimistically since some farmers made com- mitments of land or capital the following season that would not be justified except in an extraordinary year. The more important variable, yields, has also shown a dramatic upward shift due, just as with acreage, to a com- bination of factors. Among them are the Spread of dwarf 32 varieties, rapidly expanding fertilizer consumption, above "normal" water supplies in the last two years and an increas- ing understanding of modern cultural practices. The spread of the varieties has been enumerated in Tables 2.1 and 2.2 above. Benefits from the variety alone would have been minimal at best. Under reasonable water availability and no fertilizer, Mexipak has been shown to outyield dg§i_varie- ties by approximately 20 per cent.1 With this small differen— tial alone, a provincial yield of 12.81 maunds per acre com- pared to the recent historical average of 8.9 maunds would have been unlikely even if the country were completely covered with Mexipak. Furthermore it is doubtful if farmers would have shifted varieties nearly as far as they have in response to a mere 20 per cent differential. Dwarf wheats express their superiority in response to high nutrition, and as such the program was the catalyst for strong growth in fertilizer usage. Realizing the potential, a few persons connected with the wheat program campaigned vigorously within the planning and finance agencies of the government for greatly increased sup- plies of chemical fertilizer. Their success is shown by the growth pattern in usage Since 1965. It should be pointed out that even with an expansion of 60 per cent per year, it has been supply, and not demand that has been the effective 1Z. A. Munshi, et a1., "Annual Technical Report, Acceler- ated Wheat Improvement Program, West Pakistan, 1967—68," (Lahore, July, 1968), p. 30. (mimeographed) 33 constraint. What has been made available has often been sold at black market prices. TABLE 2.3.-—Use of Chemical Fertilizer in West Pakistan. (00—T Ammonium Sulphate Equivalent) Year Tonnage % Change Year Tonnage % Change 1952-53 5 1961-62 179 20 1953-54 72 1340 1962—63 197 10 1954—55 69 - 4 1963-64 325 65 1955—56 32 -54 1964-65 405 25 1956—57 44 38 1965-66 335a -17 1957-58 80 82 1966-67 549 64 1958—59 88 10 1967-68 919 67 1959-60 92 5 1968-69 1183 29 1960—61 149 38 a - Supplies curtailed by Indo-Pakistan war. Source: West Pakistan Agricultural Development Corporation, "The Demand for Fertilizer in West Pakistan" (Lahore: December, 1968). Some argue that the growth of fertilizer usage Should be attributed to the subsidy on fertilizer prices. This certainly was a factor in popularizing its early use, but the most sub- stantial growth has occurred since the dwarf wheats began to Spread (1966—67) and in the face of a declining subsidy rate. Furthermore, in 1968—69 many farmers were paying the unsub— Sidized rate for their nutrients-—the 25 per cent black market mark—up nullifying the 25 per cent subsidy. It seems clear that the subsidy is no longer as important a factor in expand- ing fertilizer use as is the opportunity to use that fertilizer productively and profitably. 34 A Catalyst for Changing Attitudes The catalytic effect of the wheat program on the farmers' increasing knowledge of improved cultural practices cannot be overlooked. Recommendations in English and local languages concerning row planting, drilling of seed and fertilizer, placement of seed, planting dates, and prOper timing and appli- cation of fertilizer and water have been an integral part of efforts to extend the dwarf wheats.l The responsiveness of the varieties to these improvements created the receptive environment in which farmers began to find it worthwhile to change. Seed rates have gradually risen from the traditional 32 seers per acre to nearly 40 on dwarf wheat. Fertilizer application to dwarf wheat has risen to approximately 50 pounds per acre average for the irrigated tracts and its demonstration effect is spilling over into higher fertilization of rice, cotton and sugar cane. The need for phosphorus is beginning to be understood. The N:P205 use ratio jumped from 12:1 in 1967-68 to 6:1 in 1968-69. Domestic industries have begun producing seed drills. Farmers seem increasingly concerned with land leveling to enhance their water use efficiency. A thriving market exists for tractors and combines. 1See for example: S. A. Qureshi and I. Narvaez, "Recom— mended Cultural Practices for Wheat Cultivation in West Pakis- tan during the Rabi Season of 1967-68," Lahore, 1967. (mimeographed) 35 In short, farmers have come to realize that change is not only possible but something to be aggressively pursued. Aresvik has summarized this point as follows: In the strategy of agricultural development the im- proved varieties have to be recognized as the bas1c catalysts for change. When théii performance'is spectacular (as with dwarf wheats) when cultivated prOperly, they take the fancy of the cultivators more than perhaps any other form of new technology. They can be adopted without any great investments in fixed capital and are within reach even for the farmers with relatively small means. They have a high catalytic effect because they require a package of improvements for the high yield potential to be realized. Such a package of improvements can rather easily be intro- duced together with the new varieties which yield a very high return while this was often nearly impos- sible--with the traditional varieties. The success and likely rate of absorption is a func- tion of the difference between the yield capacity and the return of the traditional varieties and the new ones. It is extremely difficult to get the farmers in develOping countries interested in small increases-- on the other hand, the possibility of doubling or trebling yields——will create interest and enthusiasm, and varieties of this type, according to experience with wheat in Mexico, might spread like a jungle fire. If the big yield differences are demonstrated Spec- tacularly the 'desire to change' will be lit in the farmers and create receptiveness and a 'spirit of awakening.‘ This new enthusiaSm is an important ini gredient in transforming a traditional agriculture. Enthusiasm was not confined to farmers alone. Faced initially with substantial pessimism on the part of many in the government whose leadership was necessary, the advisor to the program proceeded to plant and grow excellent crops of dwarf wheat on the farms of President Ayub, Secretary and later 1Oddvar Aresvik, "Comments on Hunting Technical Services Limited", 'Memorandum on the Tentative Comments made by the Senior Economic Advisor (Agric.) on LIP Criteria for the De- velopment of Agriculture'" (Lahore, February, 1967), pp. 21—22. (mimeographed) 36 Minister of Agriculture Malik Khuda Bakhsh and many others. Special efforts were made to have these men visit their fields during harvest and see that 40-50 maunds yields (five times the national average) were truly possible. This Simple stra- tegy paid intangible benefits many times over when crucially needed decisions were forthcoming because the policymakers had been convinced of the potential. They had finally seen that rapid agricultural progress was possible and that their own efforts and leadership could be highly productive. The enthusiasm and energy that was initially spawned by dwarf wheats has conditioned farmers and government officials alike to promote and participate in accelerated programs in rice, maize, sorghum and millets and potatoes. Additional programs following the same lines are planned for pulses and oilseeds. Significant changes have also been wrung out in the face of some opposition in the organization and conduct of wheat research. These have been amply recorded elsewhere1 and, being peripheral to this thesis, are not repeated here. It is suf- ficient commentary to note that Accelerated CrOp Improvement IPrograms as first developed for wheat are being accepted in ‘West Pakistan as one of the basic tools of agricultural progress and the concept is being rapidly expanded to other crops. lIgnacio Narvaez, ”The Accelerated Wheat Improvement IProgram in West Pakistan," (Lahore, February, 1969), pp. 29-34. CHAPTER III THE SURVEY: DESIGN AND IMPLEMENTATION Dual Purposes Sought During the winter, wheat growing season of 1968—69 an intensive survey was conducted of 115 wheat growers in Pakis- tan's central Punjab. Basically the objective was: (1) to quantify the productive differentials believed to be inherent between dwarf and dg§i_wheats, and (2) to trace the policy implications for West Pakistan. Success in the latter is pre- dicated upon a substantial measure of success in the former. If, in fact, differential productivities are discovered, these differentials Should be brought to bear on agricultural price policy now and for the near future, as long as the current generation of dwarf wheats covers the bulk of West Pakistan's irrigated wheat acreage. Policy implications of these findings form the final chapter of this thesis. Problems of Survey Research in the Farm Sector of West PakIStan There are several problems which face a survey of the ‘type done. Some are generic to traditional agriculture, others ‘to foreign countries and some peculiar to the milieu of the Itndo—Pakistan subcontinent. 37 38 Respondent Apprehension One of the most subtle, but perhaps most important is the seeming inability to elicit "straight" answers to direct questions. Two social anthropologists, after five years in a subcontinent village, aptly portrayed a typical villagers frame of reference when they wrote: Some may call our pretense of poverty, deception. Perhaps it is. But there are times when deception, as a means of self-protection, is justifiable.-- There are always hawks hovering about us. We de- liberately mislead the inquirer. We would be fools to give accurate figures, when there is a strong probability that they will be used to our disadvan- tage. In self-protection we have learned to make it almost impossible for anyone to tell who is prOSper- ing among uS. This statement was written 30 years ago and while this attitude is gradually changing, elements of it remain, partic— ularly among small farmers who even today are often at the mercy of large landlords, village shop-keepers and minor gov- ernment officials. As recently as 1967, the Department of Marketing Intelligency and Agricultural Statistics encountered the same problem in their own surveys which they summarized with the statement that, "field staff also faced difficulty in View of conservatism among the farmers because many of lWilliam and Charlotte Wiser, Behind Mud Walls, 1930- 1960 (Berkeley, California: University of California Press, 1963) I p0 1210 39 them were not willing to cooperate as they suspected their replies might put them under more financial burden."1 Local apprehension of government officials was indeed felt during the survey. The Pakistani who conducted half of the interviews was carefully instructed to stress that he was working for the Ford Foundation which is a benevolent, non- governmental, private organization interested in helping Pakistan and Pakistani farmers. Nonetheless he was at the same time clean Shaven, wearing a clean shirt and pants and polished Shoes and thus presumably was initially suspect as an exogenous element in the village scene. In addition, the author, after becoming tanned by the Punjab sun, was, on more than one occasion, initially suspected of being a tehsildar, the man who is responsible for local records, administration and taxes and who in some cases exploits this position to his own advantage. As an additional precaution, no questions were asked which directly proved the income or asset status of the re- spondent. Where these data were necessary, they were arrived at by cross compiling questions asked during different inter— views or at different points in the same interview. Analyst's Understanding As a foreigner in Pakistan, I felt it absolutely nec- essary to gain as deep an understanding of farming and 1Government of Pakistan, Survey Report on Farm Power, Machinery and Equipment, S.S.-IXIIRawa1pindi: Ministry of Agriculture and Works, 1960). I' — -' I 4O interpersonal relationships at the village level as possible. During the winter of 1967-68, I conducted a brief survey of some 52 farmers on impediments to fertilizer use. This ex- posure was just sufficient to point out quite clearly that even a great deal of academic training and on-farm experience in Europe and North America does not allow one to easily in- terpret and analyze the fine points of rural Pakistan. And without a detailed understanding, many of the responses could not be classified as data. To gain the required interpretive depth, I conducted over half of the interviews myself, traveling nearly 10,000 miles in the Punjab by car, horse cart, bullock cart and on foot and living in Villages as the guests of farmers whenever invited. It was, of course, necessary to have an interpreter along. By trial and error the year before Abdul Ghafoor, a driver for the Foundation, had been selected and he served unstintingly as driver, translator and companion through a very rigorous and exacting survey. Ghafoor's assets included thorough fluency in Punjabi and Urdu, the ability quickly to develop a close rapport with respondents, a thorough grounding in agriculture based on his own farming in Kashmir and India, and a wealth of insight into village life and customs. A, statistical research officer with the Department of Agriculture, was deputed by his government to assist in data collection. Mohammed accepted the job eagerly as a means of broadening his experience and he conducted nearly half of the 41 interviews himself. Like Ghafoor, he had proven himself dur- ing the previous survey as being a villager at heart, able to communicate with villagers at their own level and reasonably well versed in agricultural techniques. The frank comments he was able to elicit in informal discussion after the struc- tured questions helped greatly to define the decision milieu within which farmers manage their enterprises. Language It Should be emphasized that most of the interviews were conducted in colloquial Urdu or Punjabi. This strategy was consciously chosen as a means of minimizing social distance between the two discussants. In the case of the educated, larger farmers, formal Urdu proper or even English was used as seemed most appropriate on the scene. Recall and Measurement As in traditional agriculture everywhere, the average West Pakistani farmer does not keep records or accounts for his farm enterprise. Rarely even are cash expenditures re- corded or receipts retained. Unless cultural practices are enumerated as they occur they must be recalled from memory. Early cost of production surveys attempted to solve this by placing a resident interviewer in each sample village where he visited his three-six respondents daily.l There were, of lKartar Singh, Studies in the Cost of Production of Crgps in the Punjab, The Board of Economic Inquiry (Lahore: Punjab Publication No. 33, 1934). 42 course, substantial limits to the number of observations possible using this method. Given the limitations of time and manpower within which this survey operated, the best the present research team could do was to visit each farmer as soon after a given set of cultural practices had occurred as possible. Recall difficulties are compounded in West Pakistan by multiple cropping and intercropping. Operating without records, the farmer can be expected to recall details of his most re- cent land preparation, fertilization or irrigation best. With a year-long growing season, one or another of these practices is being applied to one or another crop almost constantly. One cannot wait until the end of the wheat season and then ask for a detailed elaboration of the cultivation techniques and input quantitites applied to wheat. During the wheat growing season Punjabi farmers may also be involved in harvesting, plowing, replanting and fertilizing sugarcane; planting, weeding and spraying tobacco; harvesting rapes and mustards; planting, fer- tilizing, Spraying and harvesting potatoes; irrigating each of these crops plus wheat and clover; and finally preparatory tillage in anticipation of the summer crops. It is not sur- prising that the time dimension for accurate recall is Short indeed. A related problem is caused by the fact that farm mana- gers often do not deal with their inputs in terms of precisely quantifiable units. Yields, purchased inputs and seeding rate 43 are the exception. Other inputs are less precisely quantified and remembered. Bullocks plow from before sunrise until they quit for one reason or another and the number of plowings that result may or may not be a whole integer. A heap of animal manure is spread on one or more fields until they "look good." One irrigation means the surface of a field all got wet whether that required two or Six acre inches of water. Hours Spent on a given operation were particularly nebulous. Most small far- mers do not have a watch or any reason to need one. Their reference points during the Span of a day are daylight and dark plus the various calls to prayer. 3 There is a tendency to respond with a narrow range as an approximation such as "6-8 hours," 4—5 irrigations or 10-12 plowings. For purposes of calculation the mid-point of each range was taken. For averaging costs the inherent assumption is that the errors thus included are randomly distributed and average out. The computed marginal products may be biased downward somewhat due to errors in measurement of the inde- pendent variables even if those variables are randomly dis- tributed.1 However, in most instances the ranges given were narrow and the resulting bias is probably small. Vagaries of Nature Weather often introduces a large element of uncertainty into traditional agriculture. In fact, many of the iron-clad lM. Ezekiel and K. A. Fox, Methods of Correlation and Regression Analysis (New York: John Wiley & Sons, 1959), p. 313. 44 production practices which keep traditional agriculture static in the sense of precluding growth have been purposively adopted over the centuries to keep agriculture static in the sense of precluding calamitous natural disaster. One of the twists of natural fate that affected the 1968—69 crOp was an unusual and intense heat wave during the first three weeks of March that sent temperatures into the mid—90's with some readings in excess of 100 degrees. The heat triggered the initiation of the reproductive phase of wheat growth (heading) early, presumably before the vegetative phase had produced and stored sufficient carbohydrates to sup- port abundant grain formation. In addition, evapotranSpira- tion readings indicate that the probability of moisture stress was much higher than usual in a period when there were un- avoidable Shortages of canal water in many areas. Shriveled grains resulted, and yields were accordingly somewhat reduced. Mexipak, being the higher yielder and thus requiring more carbohydrates, was proportionately more affected. Differences between the native and dwarf varieties were not so pronounced as during the preceding year when cool weather with ample rain lasted until mid—April. Consequently, the marginal product differentials that this thesis seeks will be smaller than might otherwise have been. In addition, three hail storms swept portions of the Pun- jab, including two subdivisions of the area under study. One, occurring March 25, was particularly severe with hailstones 45 reported of over one pound in size, and winds measured in ex- cess of 70 miles per hour. Hardwood trees of two feet in di- ameter were snapped off Six-eight feet above ground and blown as much as 30-40 feet before they struck ground. In the cen- ter of devastation much of the ripening wheat was destroyed. Respondents in the sample villages, including 16 farmers who had already been interviewed once or twice, suffered total crop losses and the yields of many other respondents were affected to a lesser extent. The Right Respondent When a foreigner enters a village it is quite common for several villagers to cluster around. As one begins to ques- tion a selected member of the group, there is a tendency to receive group responses. In villages with a rigid social hier- archy, a low-ranking villager may be very reticent to answer if one of his social superiors is present, or he may give answers designed to be acceptable to this revered (or feared) individual. Even in smaller, family groups there is a ten- dency for the family elder to respond to the questioner even though his sons did the actual farming. And if the old man approximates an answer wrongly, his sons will usually not embarrass him before a visitor by correcting him. The Depart— ment of Marketing Intelligence and Agricultural Statistics makes a similar observation in discussing survey difficulties: Generally farmers showed their eagerness to know the purpose of the visit and preferred to answer the ques- tions jointly. The Field Investigators also observed 46 that most cases the village headman or local leader was very active to answer the questions on behalf of the selected farmers. To overcome this we tried to identify the farmer who had actually been involved in the farming operations. Then we in- sisted on walking out to the sample fields and standing beside them while discussing their cultivation. In most cases this left the bulk of the onlookers, especially village headmen and family elders, behind in their own compounds or verandahs. The interviewees were then basically open and cooperative when alone with us in the field. Sample Design District Selection The Bureau of Statistics conducted a survey of dwarf wheat distribution in 1967-68 in West Pakistan for which they selected eight districts to represent best the irrigated wheat acreage of the province. It seemed appropriate to select a major wheat growing district from one of these for intensive study. It was also desired to select a district where agri- culture had begun to change, at least measured in terms of input usage and where dwarf wheat was fairly widespread. By selecting a more advanced district the economic implications for policy from the sample district might have longer rele- vance as provincial averages of input usage and dwarf wheat 1Government of Pakistan, Department of Marketipg Intel- ‘1igence and Agricultural Statistics (Rawalpindi: Ministry of Agriculture and Works, 1967). 47 coverage approached and passed through the levels of the sam- ple district. Table 3.1 presents the relevant criteria on which selec- tion was made. Importance as a wheat producer essentially limited the choice to Multan, Lyallpur and Sahiwal. Lyallpur and Sahiwal ranked equally according to the criteria used. However, Lyallpur is the location of the West Pakistan Agricul- tural University, the Ayub Agricultural Research Institute and headquarters for the Accelerated Wheat Improvement Program. With this locus of agricultural technology at its hub and with wheat breeders undertaking extension of both techniques and advanced seeds in the surrounding area, Lyallpur was eliminated as not being sufficiently representative. In addition, visual observation in January in the district indicated that native wheats would be very hard to find--a fact confirmed by the sub- sequent official estimate that the district was 77 per cent dwarf wheat. Multan, in addition to ranking third in progres- siveness, had the additional disadvantage that it lay some 200 miles south of Lahore which meant one lost day in transit on each end of each trip. Sahiwal, the chosen district, is a fairly good example of how the Indus irrigation network changed the face of the Punjab. Before 1915 the area was a desert described by one traveler as "rolling sand dunes patched with grass and hard, unfruitful plains glistening with salt." A major series of link canals completed under the direction of Sir John Benton 48 m .eo .m.z a s ASH seememSez H e e m SSH ensesemeem e m m m Hmm ewes umssHeem m o m o oam cummummmmst A s A m men meeomumm H H m N «mm HezHeem H m H H mam useHHeSH m N m a amp eeuHsz Axeemv Axeemc Axeemc rename re ooov poHeumHe meHReem semH .omo Ame\sech whoa mmeuemouee meIeemH mummmumm< muflmcma Enmm Hmm mmD ummnz cofluospoum HHmBmQDB HmNHHHuHmm mumBO ummnz .mqudm mUHBmHemfim m0 demam ZH mBUHMBmHD OZHzomO Bflmmz mom «HmMBHmU MBH>Hmmmm00mm QmBUmqmm wm mOZHMZ¢mII.H.m mqmdfi 49 in 1917 brought surplus water from the Jhelum river to Sahiwal district through the Chenab and Ravi rivers. The Upper Jhelum canal took water from the Jhelum to the Chenab and irrigated 350,000 acres along the way. The Upper Chenab canal picked up the increased supplies thus added to the Chenab and delivered it to the Ravi, irrigating 650,000 acres in Gujranwala and Sheikhupura districts in the process. The additional Ravi flow thus generated was diverted into the Lower Bari Doab canal and coursed 134 miles southwards changing the heart of Sahiwal and Multan districts from "hard, unfruitful plains" to the breadbasket of the Punjab. In opening up these very sparsely inhabited areas, col- onists were drawn from the best agricultural tribes of the central Punjab. Two objects were paramount. First, coloniza- tion officials sought to relieve pressure on sOme of the more heavily populated areas. Second, they sought to create villages of a type superior to anything seen in the Punjab before. To do so settlers were carefully selected on the basis of charac- ter references, village standing and proven farming ability. In addition, preference was given in Sahiwal to those who agreed to enter the production of horses, camels and mules for the army. This early selection process may well have aggre- gated an energetic population of capable farmers that would partially account for this district being among the most pro- gressive. Massive population movements at the time of parti— tion have altered the mix of racial and ethnic groups in the 50 canal colonies, but it is interesting to note, in the light of Table 3.1, that the colonist selection procedure was most rig— orously adhered to in colonizing Lyallpur, Multan and Sahiwal districts.l Sample Frame Village Selection For several years the Statistical Department of the De- partment of Agriculture has conducted an "objective" survey of wheat acreage and production. Their method involves measuring the wheat grown around more than 700 sample villages throughout the province. These villages were selected by a procedure whereby the probability of selection was proportional to the amount of wheat grown by that village. Forty-four of these villages fell in Sahiwal District and thus formed the basic sample frame. One may wonder about the propriety of studying a village that has been repeatedly studied before. However, in several villages I asked about the Statistical Department enumerator and was unable to find a farmer who remembered meeting him. Evidently the enumerator works from the village records kept by the nambardar (village headman) and seldom contacts farmers directly, hence little or no bias could be expected from his visits. 1These and other insights into canal colonies and their farmers during the period 1925-1947 are available in: M. L. Darling, The Punjab Peasant in PrOSperity and Debt (Bombay: Oxford University Press, 1947). 51 Farmer Selection No suitable lists of farmers could be found from which to draw a sample. Land records could not be used because they reflect ownership only and would have not included tenant farm- ers. They might also have identified absentee landlords or family members who hold title to the land but do not farm it. Both groups would have been insufficiently familiar with practices and inputs actually applied. Basically, the selection procedure involved arriving at the sample village and by discussions with one or more re- sponsible villagers, identifying three farmers who fit a set of predetermined selection criteria. These criteria were: 1. Grew both Mexipak and native wheats. 2. Interviewee was actually associated with growing the crop. 3. One out of five to be a farmer of more than 25 acres. Fulfilling these criteria was not always possible. The most frequent problem was finding anyone who grew desi.wheat.l In at least one village, one of the local government officials had-threatened farmers with a fine for growing de§i_wheat, so while there was a field or two in evidence, the owners could not be located. Where it was impossible to find enough re- spondents growing both varieties, growers of Mexipak alone lSubsequent data from the Statistical Department Show the district to be 65.12 per cent dwarf wheat, second only to Lyall- pur's 77.37 per cent. 52 were interviewed to add additional observations to that par- ticular set of data. When one enters a village and asks to see "a wheat farmer" he will quite often be referred to the best known, most progressive farmer in the area. But most of these men had been the first to try dwarf wheats in 1966-67 or 1967-68 and had found them profitable enough to switch completely by the year of the survey. By requiring as many growers of both varieties as possible, we precluded skewing the sample toward these men and instead obtained a fairly good cross section farmers of "average" management capabilities. VarietyfiSelection Since Mexipak-65 (White Mexipak) and Indus-66 (Red Mexi- pak) are genetic sisters except for color, they are treated as one variety in the sample. While initially any dwarf wheat was to be included in the computations, sufficient red or white Mexipak was found to make it unnecessary. Similarly, of the available research data, most would Show C-591, C-273 and Dirk to be approximately equal in Sahiwal, and in fact, the farmers grow them interchangeably and without preference. These three varieties are thus assumed to be equivalent in their yield re- sponse to inputs and are treated as one traditional variety on which the desi production function is to be based. Production Function Variables Stress was laid on a precise identification of those variables which were to enter the production function analysis. 53 The following are used for this purpose: Leger. The number of irrigations applied was asked twice, once when the grain was ripening under the assumption that irriga- tion had been completed and once after harvest. To our sur- prise many farmers irrigated once or twice more while the grain was actually turning color. Although the marginal value of these late irrigations is questionable, they are counted along with the rest. There was no way to translate an irriga- tion into an actual quantity of water applied. Irrigations thus enter the equation as integer units. Each farmer was asked about rain; however, in almost all cases, the first rain after planting was in conjunction with the March 25th hailstorm and had little effect on production. This was entirely consistent with other data which Show that pg rainfall is to be expected during a normal wheat season in Sahiwal District.l Chemical Fertilizer Farmers enumerate their fertilizer applications in terms of "bags per acre" deSpite the fact that there are nine differ- ent quantities of nutrients in the different bags. Furthermore, farmers frequently do not know what fertilizer they have 1International Bank for Reconstruction and Development, Program for the Development of Irrigation and Agriculture in West Pakistan, Annexure 9 - Agriculture, Vol. 7 (London, May, 1966), p. 62. 54 applied. The bulk of supplies are still imported and the English or German identifications on the bags are meaningless to them. Except for urea, the farmer often calls all chemical fertilizer "guara" which is the name of the plant used for green manure. Identifications such as Ameriki guara, Germany guara, Multani guara are not uncommon. Fortunately, given accurate information on bag weight, price and color, the con- tents can be positively identified as no two bags have the same combination of these factors. Price was a problem Since black marketeering in fertilizer was a particular problem this past winter season. Mark-up margins were observed as high as 25 per cent. Consequently, using Table 3.2 and the assumption that black market mark-up did not exceed 35 per cent, the responses were converted into pounds of nitrogen and phos- phate per acre. Land Preparation Initially we sought to determine the number of man-hours and the number of bullock-pair—hours that went into preparing and sowing the field. Bullock-pair-hours per acre were quite consistent between the two interviewers. However, differences appeared in the man-hours recorded per acre and in the rela- tionship between man-hours and bullock—pair-hours. The author had used a concept of "essential man-hours" in which the bullock-pairwhours formed the base (since each pair needed a driver) and additional hours were added for those jobs done by a man alone (such as sowing seed, making bunds, etc.). The 55 m N o mem.om Hmp3om m2weH om oo.m~ OHH seem uanH es mH mHMBEmoee esHeoeemIHo .N o m: mm mmHu z.eH mm oo.H~ OHH Iemum ezoue nmfiwmnm venom om cm mgmnmmonmouuwz .H mHmNHHHhHmm pcsommoo m.om oo.om OHH mmHsemum mmnm venom we I mumnmmonm IHmQSm mHmHHB .N o.mH om.m NHH mmHeemum muHQB :mwhmuo ma I mgmsmmonm Hmmsm .H mumNHHHuumm mumnmwonm m.om oo.e~ OHH mmHsemum muflc3 pcsom I we Amen mmHmHv mm“: .m m.oe oo.o~ mm meHeee.Ha muflsz pcsom I me Home EsHmeV mmHD .m o.m~ co.MH e.om mHenmsuo mufls3 mean I we Ammn HHmEmV mmHD .v o.a~ om.vH OHH mmH: Icmum mmcmuo Hmasumuufl mmnmq I om mumansm ESHSOEE< .m o.m~ os.HH m.em mmHsemnm meHes I ma mueHuHa ssHeoeee .N mamummuo m.m~ om.HH NHH muan3 mmumoo I Hm mumanSm EdHcoEE< .H mumuflafiuumm cmmouuflz A.mnHv A.mmv A.mnHV cowumflnommo w w HmNHHHuHmm mo mEmz mmn\m wmn man Hemammcm mumcmmonm cmmouuflz mucm Hmm mo Iwuusz moflum mNHm ucmmmum mucmfluusz .memH .mH Hmhsmpemm so cmumfixmm pmmz ca magmaflm>4 umuflaflunmm mo mmmhe mzowum> Mo moflumflumuOMHMSUII.N.m mqmde 56 other interviewer enumerated a considerable amount of peri— pheral man-hours where family members accompanied the farmer to the fields but were essentially unproductive while there. No way could be found to separate this social component after the fact so the two man-hour series could not be made compati- ble. The author's enumerations Showed a very close correlation between bullock-pair-hours and "essential man-hours." Hence, bullock-pair-hours (BPH) was used as indicative of the total animal and human labor required in field preparation and sowing. Seeding Rate Seers of seed per acre were recorded. There was not a great deal of variation with the traditional rate being 30 seers and the recommended rate for dwarfs of 40 seers. Most farmers fell within this range for both varieties. Farm Yard Manure (FYM) Farmers commonly apply animal manure to their native wheats as did their forefathers and less frequently apply it to dwarf wheats. Manure, straw and household sweepings are composted for several months either in a pit or in an open stack, and then taken in bullock-carts to the field. There is no common measure of quality. The only quantitative measure which most farmers could recall was the number of bullock-cart loads Spread on a given field. These figures were transformed into five maund units using the average reported weight per cartload. 57 11.912 Of course the dependent variable, yield, was carefully measured. We tried to persuade farmers to thresh the two sample fields each separately and in most cases were successful. However, in cases where the sample field was small and the total wheat of that variety was just enough for one threshing, the sample was sometimes aggregated with other plots. In these cases, the average yield for the lot was taken. Preliminary analysis indicated that the amount of unex- plained variation could be reduced by including zero-one vari- ables in the equation for three items. They were thus included with the following construction: Interviewer (zero-one) The reSpondent's fear of having the information he gives used against him has been discussed above. It is a phenomenon that still exists today. The present survey would indicate it is manifested in different degrees depending on the nationality of the interviewer. Data in Chapter IV show that the Mexican wheat yields reported to Mohammad averaged 19.8 maunds, while those reported to the author averaged 26.2. The Accelerated Wheat Improvement Program estimated Mexican wheat yields for the Province in 1968-69 at 23 maunds. Sahiwal District per- ennially achieves higher yields than the Province as a whole. Ninety-eight per cent of the wheat land is irrigated and fer- tilizer is used more intensively. Consequently, with a pro- vincial average dwarf wheat yield of 23 maunds, a figure of 58 19.8 for Sahiwal District can be labeled as inconsistent while 26.2 maunds would be within the range expected. Hence, it seems that the Mexican wheat yield for Sahiwal District was near that reported to the author. The Pakistani must have elicited de- pressed responses. The sample district was so selected and sub- divided among interviewers that there is no reason to suspect his farmers actually got lower yields. Another piece of circumstantial evidence has come to light. Farmers in some areas felt the government perceived more of a "Green Revolution" than had actually occurred. They were afraid taxes would be increased by a disproportionate amount. There is some evidence of a tacit agreement among farmers to under-report yields received in 1968-69 to the government even though they readily admitted the true levels to foreigners. For these reasons it is suggested that the discrepancy between yields reported to the author and to his Pakistani assistant was caused by an under—reporting to the latter. In estimating the production function, a zero-one variable was constructed where it was entered as one if Mohammad interviewed and zero if the author interviewed. The results for this vari- able should give a measure of the depression of the Y-intercept due to being interviewed by a Pakistani. This approach implies that the slopes of the production surfaces were not affected. Such a condition holds if one assumes that the understatement was a proportion of the whole 59 (rather than an absolute quantity) and that the proportional understatement was uniform throughout Mohammad's interviews. In this case, the elasticities would be unaffected. Such an assumption is made herein. Date of Planting (zero-one) The date of planting was the other variation in technol- ogy, the results of which were isolated by zero-one variables. The recommended planting date for Mexipak White, Mexipak Red, C-273, C-591 and Dirk was November 15th. Positive and negative deviations from that date were recorded and coded in two-week intervals. Then the following zero-one variables were con- structed: 1. D =1 if planted October 26-November 8, inclusive; D2=0 otherwise 2. D =1 if planted November 9-21, inclusive; D3=0 otherwise 3. D4=l if planted November 22 or later; D4=0 otherwise With this construction, the equation for Mexican wheat was estimated successfully. But, that for native wheats, could not be computed because of a Singular matrix problem. Hence, for native wheats D4 was dropped, leaving the effects of being planted on November 22nd or later in the constant term and D2 andD3 measuring deviations from that constant. The omitted ternIin.the Mexican wheat equation is plantation before October 26th. D D 2, 3 and D4 measure deviations from that date. 60 The effect of different dates of planting is to vary the stage in the life cycle which the plant is passing through when a given weather event occurs at some future date. The seasonal incidence of rain, heat, wind, and other factors also varies so that the results from any single year have only limited validity in predicting the effect of different planting dates in subse- quent years. Nevertheless, treating date of planting explicitly should serve to reduce the unexplained variation in yields dur- ing the sample year. Method of Sowing (zero-one) The three traditional ways of planting wheat are broad- casting, dribbling the seed by hand behind a native plow (the ke£a_method) and dribbling seed into a funnel tube attached to the plow (the pgga method) which puts seed at a uniform depth, if not a uniform rate. To these, planting with a seed drill has been added in recent years. Among this sample, broadcast— ing, kera and drilling were observed. There were, however, insufficient observations of drilled plantings to justify a separate variable. Method of sowing was, therefore, divided into two separ- ate techniques, 1ine-sowing and broadcasting. Since broadcast— ing is the most common, it was desired to measure the differ- ence in intercept that would occur with the conversion to line sowing. A zero—one variable (D5) was accordingly constructed so that it entered a zero for broadcasting and a one for line- sowing. 61 Cost Variables Those variables listed below were collected to determine the cost of production for dwarf and desi varieties. Where cost factors were expected to vary for different varieties, they were enumerated separately. Answers were not obtained from all respondents on all points of cost. Therefore, if cost data were available from less than half of the questionnaires, corroborating evidence has been sought elsewhere, and if less than 33 per cent were available, other research was used to provide the cost factor in question. Cost Item Brief Explanation Preparatory plowing and Sum of bullock-pair-hours plus an planking equivalent amount of man-hours Bund-making, sowing seed, Sum of man-hours when bullocks etc. are not involved Seed Cost of seed computed in October- November or imputed cost of home produced seed Animal manure Imputed cost drawn from other studies Chemical fertilizer Value per pound of nutrient at subsidized price Irrigation costs Computed separately for each farmer as sum of canal water assessment, plus reported pay- ments for tubewell water. Nothing is added for labor. Harvesting Actual reported payments for cutting and binding wheat (usually paid in kind) Threshing a. Bullock labor Bullock-pair-hours spent to thresh an acre b. Human labor Man-hours Spent threshing one acre 62 Winnowing Since the number of hours varies enormously depending on wind velocity, either the reported pay- ment or the average of the reported payments was used. Taxes Total of water assessment and land taxes Artisans These are charges incurred in lieu of depreciation on equipment. Land rent Actual average land rent for medium quality land adjusted by the reciprocal of the cropping intensity Miscellaneous expenditure A small amount of incidentals the amount of which is drawn from other recent studies since it could not be precisely identified using our method Wheat and straw both have economic value in Pakistan, and herein they are treated as joint products of the same produc- tion enterprise. They are assumed to be produced in fixed pro- portions with a grain:straw ratio of 1:1 for dwarf wheats and 1:1.5 for desi wheats, standard ratios for the two types. During the 1968-69 harvest, a divergence of prices for both straw and grain became markedly apparent. Desi wheat grain commanded a premium of as much as 50 per cent over Mexi- pak.just one month after the harvest was complete. And desi straw sold for an average of 66 per cent more than Mexipak :straw during the same period because the bullocks and buffaloes lireferred the soft straw of native varieties. This is an in- ‘teresting anomaly Since it was precisely to get a short, thick- tnalled stiff straw that would not lodge that the dwarf varities VNSre developed. In any event, straw and grain of the two 63 varietal types have to be considered differentiated products, each with its own distinct prices, and this has been done in computing per acre return. Implementation of the Survey Sahiwal District contains four subdivisions known as tehsils. The author and his interviewer each took responsi- bility for two. Ford Foundation provided transportation for each and the Statistical Department provided directions on how to reach the villages in their sample frame. Thus equipped, the research team was able to negotiate most of the dirt roads. Only in a few cases were horse or bullock carts needed. Surveyors in Pakistan are often cautioned to get off of the paved road if they want to see how Pakistan really func— tions. Our villages were often as much as 15-20 miles into the hinterland, all by dirt roads severely rutted by bullock carts. In several villages the author was the first expatriate to visit the village in the memory of any living resident. Our method was to visit each village at least three times. The first :round was conducted from December 15th until January 6th, and it was at this visit that we identified the respondent farmers. (Questions number 1-43, covering identification, farm character- ;istics and inputs through the time of sowing, were asked.1 With each farmer, the interviewer walked out to one field guach.of his desi and Mexipak wheat and examined the crop in the 1The questionnaire is reproduced as Appendix B. 64 field. The farmer was allowed to choose the field so it is possible that the mix actually examined may have been some- what superior to the overall average. But by allowing him the choice, it was felt he would indicate fields of which he was reasonably proud and thus more likely to give complete and accurate answers. The technique of actually inspecting the sample fields each time was a deliberate effort to focus the farmers' atten- tion and recall on that particular field and the exact prac- tices and inputs applied. Deliberate falsification was probably largely avoided as well since the farmer must have assumed that we knew a wheat crop sufficiently well to catch overt misrepre- sentations of fact concerning the field we were actually inspect- ing. In the author's Opinion, the technique helped greatly to avoid the vague generalizations one often receives when discuss- ing farm practices on the villager's verandah. The second round occurred during the last three weeks in March at which time emphasis was placed on irrigation and fer- tilization along with some questions to explore costs. These conmmise questions 44—73 in the questionnaire. March, being the gyrowing season, is a slack period and is the time when a number (Df festivals, fairs and weddings occur. Some of the respondents vwere away to one of these events at the time of the second visit. :Lf two or more were missing from a given village, the investiga- txars made an appointment and returned within the next few days. ij only one man was not available, his answers to questions 44- '73 Ivere obtained at the time of harvest. 65 The third round occurred during the month of June, the Punjab's hottest. Daily highs recorded by the author on each day in the field in June averaged 108.9 degrees. During this period the wheat had been cut, threshed and stored so there was no purpose to be served by walking out to the fields. These interviews were conducted at the man's house, frequently over a cooling glass of loessi, tea or water. This interview covered the inputs and costs associated with harvest and iden- tified the yields received. In addition, if a cursory review of previous questionnaires had revealed discrepancies, these loose ends were resolved in this final interview. By the third round, we had come to know the reSpondents quite well. In general they appeared to have become more re- laxed, trusting and open with their information. For this reason, some questions dealing with costs, returns and income were postponed until the last visit. It is also felt that the crucial yield data have about as much validity as is possible under Punjabi conditions despite the cautions of many that these are always overstated as a face-saving mechanism. CHAPTER IV WHEAT PRODUCTION TECHNOLOGY IN THE PUNJAB PART I Structure of the Final Sample One hundred fifteen farmers were interviewed during round one of the field work. Most of these were again visited during round two before the storm of March 26, 1969. On that date, hail and wind damage effectively removed 16 farmers from consideration by either severe or total destruction of their standing wheat. Two farmers reported no yield due to extreme salinity on the selected fields, another two could not be located during the final visit and four others were lost for miscellaneous reasons. Thus, more than 20 per cent of the original interviewees were lost due to unavoidable causes. As processing of the data began, answers to several of the questionnaires showed internal inconsistencies. To detect these, several points of cross reference had been built into the questions. Inconsistent answers could have resulted from erroneous interpretation of the question by the farmer, of the reSponse by the interviewer or from deliberate or inadvert- ent mis—statements of fact. Wherever a misinterpretation was obvious and could be adjusted from other entries or marginal notes, these changes were made. Where this was impossible, 66 67 the questionnaires were set aside to be used only for their collateral information. The possibility remained that there had been genuine communicative barriers with some farmers. Some reSponses were of such extreme values that they suggested a misunderstanding or mis-statement even though they were consistent within the 111 questions asked. Alternatively, accepting that they were factual responses, accurately interpreted then, they suggested a non-typical respondent. And such an outlier, representative of a different "population", by the extremity of his responses would presumably seriously affect the statistics derived. Accordingly, a computer program was developed to identify outliers objectively. For each observation, the reported values of all Yi's and Xi's were compared with computed means for each variable. The result was a printout showing, for each variable those farmers that fell beyond the mean t various multiples of the standard deviation. Those farmers who reported an input level that theoretically belonged to the outlying one per cent of the population (i 2.56 S.D.) were carefully scrutinized. Those who were outliers for more than one production function variable were again discarded. It was felt more likely that they represented errors in communications or deliberate mis- statements than true behavioral deviations. These several screening procedures eliminated an additional 10 farmers leav- ing only 81, or 70 per cent of the original sample. This experience suggests caution on the part of future sample researchers in Pakistan. Natural vagaries can arise 68 during the study and eliminate part of the sample in the case where repeated visits are necessary. A cushion of extra ob- servations is thus prudent. Furthermore, difficulty in com- munications can render some observations unusable. Adequate pretesting of the questionnaire should give an indication of the prOportion of rejects to be expected. This, plus a cushion for natural calamity if appropriate, should be factors in determining sample size. Since comparisons by Size of farm are to be made below, it is important to note the distribution of sizes in the final collection of "usable" questionnaires. The discussion that follows will classify farmers cultivating 25 acres or less to be small farmers and everyone else as large farmers. Twenty- five acres is approximately the point at which either a 35 h.p. tractor or a tubewell becomes a profitable investment. It might, therefore, be hypothesized that different input mixes would be found above and below this point. TABLE 4.l.--Size Distribution of usable Questionnaire. S h. l a 1wa Sample West Pakistan Acres Distribution Range No. of Farmers Cultivated (%) (Acres) (%) 0-25 70 4- 25 93.1 > 25 30 26-125 6.9 Originally, the sample approximated the actual land dis- tribution more closely. However, all but one of the farmers eliminated by the screening methods outlined above (excluding 69 hail) were small farmers (less than 25 acres) with the result that the distribution Shifted toward a more even representation. Perhaps it was fortunate that only small farmers were rejected since it left sufficient observations among the large farmers to estimate the various means for this subset. Input Levels by Variety Table 4.2 summarizes the overall average levels of in- puts and yields reported for dwarf and native wheats. Apparently heavier input applications are associated with dwarf wheats. Such an association could result from a conscious attempt to provide better cultural practices for the dwarfs or from the fact that those farmers who practice better techniques were also the early adopters of the dwarf wheats. To test the Significance of the differences between the means, the standard errors of the differences, 85' were com- puted using the technique for comparing means of samples with unequal variances. The statistic t' = U) QJI IQ: U) l was then derived and compared against tabular values of stu- dents t'.1 The last column of Table 4.2 Shows the probability of a Type I error associated with the hypothesis of equal means. 1Using the technique presented in: Robert A. D. Steel and IJames H. Torrie, Principles and Procedures of Statistics (New York: McGraw Hill Book Company, 1960), pp. 81-82. 7O H00. 00.0 0~.0H 00.0 00.0w eHmHS 0m.A 00.0H 00.0 00.0H 00.0 HmeHee .eH HHec ass 00. 00.0 00.0 mm.mH H0.0 mueBEmoee Ho meesoe H0. No.00 00.00 Hm.0m 00.00 emoouuHe Ho menace H0. 00.0 00.0 00.0 00.0 Lemeesec meoHummHHeH 00. 00.0 m~.0m 00.0 00.00 Hmeom\mummmc when 0000 00. 00.0w 0N.N0 00.Hm 00.00 0H00\meeoeIuHmeIHooHHsm 0N. 00.H 00.0- 0~.H 00.0 10H .>oz seem eOHpmH>me mo mxmmSV pmpcmam mama cmsa mmmq .Q.m cmmz .Q.m com: udmcH euHs emeoeflmm “Genes resume mx 0 quom peers msHuez peers Human .uofluumfio Hmzflnmm CH mnmmnz m>0vmz paw mnmzo How mHm>mq psmcH pmuuommmII.m.v mqmde 71 For instance, the difference between average seed rates applied may be read as significant at the 80 per cent confi- dence level. Interestingly the average difference reported in yields was more Significant than those of any input. This can be taken as one indication of the high yielding potential of the better variety. Approximately one extra irrigation is given to Mexican wheats. Assuming rational farmers, this suggests that dwarf wheats experience a higher response to water sup- plied than do the desi varieties, a point to be pursued more thoroughly subsequently. The remainder of those inputs whose application levels were significantly different at the 90 per cent confidence level or greater concerned fertility. Nitrogen and phOSphorus are applied in the heavier doses one would expect for a fer- tilizer responsive variety. The mean values for planting date tell us that on the average Mexican wheats are planted within seven day of November 15th (first week coded 0) and that desi plantings averages a few days earlier. Ipput Levels by Farm Size Table 4.3 shows the average input and yield levels re- ported by size of farm. Differences in cultivation practices which were Significant at the 95 per cent confidence level, appeared only for bullock pair hours spent in land preparation and in seed rate. The larger BPH/acre values among small farmers reflect the substitution of tractor power for bullocks 72 ov. wh.¢ mm.mH mv.m 0m.mH Gama» 00.0 00.0 00.0 00.0H H0.0 200 00. 00.0 00.0 00.0 00.0 000000000 00.0 00.00 00.00 00.00 00.00 00000002 00. 00.H 00.0 00.0 00.0 00000000000 00. 00.0 00.00 00.0 H~.00 0000 0000 H0. 00.0H 00.00 0H.00 00.00 0000\000 00.0 00.0 H0.0I 00.H 00.0I 00000H0 0000 IIIIIIIIIIIIIIIIIIIIIIIII mummnz ammo IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 00. 00.0H 00.00 00.0 00.00 0H000 00. 00.0 00.0 00.0H H0.0 200 00. 00.H0 00.0 00.HH 00.0 000000000 00.0 00.00 00.H0 00.00 00.00 00000002 00. 00.H 0H.0 0H.0 00.0 00000000000 00. 0H.0 0H.00 00.0 00.00 0000 0000 H00. 00.00 00.00 00.H0 00.00 0000\000 00.0 0H.H 0H.0 00.H 00.0I 00000H0 0000 IIIIIIIIIIIIIIIIIIIIIIIII mummsz mnmza IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 0H.H 00.00 00.0 00.0H 0000 0000 :029 mmmq 0 .o.m cmmz .Q.m cmmz usmcH 0003 00000000 0x I quom 00000 00 00000 00I0 .00000009 Hmswnmm c0 mNHm Eumm an mummnz Ammo cam 00030 00m mHm>mH usmcH pmuuommMII.m.0 mqmde 73 by larger farmers. Tractor hours are not included in the fig- ures in this table although for the production function analy— sis an approximate BPH equivalent value was added. In addition, according to an old Punjabi saying, additional plowing always adds fertility. The small farmer with a high labor:land ratio and managing partly by tradition might possibly follow this maxim rather than optimal resource allocation. He might plow until he had very nearly approached (if not crossed) the point of zero marginal product. Instances of 20, 23 and 25 plowings on a single field were noted during the survey, all by smaller farmers. It could also be that a pair of bullocks owned by a small farmer has a lower opportunity cost than a pair owned by a large farmer. Larger farmers more closely approximated the recommended seeding rate of 40 seers per acre than did their smaller neigh— bors. This lesson seems to have also carried over to their sowing of native varieties where they were considerably above the traditional rate of 32 seers while the small farmers were not. Differences in nitrogen application between farm Size groups were negligible for both varieties although the differ- ences between varieties are again evident. Phosphate, however, appears on the surface to be an anomaly. Large farmers applied more to Mexican wheats as might have been expected. The only two farmers to apply phosphate to desi wheat were small farmers so the difference appears reversed. 74 Two interesting and somewhat unexpected results appear in Table 4.3. First, small farmers apply more irrigations to both Mexican and native wheats than do large farmers. One possible explanation is that large farmers with their higher ownership of tubewells are able to apply a more nearly adequate water supply in each irrigation and thus need fewer of them. The small farmer with only canal or purchased tubewell water may tend to cut each field a bit short in order to irrigate as many fields as possible during his turn. In such a case he may actually apply less total water even though his average number of "irrigations" is greater. Unfortunately, as already men- tioned, there was no more precise method of quantifying water applications. Second, while large farmers realized higher yields on Mexipak, they fared somewhat worse than small farmers with the native varieties. No ready answer comes to mind to explain this dichotomy. It should be pointed out that in neither case is the difference in the means very significant. Inpgt Levels by Interviewer AS the survey progressed, it became apparent that the Pakistani interviewer might be obtaining different answers to some questions. Since both sets of questionnaires were pooled for much of the analysis, it seemed necessary to compare the reSponseS received in each case for significant differences. Table 4.4 presents such a comparison. 75 00.0 00.0 0H.0H 00.0 00.0H 0H000 00.0 H0.0 00.0 00.0H 00.0 200 00.0 0H.0 00.H 00.0 00.0 000000000 H00. 00.00 00.00 0H.H~ 00.0H 00000002 H00. 00.H 00.0 00.H 0H.0 00000000000 00.A H0.0 00.00 00.0 0H.00 0000 0000 00. 00.00 00.00 00.00 00.00 00\000 00.0 00.0 00.0I 00.H 00.0 00000H0 0000 IIIIIIIIIIIIIIIIIIIIIIII mummnz 00mm IIIIIIIIIIIIIIIIII H0. 00.0 00.0H 00.0H 0H.00 0H000 00.0 00.0H 00.0 00.HH 00.0 200 0H. H0.0 H0.H 00.00 00.0 000000000 0H. 00.00 00.00 00.00 00.H0 00000002 H0. 00.H 00.0 00.H 00.0 00000000000 0H. H0.0 H0.00 00.0 00.00 0000 0000 00.A 00.00 00.00 00.00 H0.00 00\000 00.A 00.0 00.0I 00.H 00.0 00000H0 0000 IIIIIIIIIIIIIIIIIIIIIIII mummnz MHMBQ IIIIIIIIIIIIIIIII c050 mmmq 0 .Q.m :mmE .o.m cme 0003 00000000 00 u H0.00 00000002 000000 mummnz 00mm p00 m003a 000 .00000009 0030:0m cw um3m0>umucH zoom 00 Umuuommm mHm>mH usmcHII.v.v mamae 76 For dwarf wheat the differences may be said to be sig— nificant at the 99 per cent level for irrigations and yield. Inputs of water and nitrogen to native wheats are Significantly different at the same level. In each case, the inputs enum- erated by Mohammad are higher while the yields are lower. The explanation lies in the downward bias in reported yields as a result of being interviewed by a Pakistani which was discussed in Chapter III. Mohammad enumerated water and nitrogen inputs that were moderately higher than the author's enumerations for Mexican wheat and considerably higher for native wheats. Assuming the same proportional depression, Mohammad's reported yields for dwarf wheat could be well below the author's while his degi figures, starting from a propor- tionally higher point, could be approximately equal. Applying a dummy variable for the interviewer will take account of this bias and allow pooling the observations for the production function analysis in the following chapter. To use this tech- nique requires the assumption that each of Mohammad's reSpond— ents biased his yield responses downwards by a uniform pro- portion of the total. The relationship between the variations in inputs and outputs is thus not affected. CHAPTER V WHEAT PRODUCTION TECHNOLOGY IN THE PUNJAB PART II The primary purpose of this thesis was to develop pro- duction functions for Mexican and native wheats. Parallel structure in the functions was stressed so that the resulting marginal products could be compared and tested for significant differences. A quantitative assessment of the changes in prod— uctivity of the several inputs resulting from a change in wheat variety was the desired outcome. One qualification is necessary. It was not possible to design the sample so that input levels were held identical between the varieties. In this sense then, the two functions are not strictly parallel. Had input levels not yet begun to Shift, then the differences in marginal products would have been the only necessary measure. Instead, the marginal prod- ucts estimated represent those that prevailed under actual input levels sampled in the winter season of 1968-69. Presum- ably, a partial adjustment had been made toward maximum profit input levels by growers using the new wheats. A full adjust- ment to the new equilibrium will, of course, see input usage further changed until marginal value product equals marginal factor cost. If both varieties are still grown at the end of 77 78 the adjustment period, if inputs are priced equally regardless of where applied and the two varieties are perfect substitutes at a 1:1 ratio in the market (priced identically), then equi- librium will be reached when the marginal physical products are equated. In that case, the appropriate measure would be the differences between input application levels necessary to equate the marginal products. This thesis necessarily presents both differences in application level (Chapter IV) and differ- ences in marginal products with inputs for dwarf wheat adjusted to native wheat input levels. Table 5.1 summarizes the regression estimates of the two production functions. Analysis of variance for the regression as a whole indicates that for Mexican wheat, the null hypothesis 80:81: --- =Bn=O can be rejected with the probability of a Type I error of three per cent. The coefficient of determination (R2) was 0.3296 indicating approximately 33 per cent of the sum of squares of the dependent variable can be attributed to the independent variables. For native wheat the probability of a Type I error associated with rejection of the overall null hy— pothesis was .604, while R2 was 0.1420. This sample was thus able to show that Mexican wheat yields are Significantly re- sponsive to variation in the inputs quantified, but it failed to Show such a significant relationship for native wheats. It should be added that the independent variables used represent those inputs over which the farmer has most control. Yield variation in native varieties is apparently more a function of uncontrollable factors. 79 m c00>m ou mmumc mmmnu co pmucmam mumm£3 m>0umc How pmuosuumcoo mm3 mHanHm> mGOIOHmN mumumamm 02 I .p00umm 00:0 mn009p pmucmHm 000mm xn pmuusooo m>m£ @0903 200:3 0mm» Eouw mucme Iumsmpm pHm0> ucmmmummu 05:0 mc0ucmHm mo mmump 030 Hmnuo mnu new mummHOHMmmoo mas .xwuume HmHsmch p .000:s mason 000 G0 vmusmmmz I o H . .:83000 00:» :0 c3020 uouum H maze m mo >00H0nmnoum mnu £003 pmuomwmu mn cmo 00.3 mwmmnuommn Has: msa .mmcmoflwwcmwm I Q .U xwpcmmmd 00 pmosmoummu mum 005000 manm0um> may no comm How mamme 00uumeomu I m mmm.o MHmmo.o moooo.oI «00.0 Hmhmo.o Hm0oo.oI czow mC0H I I 030H00 000 0H0.0 0H0~H.0 0000H.0 00\0H\~HI00\0~\HH 00000H0 00H.0 00000.0 00000.0I 000.0 0H0HH00 0000H.0 00\H0\HHI00\0\HH 00000H0 00H.0 00000.0 H00HH.0I ~H~.0 000~H.0 0000H.0 00\0\HHI00\0~\0H 00000H0 000.0 mmHmo.o vmomo.oI moo.o mmmmo.o vamH.oI mmEEmsoz an pm3m0>umucH AmmqméHm<> HZOIommNV mmuszAmZH mmmBO mm~.o m00mo.o 0mvvo.0I 000.o ommvo.o mNHoo.0I omuscmz mum» 80mm 000.0 Hmmmo.o mmomo.oI ~00.o mono.o 0MHoo.o mumnmmosm «0 mpcsom 00~.o 00000.0 00mo.o m0o.o vmovo.o 00m0o.o :mmouu0z mo mpcsom mow.o momo~.o 000OH.o 00H.0 mmmvm.o mHMO0.o meowummwuuH mo umnfisz m00.o vm0om.o NHHmN.OI mmm.o m0mmm.o om0mo.0I mmmm mo mummm o00.o mo0mo.o wmovo.OI Hmm.o 00000.0 Nmm0o.0I munoquwmmeooHHsm mBszH 0Ho.o 00000.0 mmmvo.H Hmo.o mmm0m.o o~0om.H Emma ucmumcou .000 uouum ucm0o0mmmoo .mHm Houum ucm000mmmoo n pumpcmum c00mmmuumm n pumpcmum 000mmmummm 00002 0>0002 ummnz cmowxmz :mowxmz How 0cowuocsm coflumzooum mo :00umE0um .mummsz m>0umz pcm m :00mmmuwmm mmnmsqm ummmq mumC0pHo mnu mo muammmmII.H.m mqmoz 5000 0000 .000050 000 an 003m0>nm000 00000 00 00O00000oo 00 000050000 00000.0 000.0- 00.00 000.0 00000.0- 00:00: 0000 0000 00000.0 000.0- 00.00 000.0 00000.0- 000000000 00 .000 00000.0 000.0 00.00 000.00 00000.0 00000002 00 .000 00000.0 000.0 00.00 000.0 00000.0 00000000000 00000.0 000.0- 00.00 000.00 00000.0- 0000 0000 00000.0 000.0- 00.00 000.00 00000.0- 00000 0000000 IIIIIIIIIIIIIIIIIIIIIIIIII 00003 m>000z IIIIIIIIIIIIIIIIIIIIIIII 00000.0 000.0- 00.00 000.0 00000.0- 00:00: 0000 0000 00000.0 000.0 00.00 000.0 00000.0 000000000 00 .000 00000.0 000.0 00.00 000.00 00000.0 00000002 00 .000 00000.0 000.0 00.00 000.0 00000.0 00000000000 00000.0 000.0- 00.00 000.00 00000.0- 0000 0000 00000.0 000.0- 00.00 000.00 00000.0- 00000 0000000 IIIIIIIIIIIIIIIIIIIIIIIII 00003 0000me IIIIIIIIIIIIIIIIIIIIIIII 002 00 00000020 0.00\.020 000: 00000000000 mo00000> mm: 0000» 000008000 0000000000 00008000m 0.000003 0>000z 000 00030 00 005000 0000 00020 00050000 00000000 00000002 000 00 000005000mII.0.m 00008 87 Table 5.5 compares the marginal products from the pre- .d_ Sd’ where - / 2 2 . Sd - SMPP + SMPP . Computed values of t were compared ceding table using the test statistic, t' = l 2 with the tabular values of student's t. Probabilities given here are the smallest tabular probabilities that exceeded the actual probability associated with the computed t'. For exam— t' t the value listed was 0.3 or 30 per ple, where t > 0.2, O.3> cent. TABLE S.5.--Approximate Probability Levels for a Type I Error Associated with Rejection of Selected Hypotheses.a . _ b . _ b . - b Input HO.MPPm—O HO.MPPn—O HO.MPPm—MPPn Bullock Pair Hours 0.3 0.5 >O.5 Seed Rate 0.3 0.5 0.5 Irrigations 0.2 >O.5 0.3 Pounds of Nitrogen 0.1 0.3 >0.5 Pounds of Phosphate >0.5 >0.5 >O.5 Farm Yard Manure >0.5 0.3 >O.5 aIn each case the actual probability was somewhat smaller than these levels. These figures are the next largest probability for which a value of students t was given in the source table. bSubscripts m and n refer to Mexican and native wheats respec- tively. MPP of Irrigation Water As mentioned in Chapter III, the most precise available measure of water applied was simply the number of irrigations a field received. "Irrigations" by a native farmer are believed to vary widely between one and six acre—inches of water depend- ing largely on the scarcity of water. Few farmers are familiar with the concept of saturation of the root zone and fewer still have the equipment to measure saturation. 88 Timing of irrigation vis a vis the growth stages of the plant is a crucial determinant of the productivity of the water applied. Wright has stated: Results have shown that the first irrigation given at about the time when crown roots begin to emerge will give highest yields. Under most environmental condi- tions the tillers begin to differentiate and the crown root or adventitious root system begins to emerge at about 21 to 25 days after sowing. If the soil is not moist at this stage, the crown root system will not develop well and tillering may be reduced. A striking observation in this experiment (in the Indian Punjab) indicated that even with synchronized tillering vari- eties many tillers will not develop until the first irrigation is given. If this first irrigation is de- layed, tillers which develop subsequently may be too late and will not contribute to yield. Wright also says that "if we are sure to irrigate at about the crown root initiation stage and also during the grain—filling period, it does not make very much difference how we irrigate during the middle period of plant growth so long as the plant does not undergo extreme moisture deficiency."2 Unfortunately without farm records, it was not possible to determine ex post when the irrigations had been applied. Many respondents could not even identify precisely the day of the month on which they were being interviewed. Consequently, one has to proceed under the assumption that errors in timing of irrigation are randomly distributed. The regression coefficient for Mexican wheat was signi- ficant at the 90 per cent level, while that for native wheat 1William G. Wright, "Critical Requirements of New Dwarf Wheat," p. 5. 21bid., p. 7. 89 was significant only at the 39 per cent level. In the latter case, the statement bi # O has an associated probability of 0.61 of being wrong. One therefore cannot accept bi for native wheat, nor the marginal product calculated from it. This comparison confirms the superior responsiveness of dwarf wheat to inputs of water. The difference reflects the different conditions and goals under which the two types were developed as well as the results of 30 years of farmer selec- tion of native varieties in an uncertain and semi—arid environ- ment. Indigenous wheats were developed in West Pakistan with the goal of widespread adaptability to that province's often under-irrigated or rainfed land. Dwarf wheats were bred and selected under fairly adequate rainfall conditions with the goal of maximum responsiveness to inputs, particularly fertil- izer. The sample gave the following estimates for dwarf wheats:l Gx. = 7.19 irrigations MPP = 1.52 maunds 1 § = 27.12 maunds Var. MPP = .845 At the point of Gxi’ one irrigation adds 125 pounds per acre to dwarf wheat yields. Table 5.6 presents a schedule of total and Imarginal products for the range of observed values. The level of one irrigation was not observed, but is included to 1Hereafter all estimated yields (Y) are under conditions of broadcast sowing, within 1 six days of November 15th, and assuming farmers were interviewed by the author. 90 approximate the Y intercept value of the curve. All other in- puts are held constant at Gx- for these calculations. 1 TABLE 5.6.--Total and Marginal Product Curves for Water on Mexican Wheat in West Pakistan's Punjab. (in maunds per acre)a No. of Irrigations Total Product Marginal Product 1 12.24 4.94 2 16.19 3.26 3 19.07 2.56 4 21.41 2.16 5 23.43 1.89 6 25.22 1.70 7 26.83 1.55 8 28.32 1.43 9 29.69 1.33 10 30.98 1.25 11 32.20 1.18 12 33.35 1.12 13 34.44 1.07 14 35.49 1.02 aOne maund = 82.286 pounds. MPP per Pound of Nitrogen Nitrogen fertilization was the second variable for which positive coefficients were estimated. The coefficients for dwarf and de§i_wheats can be accepted as significantly differ- ent from zero at the 90 and 70 per cent confidence levels re- spectively. Nitrogen fertilization of wheats has been frequently studied in Pakistan. Yet most, if not all, of the nutrient response studies have dealt with average productivity, ignoring marginal calculations. For this reason they have been of limited usefulness for fertilizer policy. 91 Marginal physical product estimates from the present equation are summarized immediately below for the point X.=G 1 xi“ Again, all other inputs are held constant at their geometric means. Gxi Y MPP Var. MPP Dwarf 28.37 27.12 .070 .00153 Desi 14.50 15.33 .040 .00132 These estimates establish the higher fertilizer reSpon- siveness of the dwarf wheats. Even though twice as much nitro- gen was applied to them, the marginal product remained 75 per cent above that of native wheat. In terms of pounds of grain returned for one pound of nitrogen applied, the marginal and average grain:nutrient ratios are as follows. Average Marginal Dwarf 17.0 5.8 Native 8.3 3.3 Specialists with the Accelerated Wheat Improvement Pro- gram routinely estimate aggregate production levels for West Pakistan using average grain:nutrient ratios of 16:1 for dwarf wheat and 8:1 for native varieties. The government's "Food Self-Sufficiency Program" used the same ratios in its projec- tions.1 The results estimated here substantiate these levels and show the difference between MPP and APP. Planning with lQureshi and Narvaez, "Annual Technical Report, Acceler- ated Wheat Improvement Program," (July, 1967), p. 39. 92 only the latter has not provided a firm basis for price adjust— ments in fertilizer policy. TABLE 5.7.-—Total and Marginal Product Curves for Fertilizer Applied to Mexican and Native Wheats.(in maunds per acre)a Mexican Wheats Native Wheats Pounds of Total Marginal Total Marginal Nitrogen Product Product Product Product 1 21.23 1.554 13.86 .523 10 25.12 .184 15.12 .057 20 26.43 .097 15.52 .029 30 27.23 .066 15.76 .019 40 27.82 .051 15.93 .015 50 28.26 .041 16.06 .012 60 28.64 .035 16.17 .010 70 28.97 .030 16.27 .009 80 29.25 .027 16.35 .008 90 29.51 .024 b b 100 29.73 .022 110 29.94 .020 120 30.13 .018 aOne maund = 82.286 pounds bNative wheat is seldom fertilized with more than 80 pounds of nitrogen. Table 5.7 gives estimated total and marginal products for the usually observed fertilization rates holding all other Si = Gxi° One conclusion drawn is that yields much in excess of 30 maunds require very heavy nitrogen in the absence of addi— tional application of complementary inputs such as water and better cultural practices. Many farmers during the 1968-69 and 1969—70 seasons expressed disappointment with Mexipak wheat. It had not given the 40-50 maund yields that they had expected after hearing initial experimental results. Many of those who 93 were disappointed had applied the recommended fertilizer dose of 100 pounds of nitrogen, yet received yields of only 25-35 maunds. Table 5.7 suggests that this was all they could ex- pect without improving several other aspects of their wheat production practices along with fertilization. Estimates for Bullock Labor Both of the coefficients estimated for bullock-pair-hour inputs of preparatory labor were negative. Confidence levels were low, 71 and 51 per cent for dwarf and dg§i_wheat, respec- tively. One is not justified in accepting the native coeffi- cient as significantly different from zero. And the dwarf wheat estimate can be accepted only with limited confidence. An earlier section of this chapter discussed the diffi— culty of interpreting negative marginal products within the Cobb-Douglas functional form. Because of the negative coeffi- cient for dwarf wheat and its low significance, no quantita- tive conclusions will be made about the marginal product. Recognizing the possibility that multicollinearity among independent variables may cause "incorrect" signs, the matrix of simple correlation coefficients may contain the explana- tion of the negative estimate. For the Mexican wheat sub- sample, selected simple correlation coefficients (r) are as follow: Correlation Coefficients Between BPH and Yield -0.01 BPH and Farm Size -0.34 Farm Size and Yield 0.05 94 This suggests that the negative correlation between BPH and farm size was the prevailing factor. Farm size is posi- tively correlated with yield, yet the larger the farm the fewer hours of bullock labor applied. The negative relation to farm size could reflect an inadequacy in the method of pooling tractor plowings (found only on larger farms) and bullock plow- ings. One tractor plowing per acre was set equal to 5.5 hours of bullock labor, the average time required to plow one acre by this method. A better measure would have dealt with soil tilth and structure but would have required analyses and in- strumentation that were not available at the time of the sur— vey. Future work should distinguish sub-samples by tractor and bullock plowing methods in order to better control the influences of each. The relationships discovered could also be explained by a lower opportunity cost for bullocks on small farms than on large ones. Unfortunately, no data was obtained that would test such a hypothesis. Estimates for Seed Rate Negative coefficients were obtained for yield with respect to additional pounds of seed per acre. Confidence levels were low, 73 and 55 per cent for dwarf and desi wheats respectively. For the reasons discussed in the section immediately above, no further quantitative statements will be made about marginal productivities. 95 As with bullock labor, the simple correlation matrix suggests the cause of the negative coefficient for Mexican wheat. Selected coefficients (r) are: Correlation Coefficients Between Mexican Native Seed Rate and Irrigations -0.40 -0.29 Seed Rate and Yield —0.15 -0.12 Irrigations and Yields 0.36 0.18 In both cases, the highest correlation measured for seed rate was with the amount of water applied and that correlation was negative. Water use, on the other hand, was positively cor- related with yield. This suggests that farmers tried to com- pensate for inadequate water supplies by heavier seeding rates. Evidently, their efforts were only partially successful since with high seed rates and a low number of irrigations, yields were still depressed. Estimates for Phosphate and Farm Yard Manure Confidence levels and the signs of the regression coeffi- cients for these two inputs were as follows: Confidence Levels and Signs ( ) Phosphate Farm Yard Manure Dwarf 0.02,(+) 0.02, (-) Desi 0.18,(-) 0.77, (-) Obviously, neither dwarf wheat coefficient can be discussed. For native wheat, the coefficient for farm yard manure is significantly different from zero, but only at the 77 per cent 96 confidence level and it is negative. Again, no marginal prod- uct calculations are made. As might be expected, manure applications are positively correlated with bullock labor use-~both reflecting the number of animals per acre. The r value for this relationship is the largest for any involving farm yard manure. Both variables, bullock labor and manure use, are negatively correlated with native wheat yields. Marginal Physical Products ’(Adjusted Input Levels) At the time of the survey, input levels were in transi- tion from optimum levels for desi to Optimum levels for the new varieties. The adjustment was incomplete however. A de- scription of the impact of dwarf wheats requires discussion of differences both in the quantities of inputs used, and the marginal products prevailing. It is of interest to know the differences in marginal productivity that would have prevailed at the traditional input levels associated with de§i_wheat. Such a calculation gives an estimate of the vertical shift in yield and in mar- ginal product when only the variety is changed. Despite the negative coefficients for some inputs, the equation in Table 5.1 was used because its coefficient of determination was higher than other available equations. Re- taining all six inputs and five zero-one variables gave a better explanation the variation in Y. Results are given only for the water and nitrogen variables. 97 TABLE 5.8.—-Comparison of Native Wheat With Dwarf Wheat Under Two Levels of Input Usage. Native Wheat Mexican Wheat Actual Actual 1968-69 Adjusted a 1968—69 Input Levels Input Levels Input Levels Estimated Yield 15.33 24.82 27.12 MPP (Irrigation) b 1.623 1.521 MPP (Nitrogen) 0.040 0.125 0.070 aInput levels adjusted to levels measured for native wheats. bAnalysis above shows this figure to be not significantly different from zero. Adjusting the input levels meant reducing them. As ex- pected, where bi>0, the marginal products were higher at the reduced levels. Of course, confidence levels increased also as input quantitites were adjusted away from the geometric means at which regression coefficients were estimated. One important conclusion concerns overall yields. Even without applying more inputs, the simple substitution of dwarf for desi seed increased yields by 9.5 maunds, a growth of 62 per cent. This should not be confused with the usual statement that substituting varieties under unfertilized conditions with four irrigations gives approximately 20 per cent more yield. The 62 per cent identified here is at input levels applied to desi wheats by the sample farmers in 1968-69 and these included a geometric mean of nearly 15 pounds of nitrogen and six irri- gations. The marginal product for nitrogen is three times as large with dwarf wheat as it is with desi under these conditions. It 98 is not surprising then that the transition to dwarf wheats brings with it a strong increase in the demand for nitrogen- ous fertilizer. Fertilizer consumption was growing at a rate of 34 per cent annually during the five years before dwarf wheats. In the three years since the Accelerated Wheat Pro- gram began, fertilizer sales have grown at 52 per cent annually and more would have been purchased had supplies been available. Finally water applications gave a statistically insigni- ficant marginal product with native wheat. Yet, one irriga- tion would add 134 pounds of grain in a dwarf wheat field, a marginal product that is significant at the 90 per cent level. These three figures, the overall yield and the marginal product comparisons for nitrogen and water, establish the su- periority of the dwarf wheats in the Punjab. Not only are they more responsive to these inputs, but also the response to any one input occurs within a much higher yield environment. CHAPTER VI COSTS OF PRODUCTION The term "costs of production" is often discussed in Pakistan. Occasionally it occurs in policy discussions cover- ing price support or input subsidy levels for agricultural commodities. More frequently it is heard from farmers or spokesmen for agriculture, often to counter revisions in price policy that have been suggested. The context in which unit cost figures are used is usually one of justifying an adjustment of the terms of trade toward agriculture. ' Lewis has observed that "unit costs of production nearly always suggest farmers are producing at a loss." He further points out rather bluntly that "the use of cost of production studies to assess economic advantage in agricultural crops was completely discredited more than 40 years ago."1 Witt goes a step further and cites seven reasons why "cost of production studies at best are only a limited guide as to trends in farm costs, and at worst are a confusing morass of conflicting statements on price policy, with pleadings for 1J. N. Lewis, "Wheat Marketing in Pakistan and Export PrOSpectsP (memo to Sarshar Ahmad Khan, Lahore, Pakistan, June 30, 1969), pp. 2-3. 99 100 special interests."1 His reasons hinge mainly on valuation problems which even in a develOped country are so serious that unit cost calculations are too subjective to be used as a basis for price policy. The problem is compounded in developing countries by poor data, in scarce supply. Cost of production estimates can be used, however, for inter-varietal comparisons as long as the figures are developed using identical assumptions. Accordingly, costs per maund of wheat were estimated using the same technique for both varie— ties. A uniform cost per input unit was multiplied by the levels of inputs reported for each variety except for seed where Mexipak and indigenous wheats were valued at 14.00 and 17.00 rupees respectively.2 The variables included are listed in Chapter III. The sum of imputed and cash costs was divided by reported yields and used to develop the frequency distributions shown as Figure 6.1. Values were either drawn from survey responses or calculated from other studies according to the criteria Speci- fied in Chapter III. Observed distributions are then cumulated into the ogive (cumulative percentage) curves shown in Figure 6.2. Using data from a single sample and year, with inputs weighted and valued identically, avoids all but one of Witt's lLawrence Witt. "Cost of Production and All That" Islamabad, Pakistan, August 8, 1969. (mimeographed) 2The price of each wheat at the beginning of November, 1968, when planting commenced. No. of farmers 15.00 13.50 12.00 10.50 9.00 7.50 6.00 4.50 3.00 1.50 0.00 NO. 101 Mexican Wheat 1)- 10.01 12.0 4. 16.0.. 18.0.L 20.0. of farmers 10.01 9.01 8.0.. Native Wheat 7.01, + 6.01 5.01 4.0,, . 3.0» + + 2.0+ + 4. +- 1.0.. ++ ++ + f 0.00 Cost per maund 4) O O O C O G) O N V KO 00 H I"! H H H 20.0 Cost per maund Figure 6.l-—Distribution of Farmers by Cost of Production per Maund. Cum. per— cent/100 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 . 0.00 1? 1:- ‘b Cum. per- cent/100 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 . 0.10 0.00 1)- db 102 Mexican Wheat -» O O O N v N N Cost per maund 14.0 16.01. 18.0 20.00 26.00 Native Wheat 1 I V V O N V‘ \O N N N Cost per maund O O 14.0 C O O O O O N \D 00 O H H H H N Figure 6.2-vCumulative Percent of Farmers with Production Costs per Maund between Zero and Indicated Value. 103 objections so long as the estimates are only compared with each other. The one point remaining is the possibility that abnor- mal weather might have had different effects on the two varie— ties. The 1968-69 season included an early spring heat wave that reduced yields through several different physiological effects on the plant.1 Water-temperature-soil relationships were the most important of the causes of yield decline. Chap— ter V showed that Mexipak is far more responsive to these fac- tors than native wheat; so in an unfavorable situation one would expect Mexipak yields to have been the more depressed. Mexipak unit costs should have been distorted upwards more than dg§i_unit costs if, in fact, such a distortion took place. Two observations are possible from Figures 6.1 and 6.2. First, the costs per maund of Mexican wheat show less central tendency than for native wheat. Current desi varieties have been grown widely since the late thirties or early forties. Farmers have had 30 years to select the best sets of cultural practices for desi wheat within the context of their own farms and prevailing price relationships. One would expect each farmer to have made gradual progress by trial and error towards an optimum allocation of his particular resources. Mexipak is a comparatively new technology in Pakistan. The successive approximations of an Optimum production environ— nmnrt have not run their course yet. Farmers are in the process of trying a variety of practices; some of which are suboptimal lMexipak Wheat Yield and Quality Seed Appraisal Committee, Mexipak Wheat Yield and Quality Seed Appraisal, 1969 (Lyallpur, September, 1969). 104 within their own farms. A wide range of suboptimal resource patterns would give the diSpersion found in Mexipak. The second conclusion is that to the extent there is a cluster, the Mexipak variety clusters around 14-15 rupees, while the native varieties cluster around l7-l8 rupees. The lower costs for Mexipak are observed in spite of the possible upward relative bias due to weather mentioned above. Under normal weather conditions, the difference might have been more pronounced. The explanation lies in the relationship between the dif- ferences in cost and yield. Mexican wheat, as grown in Sahiwal during the winter of 1968-69, had slightly higher costs per acre and substantially higher yields per acre. The upshot was a lower average cost per maund produced. Thus there do appear to be differences in the variances of the distribution of unit costs and in the level of those costs. At this point in time, the dwarfs are a lower cost, but more variable cost crop. One can easily postulate situations where farmers would have to choose between uncertainty and cost. Differences observed in the cumulative distribution curves suggest one possible use of these comparisons. One could hy- pothesize that only farmers who expect to cover at least their variable costs of production will grow grain for sale in the market. Quantities marketed would then be a function of the number of farmers covering their variable costs at a given price and the magnitude of the difference between cost and 105 price. Such a relationship would make the cumulative distri— bution curves of Figure 6.2 a useful tool for price policy. Varying the assumed price along the horizontal axis would alter the percentage of farmers who cover costs and thus help to pre— dict marketed quantities. The two curves could also be used to analyze the possible effects of differential price levels for native and Mexican wheats. CHAPTER VII SUMMARY AND IMPLICATIONS Analyses presented in Chapter V established useful re- sults for irrigations, nitrogen fertilization and each tech- nique represented by a zero-one variable except line sowing. This chapter applies these results to current issues in agri- cultural policy. Need for Objective Cr0p Estimates Reported yields of dwarf wheat differed significantly between interviews conducted by the author and those done by his Pakistani assistant. Several pieces of evidence suggested that this was due to an under-reporting to the Pakistani to minimize the apparent contribution of dwarf wheats to the re- spondent's income. Even though identical questions were asked about dwarf and native wheat, significant differences in re- sponse between interviewer occurred for dwarf yields only. Farmers apparently felt no need to understate d§§i_yields. Furthermore, several sources were cited to show that such ob- scurantive behavior is a long standing phenomenon on the Sub- continent. Provincial crop estimates are prepared by aggregating reports from local officials in every small area. These men 106 107 are Pakistanis and are government employees. To further com- plicate the matter, they are often the keeper of land records. As such, they are responsible for land tax and water rate assessment. In short, the farmer has every incentive to appear nonprOSperous in the eyes of these men. . One must conclude that provincial wheat crop estimates may become quite biased downward as dwarf wheats apread over the bulk of the wheat acreage. A similar situation may also exist for the dwarf rices. It is worth noting that for the past three years, foreign Specialists have consistently esti- mated larger crOps than shown by government figures. An objective method of estimation is vitally needed to verify the present subjective crop estimates. Without such a check, consistency in the time series data will be destroyed as dwarf wheats spread. There is presently an objective es- timate of wheat production each year using a crop cutting technique. However, government seems reluctant to accept the objective estimates, preferring instead the standard system that has been used for years. Production figures have consistently been underestimated by 20-25 per cent. Even so, the bias was uniform throughout the time series and the data could be used for some purposes. In the last two years the divergence between subjective and objective wheat estimates has increased alarmingly. Official production data today can no longer be linked to earlier years. Two conclusions can be reached. First, objective esti- mates are necessary for any crop of importance especially if 108 technological change has recently occurred. Second, government must be willing to accept the objective estimates and even to revise recent years' figures in light of them if policy is to be made intelligently. Water Pricing and Water Management Marginal products calculated for water were significant when water was applied to dwarf wheat, insignificant when applied to native wheat. This difference, while instructive in itself, limits the analysis that follows to dwarf wheat alone. Water Pricing Table 7.1 converts the marginal products estimated to marginal value products (MVP) with the assumption of a perfectly competitive product marketi’ Two columns are shown. "MVP (ex- pected)" values each maund at 16.00 rupees; the price farmers expected to get when they planted. "MVP (actual)" values wheat at 14.00 rupees per maund which is approximately what they re- ceived after the support price was lowered from 17.00 to 15.00 rupees just before harvest. Irrigation water has different prices depending on the source of supply. The cost of water from the Irrigation Depart- ment canals is assessed in one lump sum. If a field is irri— gated once from this source, a flat rate per acre is levied. Rates vary by crop; for wheat in Sahiwal it was 16.00 rupees per acre in 1968-69. Having once paid, the farmer then has 109 TABLE 7.l.--Marginal Value Product (MVP) of Water Applied to Dwarf Wheat in West Pakistan's Punjab. Marginal Marginal Value Products Number of Physical Irrigations Yield Product Actual Expected 2 16.19 3.26 45.64 52.16 3 19.07 2.56 35.84 40.96 4 21.41 2.16 30.24 34.56 5 23.43 1.89 26.46 30.24 6 25.22 1.70 23.80 27.20 7 26.83 1.54 21.56 24.64 Gxi 27.12 1.52 21.28 24.32 8 28.31 1.43 20.02 22.88 9 29.69 1.33 18.62 21.28 10 30.98 1.25 17.50 20.00 11' 32.20 1.18 16.52 18.88 12 33.35 1.12 15.68 17.92 13 34.44 1.07 14.98 17.12 14 35.49 1.02 14.28 16.32 free use of canal water for the balance of the season at a marginal cost per irrigation of zero. In view of the calculated value of water, it is not sur- prising that demand far exceeds supply. Elaborate rationing systems are found in each village with the most trusted vil- lager, often the priest, controlling a rotation schedule so that everyone receives a turn in some proportion to the amount of land farmed.’ This rotation will provide a varying number of irrigations to each man depending on the adequacy of the canal for the village(s) served. It was the author's observa- tion that farmers infrequently receive as many as seven or eight canal irrigations and almost never as many as ten. Since the geometric mean exceeded seven irrigations, most farmers must turn to other sources (primarily tubewells) for supplemen- tary water. 110 Insufficient data were obtained to estimate the cost to a tubewell owner of Operating his well. Purchasing tubewell water from neighbors is a common practice and is taken here as the appropriate Opportunity cost for owners or the actual cost to purchasers. Common practice is to sell this water by the hour of pump Operation. Rates vary and the number of hours needed to irrigate an acre vary, but on the average among responding farmers an acre required 2.8 hours at 3.60 rupees per hour for a cost of Rs. 10.08 per acre-irrigation. For those who have only a Persian well for supplemental water, the cost can be taken as 15.60 rupees per acre irrigation.1 In Sahiwal, however, this method of irrigation is becoming rare. Of course, the calculated MVP's have greatest significance at the point of Gxi. At that point an actual MVP of 21.28 ru- pees compares with marginal costs of zero for canal water, 10.68 rupees for tubewell water and 15.60 rupees for Persian well water. In all cases, there is a strong incentive to ex- pand production through increased water use.“ In each case, physical limitations on supplies available are the effective constraints.‘ Unless a farmer personally owns a tubewell, the first complaint made to a visitor is the inadequacy of water supplies{/ The present calculations suggest that this complaint Often reflects the farmer's inability to apply water to the point MVP=MFC rather than his ability to irrigate enough to lTo irrigate one acre with a Persian well requires four pairs of bullocks working 24 hours in three hour shifts. The approximate daily cost of a pair of bullock is 3.90 rupees. 111 grow a good crOpI’ The fact that the geometric mean number of irrigations (7.19) exceeds the recommended number (5-6) is also explained by these comparisons. As long as MVP > MFC, farmers will simply expand water use until they reach the physical limits of supply. They can be expected to continue to install tubewells at a rapid rate to expand the available water quan- tities. It follows that farmer investment in water deve10pment could result in more wheat than the country needs. Private and social benefits may diverge widely. There is presently no direct mechanism that would require the farmer to consider the social costs Of wheat surpluses in his investment decisions. One solution often suggested is to reduce the price of wheat, but this has proved difficult to do. After the price per maund was lowered in April, 1969 to 15 rupees, sufficient pressure was generated by agriculturally based politicians to have it again raised to 17 rupees in October. Only at the upper end of the Observed range of irrigations did marginal product fall to approximately one maund. Beyond that point, if one maund of wheat and one purchased tubewell irrigation were priced equally, farmers would cease to add irrigations. This would require a wheat price of 10.00 rupees per maund which may be too low to be politically acceptable. It should be noted that a decrease in product price would also affect the profitability of other purchased inputs notably fer- tilizer. If fertilization were reduced, total yields would 112 decline. Through the reduction of Y in MPP=bi %i, the MVP for water would also decline and MVP=MFC would be reached at a price higher than 10.00 rupees. Changing the relationship between water rates for differ- ent crops could conceivably divert some canal water away from wheat into other crops. But alternative water users in the winter are few at present--tobacco, sugar beets and sugar cane being the only important possibilities. The individual farmer has no way of diverting water from winter to summer seasons. Adjusting rates for canal water leaves the profitability of tubewell irrigation untouched. In the short run, diverting canal water to other crops means that tubewell supplementation will be needed sooner and the physical limit for tubewell sup- plies would be reached at lower average levels. The result would be increased incentive to invest in tubewells; an incen- tive derived from the possibility of producing wheat in excess of the country's needs. The question needs further research. Not all areas have exploitable ground water, and the costs of exploitation vary from place to place. But if concise analysis projects wheat production at untenable levels then the remaining alternative is to raise the price of tubewell water. Except for the small proportion of tubewells that are government-operated, this would be difficult to administer. Raising power rates would affect other uses of power. Taxing the installation process is possible but would be difficult to administer since different taxes should probably apply in different localities. 113 It thus seems that eventually hard decisions will be necessary with respect to the support price for wheat. Bur— geoning expenses on government account must cause wheat to be re-evaluated so that private and social values are more nearly in line. Water Management Evidence was cited in Cahpter V from Wright that timing of irrigations with respect to the growth cycles of the plant is very important. It is particularly critical that ample moisture be available at the time of tillering and filling Of the grain. The irrigation department attempts to allocate scarce water throughout the system by a rotation where at any given time some canals are flowing at capacity, others at partial flow and a few are dry. Yet no farmer known to the author has precise knowledge of when his canal flow will be at full ca- pacity and when it will only be a partial flow. Most of them are not certain of what dates their non-perennial canals will be turned on or Off. Yet Wright shows that moisture is needed at a very precise stage in the plant's development, between the let and 25th days after sowing. If farmers had accurate knowledge of their water pros- pects in advance, they could vary planting dates accordingly and thus gain better control of this important variable. It would seem highly advisable that the initial and final dates of flow for each water release period be announced before the 114 season. A simple index Of expected flow adequacy should also be developed and published before planting for each canal in the system. Indicating what weeks a given canal would flow, and at what strength, would help farmers plan a great deal better if the information were available early enough. A corollary recommendation is that when the wheat in a given region is most likely to be going through the tillering and grain-filling stages, canals should be discharging at their maximum rates in that region, even at the expense of some cur- tailment elsewhere in the system. Following this strategy would mean reducing discharge in the Punjab early in the fall and releasing proportionately more water in the Sind where the crOp is earlier. Over a period of about one month, the maxi— mum discharge region would be shifted northward. Assuming wheat in Punjab is planted from November lst to December 15th, maximum flows should be provided from November 20th through January 10th. Granted, there will be overlapping time periods and other problems, but this concept seems to Offer a step in the direction of more efficient resource management. The man- agement program would, however, be only partly effective unless farmers are taught the critical nature of irrigation timing. Nitrogen Fertilization Table 7.2 summarizes the MVP's calculated from the estimates in Table 5.7. Mexican wheat is valued at 14 rupees per maund and native wheat at 17.00 rupees. This price differential arose soon after harvest began in 1969 as a reflection of several .cmmomuas mo mocmom om swap OHOE m>amomu Eooamm mummcz O>aumz I 115 O mommsu oo.ha an no mwmmsu oo.ma pm In mommsn oo.va um um mm. mm. mac. oma mm. mm. omo. oaa mm. am. «no. ooa O O mm. vm. vmo. om «a. moo. mv. mm. hmo. om ma. moo. mw. Nv. one. on ha. oao. mm. mv. mmo. ow mac. mm. mm. avo. om mm. mao. mm. an. amo. ov mm. mao. mo.a mm. moo. cm I I ma.a mm. one. Amumsov axw mv. mmo. mm.a mm.a moo. om mm. owo. n I u Am>aumzv axu hm. hmo. vm.~ mm.~ vma. oa om>z mm: nm>z mm>z mm: cmmouuaz amsuod omuommxm amsuom mo mocsom ummsz w>aumz pawns cmoaxmz .mummnz O>aumz UGO cmoaxmz O» Omaammm cmmouuaz now muOSOOHm Onam> amcamumzll.m.> mamde 116 factors of both supply and demand. An expected MVP column is included for dwarf wheat as in the section above based on 16.00 rupees per maund. The reduction in support price generally affected only dwarf wheat. Native wheat prices remained above floor price levels in the 1969 and 1970 harvests. The government attempts to standardize the price Of a pound Of nitrogen at .50 rupees (50 paisa) regardless of source. During the winter Of 1968-69, critical shortages develOped and a black market in fertilizer flourished.“ Most sales were con- cluded at 23 per cent1 above the Official price or 61.5 paisa per pound of nitrogen. In addition, some small cost is in- volved in transportation between retail outlet and the farm. Under these conditions, a pound Of nitrogen may be assumed to have cost the farmer between 60 and 65 paisa in 1968-69. One conclusion then is that Punjabi farmers almost exactly equated marginal value product and marginal factor cost when fertilizing native wheatsfi Chemical fertilizer and the present desi varieties have both been available since 1953. Repeated trials have evidently given farmers a precise defini- tion Of their high profit point in combining the two. A corollary conclusion is that fertilizer applications to native wheat should be directly responsive tO changes in the price of either product or input. lThirty-two rupees per bag of Urea compared to the Official rate of 26. 117 The geometric mean of nitrogen applied to dwarf wheat was in excess Of 28 pounds. At this point, the actual marginal value product was 98 paisa. Farmers are presently in a tran- sition phase, experimenting with dwarf wheats and heavier fer- tilizer use will expand considerably before an equilibrium is reachedl The black market has disappeared and nitrogen is now selling at slightly less than the controlled rate. The total cost to the farmer including transportation is between 50 and 55 paisa. If the wheat price support is maintained at 17.00 rupees per maund, nitrogen use will expand until the geometric mean dose is between 60 and 70 pounds. Even if the procure- ment price is reduced to 15.00 rupees, more than 50 pounds will be applied. As long as MVP > MFC, usage will expand if possible.J A reduction in product price or an increase in factor price will not automatically result in using less Of the factor. However, a potential surplus problem is again evident. Acreage of Mexican wheats exceeds 7.0 million irrigated acres in 1969-70 and is expected to increase still further. If the sample represents these irrigated acres well, then the geomet- ric mean Of nitrogen applications will eventually approach 60 pounds. Estimated yields at this level exceed one ton per acre so this acreage alone will contribute over 7,000,000 tons of product.” An additional 8,000,000 acres are grown with yields of one-third of a ton. Total production could exceed 9.5-10.0 million tons. Such a production increase could seri- ously strain government's capacity to buy and store wheat if 118 it occurred rapidly. In addition, huge amounts of rupees would be needed to procure the wheat that would be offered to govern- ment.~ A related problem is the 25 per cent subsidy on fertili- zer. Each pound of nitrogen costs the government 17 paisa on the average. If 60 pounds are made available for each of 7,000,000 acres, 71,000,000 rupees will be required for fer- tilizer subsidies alone. Such resources will probably not be made available.“ Either fertilizer use must be curtailed or the subsidy eliminated, or a combination of both. These calculations indicate that the subsidy could be entirely removed without constraining nitrogen use below 50 pounds of nutrient per acre. Were the subsidy to be removed immediately, fertilization Of dwarf wheat would continue to expand until they reached this level. At the same time, there would be some reduction in fertilization of dg§i_wheat. A matrix of expected fertilization levels can be devel- Oped for various prices of nitrogen and wheat. Table 7.3 presents such a matrix. The figures within the cells are geometric means of Optimum fertilization levels rounded to the nearest ten—pound unit. Summary Insights were obtained about several variations in wheat growing techniques. Line sowing was not found to produce sig- nificant results. Date of planting variations did contribute to explaining variation in yields. In the case of dwarf wheat, 119 TABLE 7.3.--Expected Fertilization Rates Under Different Prices of Wheat and Nitrogen. Price of Nitrogen Price (paisa/pound) Wheat (Rs./Md.) 4O 45 50 55 60 65 70 13.00 70 60 50 50 40 40 40 15.00 80 70 60 60 50 50 40 17.00 90 80 7O 7O 60 50 50 the data confirmed the planting date recommendations of the government. Significant differences were found in the reporting of yields to the two interviewers which seemed to reflect a down- ward bias when farmers respond to a Pakistani from outside their village. As a result, Objective estimation methods were strongly urged for all crops which are the subject of major provincial policies. Water applied to dwarf wheats was found to be very prof- itable at present levels. Physical limitations of water sup- ply seem tO be the effective constraint to further irrigations.” Profit motives seem strong enough in light of the data herein to induce rapid development of additional water supplies, pri- marily from tubewells. It is possible that private efforts to maximize profits through tubewell construction could lead to wheat surpluses that would burden the nation. Alternative price adjustments to forestall such an event were explored. The data showed that the expansion of nitrogen fertili- zation of dwarf wheats is far from complete. Again, adjustment 120 to the high profit point (MVP = MFC) could result in unneces- sary surpluses of grain.“ Price adjustments for wheat and fer- tilizer were explored that would allow equating MVP and MFC at various fertilizer doses. Finally, the probable effects of removing the subsidy on fertilizer were briefly discussed. This discussion can also be carried forward with slight modification to estimate the effects of currency devaluation on the quantity demanded of imported fertilizer. BIBLIOGRAPHY 121 BIBLIOGRAPHY Ahmed, M. M. Socio-Economic Objectives of the Fourth Five- Year Plan. Islamabad: Planning Commission, November, 1968. Aresvik, Oddvar. "Comments on Hunting Technical Services Limited 'Memorandum on the Tentative Comments made by the Senior Economic Advisor (Agric.) on LIP Criteria for the Development of Agriculture'," Lahore, Febru- ary, 1967. (mimeographed) Asian DevelOpment Bank. Asian Agricultural Survey. Seattle: University of Washington Press, 1969} Aziz, Ch. M. A. Fifty Years of Agricultural Education and Re- search at the: Punjab Agricultural College and ResearCh7 Institute, Lyallpur. Lahore: Department of Agriculture, 1960. Bean, Lee L., et al. Population Projections for Pakistan, 1960- 2000. Monographs in Economics of Development, NO. 17. Karachi: Pakistan Institute of Development Econom- ics, 1968. Borlaug, Norman E. "Progress Report on the Accelerated Wheat Improvement Program in West Pakistan," Karachi, November, 1965. (mimeographed) Central Statistical Office. Statistical Pocket Book of Pakistan. Karachi, 1967 and 1968. Darling, M. L. The Punjab Peasant in Prosperity and Debt. Bom- bay: OxfordiUniversity Press, 1947. Department of the Army. Area Handbook for Pakistan. DA Pam. NO. 550-48. Washington, D.C., OctOber, 1965. Directorate of Nutrition Survey and Research. Prelimina_y Re- port, West Pakistan Nutrition Survey, 1964466. Islamabad, 1968. Elkinton, Charles, and Sayeed, Aziz. Agriculture in Pakistan. Karachi: U.S.A.K.D., 1967. 122 123 Ezekiel, M., and Fox, K. A. Methods of Correlation and Re- gression Analysis. New York: John Wiley & Sons, 1959. Government of Pakistan. Report of the Food and Agricultural Commission. Karachi, 1960. J . 1960 Pakistan Census of Agriculture. Vol. II. Karachi, 1963. . Census of Agriculture, 1960. Karachi, 1964. . Department of Marketing_Intelligence and Agricul- tural Statistics, S.S.—Ix. Rawalpindi: Ministry of Agriculture and‘Works, 1967. . SurveyReport on Farm Power, Machinery and Equip- ment, S.S.-IX. Rawalpindi: Ministry of Agriculture and Works, 1967. . Review of Pricing Policy IncludingiSubsidies and Su ort Prices. Islamabad: Ministry of Agriculture and WorEs, 1969. Government of West Pakistan. Quarterly Bulletin of Performance Statistics. Vol. I, NO. 1. Lahore: Planning and DeveI: Opment Board, January, 1969, Table 34. . "Second Estimate of Wheat CrOp for Year 1968-69." The Gazette for West Pakistan. Lahore, 1969. Heady, Earl D., and Dillon, John L. Agricultural Production Functions. Ames, Iowa: Iowa State University Press, 1961. Hufbauer, G. C. "Cereal Consumption, Production, and Prices in West Pakistan." The Pakistan Development Review. (Summer, 1968). Hussain, Mian Anwar. A ricultural Research at the Ayub Agri- cultural Researc Institute, Lyallpur. Lahore: Govern- ment of West Pakistan, 1960. International Bank for Reconstruction and DevelOpment. Pro ram for the Development of Irrigation and Agriculture in West Pakistan. Vol. V. London, 1966. . Program for the Development 9f Irrigation and Ag- riculture in West Pakistan. VII. London, May, 1966. 124 Khan, Mahmood Hasan. The Role of Agriculture in Economic De— velOpment: A Case Study of Pakistan. Wageningen, Holland: Center for Agricultural Publications and Documentation, 1966. Khuda Bakhsh, Malik. "Accelerated Wheat Improvement in West Pakistan." Letter D.O. No. V-ER(71)DDPAl/64 to Mr. Haldore Hanson. Lahore, May 20, 1965. (Typewritten) Lewis, J. N. "Agricultural Price Policy for West Pakistan." Lahore: Harvard Advisory Group, April 8, 1968. (mimeo- graphed) . "Wheat Marketing in Pakistan and Export PrOSpects." Memo sent to Sarshar Ahmad Khan, Lahore, June 30, 1969. Mellor, John W. The Economics Of Agricultural Development. Ithaca, N. Y.: Cornell University Press, 1966. . "Toward a Theory of Agricultural Development." Agricultural Development and Econpmic Growth. Edited by Herman M. Southworth and Bruce F. Johnston. Ithaca, New York: Cornell University Press, 1967. Mexipak Wheat Yield and Quality Seed Appraisal Committee. Mexi- pak Wheat Yield and Quality Seed AppraisalL 1969. LyalI: pur, September, 1969. Munshi, Z. A., et al. "Annual Technical Report, Accelerated Wheat Improvement Program, West Pakistan, 1967-68," Lahore, July, 1968. (mimeographed) Narvaez, Ignacio. "Accelerated Wheat Improvement Program for Pakistan." Memo to Mr. Haldore Hanson, Ford Foundation Representative in Pakistan, Karachi, November 18, 1964. . "The Accelerated Wheat Improvement Program in West Pakistan," Lahore, February, 1969. (mimeographed) National Science Council of Pakistan. Protein Problem of Pakis- tan. Islamabad, 1968. Paddock, William and Paul. Famine, 1975! Boston: Little, Brown & Company, 1967. Planning Commission. The Third Five-Year Plan (1965-70). Karachi: Government of West Pakistan, 1965. . The Second Five-Year Plan (1960-65). Karachi: Government of Pakistan, 1966. 125 President's Science Advisory Committee. The World Food Problem. I. The White House, 1967. Qureshi, S. A., and Narvaez, Ignacio. "Annual Technical Report, Accelerated Wheat Improvement Program West Pakistan, 1965-66," Lahore, August, 1966. (mimeographed) . "Annual Technical Report, Accelerated Wheat Im- provement Program West Pakistan, 1966-67," Lahore, July, 1967. (mimeographed) . "Recommended Cultural Practices for Wheat Culti- vation in West Pakistan during the Rabi Season of 1967- 68," Lahore, 1967. (mimeographed) Anonymous. "Review of Pricing Policy Including Subsidies and Support Prices." Islamabad: Government Of West Pakistan, 1962. Singh, Kartar. Studies in the Cost of Production of Crops in Ehe Pun'ab. The Board Of Economic Inquiry, Punjab, Publication NO. 33. Lahore, 1934. Steel, Robert A. D., and Torrie, James H. Principles and Pro- cedures of Statistics. New York: McGraw-Hill Book CO., 1940. West Pakistan Agricultural Development Corporation. "The De- mand for Fertilizer in West Pakistan," Lahore, December, 1968. (mimeographed) Wharton, Clifton R., Jr. "Risk, Uncertainty and the Subsis- tency Farmer: Technological Innovation and Resistance to Change in the Context of Survival." New York: The Agri- cultural Development Council, 1968. (mimeographed) Wiser, William and Charlotte. Behind Mud Walls, 1930—1960. Berkeley, California: University of California Press, 1963. Witt, Lawrence W. "Cost of Production and All That." Islama- bad, August 8, 1969. (mimeographed) Wright, William G. "Critical Requirements of New Dwarf Wheat for Maximum Production." Paper presented at the FAO/ Rockefeller Foundation International Seminar on Wheat Improvement and Production, Lyallpur, Pakistan, March 26, 1968. APPENDICES 126 APPENDIX A LIST OF INDIGENOUS TERMS AND MEASURES Units Of Measure 1. Weight Maund = standard unit of measure for agricultural produce = 82.28 pounds Seer = 1/40 of a maund = 2.057 pounds Chhattank = 1/16 of a seer = 2.057 ounces Area Square = 27.5 acres. One of the basic units in which land was allocated during settlement of the canal colonies. Rectangle = 25 acres. One of the basic units in which land was allocated during settlement of the canal colonies. Often called a square in pOpular usage. Urdu Terminology bajra - pearl millet barani - land which has only natural precipitation as a source of moisture. A barani farmer is one who farms under these conditions. bund - elevated earthen ridge separating fields, usually with a footpath on top. chapati - flat bread which is the dietary staple in West Pakistan. It is prepared by grinding whole wheat, kneading with water, pressing into a pancake shape and frying on a dry griddle. gur - raw native sugar boiled down from sugarcane juice. 127 128 jowar - sorghum. kammi - a village artisan, usually landless, who supplies his services year 'round in return for payment in kind at harvest. Examples are cobbler, barber, carpenter, blacksmith, priest and watchman. kera - sowing method where seed is drOpped by hand dir- ectly into the furrow as it is plowed. Effects line sowing but little control of seed depth. kharif - summer growing season. Roughly from mid-May to mid-November loessi - summer beverage made from home-made yogurt, water and salt and chilled if ice is available. nambardar - village headman, usually a hereditary posi- tion. Responsible for representing the village in all contacts with the government and for some taxation and assessment within the village. persian well - large diameter, hand-dug well rarely ex— ceeding 30 feet in depth. Water is ele- vated by a continuous chain of buckets or pots which empty as they pass over a drive wheel. Powered by animals. pora - sowing method where seed is dribbled by hand into a funnel on the back of a native plow and falls through a tube into the furrow as it is plowed. Some control of seed depth is Obtained. rabi - winter growing season. Roughly from mid-November to mid-May. tehsildar — administrative officer responsible for record keeping, taxation and routine administration of the lower level administrative unit known as a tehsil. tubewell - a small diameter borehole (2-6 inches) lined with sieves which may exceed 200-300 feet in depth and is powered by an electric or diesel engine. Interviewer: Date: APPENDIX B THE QUESTIONNAIRE Tehsil: IDENTIF ICATION l. Farmer's Complete Name: 2. Vil Interviewer's Remarks lage Name or Chak Number (Print): FARM CHARACTERISTICS 3. How 4. How 5. How How many acres do you cultivate? much is owned rented much of your wheat crop does the zamindar much do you get? % 6. Does the zamindar help pay for: get? % a. Seeds Yes No How much? % . Fertilizer Yes No How much? % c. Water Yes No How much? % . Land Revenue Yes NO How much? % e. Others How much? % 7. How many acres of wheat are you growing? a. This year (1968/69) Dwarf Desi Total b. Last year (1967/68) Dwarf Desi Total c. 2 years ago (1966/67) Dwarf Desi Total d. 3 years ago (1965/66) Dwarf Desi Total 8. What other crops are you growing this rabi season? CrOp Acreage Other Crgps (specify); Acreage Berseem Sugarcane NOTE: If the sum of #7A plus #8 does not equal #3, find out why and enter other land uses here. 129 DWARF WHEAT GO to one of his fields of dwarf wheat. Draw a 130 9. 10. 11. 12. 13. 14. 15. 16. map on back of previous page that identifies the field's location. What variety is planted here? Mexipak White Mexipak Red Sonora Lerma Rojo Penjamo How big is this field acres kanals? What day did you plant this field? (circle one) October: 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 November: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 17 18 19 20 21 22 23 24 25 26 27 28 29 30 In preparing this field, did you use: Bullocks (GO to #16) Tractor (Go to #13) If #13 is Tractor, then: a. What kind of tractor? Don't know b. What model number? Don't know c. How many Horse Power? Don't know NOTE: If tractor is nearby, look at it after interview to verify a and b. How many hours did the tractor actually work in this field? Plowing Hours How many plowings Other jobs Hours (specify which ones) Total Hours If tractor was hired, how much was paid to prepare this field? If #12 is Bullocks, then: How many bullocks were used in preparing this field? NOTE: If answer given is an Odd number, find out why. 17. 18. 19. 20. 21. 22. 131 How many days/hours did each team of bullocks spend in this field on the following: Team 1 Team 2 Team 3 Plowing Days or Hours Sohaga Days or Hours Karah Days or Hours Other (specify) Days or Hours Other (specify) Days or Hours Total Days or Hours NOTE: If given in DAYS, then how many hours do the bullocks work each day? How much is each bullock worth? If you sold them today, how much would you get for them? How much do you feed each pair of bullocks each day? Seers Of Bhoosa Seers of berseem or green cut corn Seers of other (specify) How many hours/days did you work on this field yourself before planting? Doing what jobs? What other peOple also worked on this field before planting? Relative (R) Hours or Days Villager (V) spent on this Outsider (O) pr Performed field (circle oné) (filliin) (fill in?— a. R V O hr/da . R V O hr/da c. R V O hr/da . R V O hr/da . R V O hr/da How many seers of seed did you use on this field? How was seeding done? Broadcast - Rabi Drill Tractor-drawn Drill 23. 24. 25. 26. 27. 28. 29. DESI 132 Where did you get the seed? a. Bought from ADC Price Md. b. Bought from shOpkeeper Price Md. c. Obtained from neighbor Price Md. d. Produced yourself last year How much fertilizer did you put on this field before plant- ing? # Of bags on Weight of Kind whole field each bag Price_per bag How many total maunds of wheat did you grow last year? Dwarf: maunds produced maunds per acre Desi: maunds produced maunds total How many maunds did you sell: Dwarf: maunds sold Desi: maunds sold How many maunds were paid to others: To whom Maunds given To whom Maunds given Did you have a crop planted on this field during the fol- lowing seasons: a. Kharif 1968 What crop b. Rabi 1967/68 What crop c. Kharif 1967 What crop Will you plant a crop on this field during next Kharif: Yes No If yes, what crop? WHEAT Go to one of his desi fields. Draw map on back Of 30. Opposite page that identifies the field's location. How big is this field? acres kanals 31. 32. 33. 34. 35. 36. 37. 133 What day did you plant this field? (circle one) October: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 November: 1 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 In preparing this field, did you use: Bullocks (Go to #35) Tractor (GO to #33) If #32 is Tractor, then: How many hours did the tractor actually work in this field? Plowing Hours How many plowings? Other jobs Hours (specify) Total Hours Who drove the tractor? If tractor was hired, how much was paid to prepare this field? If #32 is Bullocks, then: How many bullocks were used in preparing this field? NOTE: If answer given is an Odd number, find out why. How many days/hours did each team of bullocks spend in this field on the following: Team 1 Team 2 Team 3 Plowing Days or Hours Sohaga Days or Hours Karah Days or Hours Other (specify) Days or Hours Other (specify) Days or Hours Total Days or Hours How many hours/days did you work on this field yourself be- fore planting? Doing what jobs? 38. 39. 40. 41. 42. 43. 134 What other people also worked on this field before planting? Relative (R) Hours or Days Villager (V) spent on this Outsider (0) Job Performed field (circle one) (fill inY’ (fill in) a. R V O hr/da b. R V O hr/da c. R V' O hr/da d. R V O hr/da e. R V O hr/da How many seers of seed did you use on this field? How was seeding done? Broadcast Rabi Drill Tractor-drawn Drill Where did you get the seed? a. Bought from ADC Price Md. b. Bought from shopkeeper Price Md. c. Obtained from neighbor Price Md. d. Produced yourself last year How much fertilizer did you put on this field before planting? # of bags on Weight of Kind whole field each bag Pricegper bag Did you have a crop planted on this field during the fol— lowing seasons: a. Kharif 1968 What crOp b. Ravi 1967/68 What crop c. Kharif 1967 What crop Will you plant a crop on this field during next Kharif? Yes NO If yes, what crop? 135 Condition of Field: Color Stand Other Date Farmer's Name Village Name/Chak NO. Tehsil DWARF WHEAT (Return to the same dwarf field you saw before.) 44. 45. 46. 47. 48. 49. 50. 51. 52. What source of water was used for this field? Canal Tubewell Persian Wheel Rainfall Only How many times did you irrigate? (circle one) 1 2 3 4 5 6 Did you irrigate before planting? Yes NO How much do you pay for water? Rs. Per Who does the work Of watering? How long does it take to water this field? Hrs.' Did your wheat receive any rain? Yes NO How many times did it rain? Light Rain Good Rain Heavy Rain Did you have any problems getting water? How much fertilizer did you put on this field? Price Per No. of Bags Kind Bag — Rs. On This Field When Applied 53. 54. 55. 56. 57. 58. 136 How was the fertilizer applied? By whom? Broadcast Pora Drill a) Where did you buy your fertilizer? (Town) (Type of Distributor) b) How far away is this? Miles Did you have any problems getting fertilizer? Yes If Yes, what kind Of problems? NO Did you pay cash or buy on credit? Cash Credit How much fertilizer did you put on dwarf wheat? a) Last Year (1967/68) bags/acre, No Dwarf Wheat b) Year before (1966/67) bags/acre, No Dwarf Wheat When do you expect to begin harvesting this field? (circle one) March 20 21 22 23 24 25 26 27 28 29 30 31 April 1 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30. May 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 POTOO. 137 Condition of Field: Color Stand Other Date Farmer's Name Village Name/Chak no. Tehsil DESI WHEAT (Return to the same desi field you saw before.) 59. What source of water was used for this field? Canal Tubewheel Persian Wheel Rainfall Only 60. How many times did you irrigate? (circle one) 1 2 3 4 5 6 61. Did you irrigate before planting? Yes NO 62. How much do you pay for water? Rs. Per 63. Who does the work of watering? 64. How long does it take to water this field? Hrs. 65. Did your wheat receive any rain? Yes NO How many times did it rain? Light Rain Good Rain Heavy Rain 66. Did you have any problems getting water? 67. How much fertilizer did you put on this field? Price Per No. Of Bags Kind Bag - Rs. on This Field When Applied 68. 69. 70. 71. 72. 73. 138 How was the fertilizer applied? By Whom? Broadcast Pora Drill a) Where did you buy your fertilizer? (Town) (Type of Distributor) b) How far away is this? Miles Did you have any problems getting fertilizer? Yes NO P? If yes, what kind of problems? Did you pay cash or buy on credit? Cash _____ Credit How much fertilizer did you put on desi wheat? a) Last year (1967/68) bags/acre, NO Desi Wheat b) Year before (1966/67) bags/acre, No Desi Wheat ____ When do you expect to begin harvesting this field? (circle one) March 20 21 22 23 24 25 26 27 28 29 30 31 April 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 May 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 139 Name Chak NOTE: Emphasize that you want to talk just about the two fields that were selected for study. His Dwarf field was acres kanals His Desi field was acres kanals HARVESTING 74. How many people helped cut your field of dwarf wheat? 75. How long did it take them to get that field cut? hrs. 76. How many peOple helped out your field of desi wheat? 77. How long did it take them to get that field cut? hrs. 78. How much did you pay each of them? per acre/day THRESHING 79. Did you thresh your wheat with? Bullocks or Machine 80. Did you thresh each of the two fields separately? Desi: Yes NO Dwarf: Yes No (If bullocks) 81. How many pairs of bullocks did you use? 82. How many people helped with threshing? 83. How long did it take to thresh your dwarf field? “ hrs. If not threshed separately, how many hours to thresh 1 acre of dwarf wheat? 84. Did you pay these people? Yes No Cash or Wheat 85. How much did you pay them? per day/maund 86. How long did it take to thresh your desi field? hrs. If not threshed separately, how many hours to thresh 1 acre Of desi wheat? 87. Did you have any difficulty getting enough people to cut and thresh your wheat? Yes No If yes, what kind of difficulty? 140 88. (If machine), What kind of machine? How many hours to thresh 1 acre? If rented, how much did it cost? WINNOWING 89. How many people helped winnor your wheat? 90. How long did it take to winnor your dwarf field? hrs. If not done separately, how long to winnow 1 acre of dwarf? 91. How long did it take to winnow your desi field? hrs. If not done separately, how long to winnow 1 acre of desi? 92. How much were these people paid? (per maund) OTHER 93. Did any one help move the grain to your home? Yes No 94. What is the price of Bhoosa in your village? In June: Rs. per maund In March: Rs. per maund 95. What is the price for renting farm land around this village? Rs. per acre/year YIELDS (These are the most important questions) On the Dwarf field we have been studying: 96. How many acres or kanals were harvested? 97. How many total maunds did you get from this field? that means mds/acre (CALCULATE AND CHECK) On the Desi farm we have been studying: 98. How many acres or kanals were harvested? 99. How many total maunds did you get from this field? That means mds./acre (CALCULATE AND CHECK) 141 MISCELLANEOUS Last year the Government price was 17 Rupees: 100. What did you sell your desi wheat for? Rs. Did not sell _ 101. What did you sell your Mexi wheat for? Rs. Did not sell Thisgyear the Government price is 15 Rupees: Have you sold anyTWheat yet? Yes No What price did you get for? Will you sell some later? 102. Desi Rs. Mexi Rs. Yes NO 103. What price will you expect to get for? Desi Rs. Mexi Rs. 104. Where will you sell your wheat? Govt. Procurement Center Neighbors in the Village Arhti Beopari Other 105. Why don't you sell directly to the Government for 15 rupees? If the Government reduces the official price to 10-12 rupees/ maund next year: 106. How much wheat will you plant? (acres) 107. What kind of wheat will you plant? Mexi 591 273 H-68 Dirk etc. Why? 108. Will you use more or less fertilizer? More Less How much more or less? 109. Will you use more or less water? More f Less Why? When we were here last you said you had irrigated your two fields as follows: Mexipak: irrigations Desi: irrigations uni—T-.. .. V 110. 111. 142 Did you put any more water on after our last visit? Yes No If yes, how many extra irrigations: Mexipak Desi Is fertilizer still sold on Black Market since Martial Law? Yes NO APPENDIX C GEOMETRIC MEANS OF INPUT LEVELS APPLIED BY SAMPLE FARMERS TO MEXICAN AND DWARF WHEATS (on per acre basis) Mexican Wheat Native Wheat Bullock-Pair-Hours 36.412 33.475 Seers of Seed 35.000 34.078 Number of Irrigations 7.189 6.166 Pounds of Nitrogen 28.371 14.496 Pounds of Phosphate 1.328 1.107 Farm Yard Manurel 2.204 2.925 Yields 19.708 15.363 1 In units Of 5 maunds or 411 pounds. 143 $11.- ....1. .3 IVER I | W ”infill?!“ 1 16 4 6 ”'CIITI'fiHflJNIiflflfljIfl‘flW 307