"f " « ABSTRACT A MICRO-ECONOMIC ANALYSIS OF SMALLHOLDER RESPONSE TO HIGH-YIELDING VARIETIES 0F WHEAT IN NEST PAKISTAN by Refugio I. Rochin From 1964-65 to 1969-70, high-yielding varieties of wheat, characterized by dwarfness and the ability to use water and fertili- zer more efficiently than traditional (desi) varieties of wheat, spread from almost zero to forty—four percent of the wheat acreage (15 million acres) in West Pakistan. Over the same span of time, the average annual wheat yield for West Pakistan rose by 35 percent to a new record for 1969-70 of 12.9 maunds per acre, a remarkable achieve- ment commonly called the Green Revolution. According to government estimates, dwarf wheats have not spread significantly to barani (rainfed) areas which have large numbers of farm- ers with smallholdings. In 1968-69, the latest year of official re- corded figures, dwarf wheats were found on only 0.16 million acres of the 3.8 million acres of barani land. With modest wheat research on barani land and practically no knowledge of barani farming and condi- tions, it was presumed that barani smallholders would be unable to parti- cipate in the benefits of the Green Revolution, accentuating the socio- economic problems of the country. Field surveys and interviews with 226 farmers were conducted in Hazara District, a district with about 90 percent of its 1.5 million people living on small farms (averaging 5 acres in size) and producing wheat on barani land primarily fOr subsistence consumption. Compared Riel Ail ‘ “9"]; P Refugio I. Rochin to other barani areas, Hazara receives relatively higher rainfall, between 30 to 60 inches annually, with most falling during the summer monsoon. The principle objective of the surveys was to determine the extent of dwarf wheat adoption in Hazara District. In the survey loca- tions, it was found that there has been rapid diffusion of the new varieties and chemical fertilizer by farmer to farmer exchanges. Analysis of the pattern of dwarf wheat adoption shows that, overall, smallholder barani farmers are responsive to profitable innovations and will make rapid economic adjustments in resource allocation with high- yielding varieties of seed. The findings indicate that smallholders will rarely sow more than 30 percent of their acreage with a new variety during the initial year of experimentation. Analysis of the "extended family," the pattern of off-farm migra- tion and non-farm employment lead to the conclusion that the lack of credit does not appear to be a limiting constraint on the ability of the smallholders to adopt high-yielding seed, since they have other sources of farm income and migrant remittances which enable them to experiment with "risky" innovations. Respondents in Hazara adopted dwarf wheats because they (1) gave consistently better yields over desi varieties, (2) fit the cr0pping pattern of the farmer and grew in a faster time than desi varieties, (3) were subject to experimentation on a small scale, (4) were not complex in relation to desi wheat with respect to the way they were planted and cared for, and (5) had other positive features such as beards which [arotected grain against birds. All of these factors, shown in order of descending importance suggest the major characteristics that new Refugio I. Rochin varieties must have before they will be readily adopted by smallholders. Communication variables were also important in diffusing the new varieties, namely, mass media channels, interpersonal channels, and demonstration plots. The survey findings show that: (l) The farmers who become aware of innovations before others in their communities will be those who: (a) listen to the radio more often than others and who have more contacts with extension agents and fertilizer sales representatives, (b) own most of their land and have comparatively larger extended fam- ilies (12 to 15 members), and (c) own medium size farms (4 to 5 acres). Health (the combination of farm area and cash income) is not a major determinant of early awareness of innovations. (2) The farmers to adopt innovations earlier than others in their communities will be those who: (a) score high on awareness and interpersonal cosmopolite contacts, (b) own farms of 5 acres or more, have large families, and (c) are generally wealthier than others in their vicinity. Measurement of the economic impact of dwarf wheats (by way of Cobb-Douglas production functions and budgetary analysis) on smallholder barani farms has shown that: (l) Smallholders with dwarf wheats are producing on a new production function with noticeable changes being made in the factor proportions such that (2) more output is forthcoming and more labor, fertilizer, and seed is used with dwarf varieties. Also significant is the finding that the marginal product of labor is zero for both types of wheat production, desi and dwarf. (3) Those who grew dwarf wheat in 1969/70 earned approximately three times more per acre than those who grew only desi wheat. (4) Partial budgeting analysis indicated that the relative profitability with dwarf wheats (compared to desi wheats) was virtually unchanged when the price of fertilizer was in; Var Catl Refugio I. Rochin increased by 25 percent, in one case, and when the yield of dwarfs (and straw) was reduced by 20 percent, in another case, all other factors held constant. Overall, the analysis showed that the relative income position of barani smallholders improved for those who were able to use the new dwarf varieties. It also showed the potential for barani small- holders to finance agricultural improvements themselves. These findings suggest that government should design food-grain improvement programs to include farmers with barani smallholdings. The response to a program to diffuse high-yielding varieties should be significant enough to control for increasing disparities in rural income between farmers with irrigated holdings and farmers with small unirrigated plots of land. By increasing food production per capita in rural barani areas, the need to leave the farm (for the many migrants that are already doing so) should be less great, given the present growth in population. Furthermore, information on the relative importance of various types of communication channels and their impact on different categories of farmers are useful for implementing programs designed to diffuse new varieties to similar economic settings. A MICRO-ECONOMIC ANALYSIS OF SMALLHOLDER RESPONSE TO HIGH-YIELDING VARIETIES 0F WHEAT IN WEST PAKISTAN By Refugio I. Rochin A THESIS Submitted To Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1971 ACKNOWLEDGEMENTS Sincere gratitude is acknowledged to the many persons of Pakistan and, in particular, to the many farmers of Hazara District who made this study possible. In addition, the author acknowledges the cooperation received from Mr. Malik Fazel Dad Khan, Director of Agriculture Extension and Research, Dr. Abdul Matin, Chief Economist of the Planning and Development Department and Dr. C.M. Sharif, Economist of the Pakistan Academy for Rural Development, in the North West Frontier Province. Special gratitude goes to my com- panion and interpreter, Mr. Mohammed Siddiq of Village Malot Sattian, who as a barani farm migrant himself, contributed as much insight into the lives and manners of his fellow barani smallholders as into their agriculture. General acknowledgement is due The Ford Foundation for making it possible for me and my family to live in Islamabad, Pakistan, where I worked as a Training Associate in Agricultural Economics, from November, 1969 through May, 1971. I am greatly indebted for the encouragement and questionning interests of Messrs. Robert Edwards and Robert D. Havener, Representative and former Agriculture Program Advisor, respectively, of The Ford Foundation in Pakistan. To Havener goes my earnest thanks for educating me on Pakistan's agriculture. In no way, however, are The Ford Foundation, Edwards or Havener responsible for any errors of fact, logic or recommendations which enter into this thesis. The author alone is responsible for any of the views and contents presented. A very special acknowledgement goes to Dr. Lawrence W. Witt ii for his patience and "long distance" guidance via letters from the U.S. with detailed and valuable comments pertaining to the formula- tion of this thesis. Moreover, I am grateful for his kind enthusiasm shown while serving as academic advisor during my graduate study at Michigan State University. At Michigan State University, I am also sincerely indebted to others: to Dr. Carl K. Eicher for his suggestions which helped considerably in editing a first draft of the dissertation, and to Drs. Warren Vincent, Everett M. Rogers, and William Haley who read preliminary chapters of the study and made insightful comments. Last but not least, loving appreciation is due my wife, Isaura, for her patience during the year of the thesis and help in editing, and Chaguita, my daughter, who kept me cheerful even during the mdst rigorous writing days. II. III. IV. TABLE OF CONTENTS SETTING AND PROBLEM INVESTIGATED Setting Food Self-Sufficiency Program The Problem Research Agenda Objectives Chapter Brief SALIENT ASPECTS OF WEST PAKISTAN'S BARANI AGRICULTURE AND WHEAT PRODUCTION Introduction "The Barani" Physical Environment Extent of Irrigation People, Farms and Tenure Land Use Cropping Patterns and Practices Development Alternatives Wheat Performance in "the Barani" BACKGROUND DESCRIPTION OF HAZARA DISTRICT AND SURVEY LOCATIONS Introduction District Selection Major Characteristics of Hazara District Survey Locations Lora and Oghi Thanas Note on Rainfall METHOD OF DATA COLLECTION AND THE GENERAL CHARACTERISTICS OF HAZARA'S FARMS Introduction Method of Data Collection Precautionary Measures Sampling Technique and Respondent Selection Survey Implementation Handling First and Second Round Interviews Combining Lora and Oghi Samples General Characteristics of Hazara's Farms iv 42 42 42 47 49 53 54 V. DIFFUSION AND ADOPTION OF DWARF WHEATS IN HAZARA DISTRICT Introduction Historical Antecedents Sources of Dwarf Wheats in Hazara Clarification of Intent Section 1. Survey Results: General Overview Magnitude of Diffusion and Adoption Section II. Why Dwarf Wheats Were Adopted and the Risks Involved Characteristics of Dwarf Wheats as Perceived by Farmers Dwarf and Desi Wheat Yields Risk and Uncertainty with Dwarfs A New Production Function Section III. Communication and Economic Variables ,J/ Associated with Dwarf Wheat Diffusion and Adoption Channels of Communication Correlation Analysis VI. A PRELIMINARY VIEW OF NON-FARM EMPLOYMENT AND MIGRATION OF HAZARANS " Introduction Basic Considerations The Family "Family" Defined "Family" Influence Propositions Derived From Survey Findings Some Implications VII. THE ECONOMICS OF WHEAT PRODUCTION ON BARANI LAND IN HAZARA DISTRICT Introduction Section I. The Impact of Dwarf Wheats on Resource Use Model of Production Specification of Functional Form Sample Size Production Variables Overview of the Inputs Used to Produce Dwarf and Desi Wheat Estimation of Model Parameters Analysis of Variance Production Elasticities Marginal Productivities Analysis of Resource Use A Note on the Optimal Use of Inputs Land Labor 99 99 100 102 102 104 106 120 123 123 123 123 123 125 126 130 131 131 131 135 136 136 137 138 Seed Rates Fertilizer Section II. Costs and Returns to Dwarf and Desi Wheat Production Partial Budgeting Analysis Total Expenses and Input Pricing Total Revenue and Product Pricing Budget Analysis of Costs and Returns VIII.SUMMARY AND RECOMMENDATIONS Summary Background Dwarf Wheat Diffusion and Adoption Migration and Non-Farm Employment The Economics of Barani Wheat Production General Conclusions Concerning the Fourth Five Year Period Recommendations For More Immediate Consideration For Later Consideration BIBLIOGRAPHY APPENDIX A GLOSSARY OF PAKISTANI TERMS AND MEASURES APPENDIX B DEFINITIONS OF "ZERO-ORDER" AND "MULTIPLE" CORRELATIONAL ANALYSIS APPENDIX C FACSIMILE OF FIELD SURVEY INTERVIEW SCHEDULE vi 139 140 143 143 145 150 152 156 156 156 159 163 164 167 169 169 174 177 182 184 186 5.7 5.8 LIST OF TABLES CHRONOLOGY OF DWARF WHEAT DIFFUSION IN WEST PAKISTAN: 1962-1969 CULTIVATED ACREAGE, AREA IRRIGATED, AND PERCENT OF CULTIVATED IRRIGATED ACREAGE: 1967-68 WHEAT AREA AND YIELDS FOR IRRIGATED AND BARANI WHEAT IN WEST PAKISTAN TRIBAL BREAKDOWN AND LANGUAGES SPOKEN PERCENTAGE OF TOTAL AREA UNDER CROP BY TEHSIL AND SEASON IN HAZARA DISTRICT HAZARA DISTRICT'S SHARE OF WHEAT AND MAIZE PRODUCTION IN WEST PAKISTAN: 1969-70 COMPARISON OF MEAN VALUES OF ANSWERS TO THE SAME QUESTIONS ASKED DURING FIRST AND SECOND ROUND INTERVIEWS, HAZARA, 1970 DISTRIBUTION OF SAMPLE, BY TENURE AND SIZE OF AREA CULTIVATED, 1970: HAZARA DISTRICT DISTRIBUTION OF FARM SIZE IN WEST PAKISTAN DWARF WHEAT AND CHEMICAL FERTILIZER USED FOR THE FIRST TIME PER RESPONDENT BY YEAR: HAZARA DISTRICT, 1970 COMPARISON BETWEEN THE USE OF DWARF AND DESI WHEAT WITH FERTILIZER USAGE, 1969/70 DWARF WHEAT AREA SOWN AND PERCENTAGE OF TOTAL WHEAT AREA BY YEAR IN HAZARA DISTRICT: 1966/67-1970/71 COMPARATIVE YIELDS FOR DESI AND DWARF WHEAT ON BARANI LAND, HAZARA DISTRICT YIELDS 0F DWARF AND DESI WHEAT VARIETIES IN HAZARA DISTRICT, 1969/7O AVERAGE YIELDS AND INTENSITY OF ADOPTION, HAZARA DISTRICT REPORTED INPUT LEVELS FOR DWARF AND DESI WHEATS IN HAZARA DISTRICT, 1969/70 vii 18 27 34 35 36 53 54 64 66 69 70 76 79 79 80 5.9 5.1 6.1 6.2 5.6 6.7 5.5 6.9 5.1 7.1 .10 .11 .12 .10 CHANNELS OF COMMUNICATION WHICH FIRST INFORMED RESPONDENTS OF DWARF WHEATS, HAZARA, 1969/70 CORRELATES AND MEAN VALUES OF INDEPENDENT VARIABLES (BY TIME PERIOD) ASSOCIATED WITH AWARENESS CORRELATES AND MEAN VALUES OF INDEPENDENT VARIABLES (BY TIME PERIOD) ASSOCIATED WITH INNOVATIVENESS EXPLAINING INNOVATIVENESS--MULTIPLE CORRELATION MIGRANT CHARACTERISTICS DISTRIBUTION OF MIGRANTS BY SIZE OF CULTIVATED ACREAGE ON "THE FARM" SIZE OF EXTENDED FAMILY OF MIGRANT LABOR MEAN VALUES OF FACTORS FOR FAMILIES WITH MIGRANTS AND FOR FAMILIES WITHOUT MIGRANTS: HAZARA, 1970 MARITAL STATUS OF FAMILY MEMBERS WORKING OUTSIDE THE VILLAGE: OGHI AND LORA THANAS, JUNE, 1970 PERCENTAGE BREAKDOWN BY DISTANCE OF EMPLOYMENT FROM VILLAGE OF VILLAGERS WORKING OUTSIDE THE VILLAGE, JUNE, 1970 PLACES OF MIGRANT EMPLOYMENT AND DISTANCE TO VILLAGE LOCATION OF MIGRANTS FROM FAMILIES WITH TWO OR MORE FAMILY MEMBERS OFF THE FARM EMPLOYMENT STATUS AND JOBS HELD BY MIGRANTS FROM HAZARA DISTRICT, 1970 CASH INFLOW TO VILLAGER INTERVIEWED FROM RELATION WORKING OUTSIDE THE VILLAGE SUMMARY OF INPUTS AND THEIR GEOMETRIC MEAN VALUES FOR DWARF AND DESI WHEAT PRODUCTION: HAZARA DISTRICT ESTIMATED ELASTICITIES WITH THE COBB-DOUGLAS PRODUCTION FUNCTION MATRIX 0F SIMPLE "ZEROéORDER" CORRELATIONS ESTIMATED MARGINAL PRODUCTIVITIES OF INPUTS USED TO PRODUCE BARANI WHEATS: HAZARA DISTRICT, 1969/70 viii 83 9O 93 96 107 107 108 109 111 112 113 115 116 119 130 132 134 136 7.5 7.6 7.7 7.8 7.9 TOTAL AND MARGINAL PRODUCT CURVES WITH MVPS FOR NITROGEN AND PHOSPHATE FERTILIZER APPLIED TO DWARF WHEATS AVERAGE QUANTITY OF INPUTS USED PER ACRE AND THE AVERAGE PRICE PER UNIT: HAZARA DISTRICT, 1969/70 MAN-DAYS EMPLOYED PER ACRE FOR DWARF AND DESI WHEAT PRODUCTION AND PROCESSING: HAZARA, 1969/70 COSTS OF BULLOCK TEAM IN HAZARA DISTRICT, 1970 INCREMENTAL COSTS AND RETURNS FOR DWARF AND DESI WHEAT IN HAZARA DISTRICT, 1969-7O ix 142 146 148 150 153 r0 0 LIST OF FIGURES TYPICAL CROPPING PATTERN, NORTHERN ZONE RAINFALL AREA: OVERALL CROPPING INTENSITY--92% TYPICAL CROPPING PATTERN, CENTRAL PUNJAB CANAL PLUS TUBEWELL: OVERALL CROPPING INTENSITY--150% VARIATION IN ANNUAL PRECIPITATION: LORA AND OGHI, HAZARA DISTRICT, 1961-1970 AVERAGE MONTHLY PRECIPITATION FOR 1961-67 AND AUG. 1, 1969 TO JULY 31, 1970: LORA AND OGHI, HAZARA DISTRICT ILLUSTRATION OF CROSS-COUSIN MARRIAGE A MODIFIED COBB-DOUGLAS ISOQUANT 22 23 4O 41 103 129 .1. a nr $25 a Q n\~ In Ink... 1 .$ v CHAPTER I SETTING AND PROBLEM INVESTIGATED m Wheat is the most widely consumed and cultivated crop in West Pakistan affecting nearly all its people. On the average, West Pakistanis obtain 60 percent of their caloric intake from wheat products. 1! Wheat is regularly grown during the winter months from October to May on 14 to 15 million acres of land which com- prises nearly 60 percent of the season's total cropped acreage. 0f the wheat acreage, 70 percent is irrigated and 30 percent is rainfed or barani. 2! Up to 42 percent of the wheat is sown on farms with 12.5 or less cropped acres, commonly referred to as smallholder farms. In addition, 65 percent of West Pakistan's 6 million farms engage in wheat production. From the time of Independence in 1947 until 1967, wheat yields averaged around 8.8 maunds per acre in West Pakistan, according ;/ to official estimates. However, the rate of change in wheat 1/ T'Estimates of per capita cereal consumption are found in: Hufbauer, G.C. "Cereal Consumption, Production, and Prices in West Pakistan: The Pakistan Development Review, Sumner 1968; pp. 288-306. ?_/ Barani translated from Urdu means "depending on rainfall." The term is commonly used among agriculturalists in West Pakistan. This is discussed in Chapter II of this volume. 3 'lGovernment of Pakistan, Ministry of Agriculture & Works, .Agricultural Statistics of Pakistan, and Government of West Pak- listan, Season and—crop Reports. (The glossary in the ap endix provides the conversions fbr the measurements used in this study? BCI‘E HIS deca the (ON FOOT 55:71 53;! fl/ acreage, fell below the rate of increase in pOpulation (which was increasing at about 2.8 percent per annum) during the last two decades.§/ To meet deficits in food-grain supplies, West Pakistan imported 3.3 million tons of wheat from 1955-56 to 1959-60 (over the First Plan), 7.0 million tons of wheat from 1960-61 to 1964-65 (over the Second Plan), and approximately 5.3 million tons of wheat from 1965-66 to 1969-70 (over the Third P1an).§/ To make matters apparently worse, drought conditions in 1965 and 1966 lowered acreage and wheat yields in West Pakistan. This heightened the dire fore- ’ castsZ/ that it would be increasingly difficult to close the food- population gap without massive imports of food-grains. At the same time, however, the imported dwarf wheats were providing higher yields in test plots and offered hope that Pakistan's desire to remove dependence on food-grain imports could be achieved. "Dwarf wheats" refers to the new, high-yielding varieties of wheat which were developed in Mexico and imported and multiplied in 4 _/Eckert, Jerry Bruce. The Impact of Dwarf Wheats on Resource Productivity in West PakistanTS’Punjab. TUMEUbTiShed Ph.D. disser- tation, Michigan State University, 1970; pp. 26-31. 5/ T Stern, Joseph J. and Walter P. Falcon. Growth and Develop- ments in Pakistan: 1955-1969. Harvard University, Center for Inter- nationET Affairs, OccaSTfihal Paper Number 23, April 1970: pp. 8-13. 6/ Government of Pakistan Ministry of Agriculture & Works, Aaricultural Statistics of Pakistan. It should be noted that West Ristan essentially aEhTeved’SeTTisufficiency in wheat production by 1968-69. However, limited supplies were still imported to meet deficit conditions in East Pakistan. In 1969-70, West Pakistan exported wheat to East Pakistan but later turned around and imported approximately 200,000 tons of wheat to maintain a small wheat reserve. President' 5 Science Advisory Committee, The World Food Problem, The White House, Volume II, May 1967. Tall 01'] he) 5 1 West Pakistan. Dwarf wheats are characterized by: (l) relatively short genetic height (30-35 inches) which enables the plant to ab- sorb heavy applications of fertilizer without lodging from excessive vegetative growth, (2) high tillering with stiff stems which adds to the sturdiness of the plant, helps it to fight wind and heavy rain damage and at the same time facilitates harvesting, and (3) erect leaves which maximize the leaf area exposed to sunlight hence enhancing photosynthesis. Notably, the same structural character- istics were used in developing Gaines, the wheat variety which holds the world's record yield of 216 bushels per acre (about 162 maunds per acre). At the time dwarf wheats were introduced, no traditional (desi) variety of wheat in Pakistan had these high-yielding genetic characteristics. §j Food Self-Sufficiency Program In an attempt to bolster food production in West Pakistan, the Food Self-Sufficiency Program (initiated in 1967) had as its main focus the achievement of self-sufficiency in wheat production based on the cultivation of imported varieties of dwarf wheat on irrigated holdings (refer to Table l). 87* _ The general agronomical and morphological characteristics of the desi varieties are documented in a paper Dy S.A. Anwar Abidi, "The Development of New Wheat Varieties in Pakistan." Conference on Food Production Increase in West Pakistan: Problems and Effects. PakiStan Acaaemy fer Rural Developnent} Peshawar, AprTl 22-23, 1970. The more common varieties referred to as desi are C-591, Dirk and C-273. here stin Der t3 (7 "t" (,1 ) (4) ' Q: -j I 1'! -—¢, .__1 __ .4 1 CD The specific strategy called for a combination of factors, heretofore untried in West Pakistan history. Point by point it 9/ stipulated:'— (a) Bringing 4 million acres (out of the present total irri ated wheat acreage of about 8 million acres) unaer MExican wheat, giving it an average fertilizer application of 4 bags per acre in terms of ammonium sulphate and providing each acre one to two additional waterings (average increase in water delta of 33%). (b) Providing additional fertilizer for two million irrigated acres under indigenous wheat at the rate of one bag per acre in terms of ammonium sulphate. (c) Increasing irrigated wheat acreage by 1.5 million acres. (d) Making organizational changes, increasing credit availability and devising a system of incentives and subsidies to ensure that input targets provided , above are in fact met. (e) Allocating funds in A.D.P. for water and agricultural program of third plan on a priority basis, specially encouraging schemes linked to wheat self-sufficiency. Furthermore, production targets were set for an annual three percent increase in wheat production in the districts of the Indus T_D_/ Plains which are largely irrigated. In a period of five years, beginning in 1964-65 to 1969-70, dwarf wheats spread from almost zero to forty-four percent of the wheat acreage in West Pakistan. Over the same span of time, the 9/ . T Government of West Pakistan, Planning and Development Department of (Project Wing). Implementation Plan for the West Pakistan Food Self-Sufficiency Pro ramme, Printed by the Superintendent, Government ‘Printing Lahore, West PaEistan. August 1967, Chapter I (underlined Dy the author). A.D.P. apparently stands for Agricultural Development Program. . 351/ The districts are namely: Sargodha, Lyallpur, Gujrat, Jhang, Muzaffargarh, Sheikhupura, Sialkot, Gujranwala, Multan, Sahiwal, Lahore, Bahawalpur, Rahimyar Khan, Bahawalnagar, Khaipur and Nawabshah Districts. ' (IVE an average annual wheat yield for West Pakistan rose by 35 percent to a new record for 1969-70 of 12.9 maunds per acre. This record of 1 / events, as shown in Table 1, is commonly called the Green Revolution[__ The Problem According to official releases, rainfed (barani) wheat areas have been unable to substantially adopt the high-yielding dwarf wheats. In 1968-69, the latest year of official recorded figures, dwarf wheats were found on only 0.16 million acres of the 3.8 million acres of barani wheat land.l§/ Most conclusions about the inability to diffuse dwarf wheats on barani acreage were drawn from speculative reasoning based on limited knowledge about the farmers who cultivate barani land and 13/ their resource endowment. In the words of a noted national planner;-—' 11f More complete treatment of the diffusion of dwarf wheats in the Indus Plains of West Pakistan is found in: Eckert, Jerry B. Ihg_ Impact of Dwarf Wheats on Resource Productivity in West Pakistan's Punjab. Op, 915, Chapter II; pp. 12-36. 2/ __'Source: Government of West Pakistan, Planning and Development Department, Bureau of Statistics, Rabi 1969-69 Crop Acreage of West Pakistan: 8y Districts and Tehsils, Table I. This is the only known publication, which Ts compiled By the village revenue collectors (Patwaris), that gives an estimate of irrigated and barani acreage covered Dy dwarf wheat. In addition, The Fifth Annual Technical Report: Accelerated Wheat Improvement Program in STnH, Punjab & N.W.F.P. 1969-70, (Dy Munshi, et. a1.) states that no more than 0.2 million acres were covered with Mexipak and other short varieties during 1969-70; p. 11, Table 2. As of June 1, 1970, West Pakistan was sub-divided into four Provinces: The Punjab, The Sind, Baluchistan, and the North West Frontier (N.W.F.P.) This break-up of the one unit temporarily curtailed official releases on crop acreage and produc- tion. At the time of the writing, no further reports were available on the spread of dwarf wheats in West Pakistan. 13/ —_'Aziz, Sartaj. "Problems and Prospects of the Green Revolution. TheiPlanning Conmission, Mimeographed, May, 1970. Islamabad, West Pak stan. tF~**>.‘m1-*3 kc ‘ca.igac U: ks1i~ IQ‘WE ‘\A‘LQ\¢CI\ ccxl.‘ ‘\qI“CEEh \u‘5tu‘cflfi‘c \‘qu 6? 6L .0) 5.0... u m t- ... c . mum SOL U m. .3: m I - ll ,- KillllllllllllllKlTl 5151*! 9125! ll... . -l ----l kl ELI (NT - 1E1 I r 1 I, ii stay - IMwAUN. - u z< - an - .«Ll‘nh ~ «mi-3 Z n ZAU~ ....\d § \ VN- s 9‘ fies! \V\p\\‘- \qu L‘s-s VNJER v~\\\ In I \ ~u~o.\\ A.mn_ ocm.mv cop one u mncams www.mm xpmpmerxogaqo use .mn_ www.mm u teams mco « .oupxmz humblwm cowxm: .ummsu can m~wmz mo mupmw>imcwmmmgocH ugmzoh mmmgmoca co "mo-momp Hmommm p>zzmu mecou acmem>ogaEH news: new «Nsz Fucoppmcgmpcm .N.N as: .3 .a 2:. _.N 223 .8 .a 31.0. .mm .238 .3. ”883m mm.m_ o.¢¢ ooo.ovw.m ON\mmmP u mom— mo.F~ o.om ooo.~mm.m mo\mom~ I mom. mo.P_ o.m_ ooo.mmm.m mo\~mmp meow owgume omn._¢ Romp om.w m.o ooo.omm mo\momp mcou uwgume om comp e~.m _.o coo.mp mm\momp meow creams omm mmmp mm.m _wc o_ mm\¢mm_ . vamp mm.m pw: PP: ¢m\momp Amopwx oomv mmpaEmm acmewcmaxm mmm_ 8.8 7:. E. 85.2 A883 8,83 2823: $2 “chummwcsq moc< poms: Peach pawn: me> ourxmz Soc; gum» upmw> yawn: mo ucmocmg m< mcuzo cuvz umgm>ou qogu nmuLonEH auwucmao vmmm mmmgm>< mmc< «was: Mmmzo mmmmcu< vmumepmml momp-moap nzuoqozom=u H ~4m

M~J~ 0 PD 45 40 35 FIG-UT LEVEL OF PRECIPITATION IN INCHES 4O 10 Year Ave. ---4> Wheat "3‘7‘1'6' 35 ,—-/' 35 93 - period 1969/7O I I A I I I I I I I 61 62 63 64 65 66 67 68 69 7O LORA 6° ' 58.59, 56.76 55 - 8 Year Ave. "" ' ----- 4- 50 b 45 F 2.82 40 - 36.17 35 . wheit/ a. .......... g eriod 969/70 I I I I I . . j L I 61 62 63 64 65 66 67 -68 69 60 FIGURE 3.1 VARIATION IN ANNUAL PRECIPITATION: LORA AND OGHI, HAZARA DISTRICT, 1961-1970. nae-.UE~ E. {deaudwlvuflkl \fihquOE ODOLU>< Average Monthly Precipitation in Inches 41 9 r 8 Kauai... ,,___1. 7 h 5 . FL OGHI ,ZAQQ, ‘ ' «I 72 g ....... 4 1 s 9 53 '1' n6". 1.33 O l 14. . .4. 4 l . . - ”9&9; J F M A M J J A S 0 N D 10 ’ 9 ’ 5 L936. 3 r "Zanh 7 ' 6.34 p. ...... .4 5 .5.5.4. 5 __1 4 3 , F"— ' 2 1 t . 0 ‘ . $0435. 1.0.»3? ‘ :9:.3.Q: ‘ Ja'A6’: 0&9 J F M A M J J A S O N D FIGURE 3.2 AVERAGE MONTHLY PRECIPITATION FOR 1961-67 (in solid lines) AND AUG. 1, 1969 TO JULY 31, 1970 (in broken lines): LORA AND OGHI, HAZARA DISTRICT CHAPTE R IV METHOD OF DATA COLLECTION AND THE GENERAL CHARACTERISTICS OF HAZARA'S FARMS Introduction The central purpose of thise chapter is to discuss the method of data collection and the precautionary measures which were taken to encourage reliable answers from the sample of respondents who were interviewed. The sampling technique which was employed and the way the survey was implemented are also covered in this chapter. The final portion of this chapter depicts the general characteristics of Hazara's farms. Method of Data Collection Interviewing and data collection on the micro-level in Pakistan is an art in itself, involving new and different terrains and socio- cultural factors from one place to another. Eckert, in his recent study on the impact of dwarf wheats in the Punjab of West Pakistan found problems with: (l) eliciting "straight" answers to direct questions, (2) dealing with the intricacies of Punjab agriculture, (3) learning to question farmers who had difficulty recalling and measuring inputs and outputs and (4) finding the "right respondent" in a society which1/ tends to make spokesmen of the better infOrmed and wealthier farmers. Such difficulties emphasize the need for well planned and executed I! J.B. Eckert, gg, cit., pp. 37-46. Eckert refers to the "right respondent" as the farmer who actually worked the land and managed the farm operations, whether an owner or tenant. 42 43 methods of data collection. They also stress the importance of knowing the culture well enough to conduct survey work on a proper footing. In Hazara, for example, village muslim women practice gargah, which means that they conceal themselves against all outsiders and sometimes against local villagers after having reached puberty. Pg:§ah_was said to have originated with the prophet Muhammed and continues strongly to this day in both the urban and rural areas of West Pakistan. However, the practice of gargah_makes it difficult for the outsider or foreigner to move about freely in and around village communities until the women have been warned and/or rushed into their quarters. Getting around the problem requires someone familiar with the area to make an announcement within the village that a visitor is coming. This is easily done by telling male adults or children that the village will be entered. Failure to understand this custom can make interviewing in a Hazara community rather difficult and uncomfortable. However, once one gets by the fOrmality, movement within the village is unrestricted. Another custom that prevails throughout Hazara is household courtesy. Most visitors will be invited into a house for a cup of tea. "Distinguished" visitors or burra sahibs will in most cases be given a complete dinner. Often the interviewing team went through as many as 15 cups of tea a day. Acceptance of a cup of tea is an important way of showing respect and easing the relationship between the guest and the host. It does, however, take time and one must be prepared to sit for an hour or more if necessary. Therefore, the survey method of collecting data had to be adapted to fit the socio-cultural context of Hazara. Since data collection is a major component of research of the type presented here, it is worthwhile Ten tail 4:0 ti- )1. Mu. I \J (*1 LA. ) (I) (A) - I I <4 44 reviewing the precautionary measures that were taken fOr the implemen- tation of the field survey. Precautionary Measures Before any interviews were conducted, all of the above caveats were taken into consideration. In addition, the author consulted with members of the Pakistan Academy for Rural Development in Peshawar who have field research experience in and around the North West Frontier Province. They advised the author on problems that could be encountered in local field surveys and ways to get around them. All things considered, it seemed imperative that one of the first musts would be to find ways to avoid being mistaken for a revenue collector, a policeman or a government official with the intent of seizing land, for example. In short, special measures called for ways of reducing heterophily between the inter- viewers and the prospective respondents in order to gain cooperation and possibly reliable and valid data. Heterophily is defined as the degree to which two or more people who interact are different in certain attributes such as beliefs, values, education, or social status. The more different two people are, the more heterophilous. Previous field experience in villages has shown that communication between individuals is less effective when a high degree of heterophily is present.g/ “Less effective" here means that it is difficult to get clear-cut answers to the interviewer's questions and to have the respondents' full attention and cooperation. Measures g/ . Ro ers, Everett M. Modernization Among Peasants: The Imgact of Communicat on. Holt, Rinehart and Winston, Inc. , pp. , , 233-237, 5T,— 264. 45 were thus taken to reduce heterOphily by way of (l) pretesting the interview schedules, (2) conforming the manner and dress of the interviewing team to local customs and (3) explaining the interviewers' intent. In more detail: (1) Pretesting Interview Schedules Having a clear and concise interview schedule can serve to reduce heterophily between the interviewer and the respondent. A com- plex and ambiguous schedule, on the other hand, is indicated by a large number of refusals to answer questions, or rather obvious wrong or inconsistent answers from farmer to farmer. The usual result is to have a consted and perhaps embarrassed respondent. Two interview schedules were used in this study and are reproduced in the appendix. Both schedules were pretested within two weeks of the main interviews, in areas near Lora and Oghi Thanas but outside the sampled thanas. Ten farmers were interviewed with each of the two pretest interview schedules. The results proved helpful; useless and time consuming questions were eliminated, some questions were clarified, the interview schedule was reorganized in order to get continuity and flow in the answers. At all times an attempt was made to structure the interview in such a way that the interviewer would answer the easier "yes" and "no" questions before more difficult questions were asked. The pretest interviews also helped to enhance the finesse of the interview team. In addition, both pretests came out with large variabil- ity in the crop production and input data which suggested that we would need to draw from as large a sample as possible. 46 (2) Conformity of the Interviewing_leam All interviews were conducted by the author with the aid of an interpreter. The interpreter, Siddiq, was born and raised on a barani farm of Murree Hills which has conditions quite similar to Lora and Oghi. Siddiq could not only speak the farmers' language but he was highly respected as a Muslim follower; his neatly kept beard was a strong Sign that he practiced his religion faithfully. The author, on the other hand, was often mistaken for a local because of his looks and compliance with local customs at community gatherings. Personal appearances of interviewers can have imposing effects on respondents. A man in a suit, for example, commands respect and special courtesies, aside from being rather obvious in a rural setting. In this study, both interviewers wore the native shalwar-kameez at all times in the village areas. Furthermore, the interview team stayed in the village areas, eating, sleeping, and participating in some local activities. The generous hospitality and many invitations to have dinner and discuss farming were signs of reduced heterophily. (3) Explaininngnterviewer's Intent Explaining interviewer's intent is a very important step 3/ towards reducing heterophily. Rogers writes that: In one Latin American community local rumors were spread that the researchers conducting the field interviews were really looking for farm children to send to a sausage factory. In the case of the Colombian study, an increase in local property taxes was announced Dy the Government during the data collection. The farmers were asked as part of our interviews, how many animals they owned and how many acres they operated. Naturally, they assumed a direct causal link between the information they were giving and the tax increase. Interviewing was brought to a halt by .3! Rogers, ibid., p. 367. 47 noncooperation until the community leaders, who fortunately understood the nature of the investigation, had an opportunity to explain the situation to their neighbors. Subsistence crop production occupies by far most of the barani smallholders' time and energy. They are thus indeed concerned with anything which affects their farm management decisions and practices. Hence, each and every respondent was told (1) that the interviewers were interested in learning all they could about barani crop cultivation, (2) that we were not working for the Government of Pakistan, (3) that the information obtained would be instrumental in helping to improve barani agriculture, and (4) that we had the welfare of the Hazarans in mind at all times. We were never asked if we had ulterior motives and seemingly had full cooperation. Sampling Technique and Respondent Selection Three sampling techniques were considered before a fourth modified technique was chosen for the survey. The first consideration was to draw random samples of farmers from village land records which are held by the patwaris or revenue collectors. Such a sampling technique proved unsuitable for obvious reasons: (1) it would have involved using revenue records and perhaps giving the wrong impression of interviewer's intent, (2) the sample would have included owners of land to the exclusion of tenants, and (3) it would have been difficult to locate selected farmers in the very mountainous area of Lora and Oghi since many homes are widely scattered. The sheer physical feat of reaching the right res- pondent would have been enormous. Second consideration was given to registering all cultivators in a village and drawing random lots from those registered. This sampling 48 technique had the following drawbacks: (1) just going into villages would have taken time and numbers of interviewers and, hence, would have made the survey work costlier, and (2) it was believed that dwarf wheat diffusion would be location specific. Namely, some villages would be saturated with the wheat variety and other areas would not. And as noted above, a large geographic area like the thana was required in order to get an idea of the spread of the new variety. Third, using maps from which to draw sample villages was not possible because none were available. Then again, the broken terrain with its insurmountable obstac1es to traverse would have made it a for- midable task to reach a number of villages. The sampling technique which was relied upon involved the fOllowing steps in selecting respondents: (l) Equipped with a Land Rover, the interviewing team would drive to an area at random without notifying the villagers in advance and would stop at places within sight of a village or a group of homes, (2) the interview team thus walked to the homes or village and whenever a farmer was encountered, usually in the field or on the trail, we asked him the following questions which formed the criteria for an interview: (i) whether or not the man grew wheat (during the first round) or maize (during the second round), (ii) whether or not the man actually cultivated and worked the crop, and (iii) whether or not the man had ten minutes to spare for general discussions on the crop. If the man answered affirmatively to all three questions and agreed to be interviewed, then the interviewer's intent was stated before we pro- ceeded with the actual interview. After the first interview we'd ask the farmer to take us into the village. In those cases where we did not encounter an adult male on the trail, we would ask local children to take us into the village to the headman's home which is the general 49 receiving area. These latter precautions were strictly adhered to in order to avoid the problem of interferring with the practice of Egrgah, Such a selection procedure is subject to some bias (1) if there are a large number of refusals, and (2) if the interview team consciously selects "better looking" respondents out of a group. In answer to the first part, there were no refusals from those who fit the criteria. As to the second part, attempts were made to interview as many of the farmers that we encountered, one by one. During some encounters there were at most 3 or 4 farmers to deal with at a time. Those listening in at the interview and not directly interviewed were asked to refrain from making side comments during the interview. This request was respected by all and none seemed embarrassed by being asked to sit silent until their turn. It should also be noted that there did not appear to be a "parrot effect" involved whereby those waiting to be interviewed repeat the same answers as those already interviewed. 0n the contrary, farmers were clear in pointing out when they did some things differently from others who were interviewed in their presence. Overall, the selection procedure appeared to work exceptionally well. Farmers were COOperative and seemingly interested in the questions. Some even commented that the questions made them think more about their present farming practices. Furthermore, the high response to the questions asked gives reason to believe that the farmers attempted to answer all questions to the best of their ability. In some cases, farmers said they weren't sure about their answers. This seemingly indicated a desire to avoid giving just any answer as a courtesy to the interviewing team. Survey Implementation Two field surveys were conducted during two time periods in order 50 to coincide with the post-harvest periods for wheat and maize; May/June and October/November, respectively.fl/ The interview schedule used during the first period was aimed specifically at questions on MexipakE/ wheat adoption and impact. In addition a number of questions on communication, farm size, resource use in wheat production, family size, and non-farm income were asked. The second period interview schedules focused on maize production and resource use, and further questions on dwarf wheat adaption. It was also designed to get information on farmers' use of fertilizer and off-farm employment. The second round of interviews created a problem for the interview team. Namely, should those interviewed during the first period be interviewed again to round out the information dealing with the same farm units? Or should we work with a sample drawn randonfly as before and accept the possibility of interviewing some of the respondents interviewed earlier? Because of the extreme difficulty in relocating farmers, added cost, and the limited time available to the interview team, the second alternative was taken. In the first round of interviews, the heads of household represented a cross-section of 45 villages spread throughout Lora and Oghi areas -- 71 villagers represented 27 villages in Oghi and 72 villagers represented 18 villages in Lora. In all, the first sample was composed of 143 respondents. In the second round, 98 farmers were interviewed -- 50 villagers represented 26 villages in Oghi and 48 villagers represented 13 villages 4f _'Both interview schedules are reproduced in the Appendix. E/ ‘ Mexipak is the common term used fOr all dwarf wheats in Hazara. Few farmers are able to distinguish the differences in the new dwarf varieties. 51 in Lora. Portions of one farmer's answers to the questions on production had to be dropped from the sample in Oghi because all his land was farmed by tenants. However, the rest of this interview schedule was retained to add to the information on the questions dealing with off-farm employ- ment and family size. Handling First and Second Round Interviews As it turned out, there was indeed some overlap of farmers during the second round who were also interviewed during the first round. In Oghi thana, two farmers were interviewed both times and in Lora thana, thirteen. Handling the overlap presented another problem in compiling and organizing the data, but it also presented a way of checking the reliability of the data since some questions were repeated. For example, both the first and second interview schedules had the same questions on total farm area owned Dy the respondents, the farm area cultivated and the size of family. Given the nature of these questions we would generally presume that the answers would be the same during both interviews. However, some farmers gave apparently different answers in every case. Out of the 15 who were interviewed twice, there were only 4 farmers who answered that they owned the same amount of land. Only 3 farmers repeated the same answer on cultivated area and surprisingly, only 8 farmers mentioned the same family size. Thus, it became imperative that a test be made to see whether the two answers between first and second inter- views were significantly different, and what the implications would be for the reliability of the survey data. The procedure involved taking second period answers and subtracting them from first period answers for each of the 15 respondents. The null 52 hypothesis tested by the Students-t was that the mean value of the answers given during the first period interview were equal to the mean value of the answers given during the second interview in response to the same questions asked on "owned land", "cultivated land" and "family size". Table 4.1 illustrates the results of the test. Notable are the rather large standard deviations which are shown in parenthesis; area owned per respondent ranged from 0 acre (in the case of one tenant farmer) to 35 acres. Area cultivated ranged accordingly. Families ranged from 4 to 18 members per respondents. Also evident, however, is that there is no significant difference with any of the mean values. The levels of significance are the same for "owned land" and "family size". The mean value of the area cultivated apparently changed the most. One possible explanation may be that farmers associated the area cultivated with the area they had just harvested. Since wheat was just harvested at the time of the first interview and since it occupies proportionately less land than the summer crops, farmers may have had easier recall of that land they just tended and correspongingly gave a smaller number for cultivated acreage. 0n the other hand, the second interview was taken after the maize harvest which occupies proportionately more land. Even with the apparent descrepancies, the data appears to be fairly reliable. The noticeable errors associated with the variance of answers between interviews indicates that they are equally distributed about the mean. Thus, with a larger sample, we should not be too concerned that farmers could not recall information precisely since the mean values of the observations appear to be without significant bias. 53 TABLE 4.1 COMPARISON OF MEAN VALUES OF ANSWERS TO THE SAME QUESTIONS ASKED DURING FIRST AND SECOND ROUND INTERVIEWS, HAZARA, 1970. ( n = 15 ) Number Mean,Va1ue of Answers From: Ho: U1 = U2 of: First InterVTew Second InterVIew No Significant (May/J une) (Oct/Nov) Difference With: Acres 10.18 10.68 at= 0.50 Owned/Farm (11.14) (11.56) Acres 3.61 4.28 a = 0.30 Cultivated/Farm (2.65) (4.87) Family 9.53 8.93 a = 0.50 Members/Farm (3.64) (2.99) Combining Lora and Oghi Samples In the analysis that follows, the data from Lora and Oghi will be combined in order to develop a representative sample for Hazara District and the Himalayan barani. No attempt is made to compare the production functions and decision environment between Lora and Oghi, Instead, generalizations will be made about the population of Hazara District's agricultural sector on the basis of the data drawn from the two samples. Since the two interview schedules adduced a number of different issues, the reader is cautioned against thinking that the sample size is uniform ’for a number of the relationships that are tested herein. For instance, in the first round of 143 interviews, questions were asked on the commun- ication factors which alerted the farmers to dwarf wheats. Thus, the sample size is 143. During both interviews, questions were asked on aspects of off-farm migration so the sample size is larger. For the farmers who were interviewed twice, infOrmation is averaged on the 591" n00 if ‘A r hi Ut for 54 pertinent questions, in order to adjust for the discrepancies in answers noted above. General Characteristics of Hazara's Farms Table 4.2 below shows that approximately 95 percent of the sample of respondents cultivate less than 15 acres; 87 percent cultivate less than ten and 64 percent less than five. About 82 percent of the res- I pondents own their land, 4 percent both own and rent land and the rest are tenants. However, 34 percent of all the respondents cultivate less than 2.5 acres (not shown in table) and 97 percent of the tenants cul- tivate less than 10 acres. TABLE 4.2 DISTRIBUTION OF SAMPLE, BY TENURE AND SIZE OR AREA CULTIVATED, 1970: HAZARA DISTRICT TENURE Cultivated Owner Owner-cum- Tenant All Farms Area (n = 185) Tenant (n = 30) ( n = 225) (acres) (n = 10) % % % % <5 59.47 70.00 90.00 64.01 5<10 25.95 30.00 6.67 23.57 10<15 9.18 0 3.33 8.00 15<20 2.70 0 O 2.21 20> 2.70 0 O 2.21 Total 100.00 100.00 100.00 100.00 The average size of holding that is cultivated by 225 interviewees is about 5 acres; 97 percent of this cultivated acreage is barani. What little is irrigated is sown with rice during the summer season. Water for irrigation,however, is not perennial; it too comes from melting snow 55 and monsoon rains and is channeled long distances around the mountains. Nearly all of the farmers struggle with small and widely separated plots 6/ that multiply the difficulties of efficient use of production factors{_ Each plot averages around a quarter of an acre in size and farmers cul- tivate on the average as many as 15 individual and separate pieces of land that are generally terraced and scattered. Government attempts, 7/ however, to consolidate holdings have met with little successf— Most families live humbly in homes made of rock covered over with hand-molded mud. The average family has a few trees spotted around the house, a buffalo and/or goat for dairy products, some chickens, a dog, and a pair of bullocks for plowing. Some homes have a private well for drinking water or a nearby well shared by village neighbors; none of the village homes have indoor plumbing. Village homes are without electricity; the kerosene lamp is now replacing the candle. Wood and fuel oil are used for heating and cooking. Latrines are not to be found. Few homes are located near an all weather road. They are generally nestled against protective mountainsides and provisions and building materials are carried Dy hand, sometimes by rented donkey for a small carrying fee. 6/ '_ Fragmentation is notably a function of family size and inheritance practices. Muslim families continue the tradition of dividing farm land among sons and daughters; usually in the proportion of 2 pieces per male to 1 piece per female. The more off-spring there are the more fragments that develop. In some cases, Off-spring pool their resources as farms become unproductive in terms of feeding all family members, male members decide who will farm and the others will leave to seek emplqyment else- where. The remittances from off-farm employment are sometimes pooled in the family income. Evidence of this is discussed in Chapter VI. 7 . TlRizvi, S.M.A., et.al. ConSolidation of Holdin s: A Study_of the Process of Consolidation of AgriCUTtural HOldings in electedVillages in Peshawar District, Pakistan Academy for Rural Development, PeShawar, West Pakistan, January 1965. 56 'The heads of household interviewed in the survey were about 47 years of age. Forty-five percent said they were literate. However, they appeared to read with considerable difficulty. More and more the younger males are sent to rudimentary primary schools, mostly in anticipation of leaving the farm by the time they reach 20 years of age (discussed in Chapter VI). The extended families which live in one household, average between ten and eleven people; about 2 to 3 per cropped acre. Most families have a relation employed outside the village area, often working in Karachi city, a thousand miles away, but keeping their wives and children on the farm. §/ The typical family's total income is about $235 a year. About $150 is the value of the farm-produced crops: wheat, maize, a little rice, with their straw and hay cut from the surrounding hills. Another $25 is earned Dy the farmer at other jobs in the village. The rest of the income, $60, comes from relations employed away from the village who send either money orders or cash to the family on the farm. Little of the grain produced reaches the market. If a farmer sells any, it is only when debts fall due or when cash is needed to cover costs of medicine or wedding ceremonies. The poorest families consume practically nothing besides the grain they produce. Overall, these barani smallholders of Hazara District do not appear to be likely candidates to adopt innovations. They simply cannot afford huge losses in their crop production. Yet without agricultural advances 9! Estimated on the basis of the international exchange rate of ten rupees per U.S. dollar. 57 and with a continuing growth in population, Hazarans will have no alternative but to add to the mainstream of people joining urban concen- trations, a prospect the country can ill afford. CHAPTER V DIFFUSION AND ADOPTION OF DWARF WHEATS IN HAZARA DISTRICT Introduction The main purpose of this chapter is to examine and analyze the process of dwarf wheat diffusion and adoption as experienced Dy a sample of smallholder farmers in Hazara District. Particular concern will be with (l) the extent to which barani smallholders are using the new technology and (2) those factors involved in prompting barani small- holders to substitute a new technology (a new variety of wheat) for one with which they are familiar. Concomitantly, at the root of this investigation lie the following questions: Do barani smallholders represent an economically inert peaSantry or do they represent farmers responsive to economic incentives? Will they make rapid economic adjustments in resource allocation with high-yielding varieties of wheat (and maize) which are neutral to scale, labor intensive and suitable to barani farming practices and conditions? Where do farmers obtain their first information on new varieties? What variables correlate significantly with the decision to try a new variety? What factors facilitate the diffusion and adoption process? How can new agricultural practices be put into practice as quickly as possible? In this study, an innovation is defined as the introduction of a new factor of production or agricultural practice. The innovator -- the farmer who is earlier than others in his community to apply the innovation -- must be willing to take the risks involved in trying an innovation. In many cases the risks are high, especially when much more 58 59 is involved, such as the accompanying use of fertilizer with a new variety. By obtaining needed information regarding the actual performance of the innovation, the innovator plays a vital social role. With a proper understanding of the human communication that follows early adoption and the complex social and economic interactions of farmers in a barani environment, we move closer to knowing why some peOple change and how social systems can be made dynamic. Furthermore, by focusing on these particular aspects of change, it may be possible to answer important questions raised in recent discussions on the relative usefulness of the mass media, extension services and other channels of communication as they serve to diffuse new technology.l/ InfOrmation on these points would be important for implementing programs designed to diffuse new varieties to similar economic settings. After a look at the historical antecedents of dwarf’wheats in Hazara, the latter portions of this chapter are divided into three main sections: Section I includes a general overview of the findings dealing with the magnitude of dwarf wheat diffusion and adoption. It should be noted that during the first round of interviews, questions were asked primarily on the use of dwarf wheats. As farmers were interviewed it became evident that fertilizer was also being used Dy many in the sample for the first time. In order to incorporate information on this innovation, fertilizer, farmers were asked questions during the second 1 TJSee Part III, "Food Production Increase and Role of Administrative and Extension Services" in Sharif, C.M. (ed.), Food Production Increase in West Pakistan: Problems and Effects, Pakistan Ahademy fOr Rural Development,_Peshawar,June, 1970, pp. 179-239. 60 round of interviews on when they first started using both chemical fertilizer and dwarf wheats. The answers provide an interesting parallel with those of dwarf wheats. However, due to the relative smallness of the sample included in the second round of interviews, the information gathered on fertilizer use will be incorporated as supplementary material to that on dwarf wheat adoption. Section II will cover the reasons behind dwarf wheat adoption, such as, the unique characteristics of dwarf wheats which were singled out Dy the farmers interviewed and the average yields with desi and dwarf wheats. In addition, some attention is given to the risk and uncertainty associated with the new varieties of wheat and the new factor proportions employed. Section III presents the channels of communication most instrumental in informing barani smallholders about dwarf wheats for the first time. For the major part of this section, zero order correlation and multiple correlation analyses are utilized to analyze the empirical relationships between several "economic" and "communication" variables related to the adoption of dwarf wheats.g/ 3/ Historical Antecedents— Sources of Dwarf Wheats in Hazara The first dwarf wheats to enter Hazara District were used by the field staff of the Regional Department of Agriculture, the Agricul- tural Development Corporation, and some individual farmers. ’27 — Those unfamiliar with correlational analysis should refer to Appendix 8. 33! Historical coverage of dwarf wheats in West Pakistan is found in: Eckert, Jerry B., The Imgact of Dwarf Wheats on Resource Productivity in West Pakistan's’PUnjab, 92, 51:}, Chapter II, pp. l2-36. 61 (l) The Department of Agriculture and the Agricultural Development Corporation (ADC). According to the Extra Assistant Director of Agriculture of Hazara District, 1966/67 was the first year that dwarf wheats entered the District under the auspices of the Regional Department of Agriculture. The first inshipment of 400 maunds (enough for at least 400 acres) was distributed to different areas of the District for both commercial sales and to some extent to be sown by the field extension staff on a number of controlled and carefully selected "demonstration plots." TWelve maunds of the first shipment were distributed in Mansehra (where Oghi is situated), twenty maunds in Abbottabad (where Lora is situated) and the rest in Haripur (the area with most of Hazara's irrigated farm- land). At that time, each bag of dwarf wheat cost the government Rs. 54 per maund, compared to Rs. 20 per maund for the best desi varieties. 'In 1967/68, nine thousand maunds of dwarf wheat were commercially available to the farmers through the Agricultural Development Corpora- tion. The price per maund dropped to Rs. 36 but this price was still far above the price of desi varieties. By 1968/69, dwarf wheat seed reached a significant number of the farmers' fields. The Agricultural Development Corporation (ADC) sold a smaller amount (5,500 maunds) than the year before and at a lower price of Rs. 22 per maund. In 1969/70, the year of this survey, enough seed was apparently trading hands from farmer to farmer and relatively little was sold by the ADC. (2) Individual Farmers It's extremely difficult to estimate the number of individual 62 farmers who brought dwarf wheat seed into Hazara District. However, one person was encountered in Oghi who brought dwarf seed to his farm in 1966/67. Mr. Khan read about Indus 66 in the English newspaper and learned of its tremendous yields over desi varieties grown under similar con- ditions. In 1966/67 he was able to purchase enough seed (through his personal contacts) for 6 acres of strictly rainfed land. Khan harvested 32 maunds per acre, a record never before witnessed in Oghi area. The same year that he purchased his Indus 66 happened to be Mr. Khan's first year as a full-time farmer. Before then he was a lawyer in Lahore. What prompted his turn to farming was his dissatisfaction with the legal profession and the restoration of some 200 acres of his land that was originally held in abeyance by the Land Reform Movement of West Pakistan. This serendipitous group of events revolving around Khan's decision to farm produced one of the major demonstration effects in Oghi thana. The fOllowing year, according to Mr. Khan, several farmers approached him fOr some of his seed which he willingly exchanged for an equal amount of desi wheat, since he preferred the taste of the latter. Mr. Khan's first experience with dwarf wheats should not be passed over lightly. Under similar circumstances, Huke and Duncanfl/ conclude from a recent field study in the Philippines that: 4/ T'Huke, R.E. and J. Duncan, "Spatial Aspects of HYV Diffusion," mimeo, International Rice Research Institute, The Philippines, Summer, 1970. 63 The pattern of spread of high-yielding varieties in Gapan owes much to the fact that three major landowners had sufficient political or economic power to secure 1R8 seeds (rice seed from IRRI) in July 1966. These three landlords caused the seed to be planted by selected tenants in barrios (villages) close to oblacion (major cities) . From this modest beginning, the use 0% high-yielding varieties spread unevenly but with great speed to all corners of the municipality. Clarification of Intent The "Mr. Khans" located in the agricultural sector of West Pakistan tend to be more advanced and wealthier than the majority who farm the land. With all other things equal, such individuals, in turn, would serve as important conduits of output-increasing innovations in more progressive areas. But in Hazara and other parts of West Pakistan, all other things are not equal. Land is unevenly distributed with major proportions of the acreage held in a few hands and with large numbers of smallholders holding proportionately smaller fractions of the cultivable acreage (see Table 5.1). Wealth, as well, is also unevenly distributed between large and small farmers. Considering these diff- erences in equality and due to the "chance" circumstances of Mr. Khan's success, his exceptional education and relatively large farm (even in today's context), he is not included in the analysis that follows. Strict analysis of the aspects of change exclusively among smallholders serves to select variables which identify their characteristics the best. From this focus, we seek to recommend strategies of change which can be implemented with some assurance of access to smallholder barani farmers. 64 TABLE 5.1 DISTRIBUTION OF FARM SIZE IN WEST PAKISTAN Area (acres) No. Farms (000) Percent <1 742 15 1 <2 1/2 856 18 2 1/2 <5 806 16 5 <7 1/2 581 12 7 1/2 <12 1/2 759 16 12 1/2 <25 729 15 25 <50 286 6 SO <150 88 2 150 > 14 <.5 Source: 1960 Pakistan Census of Agriculture Section I. Survey Results: General Overview fagpitude of Diffusion and Adoption In the sampled areas of Hazara it was found that change is taking place at a very rapid rate with the diffusion and adoption of new dwarf wheats. The pattern of change is recorded in two ways: (1) the number of dwarf wheat users over a period of time and (2) the area sown with dwarf wheats over a period of time. (1) Dwarf Wheat and Fertilizer Users Over Time Research on the diffusion of innovations among agricultural pro- ducers indicates that typically the innovation is accepted slowly at §/ first. Gradually, the rate of acceptance begins to speed up until if Rogers, Everett M., Diffusion of Innovations, New York, Free Press of Glencoe, 1962. 65 a large proportion of farmers have accepted the innovation. After wideSpread acceptance, the number of potential adopters is much less and the rate of adoption begins to decline. Plotting this pattern of acceptance against time gives an "S" shaped curve of the normal dis- tribution. Table 5.2 shows the number of respondents who were using dwarf wheats and chemical fertilizer for the first time. A different pattern of use if discernible between the two innovations. In line with the observations made in many research studies dealing with the diffusion of innovations, Table 5.2 shows that initially, in 1966/67, less than one percent of the sampled barani smallholders were using dwarf wheats. During the same year, a slightly larger fraction of farmers used chemical fertilizer. In subsequent years, looking at the data from 1967/68 to 1969/70, more and more farmers used both dwarf wheats and chemical fertilizers. However, dwarf wheat use noticeably surpassed the use of chemical fertilizer. During 1969/70, the rate of change in both cases began to slacken as the majority of barani smallholders had already tried dwarf wheats and fertilizer for the first time. 66 TABLE 5.2 DWARF WHEAT AND CHEMICAL FERTILIZER USED FOR THE FIRST TIME PER RESPONDENT BY YEAR: HAZARA DISTRICT, 1970. DWARF WHEAT 9/ CHEMICAL FERTILIZER Eff (n = 226) (4n = 95) Growing No. of New No. of New Period for Users Each Cumulative Cumulative Users Each Cumulative Cumulative Wheat Year Number % of "n" Year Number % of "n" 1966/67 2 2 0.88 3 3 3.16 1967/68 28 30 13.26 5 8 8.42 1968/69 45 75 33.17 20 28 29.47 1969/70 75 150 66.35 30 58 61.05 1970/71 9/ so 200 88.47 11 59 72.63 2/Respondents' anticipated use. E/Data collected from first and second surveys. E/Data collected from second survey. Why do patterns of acceptance of an innovation generally show an "S" shaped surve of the normal distribution? It is often thought that if acceptance were essentially in response to economic profitability of the innovation, then the diffusion curve would be expected to take the shape of a vertical straight line. The main reason for expecting an "S" shaped curve is attributed to the network of interpersonal communication which influences the potential adopters. Havens and Rogers 517 state: §/ Havens, Eugene A. and Everett M. Rogers, "Adoption of Hybrid Corn: Profitability and Interaction Effect," Rural Sociology, Vol. 26, No. 4, December, 1961, pp. 410-11 (underline added). 67 It is our contention that once an innovation has fulfilled minimum considerations of profitability, it is largely the amount of interaction between individuals who have or have not adopted the innovation that determines the rate of adoption for individual farmers ..... The interaction effect is the process through which individuals in a social system who have adopted an innovation influence those who have not yet adopted. However, it is clearly evident that there were also shortages in the availability of new seed during the first two years of dif- fusion in Hazara. Hence, it is difficult to separate the "interaction effect" between individuals from the "economic response" of farmers to adopt higher-yielding varieties. This issue deserves special consideration and in Section III an attempt is made to separate the two by way of multiple correlation analysis. By focusing on the use of dwarf wheats for the moment, it is quite plausible to believe that as more and more farmers adopt an improved variety of seed, they become active agents in demonstrating the outcome of their decision. Festinger's dissonance theory lends support to the possible interaction effect.Z/ It generally states that once people have made a decision to either accept or reject an innova- tion they go out of their way (1) to make sure they have made the right decision, (2) to learn more about the innovation and perhaps unconsciously, to selectively expose themselves to further messages which were con- sistent with their decision. Simultaneously, they new (3) try to block out negative responses to their decision and (4) try to get more people to follow their decision. In line with the last statement is Z/ Leon Festinger, "Behavioral Support for Opinion Change," Public Opinion Quarterly, 1964, 28: 404-18. 68 the finding that some farmers were persuaded to adopt dwarf varieties even though they never saw the wheat in the field before trying it for the first time. The pattern of fertilizer acceptance looks similar to that for dwarf wheats. In actuality, it is more difficult to interpret. Table 5.3 shows the use of chemical fertilizer on dwarf and desi wheats. According to the respondents, nearly 19 percent were using chemical fertilizer before they experimented with dwarf wheats. They started fertilizer before dwarf wheats were generally available. Another 39 percent tried fertilizer and dwarf wheats together for the first time for each. Another 9 percent were using dwarf wheats but had never used chemical fertilizer. On the other hand, about 9 percent of the sample were using only desi wheat and fertilizer. A larger proportion (nearly 14 percent) who were using the traditional variety exclusively have never used fertilizer. These findings are consistent with a number of possible explana- tions: (i) dwarf wheats may do well without fertilizer, (ii) farmers cannot afford fertilizer but want to experiment with new varieties, (iii) farmers experiment with chemical fertilizer on desi varieties before switching to dwarfs, and (iv) farmers believe that fertilizer must accompany dwarf varieties. Though some of the propositions look contradictory, there is little supporting evidence to test them out. On balance, though, it looks as if different farmers have reacted differently to the use of dwarf wheats and fertilizer and that adjust- ment to these innovations is still taking place. Another indication is that farmers have not yet decided, one way or the other, what to do about fertilizer. 69 TABLE 5.3 COMPARISON BETWEEN THE USE OF DWARF AND DESI WHEAT WITH FERTILIZER USAGE, 1969/70* (n = 95). Percentage of Farmers Who Used Chemical Percentage of . . . . Farmers Who Fertilizer for the First Time. Have Not Used Percentage of Before Using Same Time as After Using Chemical Respondents: Dwarf Wheat Dwarf Wheat Dwarf Wheat Fertilizer Using Dwarf Wheat 18.95 38.95 9.47 9.47 Using Only Desi Wheat 9.47 Not Not 13.68 Applicable Applicable *Figures are in terms of percentage of "n". (2) Dwarf Wheat Acreage Over Time Table 5.4 gives the average area sown with dwarf wheat and the percentage of total wheat area per farm in the Lora and Oghi areas. In 1966/67, less than one acre was sown with dwarf wheat Dy two farmers in the sample. From a modest beginning, barani smallholders began to sow more and more of their area with dwarfs. In fOllowing years other farmers went through a period of experimentation and acquaintance with the variety. Few, if any, sowed 100 percent of their wheat area to the new variety during the first two years. Between 1968/69 and 1969/70 growing periods, more farmers with smaller holdings began to use dwarf wheats. But due to the increasing proportion of smaller farms among the users, dwarf wheat area as a percentage of total wheat area did not increase substantially. For 1970/71, the respondents anticipate sowing 7O 72 percent of their wheat area to dwarfs. Corresponding to the above, the average area sown with dwarfs has increased on each farm. TABLE 5.4 DWARF WHEAT AREA SOWN AND PERCENTAGE OF TOTAL WHEAT AREA BY YEAR IN HAZARA DISTRICT: 1966/67 - 1970/71 Cumulative Growing No. of New Average Number of Average Intensity Period Users Each Dwarf Wheat Acres of Adoption 5] Year per Farm 1966/67 3! 2 0.375 30.00 1967/68 30 0.775 34.02 1968/69 75 1.375 65.04 1969/70 150 1.412 65.90 1970/71 9/ 200 1.725 72.10 a/ "'Unreliable figures for comparative purposes due to small number of respondents. 91 Anticipated area. Cl T'Ratio of dwarf wheat area to total wheat area. Out of the total number of dwarf wheat users, only one farmer tried dwarf wheats (in 1968/69) and did not use them the fOllowing year (1969/70). The stated reason fOr this rejection was the "bad taste and quality" of the unleavened bread (chapatti) made from the new 71 wheat; he had a variety with a red-grain which is considered inferior to white-grain. However, the same farmer said he saw some white-grain types (Mexipak-65) in the village and would attempt to acquire enough seed to sow his entire wheat acreage with it. In the group of res- pondents, two farmers said they had never heard of Mexipak until the time of the interview. xTheir numbers represent less than two percent of the sample. They do, however, point out the need to know more about the way other farmers became aware of the dwarf varieties of wheat (discussed in Section III). Overall, findings on the micro-level are significant enough to Show widespread use of dwarf wheats in the unirrigated portions of Hazara District. This leads one to suspect the validity of the official estimates regarding the actual spread and adoption of dwarf wheat to barani areas, at least in those areas where rainfall appears to be adequate. In addition, it is believed that the impact of the dwarf wheats would have been much greater had the rainfed areas been treated equally in the initial food-grain self-sufficiency program which was primarily directed to attaining dwarf wheat adoption on irrigated farms. Section II. Why Dwarf Wheats Were Adapted And The Risks Involved The dominant failure in programs aimed at diffusing technologically improved varieties has been the lack of understanding about (1) the important characteristics of the innovation and (2) "the relationship between the expected variability in yields using current varieties and practices; the expected variability of yields using new varieties and 72 practices; and the relationship of these to the absolute levels of living of the farmer."§/ The first point should be of more concern to those who develop new varieties for diffusion and the second point is generally of concern to the farmer confronting a possible innovation. Characteristics of Dwarf Wheats as Perceived Dy Farmers Aside from the relatively higher yield achieved with dwarf var- ieties (discussed below), farmers frequently named a few other important characteristics which influenced their adoption behavior. Namely, it was found that: (l) Dwarf wheats fit the cropping pattern of the farmer and grow in a faster time than desi wheats. Respondents mentioned that desi wheats would not do well if sown after November. On the other hand, they said that dwarf wheats still performed well if sown before the end of December. This is crucial since late rains result in late sowing. With dwarf wheats the farmers have a longer decision period to work with. In addition, dwarf wheats generally mature 10 to 20 days faster than desi wheats, thus giving the farmer time to clear the field for maize. Longer-growing desi wheat varieties interfere with multiple cropping and the collection of residual moisture; much land is fre- quently left fallow as a result of late maturation. 787 * T'Wharton, Clifton R., "Risk, Undertainty and the Subsistence Farmer", War on Hunger: A Report from the Agency for International Development, May 1969, p. 15. 73 (2) Dwarf wheats are not complex in relation to desi wheat and are subject to experimentation by individual farmers. A number of tasks are still done in the same way for both desi and dwarf wheats: ground preparation, broadcast sowing, weeding, harvesting and winnowing. The only new difference is the addition of fertilizer and, as we have seen above, both growers and non-growers of dwarf wheats are using it. Dwarf wheats are neutral to scale meaning that any size farm can use them. It is evident that barani smallholders initially experimented with handfuls of seed, saw good results with their own experiments and in- creased the area sown with dwarf wheats the following year. Without this neutrality, it is doubtful that smallholders would have experimented and adopted the variety so willingly. (3) Some respondents liked the bearded features of dwarf wheats because the beard gave some protection against birds. Other respondents stated that they liked the taste of the white variety (Mexipak-65). (4) In addition to the above, and perhaps the most important characteristic, is the ability to see the differences between old and new innovations. Dwarf wheats are short-statured, with high tillering, and dark green compared to desi varieties; the physical contrasts between dwarf and desi wheats are very noticeable. During the second round of interviews, farmers were interviewed right next to demonstration plots with new, high-yielding varieties of maize and asked if they had ever heard of or seen any improved varieties of maize in the area. Sur- prisingly, many said no, even though a field of maize, which gave thirty percent more yield, was sitting in front of the respondents. What this illustrates is that new varieties need something to attract 74 attention, whether a visible symbol or a visible characteristic such as dwarfness. One complaint, however, was that dwarf wheat gave less fodder. However, barani smallholders appeared willing to substitute less fodder for a higher grain yield. Overall, the barani smallholders' perceptions about dwarf wheats indicate the need for a careful consideration of the characteristics built into a new variety. Such a consideration of the characteristics requires that due weight be given to the communicable variables involved. It should be clear that technological improvement alone, vital though it is, can only go so far in attaining widespread adoption of a new variety. In general, the innovation, in order to be usefully applied to barani agriculture, should be suitably modified to satisfy the interests of potential adopters. Dwarf and Desi Wheat Yields All farmers in the sample know the size of their cultivated acreage and measure grain that has been sun dried on the threshing floor with a y9g1_(a wooden or metal measuring bowl). ngig, which are treated like heirlooms, are fairly uniform in size and are filled until grain spills over the sides. Each farmer kggws how much wheat, maize and rice weigh in his wodi in terms of seers._' In addition to counting the number of wodis of each harvested crop, barani smallholders seem capable of 277 One seer equals 2.057 lbs. of 1/40th of a maund. One maund equals 82.286 lbs. One wodi holds approximately 5 1/2 seers of wheat grain. 75 recalling production for at least three years. Customarily, they frequently discuss production over the hgkka_(smoking pipe) in casual gatherings. Thus, the yield data appear to be reliable. Table 5.5 gives comparative yields on a per acre basis between desi and dwarf wheats. It can be seen that for each and every year dwarf wheats out-yielded desi wheats by a consistently wide margin, nearly 100 percent on the three year average from 1967/68 to 1969/70 (the first year, 1966/67, has been deleted because of the small number of growers). 1967/68 is considered one of the best years for wheat in West Pakistan; both temperature and rainfall were within the range conducive to good yields with the dwarf wheats. This seems in line with the exceptional dwarf wheat yields reported in Hazara for that year. In subsequent years, dwarf wheats were grown on more and more acreage which apparently included a mix of factors resulting in reduced yields: (i) poorer land under dwarfs, (ii) poorer farm managers growing the new varieties, (iii) less ideal weather, and (iv) less fertilizer per acre on dwarfs. Yet, dwarf wheats yielded comparatively more than desi in all periods. Pertinent at this juncture is a comparison of wheat performance on irrigated holdings. A year before this study, J.B. Eckert conducted a study of 115 farmers (from which he obtained 86 useable interview schedules) for making yield and input comparisons between dwarf and desi wheats.lg/ Among Eckert's respondents, dwarf wheats averaged 22.80 maunds per acre and desi wheats 16.20. That is, dwarf wheat H. -19]- Eckert, pp, 913,, p. 72. 76 TABLE 5.5 COMPARATIVE YIELDS FOR DESI AND DWARF WHEAT ON BARANI LAND, HAZARA DISTRICT* Year Number of Yield/ Number of Yield/ Differ- as % of Growers Acre Growers Acre ential Col.(5) Questioned (mds) Questioned (mds) (3)-(5) (1) (2) (3) (4) (5) (6) (7) 1967/68 26 23.92 17 9.08 14.84 163.4 1968/69 60 17.52 28 10.24 7.28 71.1 1969/70 98 15.12 62 8.48 6.64 78.3 3 Year Ave. 67/68 -69/70 61 18.85 39 9.27 9.58 103.3 * Table 5.5 reflects the structure of the interview schedule and should not be interpreted to mean that the number of desi wheat users is increasing. averaged 40 percent more than desi wheat on the larger well-irrigated farms. Another study of 502 farmers conducted during the year of this study by Mushtaq Hussain found similar results between desi and dwarf wheats in districts of the Sind and Punjab where dwarf wheats averaged 1 / around 23 maunds per acre and desi wheats around 16.5 maunds per acre. l/ TT'Hussain, Sayed Mushtaq: "Price Incentives for the Production of High-Yielding Mexican Varieties of Wheat", The Pakistan Develqpment Review, Vol. X., Winter 1970, pp. 448-468. 77 Notably, yield differentials between dwarf and desi wheats were reportedly greatest in the Sind and Rawalpindi District of the Punjab. Comparatively speaking from this small body of research results, Hazara's farmers experienced the greatest percentage increase in yields over desi varieties in West Pakistan. But in absolute terms, irrigated land out-performed barani land in both dwarf and desi wheat yield averages; an indication that barani smallholders still have a long way to go to increase their relatively lower levels of output. Risk and Uncertainty With Dwarfs Wharton has studied the interaction of risk, uncertainty and sub- 12/ sistence on technological innovation and states that:‘“‘ When the subsistence farmer confronts a possible inno- vation, he will be concerned with two questions: (1) Will the new method, taking its probable costs into account, produce an expected yield appreciably higher than his old method? (2) Is there a reasonable probability that something will go wrong, and that the new method will result in a net yield below his minimum subsistence level? Even if the answer to (1) is yes, he will not change his method unless he can also answer (2) in the negative. The closer a smallholder's current output is to his minimum sub- sistence level, defined as just eggggh_to feed his family over good and bad years combined, the more conservative he is likely to be. But, "if he can be convinced that the new method is not only better but 12/ '_7 Clifton R. Wharton, Jr. "Risk, Uncertainty, and the Subsistence Farmer", Development Digest, Vol. VII, No. 2, Agency for International Development, U.S. Department of State, Washington, D.C., April 1969, p. 7. Excerpted from "Risk, Uncertainty and the Subsistence Farmer: Technological Innovation and Resistence to Change in the Context of Survival," a paper presented at the Joint Session of the American Economic Association and the Association for Comparative Economics, Chicago, 28 December, 1968. 78 reliably so, and that is probable negative variability (in terms of worst possible yield) will still leave him better off than he was before, then he is most likely to make the change."l§/ It is clearly evident to the respondents that dwarf yields are greater than desi yields. Respondents also claimed that there is less risk and more certainty in sowing dwarfs, they at least "got their seed back.“ On the other hand, farmers said that many times they had to feed their desi wheat as fodder to the animals, they "got no seed back." How close the farmers were to describing the relative amount of risk involved is illustrated by the contrasts in Table 5.6. It should be pointed out that the data comes from 71 farmers who grew only dwarf wheats and 35 farmers who grew exclusively desi wheat. Those who grew dwarf wheat obviously experienced the greatest variability in yield as indicated Dy a range in output from 6.5 to 26.6 maunds per acre. Desi growers, on the other hand, who have grown their traditional wheat for several years, experienced a range in output from 4 to 13 maunds per acre. A farmer who shifted entirely to dwarf wheats would, by all indications, run some risk of getting lower than average yields usually obtained for desi varieties. Looking at the positive side of the spectrum, given all the favorable input combinations, a high of 26 maunds per acre could be expected with the new varieties. This helps to explain part of the reason behind dwarf wheat adOption. 131 Wharton, Ibid., p. 7. 79 TABLE 5.6 YIELDS OF DWARF AND DESI WHEAT VARIETIES IN HAZARA DISTRICT, 1969/70 Type of Negative Average Yield Positive Sample Wheat Standard (mds/acre) Standard Size Deviation Deviation (n) Dwarf 6.50 16.56 26.62 71 Desi 4.07 8.54 13.01 35 Some risk is evident with dwarf wheats and one should look more closely to the intensity of dwarf wheat adoption; the ratio of dwarf wheat area to total wheat area per farm. An analysis of the intensity of adoption is essentially a study of farmer's decision-making under risk and uncertainty, for the ratio indicates how cautiously farmers move from experimentation to full use of the new variety. The comparisons shown in Table 5.7 are drawn up from the data collected during the first round of interviews. Farmers who began using dwarf wheats between 1966/68 got the highest yields among the respondents in 1969/70. Their mean value of adoption intensity was 89.7 percent. Those who began using dwarf wheats in 1968/69, had lower yields, similar standard deviation and similar intensity of adoption with dwarf wheats in 1969/70. TABLE 5.7 AVERAGE YIELDS AND INTENSITY 0F ADOPTION, HAZARA DISTRICT Year of’ Number of’ Average *Average 2/ Adoption Growers Yield,1969/7O Intensity of Adoption, 1969/70 - (n). (MS/acre) (76) 2* w‘ man SOD. 1966/689/5 30 18.48 10.24 89.7 1968/69 45 16.43 10.60 89.3 1969/70b 75 14.17 7.13 79.0 1970/71—/ so -- -- 31.7 3! 1966/67 and 1967/68 are combined to increase "n". EfAnticipated 5] Ratio of dwarf wheat area to total wheat area. 80 Finally, those who grew dwarf wheats during the year of the survey, began Dy sowing nearly 80 percent of their wheat acreage with the new varieties. Their yield (and standard deviation) dipped lower than those of earlier adopters. Farmers who anticipate growing dwarf wheats for the first time in 1970/71 indicate a desire to sow 32 percent of their wheat fields with dwarfs. A New Production Function Variability in yields reveal only a portion of the risks involved with the new varieties. Farmers' decisions have both a risk and un- certainty dimension which relates to year by year variability (i) in costs of production, determined Dy the factor proportions required for pro- duction, and (ii) in price of the product. The former is more important than the latter among subsistence farmers who market little or nothing of their total product. Table 5.8 clearly shows that factor proportions have changed to a considerable extent with the use of dwarf wheats. In essence, the farmers have shifted from one production function to another, and in the process of change the dimensions of risk and uncertainty also change. TABLE 5.8 REPORTED INPUT LEVELS FOR DWARF AND DESI WHEATS IN HAZARA DISTRICT, 1969/70 Input (A) Dwarf Wheat (8) Native Wheat ‘HozxA = X3, ("771) iihz3s) Rejected with Mean 5.0. Mean S.D. Less than: Man-Hours/Acre 106.8 123.2 68.4 38.3 .01 Seed Rate (Seers/Acre) 36.5 8.4 32.2 9.1 .01 Pounds of Nitrogen/Acre 40.7 44.2 6.0 19.1 .001 Pounds of Phosphate/Acre 28.7 46.3 8.3 23.1 .01 Yield (mds/Acreli 16.6 10.1 8.5 4.5 .001 * A test for the difference in mean values was made with the Student t. The figures shown indicate the probability of making a Type I error. 81 The costs and returns deserve much more attention than given in this chapter. A section is devoted to examining expected changes in net profits associated with cost variability in Chapter VII. Three important points should be drawn from the above: (1) The variability in dwarf wheat yield is relatively large vis-a-vis the variability in desi yields. More attention should be given to finding ways to help barani smallholders move towards the optimum allocation of resources used to produce dwarf wheats in order to reduce the yield variability. (2) Stemming from the first observation is the fact that barani smallholders are still not committing their full wheat acreage to the new wheat because they are apparently experimenting under their own farm conditions. (3) If the supply imperfections and distribution defects are uniform for all smallholders in Hazara, why do some farmers tend to be relatively quicker than others in introducing an innovation? This last point is a basic question underlying Section III. Section III. Communication and Economic Variables Associated With Dwarf Wheat Diffusion and Adoption In this section of the chapter an examination is made of the communication and economic variables (which will be defined below) associated with dwarf wheat diffusion and adoption among barani small- holders. The section is divided into two parts. First, an overall perspective is presented with regard to the way farmers were first made aware of dwarf wheats, namely, the channels of communication. Second, zero-order and multiple correlational analysis is used to deal with a number of questions so far left unanswered: What factors are most important at the farm level in creating awareness and adoption? What farmers are the first to become aware of and adopt innovations in a smallholder setting? What relative importance can be attached to economic and communication variables in motivating smallholders to adopt innova- tions? Finally, knowing that a dynamic agricultural sector is identified by its rate of change, what variables can "best" promote change? For the government of West Pakistan, which has direct control over some comm- unication and economic variables, answers to these questions are of para- mount importance. In addition, there is practical usefulness for agri- cultural extension agents if they can identify potential innovators among smallholders, and then use different strategies of change with different categories of smallholders. Channels of Communication During the first round of interviews, respondents were asked to name the first source of information telling them of the new dwarf wheats. Generally, there are two main channels Of communication, interpersonal and mass media channels.lfl/ Frequently mentioned, however, was the demonstration plot. Because of its importance as a media for information and because it has been used in Pakistan for years Dy the Department of Agriculture, a separate sub-heading has been made for the demonstration plot. In particular, it has been grouped with the category for inter- personal channels since demonstration plots serve primarily as topics 111 Both channels function in different ways and their effective- ness also differs according to the way they are used. Their more distin- guishing characteristics are discussed in? Rogers with Svenning (1969) p. 125. 83 for discussion between farmers. In the following, a brief overview will be presented on the channels of communication which first informed respondents of dwarf wheats. The findings are shown in Table 5.9. More detail is explained below. TABLE 5.9 CHANNELS OF COMMUNICATION WHICH FIRST INFORMED RESPONDENTS 0F DWARF WHEATS, HAZARA, 1969/70. Percent of Total Channels Number of Respondents Respondents Mass Media 33 23.78 ’Magazine (in Urdu) _l. 0.70 Radio 32 23.08 Interpersonal 108 74.82 Localite ‘TPT 35.66 Cosmopolite 33 22.38 Demonstration Plot 21 16.78 Not Aware of Dwarfs* 2 1.40 TOTAL 143 100.00 * Not aware at the time of interview. (1) Mass Media Channels Mass media channels refer to the radio, television, film, news- paper, magazines -- anything with a capacity to reach large audiences quickly over great distances. Agricultural programs are broadcast daily over the radio in West Pakistan. Many are coordinated with the Bureau of Agricultural Information as a part of an Education Extension com- ponent. In addition, the Bureau publishes a monthly calendar of the radio programs for their respective areas. Radio programs are presently 84 beamed from Lahore, Rawalpindi and Peshawar. The first two stations broadcast in Urdu/Punjabi and the third in Pushto. It was found that the radio was the most effective mass media channel to inform the smallholders of dwarf wheat performance and availability. As shown above in Table 5.9, 23 percent of the respon- dents interviewed during the first round became aware of dwarf wheats over the radio. But 56 respondents out of the 143 (39 percent) owned radios. Ten people (7 percent) who did not own radios stated that they were first made aware of the dwarf wheats over this media. This tends to indicate that even families without radios were as apt to listen to the radio as much as owners in Hazara's rural areas.l§/ Altogether, only one farmer learned of Mexipak from written media, a magazine written in Urdu. No other type of mass media channel was mentioned by the respondents as a first source of information on the new wheat varieties. (2) Interpersonal Channels There are essentially three types of interpersonal channels which informed the barani smallholder of dwarf wheat yields: (i) interpersonal localite or those originating within the social system of the receiver, i.e. the neighbors, village shopkeepers, etc. (ii) demonstration_plots or localite visual field displays of agricUltural innovations that lead to some discussion among farmers. Both Lora and Oghi areas had the same number of demonstration plots installed on farmers' fields by the Field Assistants; six plots in each area in 1967/68 and 15/ '__ It was often said that the war with India in 1965 created a tremendous surge in the demand for radios. Furthermore, Hazara is rela- tively near to the trade routes for contraband radios which come in from Afghanistan and it should not be too surprising that the smallholders possess radios. 85 five plots in the following two years. For 1970/71, the number has been reduced to one each. Their locations have all been near the market centers of Lora and Oghi. (iii) interpersonal cosmopolite or those Channels which have their origins outside the immediate social system, i.e. agricultural extension personnel and distributors of farm supplies. It should be noted that both Lora and Oghi cities have offices of the Department of Agriculture, each headed by an Agricultural Assistant (usually a man in his 30's, with a bachelor's degree in agriculture from Peshawar University). Each Agricultural Assistant, in turn, supervises 3 or 4 Field Assistants (usually men in their late 20's who matriculated in second or third division and who completed a one-year certificate course in the Agricultural Training Institute in Peshawar). "A Field Assistant is expected to be the Government's principal contact with farmers in the area of one or two Unions, which means 10-20 villages, or 10,000-25,000 people."l§/ The interpersonal localite channels had the largest impact on the farmers. Approximately 35 percent of those who know about dwarf wheats (Mexipak) first heard of them from interpersonal localite channels (shown in Table 5.9). Another 17 percent of the respondents said they first saw dwarf wheats in demonstration plots and later asked the farmers growing the variety about the crop. Overall the findings show that interpersonal exchanges between barani smallholders carry the most messages and that the dwarf wheat demonstration plots (which showed striking differences next to desi plots) were effective transmitters of the dwarf yield message. Approximately 22 percent of the respondents heard the message initially from interpersonal cosmopolite sources such as the Government extension peOple, the ADC agents, and other fertilizer dealers. It 16/7 __ Davy, Dorcey, F., "Improving the Training of Field Assistants in the Agricultural Training Institutes of West Pakistan," mimeo, The Ford Foundation, March, 1967, p. 4. 86 should be noted that Field Assistants (lowest level extension agents of the AGriculture Department) were very instrumental in diffusing dwarf wheat varieties. Aside from personally informing farmers of dwarf wheat potential, they were responsible for the installation of many of the demonstration plots on farmers' fields which, in turn, were 12] catalysts in dwarf wheat diffusion. Correlation Analysis Dependent Variables The dependent variables in this section of analysis are "aware- ness" and "innovativeness." Awareness is defined as the degree to which a barani smallholder first hears of or sees dwarf wheats. The following operational numbers are used to measure the degree of awareness among smallholders, i.e., those with a number of 6 were the first to become aware of dwarf wheats among the sample of respondents. Operational Sample Numbers Explanation Size 6 First heard of or saw dwarf wheats in 27 either 1965/66 or 1966/67 5 First heard of or saw dwarf wheats in 1967/68 71 4 First heard of or saw dwarf wheats in 1968/69 41 3 First heard of or saw dwarf wheats in 1969/70 2 2 First heard of dwarf wheats at time of 2 interview, 1970 L7] Since 1966, a group of 155 farms located in Lagana Province of the Philippines have been surveyed annually to observe the changes taking place with the introduction of new, high-yielding rice varieties. One observation was that "Over 50 percent of the farmers surveyed reported that the extension worker was the major source of their information regarding new varieties. The second most important was the neighboring farmers." The International Rice Research Institute, Annual Report, Los Banos, The Philippines, 1968, p. 311. 87 Innovativeness is defined as the degree to which a barani small- holder is relatively earlier than others in his social system to adopt dwarf wheats. The following Operational numbers are used to measure the degree of innovativeness. Notably, the larger the number the higher the degree of farmer innovativeness. Operational Sample Numbers Explanation Size 6 Began using dwarf wheats in 1966/67 2 5 Began using dwarf wheats in 1967/68 28 4 Began using dwarf wheats in 1968/69 45 3 Began using dwarf wheats in 1969/70 75 2 Anticipates using dwarf wheats in 1970/71 50 1 Has not yet decided to try dwarf wheats 26 For the analysis presented below, operational numbers 3 and 2 of awareness will be combined as will operational numbers 6 and 5, under innovativeness, in order to increase the size of sample for these cate- gories for farmers. Independent Variables Awareness and innovativeness are hypothesized to depend on several independent variables. A summary of the independent variables are presented in the following paradigm. Each, in turn, is identified by a code name as shown. Since awareness precedes innovativeness, it too becomes an independent variable of innovativeness in this particular instance. 88 PARADIGM OF CORE VARIABLES RELATED TO AWARENESS AND INNOVATIVENESS Dependent Variables Independent Variables (Yi) (Xi) I. Economic Variables a. b. c. d. e. f. 1. AWARENESS g. Total area cultivated per farm (acres) Percentage of area owned by respondent (%) Size of family on the farm Cash earned in the village (rupees/year) Total cash income (CASH + MREMIT) Income per capita (INCOME %- FAMILY) Dwarf wheat yield per acre 2. INNOVATIVENESS II. Communication Variables a. b. c. d. e. Awareness Mass Media Contact Interpersonal Localite Contact Interpersonal Cosmopolite Contact Demonstration Plot III. Intervening Variables a. b. A few things should be noted with regard to the independent Migrant remittances Absence of male family member variables and Tables 5.10 and 5.11: Code Name AREA OWNER FAMILY CASH INCOME Y/CAP YIELD AWARE MMC ILC ICC DEMO MREMIT MIGRANT (1) Some barani smallholders earn cash in the village by selling milk, butter and eggs. Others have odd jobs like carting packages and bundles. A large percentage of the farmers interviewed receive military pensions. This last source of income in further explained in the next chapter. Altogether, these forms of income constitute CASH. (2) All of the communication variables, save that of AWARE, are mutually exclusive. For example, a farmer either became aware of dwarfs Dy MMC, ILC, ICC, or DEMO, but not any two. For each farmer interviewed, the value of one is given to the first source of information on dwarfs 89 and a value of zero for the rest. The number of observations on these variables are 143. Hence, under the awareness component the mean values for each category of farmers sums to 100. Under the innovativeness component the sum is less than 100 since the observations of 143 farmers are spread over 226 respondents. (3) The variable, MIGRANT, also enters as a zero-one variable. The value of one is given to each family that has at least one family member working away from the farm, and zero for all other families. Concomitantly, the last set of families receive no migrant remittances. Chapter VI clarifies the definitions for "migrant" and "migrant remittances." Both variables are "intervening" because migrants serve as sources of information and income simultaneously. (4) It should be explained that YIELD represents the actual yield per acre with dwarfs in 1969/70. Conceptually, this variable is inserted as a proxy for the joint effect of two variables which are ordinarily difficult to measure: (i) It is a proxy for "management" in that t e best farm managers are usually distinguished from the worst managers by their higher levels of farm output per unit of input. Such farmers may be expected to be more efficient and innovative than others and we would expect to find a high correlation coefficient between a. measure of innovativeness and management. (ii) It is a proxy for land quality. It is felt that the best farm managers also have the best land in terms of location, amount of farm yard manure applied and nutrient content of the soil (or soil fertility). (5) Tables 5.10 and 5.11 are drawn up expressly fOr the purpose of showing (i) the zero-order correlation coefficients between "aware- ness" (or "innovativeness") and each of the identified independent variables and (ii) whether or not the correlation is linear or not. For example, on Table 5.10, the correlation coefficient between FAMILY and "awareness" is 22.82 percent and the relationship is statistically significant at the 0.02 level. If we concentrate on FAMILY for the time being, Table 5.10 also shows that during 1965/67, twenty-seven farmers out of 143 became aware of dwarf wheats for the first time. The same farmers had an average of 15.1 family members living in their households. In 1967/68, seventy-one farmers became aware of dwarfs and the size of their extended families averaged 12.4 persons per household. Forty-one farmers learned of dwarf wheats in 1968/69 and their families averaged 10.2 persons. Four farmers out of 143 first heard of dwarfs in 1969/70 and their families numbered 8.3 persons on the average. Scanning the data once more, it can be seen that family size decreases from 15.1 to 12.4 to 10.2 to 8.3 over the years from 1965/67 to 1969/70. Clearly, the relationship is linear and highly significant. Correlates of Awareness The dependent variable "awareness" was correlated across 143 respondents with most of the variables shown in the paradigm. Table 5.10 presents the results. Of the economic variables, farmers who owned 1 list of ti han other linear re' correlata TiBLE 5.1 Indefiende iarlabl es \ . ECCYOVIC DRYER 1 “NH CASH (l INCOME WW I 90 most of their land and those with large families were relatively sooner than others in becoming aware of dwarf wheats. The mean values indicate linear relationships for these two variables and each is significantly correlated with "awareness." TABLE 5.10 CORRELATES AND MEAN VALUES OF INDEPENDENT VARIABLES (BY TIME PERIOD) ASSOCIATED WITH AWARENESS (N = 143) MEAN VALUES 0F VARIABLES ZERO-ORDER AND TIME OF AWARENESS Correlation Independent 1966767 I967768 1968769 l969770 Coefficient Variables (n=27) (n=7l) (n=4l) (n=4) 9] (Percent) ECONOMIC AREA (acres) 7.0 5.2 4.5 8.1 12.17 OWNER (%) 93.59 91.93 79.68 77.25 19.17*** FAMILY (No.) 15.1 12.4 10.2 8.3 22.82** CASH (Rs) 1,149 862 812 165 11.02 INCOME (Rs) 1,913 1,822 1,645 920 7.85 Y/CAP (Rs) 145 159 163 124 -1.53 COMMUNICATION MMC (i) a] 30 26 15 0 15.77**** ILC 2%) a] ll 37 56 100 -37.21* ICC % a] 44 24 10 0 28.33* DEMO (%) _a_/ 15 l3 19 O -1.18 INTERVENING MIGRANT (%) a/ 63 61 61 75 -O.42 MREMIT (Rs.)—' 764 960 833 765 0.46 2[Figures shown are percentages of the respective values of n. EIUnreliable sub-sample figures due to small number. * Statistically significant at the .01 level. ** Statistically significant at the .02 level. *** Statistically significant at the .05 level. **** Statistically significant at the .10 level. 91 None of the other economic variables are significantly correlated with awareness. However, the variable AREA is obviously non-linear and to some extent it appears that farmers with medium size farms score relatively higher in terms of awareness than farmers with large farms. Though the mean values for INCOME and CASH show linear relationships with awareness, they are not significantly correlated since the actual observations record rather large standard deviations from the mean values (not shown in the table). Judging from the zero-order correla- tion coefficients, it is quite probable that "wealth" (the combination of farm area and income) in a smallholder environment does not have a significant association with awareness. Two variables under the heading of communication are significantly correlated with awareness at the .01 level, and one variable at the .10 level of significance. Namely, both interpersonal cosmopolite (ICC) and mass media contacts (MMC) were important for creating awareness among smallholders. Those farmers who scored low on awareness became informed of innovations primarily through interpersonal localite contacts (ILC); a very significant association. The correlation coefficients for migrants (MIGRANT) and migrant remittances (MREMIT) are not significantly different from zero. How- ever, the mean values for migrant remittances show a curvilinear relation- ship over time. Since the percentages shown for migrants are uniform over time, we should not expect to find any significant correlation. However, it is apparent that many respondents have family members working outside the village community (the subject matter of Chapter VI). In sum, it can be seen that farmers who become aware of 92 innovations before others are generally those who own most of their land, have large families, frequently listen to the radio and have more contact with change agents and sales representatives of farm inputs (the interpersonal cosmopolite channels of communication). Correlates of Innovativeness "Innovativeness" was correlated across 226 respondents with all of the variables shown in the paradigm. It should be remembered that YIELD represents, conceptually, a proxy for land quality and a proxy for "management" in that the best farm managers are distinguished by their higher levels of productivity with dwarfs. As such, we would expect to find a significant correlation with innovativeness. As seen in Table 5.11, this is the case and the relationship between YIELD and “in- novativeness" turns out to be the strongest among all the variables considered. Most of the other variables which are significantly correlated with innovativeness do not need particular explanation except to say that the economic variables predominate in explaining innovativeness. Judged by the correlation coefficients, income per capita, mass media channels, demonstration plots, percentage of migrants and migrant remittances seem to have the least influence on farmers' innovative behavior. Respondents with an average size farm of 5 acres and above, large families, and relatively higher incomes are more innovative than others. Similarly, farmers who score high on awareness and interpersonal cosmopolite contacts are more likely to adopt innovations before others in their community. Finally, it should be pointed out, that all of the variables which n9 ft. 712' I a" 93 were significantly correlated with innovativeness show linear relation- ships among the mean values. As such, the zero-order correlation co- efficients appear to be fairly reliable indicators of the relationships between the variables identified. TABLE 5.11 CORRELATES AND MEAN VALUES OF INDEPENDENT VARIABLES (BY TIME PERIOD) ASSOCIATED WITH INNOVATIVENESS (N = 226) MEAN VALUES OF VARIABLES AND ZERO-ORDER TIME OF INNOVATIVENESS Correlation Independent 1966768 1968769 1969770 1970771a7 None57 Coefficient Variables (n=30) (n=45) (n=75) (n=50) _' (n=26) (Percent) ECONOMIC acres) 8.5 5.6 4.2 4.6 3.1 24.48** OWNER (%) 95.13 86.67 88.00 80.68 69.58 10.63 FAMILY (No.) 17.8 12.1 10.4 9.1 8.3 33.54* CASH (RS) 1,691 765 854 533 408 24.85** INCOME (RS) 2,965 1,548 1,537 1,206 995 27.03* Y/CAP (RS) 171 155 148 148 130 5.22 YIELD Mds) 18.48 16.43 14.17 NA NA 56.63* COMMUNICATION core) 4.8 3.7 3.0 2.8 0.5 46.33*‘ MMC (%) 20 27 9 16 O 8.16 ILC (%1 1O 16 37 30 12 -35.57 ICC (% 4O 24 8 6 4 29.07* DEMO (%) 17 9 12 8 0 4.73 INTERVENING MIGRANT (%) 7O 4O 59 64 58 -3.58 MREMIT (RS) 1,274 783 683 673 587 12.32 é-/Anticipated use of dwarfs. 9(Have not yet decided to try dwarfs. * Statistically significant at the .01 level. ** Statistically significant at the .02 level. NA = not applicable. 94 Multiple Correlation Analyais The zero-order correlation analysis made it possible to detect strong associations between "innovativeness" and several economic and communication variables. Earlier in the chapter we mentioned the poss- ibility of separating the influence of the economic and communication variables in order to see which sets of variables explain innovative behavior the best. For this purpose, multiple correlation analysis has been employed. To explore this possibility, the variables highly correlated with innovativeness were used as independent variables in two different models.l§/ In the first model, innovativeness is a function of YIELD, AREA, FAMILY, OWNER, INCOME and AWARE. In the second model, innovative- ness is a function of YIELD, AREA, OWNER, INCOME, AWARE, MMC, ILC, and ICC. Inadvertently, FAMILY was not included in the second model. The exclusion, however, does not alter the results in any significant way. Purposely, only three communication variables are added to the second model and DEMO is not included in order to avoid a singular matrix problem in the computer calculations. Although inextricable, the con- stant term holds some of the variation in the dependent variable which can be attributed to the DEMO variable. Table 5.12 presents the results of the multiple correlation analysis. In the first model, 47 percent of the variation in the Bf - Description of the model is in Appendix 8. Ordinary Least Squares techniques were used to estimate the functional relation- ships between variables. 19391 ‘ndpf ~1- méel .‘A )4 ‘8‘ ——1—1 95 dependent variable is explained by the variation in the identified independent variables, i.e., R2 = .47.12! Analysis of variance for the model as a whole indicates that the regression is highly significant at the 0.0005 level. The regression coefficients for YIELD and AWARE are statistically significant at the 0.0005 level (shown in column 4). The regression coefficient for FAMILY is also significantly different from zero at the 0.029 level. The reader should note that the same var- iables contributing to the explanation of "innovativeness" in the multiple correlation are the same as those found significantly related in the zero-order correlation tests. On the other hand, the variables for income, ownership of land and cultivable area appear less significant in explaining innovativeness. Attention should be focused on columns (5) and (6) of Table 5.12. Column (5) shows that R2 which would be obtained if Xi were deleted from the least squares equation and the equation were recalculated. In more precise terms, R2 delete i, "is the proportion of the sum of the squared deviations from the mean of the independent variable which can be accounted for by all of the independent variables [including the constant variable (mean of the dependent variable )] except variable Xi."gg/ Column (6) shows the difference between each value shown in 2 column (5) and the R for the regression model. Quite clearly, column (6) 19/ __'The explained variance compares favorably with other studies of a similar nature dealing with innovativeness and multiple correlation analysis. For a concise summary of other studies see: Rogers, Everett M. Modernization Among_Peasants: The Impact of Communication, 92, 513,, pp. 3025303, Table 13.4. -_'Ruble, William L., Donald Kiel, and Mary E. Fafter, "Calculation of Least Squares (Regression) Problems on the Least Squares Routine," Agricultural Experiment Station, Michigan State University, November, 1969, p. 38. 71M 5. Independ 96 TABLE 5.12 EXPLAINING INNOVATIVENESS--MULTIPLE CORRELATION (N = 143) Model No. 1 Independent Variables R2 = .472; R2 = .448 Sig. < 0.0005 Code Names Re .Coef. Std-Dev. Sig.= m R2 Deletes Rz-R2 Deletes (l) 2) (3) (4) (5) (6) Constant 0.4713 0.4482 0.295 -- -- ECONOMIC YIEED 0.0469 0.0069 <0.0005 0.293 0.179 AREA 0.0118 0.0168 0.486 0.470 0.002 FAMILY 0.0232 0.0105 0.029 0.453 0.019 OWNER -0.0023 0.0024 0.348 0.468 0.004 INCOME 0.0000 0.0000 0.217 0.466 0.006 COMMUNICATION AWARE 0.4309 0.0933 <0.0005 0.389 0.083 Model No. 2 Independent Variables R2 = .465; R2 = .433 Sig. < 0.0005 Code Names Reg.Coef Std-Dev Sig. = a R2 Deletes RZ-R Deletes (1) (2) (3) (4) (5) (6) Qggstggt__ 0.8522 0.5202 0.104 -- -- ECONOMIC A YIELD 0.0459 0.0071 <0.0005 0.297 0.168 AREA 0.0233 0.0160 0.148 0.456 0.009 OWNER -0.0016 0.0024 0.506 0.463 0.002 INCOME 0.0000 0.0000 0.132 0.456 0.009 COMMUNICATION AWARE 0.4063 0.1004 <0.0005 0.399 0.066 MMC -0.0616 0.2372 0.796 0.465 0.000 ILC -0.2718 0.2189 0.217 0.459 0.006 ICC 0.0573 0.2448 0.815 0.465 0.000 97 shows that if the economic variable YIELD were excluded from Model No. l, R2 would be reduced by .179. Furthermore, if AWARE were dropped, R2 would be reduced by .083. The deletion of any other variable does not alter the value of R2 substantially. Relative to the other variables, the economic variable YIELD and the communication variable AWARE add more to explaining innovativeness than any other variable in the first model. Model 2 should be fairly self-explanatory. 0f importance, how- ever, is the insignificance of the three variables MMC, ILV, and ICC in explaining innovativeness. Columns (5) and (6) highlight this asser- tion. As in Model 1, YIELD and AWARE explain most of the variation in the dependent variable, "innovativeness." In sum, there is evidence that multiple correlation methods using economic and communication variables can be used to explain the varia- bility in innovativeness. An R2 of over .45 was obtained by each of the two models. The analysis, however, draws some rather simple conclusions: (1) Those farmers who were characterized by a relatively high yield with dwarf wheats in 1969/70, are more likely to be innovators than others in their community. The same farmers, in turn, may be identified as the best farm managers with better quality land, vis-a-vis the other smallholders. (2) The sooner farmers become aware of yield increasing innovations which are neutral to scale, the faster they may adopt the innovation. (3) If we accept the first two points, then it becomes evident that any program designed to diffuse innovations must first concentrate on creating awareness among smallholders. Aside from the important role that the radio can have in diffusing innovations, it behooveS'Uuaextension staff to select those smallholder farmers, who exeml i fy :1 furthe separate cation va are, it :11 variat and seque 98 exemplify the "best" farm management practices, to serve as demonstrators of further innovations. (4) Finally, it does not appear possible to separate with any success the relative importance of economic and communi- cation variables catalyzing the process of diffusion and adoption. Further- more, it does not appear to be a useful argument to pursue; that one set of variables is more important than another. The two are interdependent and sequentially different from one another. CHAPTER VI A PRELIMINARY VIEW OF NON-FARM EMPLOYMENT AND MIGRATION OF HAZARANS Introduction The purpose of this chapter is to summarize the findings from the two surveys on the salient aspects of off-farm migration and non-farm employment. The surveys of Hazara's smallholders sought answers to the following questions: (1) What importance do prior familial ties in the rural areas and in the urban areas of destination have on where a person migrates, and what steps does the migrant take to make his decision reversible if things don't work out for him? (2) What are the particular characteristics of the off-farm migrant, family and the village farm that they leave? (3) What are the important country and zonal differences in rural-to-urban migration -- is the migration one way between two points or two ways, a back and forth movement from village to city? Or, is off—farm migration taken as a once-and-for-all movement to the large towns? (4) How well are migrants from various source areas with different characteristics prepared for employment, i.e. what jobs do they take, how do they maintain themselves? (5) Does the rural-urban migration typically lead to a transfer of unemployment from traditional agriculture to traditional services in the cities? (6) Is there a return-flow of cash and goods from urban areas to rural areas -- is this economically significant, how does it affect farming decisions 99 100 of those family members who remain in agriculture? Answers to the above questions come from respondents in Lora and Oghi thanas who were asked about the family members who had left the farm to work outside the village communiiy. Infor- mation provided was on age and years of schooling, ability to speak Urdu and/or English, place of employment, length of employ- ment, why people left the farm and a few other questions. Overall, the answers come from 226 respondents. In the following, some basic considerations on the prospects for employment in the Northwest Frontier Province will be pre- sented. Next, some attention will be given to defining the extended family that seems to influence the pattern of migration that occurs. Then, the findings of the survey will be reported as a set of propositions regarding the salient aspects of off- farm migration. This is done with the feeling that more research is needed on this subject and that the findings are preliminary in nature. As such, the propositions serve as reference points for further enquiry and analysis. In the last section of this chapter, the implications of migration will be presented. Basic Considerations According to official estimates, the North West Frontier Province (N.W.F.P.) is a net contributor to the labor force in the rest of West Pakistan. Moreover, the magnitude of outmigration from the N.W.F.P. is reported large: (1) In 1951, the population of Peshawar Division and the adjoining tribal areas was 4.9 million. By 1961, the population had increased to 6.4 million -- an increase of 28.2 percent for the ten year period. During the 101 same period, 1951-1961, there was also a net out- migrationl/ of 77 thousand people from the same area. (2) It is estimated that the population in the same area for 1970 is 7.9 million people. Assuming the rate of out-migration to continue from 1961 to 1970 at the same rate as between 1951-1961, the et exodus is estimated at 86 thousand for the periodvg What makes the estimates particularly revealing is that the prospect for absorbing labor into the industrial sector of the N.W.F.P. does not look too promising. With a limited industrial base of 115 registered factories in the N.W.F.P.--compared to 1,890 factories in Karachi and a balance of 4,240 units in the rest of West Pakistan--it has been estimated that the industrial sector of the N.W.F.P. does not even hold 2 percent of the labor force (of about 3 million persons) in the area.§/ In addition, it does not appear that the large manpower exodus adversely affects the industrial labor supply in the northern Province. On the contrary, one of the Frontier's major resources appears to be labor. 1/ Net out-migration refers to the difference between the number of persons "born outside-counted in Peshawar Division" and "born in Peshawar Division/counted outside" as recorded in the 1951 and 1961 Census of Population. For a summary of the Census data, the reader is referred to: Matin, Abdul, 33, al., Resource Base and Economic Progress of the Peshawar Valley, Stfiay conducted by the Department of Economics, University of Peshawar, June 1970. 2/ Khan, M. Ahmad, "Population and Migration Pattern: N.W.F.P. and Adjoining Tribal Areas". Paper prepared for the Project Office, Regional Development Plan, University of Peshawar, North West Fron- tier Province, January 1971. 3] Matin, Abdul, Industrialization of the N.W.F.P., Board of Economic Enquiry, July 1970, See Chapters 1 and 2 for a general description of industrialization and labor resources in the N.W.F.P. 102 The Family “Family" Defined "Family" is defined differently among societies and arises primarily out of the many complex sets of alternative marital arrangements that can be made. In Hazara District, where most persons are Muslims of the Sunni Sect, the villagers adhere to the "prescriptive" principle, that is they indicate whom a person shall marry. The most widely practiced arrangement of this kind is known as "cross-cousin marriage" which is illustrated in the following paradigm.5/ If the prescriptive principle, coupled with the practice of cross-cousin marriages, is followed closely -- and it usually is -- no marriages will take place between tribes. In turn, it is common to find whole villages or areas made up of one tribe, such as the Abbasis or Gujars, etc., with extremely close and hence closed relationships. The nature of the relationship between a villager interviewed and the villager working outside the village helps to explain what the composition of a family is in a Pakistani society. The greatest proportion, or 56 percent, of those working outside were sons of those interviewed. The next largest group, or 30 percent of those working outside were brothers. The rest were nephews (8 percent), cousins (2 percent), fathers (1.7 percent) and uncles (.4 percent). 4 T/Terminology and definitions for this discussion are taken from Felix M. Keesing, Cultural Anthropology: The Science of Custom, Holt, Reinhart and Winston, New YOrk, May 1963, pp. 255-286. 103 A Male 0 Female Brother-Sister u 11 U (:5: C) Husband - Wife a1 ._. Figure 6.1 ILLUSTRATION OF CROSS-COUSIN MARRIAGE I! L_|I The type shown is a symmetrical or double one. By following the lines of relationship and the marriage unions it will be seen that a man, in marrying his mother's brother's daughter, is also marrying his father's sister's daughter. Note that lines indicate blood relationships, and equal signs indicate marriages. Source: F.M. Keesing, ibid, p. 259. Locality and residence in marriage also defines the family. The most common practice in Hazara District is for the bride to go live with, or in the same locality as, the husband's people, so-called patri- local ("father-place“) residence. But present day land pressure, as found in the sampled areas, may require a home apart from the parents of both, so-called neolocal ("new place") residence, which is rare but can be expected to increase in the not too distant future for the people of Hazara District as the man to land ratio rises. 104 The Pakistani family is a consanguineal type of family in which ties of “blood", or descent, dominate. The "family“ is thus an extended family with ties running across and among cousins, to grandchildren, to anyone associated by a blood tie. This strong bond among members is sometimes loosened to include close adult friends who are called uncle and auntie (in the local language) in front of the children. This treatment of friends tends to reinforce the familiness of the Pakistani culture. It also serves as a way for female members of the family - practicing purdah?/- to know non-family members as part of the family without breaking traditional family ties and practice of purdah, However, the number of non-family friends introduced to the whole household is rather limited. Many male friends can spend a lifetime together and never see each other's wives. "Family" Influence With such large extended family groupings there are metic- ulous definitions and rules governing the behavior of individuals. Men and women fit into well-defined divisions of labor where the common household chores are divided among young and old. Sometimes the work allocations appear skewed in favor of the men as the women work during all moments of the day. They are the first to rise and the last to sleep. They do most of the cooking - except during large wedding feasts when professional village cooks é/ "Purdah" is the practice of female seclusion. The word purdah literally means veil or curtain. 105 prepare the meal - make the butter in seemingly secret fashion, wake the men, feed and clothe the children, weed the fields of greens which they feed to the animals, collect the fire wood, get the water, etc. The men manage the house, prepare the fields, harvest, go to the market, handle the cash, and smoke the hukka_ (water pipe) most of the evening while exchanging words with the neighbours. Young parents take their place among the number of persons working collectively, and are subject to the direction of elders. A man may not be the head of his household as his father, uncle, or older brother can "command" the roles of his family. In cer- tain respects, there is no personal freedom and initiative, little privacy, and slight authority over their own children if an elder lives with the family. On the other hand, the extended family system shows unity of effort in economic matters and creates an umbrella of security for all to share. At times of crisis, such as sickness, death and indebtedness (easily brought on by marriage and the "bride-fee"), the situation is likely to be met by all members of the family with little fuss and fumbling, for such events are common and there are many hands to help. The ex- tended family also has a continuity of life. As the old people die off, the young take their place. Everyone experiences a change in role with age; they gain authority, responsibility, and prestige which everyone looks forward to and receives in turn. Most often, family males make the economic decisions in a Pakistani household and determine the destiny of the family. Indirect questioning, without an interview schedule, revealed that family males decide which member (or members) will continue farming and 106 which members will seek off-farm employment. The family elder, of course, has the final word. Those who remain to farm are responsible not only for the operations of the farm but the wel- fare of the whole family. 01d males are almost certain to stay and a younger male (or males) may also stay to do the hard labor. It is believed that those who leave 90 to places where other family members or village members are present. These forerunners of outmigration serve as the contacts for the newer arrivals. How the first outmigrants of a family adjust to off-farm work and conditions is a question largely unanswered. Propositions Derived From Survey Findings (1) Migrantsé/ are_generally4young_men with some schooling and they are usualLy from small farms with "medium size" extended families. Table 6.1 provides a general overview of migrant characteris- tics. Those who have migrated presently average about 28 years of age. 0n the average, most have been working away from the farm about 6.7 years. The difference between the two indicates that migrants leave the farm at about 21 years of age. When they leave the village they have approximately 4.9 years of village schooling, some ability to read and write Urdu. It should be noted, however, that there is some variability in the data since some migrants are much older and cannot read or write. But we can generally expect these characteristics to be representative of most of Hazara's migrants. 6 . "[The word migrant is used rather loosely in the following discussion. Broadly, it refers to those who have permanent jobs (or seek such jobs) outside the village community and who reside in a house away from the village family. 107 TABLE 6.1 - MIGRANT CHARACTERISTICS Average Average Number Average Age Years of Age of Years Working At Time of Schooling (1970) Outside the Village Migration Lora 27.29 6.06 21.23 5.13 Oghi 29.46 7.33 22.13 4.66 Both 28.38 6.70 21.68 4.90 Table 6.2 indicates that all migrants are from farms which have less than 20 cultivated acres. Ninety-six percent of the migrants are from farms with less than 15 cultivated acres, 88 percent from those with less than 10, and nearly 57 percent from 5-acre farms. This suggests linear relationship between farm size and out migrants directly, at least at this level of aggregation. TABLE 6.2 - DISTRIBUTION OF MIGRANTS BY SIZE OF CULTIVATED ACREAGE ON "THE FARM" Average Number of Migrants from; Percentage of Both Cultivated . (86535) Lora Oghi Both % Cumm. % < 5 90 46 136 56.90 56.90 5 ‘ 10 23 52 75 31.38 88.28 10 < 15 7 13 20 8.37 96.65 15 < 20 1 7 8 3.35 100.00 Total 121 118 239 100.00 - 108 Table 6.3 shows that the size of extended family varies to some extent from less than 5 family members living in the same household to more than 40. The majority (54%) of the migrants is from extended families with 6 to 13 members. In families with 5 or fewer members (which probably means only one adult male) the men and children are apparently needed to manage the farm and cannot be released for outside jobs. However, not much can be said at this time as to why some larger families do not have any migrants. A cursory look at the data suggest, though, that large families are found on large farms. TABLE 6.3 - SIZE OF EXTENDED FAMILY OF MIGRANT LABOR Size of Number of Migrants From Percentage of Both Extended Family Lora Oghi Both % Cumm. % 5 or less 15 12 27 11.30 11.30 6 - 9 43 21 64 26.78 38.08 10 - 13 32 33 65 27.20 65.28 14 - 17 10 19 29 12.13 77.41 18 - 29 16 21 47 15.48 92.89 30 and more 5 12 17 7.11 100.00 Total 121 118 239 100.00 - All of the above, however, pertains to those families with off-farm migrants. Table 6.4 compares the mean values of families with migrants with those families who have no relation working out- side the village. In brief, compared to families with migrants, families without migrants are smaller, earn more cash in the village, 109 sow more acreage with wheat, have slightly larger farms (though not significantly larger) but have lower incomes per capita, primarily because they do not receive any migrant remittances. With smaller families, they seemingly do not have males of working age and in excess (over the labor required) to leave the farm. Families with migrants, on the other hand, with comparatively more people in the household, can "afford" to send family members to urban centers in search of jobs. TABLE 6.4 MEAN VALUES OF FACTORS FOR FAMILIES WITH MIGRANTS AND FOR FAMILIES WITHOUT MIGRANTS: HAZARA, 1970.* Families With Migrants Families Without Migrants Factors (n=l30) (n=96) Mean Std. Dev. Mean Std. Dev. Percentage of Land Owned (%) 88.346 30.434 80.333 39.056 Dwarf Wheat Area (Acres) 0.919 1.084 1.407 2.106 Total Wheat Area , (acres) 1.559 1.201 1.978 2.164 Total Cultivated Area (acres) 4.865 4.521 5.215 6.054 Cash Earned In Village/yr.(Rs) 593.72 1,037.93 1,227.63 1,489.87 Migrant Remittances per yr. (Rs) 1,269.42 1,521.34 0.00 0.00 Total Cash Income A per yr. (Rs) 1,863.14 1,861.91 1,227.63 1,489.87 Size of Family (#) 12.13 7.74 9.90 7.53 Income per Capita (Rs) 188.05 199.57 130.15 140.70 *The factors in this table are not normally distributed since the mean value minus one standard deviation for some observations are negative. This suggests that the largest proportion of the farmers under study fall below the norm indicated by the mean value. 110 (2) Mostly adult males leave the village family. Married men usually staygaway from their family for a year or more while they work in other places. Table 6.5 shows the marital status of family members working outside the village. Of the 239 males working away, 73 (31%) were single and 166 (69%) were married. Of the married ones, 137 (57%) had their immediate family (wife and children) living in the village under the custody of the extended family. The rest, 29 (12%), of the married men were reported to have their family with them in the designated place of employment - in other words, living outside the village. Persons interviewed were asked how often the family members working outside the village visited the village. The majority of those outside (65%) visited their family in the village only once a year. There appears to be, as one might expect, a strong relation- ship between the distance from village to place of employment with the number of village visits per year. This will become evident when we examine the distance to the places of employment for most Lora and Oghi villagers below. 111 TABLE 6.5 - MARITAL STATUS OF FAMILY MEMBERS WORKING OUTSIDE THE VILLAGE: OGHI AND LORA THANAS, JUNE 1970 MARITAL STATUS* Single Married with Married with Total Family in Family in Place Village of Employment Lora 46 67 8 121 (.380) (.554) (.066) (1.000) Oghi 27 70 21 118 (.229) (.593) (.178) (1.000) Both 73 137 29 239 (.306) (.573) (.121) (1.000) *Figures in parenthesis show proportion of total. Total number of villagers interviewed who have relations working outside the village = 130 Lora (63) + Oghi (67) thanas (3) Male family members_go to such far-off places as Karachi to find various forms of employment. One-third of the family migrants were employed within a 200 mile radius of their village; mostly within Peshawar and Rawalpindi Districts. However, approximately 50% were working in Karachi, about 1,000 miles away from either Lora or Oghi thanas. The actual distribution for the distance travelled by the migrants is shown in Table 6.6. 112 TABLE 6.6 PERCENTAGE BREAKDOWN BY DISTANCE OF EMPLOYMENT FROM VILLAGE OF VILLAGERS WORKING OUTSIDE THE VILLAGE, JUNE 1970 Approximate Number of Villagers Percentage Distance from Working Outside the Distribution Village to Place Village From: of Total of Employment (miles) Lora* Oghi Total 1 - 100 22 36 58 24.26 101 - 200 8 14 22 9.20 201 - 300 6 9 15 6.28 301 - 400 4 3 7 2.93 401 - 500 O 0 0 0 501 - 600 3 4 7 2.93 601 - 700 0 O 0 0 701 - 800 O O 0 0 801 - 900 0 0 0 O 901 - 1000 75 51 126 52.72 1001 - 2000 0 2 2 0.84 on the road* 2 O 2 0.84 Total 120 119 239 100.00 *Two persons were reported working and sleeping in the one truck that they drove between Lora and Karachi. They visited their own village and families in Lora only once every three months according to the villager interviewed. 113 Table 6.7 gives a breakdown of the major cities employing Hazara's migrants. It indicates again that the overwhelming pro- portion of migrants ends up in Karachi. However, an apparently large number is in the Rawalpindi/Islamabad area which is particularly close to Lora. In addition, it is important to note that Islamabad, the national capital of Pakistan, has shown a remarkable growth in populace and buildings in the last decade. In the future, it may well attract increasingly larger numbers of migrants than indi- cated by the present data. TABLE 6.7 - PLACES OF MIGRANT EMPLOYMENT AND DISTANCE T0 VILLAGE Major City of Approximate Distance to: Number of Migrants Employment Lora Oghi Lora Oghi (in miles) Rawalpindi/Islamabad 40 8O 21 4 Peshawar 120 110 6 8 Lahore 215 290 9 3 Karachi 1,000 1,090 49 71 Most villagers travel by bus or train to their place of employment. It is not known whether any travel by air also. A one— way trip from Rawalpindi to Karachi takes 26 hours and costs anywhere from Rs.30 to Rs.170. This is relatively inexpensive travel even by Pakistani standards. Two things are noteworthy: (i) Karachi is the largest cosmopolitan city in Pakistan with the largest centralized industrial base. Wages are notably higher than in other cities with jobs available in a variety of fields. (ii) Lahore is also large and industrialized, within approximately 250 miles of the sampled thanas, yet relatively few villagers 114 were reported to be working there. The reasons, for which there are no immediate answers, may be: a. Lahore's wage rate structure may be relatively lower than that of Karachi's due to any number of factors like large numbers of available laborers, relative to the number of jobs; skilled or semi-skilled jobs for which members of Lora and Oghi thanas are not qualified to take; etc. b. Lahore is the center of the Punjab and the members of Lora and Oghi belong to the N.W.F.P. The people of Oghi, in particular, speak Pushto and have closer ties with their Pushto speaking brethren. Punjabi and Pathan cultures have historically not mixed comfortably. But the villagers of Lora, for all practical purposes, speak Punjabi and look and act much more like Punjabis than the people of Oghi. Yet, they have a small number of villagers working in Lahore. Karachi, on the other hand, being more cosmopolitan, has a great diversity of people, languages and traditions with which a great many people can identify. Hence, it might be an easier place for migrants with cultur- ally different traditions to settle in. c. It could simply be that the people of Lora and Oghi have had more contacts over longer periods of time in Karachi. This may be perpetuated by common-law leases on particular jobs which certain families seem to hold. With regard to the last possibility, it was found that 65 (28%) of those interviewed had two or more family members working out- side the village at the time of the interview. As shown in Table 6.8, 77percent of the 65 were working in the same city. And, out of 65, about 44 percent were working together in the same job. Hence, there is reason to believe that kinship ties are important for finding jobs off the farm. Ville 861511 110.1 115 TABLE 6.8 - LOCATION OF MIGRANTS FROM FAMILIES WITH TWO OR MORE FAMILY MEMBERS OFF THE FARM Village In Same City In Different Total Number Cities of Families . With Two or Sub- In Same In Different . Total Job Jobs More ”'grants Lora 24 19 5 ll 35 Oghi 26 10 16 4 30 Both 50 29 21 15 65 Total 76.92 44.62 32.30 23.08 100.00 (4) Jobs taken by_barani smallholders are varied but the generalgjob is lowjpaying_and manual. Conventional wisdom has it that a village outmigrant from the N.W.F.P. is limited in the types of jobs he takes; namely, most are thought to be chowkidars (night watchmen) and simple traders in the city markets. However, Table 6.9 indicates no such stereotype. The sample survey revealed that a wide range of jobs are taken out- side the village by Hazara District villagers. Such jobs have been classified in six separate groupings in the table below with an extra group for those unemployed. Some of the job skills were arbitrarily included in a certain classification. For example, all drivers were classified as "business or household servants". In one known case, the driver was the owner of his own car and could easily be classified in the "self-employed" group. The notes to the table will help the reader do his own reshuffling of the jobs included under the various classifications. 116 .AO-FV xeaeu seameas awn: .Ao-_v comm . ucwaugmamo mxgox ovpnza .Agum .oupv pcmucmuu< xgmmgzz ucm tease . ucmspgmawo ummgom .Aoumv memcms mace; .A4-~ .onpv acmasou uwsuumpm .Aoupv pcmummmm< upmwu - mgzapauwsm< mo acmEHLmamo .Aoapv :memEQWF .Agap .OImV mgmgummp Foogum .A41mv coumwxmm we >U>L=m .A4-~ .01Fv cos loavp gammePmp vcm mconampmp .Aonmv mgmxgoz Amzpwms .A41m .ouev wuwpoq whoumv muWFoa Lawson .ngnm .oumv mmcw_sw< PocowpmcgmucH :mpmwxma saw: umxcpasm mmocu op memeoz "souumm uwpnaa Lo u:m&cgm>oo “my ahxquz ou we meme cowuucwpmwu ocv z>mz Lo .musomew< .xEL< :mpmwxoa on» cw smcpvm omega op msmmmm uxgmpvaz chowumz ANV .mpmw 3mw>gmucw mg» mo mcucoe Lace segue; mom—_r> on“ ummp xPco we; use weep awe?» mcp sow mmmppw> on“ mo uzo mom: Aws04 Eocmv snow pan ppq .ucmsxopqsm .mm>wumFmg mw: mo macaw as» zocx go: new umzmw>gmpcw smmmppw> asp .nmxopasmca mo mmmmo .mnow m>mg go: on use ucwexo_q5m wwwmuao Loe mmmp_w> use weep m>ms on: mFQoma ”umAOEmEmca AFV —Fm :H “Ac n acme new A n oL04 mews: xgommumu 56mm :w moon23: szpum 0p cemugma mwmmgucmgmq cw mmgzmw+v mesa—cu co mmpoz a coco.P NFNN. mNPN. mmmo. _mmp. mk_m. _kp_. eemo. peaaeaa omN mm mm m mm mm mm mp spam me_ am ON m cm mm o, 5 _;mo _N_ KN mm N mp mm m. m ago; Amy ARV on Amv Aav Amy ANV AFV acm>gmm mmmcwmzm Louumm mpm>wgm Fp_z Louomm uwpazm xgmpwpwz Page» so upocmmzo: so umonaEmnmpmm Logonme mpwuxm» so ucmEch>oo Pocomumz umxwhmEm mmwwwmn um< mo zefiz om>osazm chap .euHmemHo m as”: mmoq oz< m=e03azm - m.m m3m .A4uav ucm>emm Pmuoc .A4-_ .olmv amazoumz “cow: .Ao-_v xonlcm V Seascapes Pageamo; .A3-N .o-_v peaseam aeoeamsa; .apnmzwonsev measeew .Ks-a .o-_~ :wzo__om mg» a? msmzuo Lee mcwxeoz omega op msmwmm “mu:u>emm mmmcvmzm so upocmmso: -gmmz .A4-F . ._ mememmm "mach m .Ao-~v muwpcwgqam aogmxgoz .Aoupv suntan .A4-¢ emammxaocm .Ao-_v Lounges xcmn .Ao-mv smmmcme mmpmm .Ao-pv cpwmeumFQ .A4umv mama to Loves“ .Ao-m Lopwmp .A4-¢V gmcxo prpmuomu .Ao-_v uwcmsome o-mv Lopomepcou mchFwsn .Aoupv .ou ovpmo mmxopqsm .Ao-~v Loungmao «genome .A41m .oumv me—u . Jupv gmgmwpoq _vmcmu: .A4-_V smpmmn Lone?“ .A4-_V mmzo; emzoa smmcwmcm ucmcwmmg .Ao-—v LmFown mow .Moupv coma .A4-~ .oumv ucmvcmpum co_ampm mam .Agumv memn "Louumm mpm>wea on» mo mgmcuo so» so mm>meEmsp so» gmguwm mach mo “mm msomcmPFmUm_E cw xeoz on: mmocu op mgmwmm "Louumm mum>eea Lo vmxdpaEm epmm .Acumv cacao ucm Aoqu xgopuoe warn .A4upv mamaeou Pee .Ao-PV_F.rerwmmw .Ao-FV mcwcmugmm .A4-Pv mcvxuma .Ao-pv covpuagumcoo cw mxmmu Fences mcwov mmocu op mgmmmm “gmgonmg .wcumgmx cw emumUOF x—umos mom cows: mp—ws mpwuxmu cw mcvxgoz omega 0» mgmwmm "ppwz upwuxmp Anm::wucouv I m.© m4m1y‘means of objective tests of performance rather than on the basis of family emolo: by d1 tween count effec any 1 certg WEfit Unti' caoa Seal 122 family and status. The fourth implication is that off-farm migration and non-farm employment serve to reduce the disparities within Pakistan conditioned by differential rates of natural increase in the population and be- tween the income earned by the many different groups which serve the country. It should be made clear that there is an income generatinn effect going on with the free flow of labor from the N.W.F.P. Namely, the poor people of Hazara District are able to work in order to feed their poor families. The work they do fills needed positions in Karachi and other large centers. The money earned in turn goes to purchase grain from the Punjab and keeps the more productive farm workers of the Punjab employed by an effective demand for Punjabi goods. Break any link in this chain and more problems are created to deal with. In certain respects, the N.W.F.P. should be concerned with the employ- ment situation in other areas holding large numbers of its migrants until such time that the Province can develop its own productive capacity to employ more people. Remunerations back to the farm have apparently aided some barani smallholders to be rapid adopters of the new, high-yielding varieties of wheat. They probably wouldn't have adopted without off-farm sources of income. In essence, the extended family provides some degree of "risk insurance" against penalties from crop loss due to innovation; especially, when family members have some assurance of non-farm employment. The fifth implication is that migration opens the rural areas to outside information. Horizons are widened and aspirations enhanced. Farmers become more cosmopolitan (worldly) and experienced in ways which can move them to greater opportunities, THE E duction wheat v1 NNam- Droduct meats CHAPTER VII THE ECONOMICS OF WHEAT PRODUCTION ON BARANI LAND IN HAZARA DISTRICT Introduction This chapter is divided into two sections. In Section 1, pro- duction relationships between inputs and outputs for dwarf and desi wheat varieties are specified with the Cobb-Douglas production function.11 Comparisons between the estimated regression coefficients and marginal productivities are made in order to depict the economic impact of dwarf wheats on resource productivity on Hazara's barani land. In Section II, the costs and returns of dwarf and desi wheat production are analyzed with partial budgeting analysis in order to assess whether or not the relative income position of barani smallholders has been able to improve for those who use the new high-yielding varieties of wheat. Section I. The Impact of Dwarf Wheats on Resource Use Model of Wheat Production Specification of Functional Form The objective in fitting the Cobb-Douglas production function to the set of cross-section data (discussed below) is to obtain estimates of the marginal contribution of each input to output of either dwarf or desi wheat. Such estimates make it possible to evaluate the perfbrmance l! The term production function refers to the physical relation between a farm's inputs of resources and its output of a particular commodity per unit of time, exclusive of prices. 123 oi the used wi {ebb-Dc oartial n CS. study 1 I119 fu 13 C01 least 124 of the sample of barani farms and to compare the productivity of inputs used with dwarf or desi varieties. Moreover, the common use of the Cobb-Douglas production function in many studies of a similar nature partially attests for its general usefulness as a tool in applied econo- mics.g' By following past experience, the results contained in this study may lend themselves to comparative analysis with other studies. The function has the general form: b1 xb b- Y.=AX 2...x.1u , j 1j 2j 13 j is converted to linear logarithmic form, and is fitted by ordinary least squares procedures. In the function, Yj = output of farm j, A = some constant, Xij= amount of input i used by farm j, bi = the elasticity of production of input i, Uj = a stochastic term. The estimates with the above functional form are notably subject to a variety of possible errors which have been discussed and analyzed §/ elsewhere. The major limitations of our specified model are that: 2[For example, see a more recent study and its references: Yotopoulos, Pan A., Laurence J. Lau and Kuttu Somel "Labor Intensity and Relative Efficiency in Indian Agriculture", Food Résearch Institute Studies in Agricultural Economics, Trade and Development, V61. IX, No. 1, 1970, pp. 43-55. g] Massell, Benton F. and R.W.M. Johnson, Economics of Smallholder Farming in Rhodesia: A Cross-Section Analysis of Two Areas. *Food”Research Institute, Stanford University, 1968, Chapter 7, "Problems of Statistical Estimation", pp. 38-44. 125 (l) we have a single equation model of production; (2) some variables ,will undoubtedly be missing and others will be unqualified, such as the flow of capital and the quality of soil; and (3) some sample variables are zero for some farmers, i.e., chemical fertilizer.fl/ But for our purposes, these errors should not present any major limitations on the analysis of the functional relationships and the results should be useful in explaining the general pattern of resource use. The results should also be useful in identifying inputs which pro- mise the largest increase in output from greater or lesser employment. Sample Size The first round of interviews during May/June, 1970, collected the raw data for the production function specified above. Out of the 143 farmers who were interviewed, 71 exclusively grew dwarf wheat, 35 desi wheat and the rest of the sample grew both dwarf and desi varieties. Those growing both varieties are excluded from the production function analysis below. Due to the restrictions placed by a small sample, no attempt is made to compare the production functions of Lora and Oghi. The data are combined in order to serve as a sample of representative observations on Hazara District as a whole. Although four outliers appeared in the dwarf wheat sample of respondents and two in the desi wheat sample, all observations are kept intact. 5.[The reader is referred to: Johnson, S.R. and Gordon c, Rausser. "Effects of Misspecifications of Log Linear Functions When Sample values are Zero or Negative". American Journal of Agricultural Economics, Vol. 53, No. 1 February 1970, pp. 120-124. I? 126 Production Variables Six production variables are considered in this study: output (the dependent variable), land, labor, seed rate, nitrogen and phosphate (the inputs). Output (Y3) The dependent variable, output, is measured in terms of maunds harvested; one maund equals 82.286 lbs. Conversions were made from the village unit of measurement, the wodi, to the maund during the interview 5/ periodf_ Land (X1) Land is measured in terms of acres planted with each wheat variety during 1969/70. The standard village unit of measurement for land is the kanal. Eight kanals equal one acre and the conversion was made accordingly. The quality of land is subject to some variation. Barani small- holders can distinguish between bgfljgr_(or 53553:), very poor pebbly soil, and maira, improved soil near the home which customarily receives farm yard manure. In almost all cases, wheat was sown on mgira land. No further attempt was made to qualify land in the sample. Labor Utilization (X2) Labor utilization is measured in terms of man-hours used for field preparation and sowing. Since each pair of bullocks used to pre- pare the land needs one driver, man-hours were derived by asking the é[The wodi is described in Chapter V along with a statement re- garding the reliability of the data on yield. respond the tin number of bull field 1 to est' bionn record- majori Dripar ered New MUN Seed | \ itrea equal D10110 Tate" '0 wt the . 127 respondent: (i) how many times each wheat field was plowed, including the time for sowing, (ii) the number of bullock-pairs employed, (iii) the number of days it took to plow the wheat fields one time using X number of bullock pairs and (iv) the approximate number of hours spent in the field plowing each day. From answers to these questions, it was possible to estimate the number of man-hours used per farm. Family males are charged with all aspects of wheat production. Information on labor input -- both hired and family labor -- was recorded separately in the course of the interviews. The overwhelming majority of the respondents hired no labor for any aspect of the field preparation and wheat processing. 0f the 143 farmers who were inter- viewed with regard to wheat production, only two hired labor. This presents a problem for valuing labor as we will note in the cost and returns section of this chapter. Seed Rate (X3) "Seed rate" is measured in terms of total seers applied to wheat acreage per farm. One seer is equivalent to 2.057 lbs. and 40 seers equal one maund. In a barani environment, high plant populations use proportionately more residual moisture. Hence, it is thought that "seed rate" should contribute to yield variability; it is, in essence, a measure of response to drought stress. Nitrogen (X4) and Phosphate (X5) Fertilizers Nitrogen (N) and Phosphate (P205) are measured in pounds applied to wheat acreage per farm. Since fertilizer use is relatively new to the farmers, they can easily remember at least two or three characteristics 128 of the type applied; i.e., bag weight, price, color, place of purchase. Any two of these can be used to identify the type and amount of fertilizer used. Farm yard manure could not be measured nor included in the inputs because: (i) size of livestock herd differs from farm to farm, (ii) some families use cow-dung for home fires and (iii) the farm yard manure is spread until it "looks good“ in the field. As such, it is never weighed nor evenly distributed among fields. Hence, the obvious difficulty in measuring a suitable rate for farm yard manure precluded all attempts to include it as a variable. As discussed in Chapter V, not all farmers used chemical fertilizer. In a Cobb-Douglas function, no output is obtainable if any input equals zero (since the log of zero is not defined). However, we know that some output is obtained without N or P205. Thus, these inputs need to be treated differently from the conventional inputs of land and labor. In order to handle the zero values, we let M = M' + C where M' = number of pounds of fertilizer applied and where C is some constant which was arbitrarily set equal to 0.01. In this way, the log of M is defined for all levels of output. The example of Massel and Johnson illustrates the conceptual changes that must be taken into consideration.§/ In Figure 7.1, fertilizer is measured along the vertical axis, 0E, and land along the horizontal, OT. Curve CD is a Cobb-Douglas isoquant which shows the trade-offs between fertilizer and land for a specific level of output, holding all other gfgp.9_i_t_. pp. 35-36. 129 A C L. (I) .§ 2: L .2 0: ::>'T : 0 Land at ----------------------------- p-S FIGURE 7.1 A MODIFIED COBB-DOUGLAS ISOQUANT inputs constant. Instead of being asymptotic to 01, the isoquant is asymptotic to RS. The distance R0 = C is a measure of natural soil fertility.zj By using the above measure of M in the Cobb-Douglas production function, the regression coefficients for M are not equal to the production elasticities. The elasticity, however, is easily obtained by multiplying the regression coefficient by (X-C)/X wgere X = the value of the variable with the constant and C = the constant. The calculations in what follows were made at the gemetric means of the respective variables. 1/ Ibid., p. 36 §-’ Ibid., p. 47 fiver EXEI trat W‘ 3905 lot 130 Overview of the Inputs Used to Produce Dwarf and Desi Wheat The contrasts between dwarf and desi wheat production are exemplified by the input/output relationships shown by the geometric means in Table 7.1. A comparison of the geometric mean values illus- trates that for all input categories, substantially more resources go to produce dwarf wheats than desi wheats. In turn, the yield differ- ential is also substantially higher for dwarfs. In a purely physical sense, innovation with dwarf wheats is both factor using and output increasing; the magnitude of which is analyzed below. The comparisons also indicate that the production functions for dwarf and desi wheats, though computed with the same functional form, are different from each other. Analysis of the significant differences between dwarf and desi wheat production follows. TABLE 7.1 SUMMARY OF INPUTS AND THEIR GEOMETRIC MEAN VALUES FOR DWARF AND DESI WHEAT PRODUCTION: HAZARA DISTRICT, Output and Geometric Mean Values* . Input Category Dwarf Wheat (n=71) Desi Wheat (n=35) Y Yie1d (mds) 14.436 (13.59) 7.764 ( 7.48) X1 Land (acres) 1.062 ( 1.00) 1.038 ( 1.00) X2 Labor (man-hours) 76.926 (72.41) 61.712 ( 59.48) X3 Seed Rate (seers) 37.733 (35.52) 31.954 ( 30.80) X4 Nitrogen (lbs) 4.529 ( 4.26) 0.036 ( 0.04) X5 Phosphate (lbs) 0.434 ( 0.41) 0.034 ( 0.03) * Geometric mean values are average estimates derived from computations with the Cobb-Douglas production function. Figures in parenthesis show conversions to a per acre basis. Giro id 131 Estimation of Model Parameters Analysis of Variance An F test was used to test for the significance of interfarm diff- erences in dwarf and desi wheat production. For both dwarf and desi wheats, the null hypothesis that all bi = 0(i = 1,2,...5), can be rejected with a Type I error of less than 0.05 percent. The coefficient of determination (R2) for the regression model as a whole for dwarf wheat was 0.7155 (standard error = 0.24) and for desi wheat, 0.6803 (standard error = 0.24). Both Rz's indicate that approximately 71 and 68 percent, respectively, of the sum of the squared deviations from the mean of the dependent variables are accounted for by the independent variables. In other words, both dwarf and desi production models show that the wheat yields are significantly responsive to variation in the inputs quantified. The remaining 28 and 32 percent, respectively, of the unexplained variance might be attributed to such factors as management, soil fertility, weather, and other factors which are not quantified but are assumed to be randomly and normally distributed among the sample of observations. Production Elasticities Table 7.2 shows the estimated elasticities of production together with their standard errors (3.0.) and levels of significance. Elas- ticities are important because they indicate the percent change in output which would, on the average, be associated with a one percent increase in the input concerned while all other inputs are held constant. The elasticities are also needed to estimate marginal productivities of the inputs. 7"." ~ Regn cons- 132 TABLE 7.2 ESTIMATED ELASTICITIES WITH THE COBB-DOUGLAS PRODUCTION FUNCTION WHEAT TYPE INPUT DWARF (n=71) a/ DESI (n=35) a/ Elasticity S.D. Sige Elasticity 5.0. Sig: Constant 0.660 0.472 0.167 1.334 0.524 0.016 1. Land 0.339 0.298 0.260 1.091 0.372 0.007 2. Labor -0.020 0.081 0.810 -0.102 0.188 0.590 3. Seed Rate / 0.318 0.281 0.262 -0.152 0.330 0.648 4. Nitrogen _b/ 0.067 0.019 0.001 0.012 0.035 0.629 5. Phosphate —- 0.050 0.017 0.003 0.014; 0.032 0.603 EISignificance -- The null hypothesis bi = 0 can be rejected with the probability of a Type I error (two-tail test) shown in this column. 9(Regression coefficients for N and P 05 have been corrected for the constant value which was added to aIl observations. A comparison of the two sets of production elasticities reveals consistently the following results: (1) An important factor in explaining interfarm differences in output is land. The estimated elasticity for land sown with desi wheats is significant at the 0.7 percent level of probability. The estimated elasticity for dwarf wheat land, however, has a much higher probability of a Type I error associated with it; that is, in 26 cases out of 100, if functions were fitted to different Samples from the same population, the production elasticity for land lVith dwarfs would be as large or larger than the present estimated Cnaefficient. (2) The production elasticities for labor are negative but not significantly different from zero. As such, in the production Of'both wheats, additional man-hours of labor contribute almost nothing to out ooerai 511/91; orodu cutou meat an in 133 to output. A priori, with large extended families and small scale of operation, we would expect "cheap" labor to be used very (too?) inten- sively; something we discuss in more detail below. (3) All of the production elasticities for fertilizer indicate positive influences on output. However, only the elasticities for fertilizer applied to dwarf wheats are significant. In the case of nitrogen applied to dwarf wheats, an increase in N by one percent increases total output by 6.7 percent. An increase in P205 by one percent increases total output by 5 percent, all other factors held constant. Another important factor to explain changes in dwarf wheat output is the number of seers of seed used per farm.) However, a negative elasticity is evident for the input of seed used for desi wheat; as are the signs on the elasticities for the inputs of labor. The negative sign with the coefficient implies that total output can increase with less of the input. Negative coefficients may result from high intercorrelations between variables. Among the sample of observations, land, labor and seed rate are positively correlated with one another to a large degree as shown in Table 7.3. Such intercorrelation comes about as the farms employ more labor and use more seeds in accordance with expanding wheat acreage. Complementarity of these inputs, however, tends to obscure ‘their separate contributions to output since their stanUard deviations are also large. In essence, overestimation of one production elasticity tends to necessitate some underestimation of one or more of the other Production elasticities. Hence, reduced reliability can be placed in the accuracy of the estimates presented. Aggregation of variables is suggested when high complementarity is evide because lidual C 119W 1 134 9/ is evident.—' However, inputs were not combined in the present study because of the difficulty in finding another way to indicate the indi- vidual contributions of the inputs, i.e., land, labor and seed rate. TABLE 7.3 MATRIX 0F SIMPLE "ZERO-ORDER" CORRELATIONS Dwarf Wheat Model Yield Land Labor Seed N P20 5 Y Yie1d 1.0000 X1 Land 0.7524 1.0000 . X2 Labor 0.5968 0.7332 1.0000 X3 Seed 0.7699 0.9636 0.7052 1.0000 X4 N 0.6211 0.4629 0.3993 0.4918 1.0000 XS P205 0.4949 0.2902 0.3894 0.3306 0.3420 1.0000 Desi Wheat Model Y Yie1d 1.0000 X1 Land 0.8176 1.0000 X2 Labor 0.6182 0.8028 1.0000 X3 Seed 0.7559 0.9508 0.8098 1.0000 X4 N 0.4198 0.4775 0.4146 0.4719 1.0000 X5 P205 0.0852 0.0470 0.0932 0.0066 -O.l652 1.0000 g/Bradford, L.A. and Johnson, Glenn L., Farm Management Analysis, New York, Wiley, 1953, pp. 143-45. targir (“PH HPP fl ticit. means where and 1 from 010: 11Mr ”EST tho: 1300 91v 135 Marginal Productivities Table 7.4 shows the estimated marginal physical productivity (MPP) and the marginal value of productivity (MVP) for each input. The MPP for each input was calculated as the product of the input's elas- ticity of production and its average product calculated at the geometric means of output and input, i.e.: _ E . (G -) MPP)“. - i.) y.) i=1.2. ..., 5. G xij where Eij = the elasticity of factor i in the production of crop j and G = the geometric means of outputs Yj and inputs xij, accordingly. The MPP's for both dwarf and desi variables are multiplied times Rs. 20.00, the average price for both dwarf and desi wheat in Hazara during the time of the interviews, in order to derive the MVP's. Theoretically, MVP's are the net earning powers (or net returns) of the different inputs. It should be noted, however, that there are limitations arising from the method of estimating MVP's. Notably, MVP's are derived from production elasticities. Any bias in the latter also results in the MVP's being biased in the same direction as the elasticities used. The results can be useful, though, in comparing the relationships between those inputs which are significant and for which the farmer has more control; namely, the amount of seed sown per acre and the amount of fertilizer applied to wheat. In the following, an analysis will be given for each factor of production used in producing dwarf or desi wheat. IIBLE 7.4 IiP‘Jl 3. Land 2. Labor 3. Seed L. 1110‘: 136 TABLE 7.4 ESTIMATED MARGINAL PRODUCTIVITIES OF INPUTS USED TO PRODUCE BARANI WHEATS: HAZARA DISTRICT, 1969/70 WHEAT TYPE DWABE_(n=7l) DESI, (n=35) INPUT MPP MVP* MPP MVP* 1. Land 4.601 92.02 8.160 163.20 2. Labor -0.004 -0.08 —0.013 -0.26 3. Seed Rate 0.122 2.44 -0.037 -0.74 4. Nitrogen 0.024 0.48 0.008 0.16 5. Phosphate 0.028 0.56 0.018 0.36 *MVPi = Rs 20 ' MPPi (where i = l...5). Analysis of Resource Use A Note On the Optimal Usefiof Inputs In perfectly competitive markets, where factor and product prices are independent of the actions of any single buyer or seller, inputs are employed in the most profitable quantities when their MVP's are just equal to their prices. That is, for optimal input usage, the cost of employing the last unit of the factor (marginal factor cost of MFC) should just equal the value of the additional product (MVP) which is produced as a result of adding that last unit of the factor. With a perfectly elastic SUpply of the factor, MFC = factor price. Thus, when the MVP of an input exceeds its price (or MFC), it becomes more profitable to use more of the input. The converse is true when the MVP of an input is less than its MFC or price. Knowledge of the way MVP's relate to factor prices indi cat 'and DECaUSE ‘and is oenerar borrow 35 man Qualit ”Ezara EXSect ifact the e 137 indicate whether more or less of a resource should be used. L209. Very little is known about the rates of return to land in Hazara because few farmers actually sell land. This is an age old practice since land is kept in the family and is usually handed down from generation to generation of family members. The most corrmon practice is for farmers to borrow money with land used as collateral. The interest paid on the loan is equal to the cash value of the crops produced on the land (until the debt is paid) by the lender. The production function, however, yields marginal returns to land as many rupees per acre per wheat crop. (The returns are for average quality maira land.) Judging from the relative scarcity of land in Hazara District and the exodus of manpower from the area, it should be expected the rate of return to land would be relatively higher than those of other inputs. The fact that the MVP of land for dwarfs is lower than that for desi wheats should not be construed to mean that farmers should produce more of the latter since the two varieties produced are from completely different production functions as noted above. More important is the conclusion that the marginal returns to land are generally con- sistent with what is known about the scarcity of the input. Furthermore, it indicates that leaving land fallow during the winter period is a costly practice for barani smallholders to continue and that each acre sown with Wheat during winter can be expected to add at least Rs 90.00 to the farmer's total income. Furthermore, this information also gives us an idea on the approximate price to attach to barani land in evaluating net profits For dw Labor :roduc tities two on 0f 39' to uh I: 011 138 for dwarf wheat production. Lem: An interesting conclusion from the results is that the marginal products of labor for both varieties of wheat are zero and with elasti- cities not significantly different from zero. This may be attributed to two main factors: the first is the prevailing labor-intensive techniques of agricultural operations and the second is that labor availability is high relative to available cultivable acreage. The latter is contingent upon the prevailing pattern of off-farm migration. As noted previously, migrants are generally those between 20 and 40 years of age. Thus, a growing cadre of young and a large proportion of old males are left to till the land. The opportunity costs for those on the farm are apparently low in Hazara District and the chances for long distance mobility are less for these two age groups of males. Under- standing the general pattern for off-farm migration plus the fact that many men were still available on the farm, it was not too surprising to find that farmers were plowing more often for dwarf wheat; even to the extent that the marginal product of labor approximates zero. Generally, with dwarf wheats, the number of plowings increased 3 to 4 times over the usual 2 or 3 plowings for desi wheats. Some farmers plowed as many as 6 times for dwarfs. Aside from believing that more plowings would add more fertility to the soil, they generally thought that the new varieties "required" the extra plowings in order to give good yield. The point is, there was labor available to plow the dwarf wheat fields more often than is normally done for desi wheats. Considering all aspects of the labor situation, the findings on labor in the tech dence th increase is avail large an higrant the old for the tradita attach! orobab zero f is apt above form 1 Seed \ f181l DrODI 139 labor indicate that: (1) Even though the labor input used to cultivate the technologically improved varieties has increased, there is no evi- dence that the marginal returns to labor under present circumstances have increased substantially enough to alter the conclusion that surplus labor is available in Hazara District.lQ/ (2) As long as family size is large and families can sustain themselves with partial support from migrant remittances, there will be others on the farm (the young and the old) who will be willing to accept lower than "non-farm" returns for their labor. (3) As long as families are "held together" by tradition in the present cultural context, there will continue to be an attachment to the farm family. With these considerations, it is quite probable that the MVP for labor in the production of wheat is close to zero for all phases of crop production. Hence, the productivity of labor is apt to remain low until farm production can be raised substantially above present levels or until the size of land holding can expand for the farm families. Seed Rates It should be recalled (from Table 7.1) that the geometric mean values of seed rates were 35.52 seers per acre fOr dwarf wheats and 30.80 seers per acre for desi wheats. At this level of "seed rate", the MPP for desi varieties is not significantly different from zero. 0n the other 19lCorroborative evidence of the possibility of an existing agric- ultural labor force with almost zero returns to output is cited in: Mian, Nural Islam, Disguised Rural Unemployment in the N.W.F. Region, Board of Economic Enquiry, 1965, pp. 8. In partiCUlar, it is noted from field survey data that rural workers were engaged in work for only 712 hours during the year (or about 89 days per year). "Agriculturalists proper" were employed only 87 days, according to the study. hand, arent :ontr SDdT' ONEF .1, I .41..- that rats (J (7 140 hand, the MPP for dwarf "seed rate" is positive and significantly diff- erent from zero, according to the production elasticities. The findings contradict the hypothesis that barani farmers would use dwarf wheat seed sparingly with fertilizer in order to effectively use the residual moisture for good yield. The relatively higher seed rate and yield with dwarf varieties make it difficult to recommend one way or the other the amount of seed that should be used per acre. When we view desi wheats separately, it is plausible to explain that farmers (who have used desi varieties for years) are using a seed rate that pushes output to the maximum possible level. Hence, at a rate of around 30 seers of desi seed per acre, we should expect the MPP of this "seed rate" to approximate zero. In the case of dwarf wheats, the higher seed rate and larger output per acre suggest that dwarf wheat seed, especially Mexipak-65, is a relatively more efficient user of residual moisture, even when combined with chemical fertilizer. Tied to this is the realization that farmers plowed their fields more often for dwarfs. The net effect may be that the soil's residual moisture was actually higher on the land plowed more often than on the land plowed for desi varieties. These are in no way conclusive findings, but they do indicate that more research is needed to test the resistance of dwarf varieties to various levels of drought stress. Fertilizer As noted in Chapter V, few desi growers used chemical fertilizer. 0f the 35 farmers included in the sample for desi wheat, five applied nitrogen and another five phosphate; none used both chemical fertilizers on their desi wheat which explains the negative simple correlation 141 for rx5 x6 shown in Table 7.3. Because few farmers who grew desi wheat applied fertilizer and because neither N nor P205 were significant in the case of their production elasticities, there is little that can be said about the use of fertilizer on desi varieties. Hence, attention is focused on fertilizer applied to dwarf wheat. Positive production elasticities were estimated for both N and P205 applied to dwarf wheats and both can be accepted as significantly different from zero at the 95 percent confidence levels. Table 7.5 gives estimated total (TPP) and marginal physical products (MPP) and MVPs for the usually observed fertilization rates of N and P205, holding all other inputs constant at their geometric means. Estimates calculated for nitrogen fertilizer application, indicate that at least 9.56 maunds per acre of dwarf wheat can be achieved with 0.01 pounds of N. With the same amount of phosphate, the minimum expected output can approximate 12 maunds per acre. It should be remembered, though, that the calculations hold when all other inputs are measured at their geometric means. One other observation is that yields in excess of 15 maunds require heavy fertili- zation with N and P205. In addition, the marginal returns to 10 pound units of fertilizer drop very quickly from Rs 1.54 to Rs 0.33 for both fertilizers. Sample respondents who purchased and used chemical fertilizer were asked during the interview the price paid for each bag of fertilizer. Farmers paid, on the average, about Rs. 0.54 (or 54 paisa) per pound of nitrogen and 44 paisa per pound of phosphate. The price of each includes the transportation cost between wholesale outlets and village merchants and the latter's profit. Adding a small amount to account 142 for the cost of transporting fertilizer from the village to the farm, a pound of nitrogen may be assumed to have cost the farmer about 56 paisa and about 46 paisa per pound of phosphate. These costs can be compared with the MVP's estimated with the production functions in Table 7.5. TABLE 7.5 TOTAL AND MARGINAL PRODUCT CURVES WITH MVPS FOR NITROGEN AND PHOSPHAEE FERTILIZER APPLIED TO DWARF WHEATS (ON A PER ACRE BASIS ._ Pounds of Nitrogen (N) Phosphate (P205) Fert'1129r TPP MPP MVP TPP TMPP iNVP (mds) (mds) (Rs) (mds) (mds) (Rs) .01 9.56 NA NA 11.93 ‘ NA NA 0.4191 NA NA NA 13.59 1.630 20.56 4.269/ 13.59 0.152 3.24 NA NA NA 10 14.59 0.076 1.51 16.01 0.770 1.54 20 15.11 0.035 0.69 17.51 0.042 0.84 30 15.39 0.026 0.52 17.77 0.028 0.56 40 15.67 0.020 0.39 18.02 0.022 0.43 50 15.97 0.017 0.33 18.25 0.017 0.35 NA = not applicable. 2[Estimates are approximate due to rounding and extrapolation between some values for TPP. 9-/Geometric mean for P205. S(Geometric mean for N. The geometric mean values for phosphate and nitrogen applied to dwarf wheat were 0.41 lbs. and 4.26 lbs. (per acre), respectively. At these points, the MVP's were much higher than the costs of both. We 143 should expect, that as long as MVP > MFC, usage of N and P205 will expand by barani smallholders if possible. In particular, Optimal use of N at present prices should be between 20 to 30 pounds per acre. Approximately 30 to 40 pounds of P205 would come close to an optimal use of phosphates at the given price of 46 paisa per pound. With the application of these rates of fertilizer, the variability in dwarf wheat yield may be reduced substantially for barani wheat production. Section 11 Costs and Returns to Dwarf and Desi Wheat Production The preceding discussion focused on the patterns of resource use and the ways to produce either dwarf or desi wheat at least cost. Be- fore dwarf wheats are substituted for desi wheats, they must show the ability to raise net profits above former levels for sustained periods of time. What happens if the price of fertilizer or another input rises? Will smallholders still find it profitable to continue dwarf wheats? This question and others can be answered by partial budgeting analysis which is a useful technique for choosing between alternative packages or technologies of wheat production. In addition, it gives policy makers a picture of the input/output relations and the associated values under different prices. Partial Budgeting Analysis Bradford and Johnson define a budget as a written plan for future actions which is composed of two main figures, total revenue and total 11/ expenses.__’ The difference between the two is called net profit or net ll/Bradford, Lawrence A. and Glenn L. Johnson, 92,513,, pp. 328-329. 144 loss depending on whether total revenue is greater or less than total expenses. Partial budgeting consists of altering one assumption of a budget, such as factor cost in order to predict its effects on net return or net loss. As such, it forms a part of static microeconomic analysis and like regression analysis, partial budgeting analysis assumes the conditions of perfect competition and divisibility of inputs. In the present context, involving the comparison of a "new package" (the factor requirements for dwarf wheats) and an "old package" (the factor require- ments for desi wheats), these assumptions do not pose major limitations. The usefulness of partial budgeting analysis can be illustrated with the following example: Assume that a farm is producing 1000 rupees worth of wheat with a total input cost of 400 rupees. The difference leaves 600 rupees to pay for other incidentals of the farm family. Thus, the ratio of net returns to expenses in wheat production equals 1.5 to 1. Now assume that the farmer tries a new package of inputs but now spends 650 rupees (or 150 rupees more than the former expenditure). However, the new package produces 2,000 rupees worth of wheat. In the second case, the net return is 1350 rupees and the ratio of net returns to expenses is about 2 to 1. As soon as the farmer becomes certain that he can produce 2 rupees worth of wheat for 1 rupee of additional ex- penditure, on the average, he will seek to do so by using the new package of inputs. The farmer is better off than before. However, if input and output prices change, it is quite possible that the relative profitability can change such that the old package results in a more favorable ratio of net returns to costs. If this happens the farmer may shift back to the old package of inputs to the exclusion of the new. Partial budgeting analysis (which compares the results of price changes) is thus important 145 in determining whether the barani smallholders will continue employing the new package of inputs compared to the old. In the following, the components of total expense and total revenue will be examined separately before comparing costs and returns for dwarf and desi varieties. Total Expenses and Input Pricing Total expenses are the summation of all items used in production times their prices for a given time period. There are notably two factors which can change in the evaluation of total expenses: the quantity used of each item and the price of each item. Table 7.6 summarizes the prices paid by farmers for materials and 1 other items to produce wheat during 1969/70. The table also summarizes the average quantity of inputs used per acre to produce the wheats. The quantities shown are the "actual" average levels of inputs used by the sample of respondents and should be distinguished from the "geometric mean“ levels which were estimated from the Cobb-Douglas production function (see Table 7.1). In actuality, the average fertilizer dosage was more than the amount estimated by the model of production. All other "actual" figures correspond closely to the geometric mean values. The following discussion will elaborate on each of the inputs quantified in Table 7.6, with focus on input pricing. 146 TABLE 7.6 AVERAGE QUANTITY OF INPUTS USED PER ACRE AND THE AVERAGE PRICE PER UNIT: HAZARA DISTRICT, 1969/70. Ave. Quantity of Inputs Used For Ave. Price Item Dwarf Wheats Desi’Wheats 1969/70 (n=71) (n=35) (Rupees) Land 1.0 1.0 Acre 70.00/Acre Human Labor 31.5 21.1 Man-Days 0.50/Day Bullock Team 13.1 8.8 Days 2.50/Day Seed Rate 36.5 32.2 Seers 0.50/Seer Nitrogen 40.7 19.2 Pounds 0.56/lb. Phosphate 28.7 8.3 Pounds 0.46/lb. 1. Land For the price of land, an intermediate price between the value of the marginal product estimated by the dwarf wheat model (refer to Table 7.4) and the opportunity cost (what one acre would earn if used for the next best crop) was selected. In Hazara District, where land is not readily sold, it was thought that the MVP of land would reflect its real worth in present employment. The next best alternative crop for the sampled areas is barley. Before wheat, barley was grown more extensively in Hazara (and still is in a few areas) during the winter months, matured in the same period as desi wheat, did not receive any ferti lizer and yielded about 8 maunds per acre. The average wholesale prICE! of barley grain has generally been about 13 rupees per maund, and Its rust profit per acre of rainfed land has recently been estimated at163o 56.00 per maund.lg/ The intermediate price, Rs. 90.00, appears ~12/Department of Agriculture, West Pakistan Peshawar, Report on g—LOP-LIV'LLPattern for Peshawar Regigg, published by the Director of gmc” ture, Peshawar RegTon, Peshawar, 1969, p. 47. 147 to be an adequate reflection of the price of land used during the wheat cultivation period. 2. Human Labor Labor poses an interesting anomaly for two related reasons: (i) it has a positive wage rate when actually employed but shows up as a factor with zero returns, in the MVP's, for dwarf and desi wheat production, and (ii) it is primarily family labor as reflected by the fact that 99 percent of the labor used in wheat production among the sample of respon- dents was family labor. Accordingly, the problem of valuing labor becomes one of (i) deciding whether or not to attach a positive wage rate to the services of labor and (ii) deciding whether or not to consider the services of labor as either an expense or a return to the farm family. The final decision obviously has strong bearing on the net profits per acre of dwarf or desi wheat and is not easy to deal with. To simplify matters, it is assumed that labor has a positive opportunity cost of 50 paisa per day. Labor kept on the farm is also considered an expense undertaken by the farm family to maintain its food supplies. It has been noted that the labor used for dwarf wheat field pre- Paration and sowing has increased by about 50 percent for the same chores done for desi varieties. Though dwarf wheat yields have increased by as "WC?! as 100 percent over desi yields, the man days used for harvesting, t1"‘eShing, winnowing and storage have only increased by around 50 Percent, \ 137' TT'One man-day is equivalent to eight man-hours of work. 148 TABLE 7.7. MAN-DAYS EMPLOYED PER ACRE FOR DWARF AND DESI WHEAT PRODUCTION AND PROCESSING: HAZARA, 1969/7O (n=143) Field Preparation Clara 0 Winnowing Wheat and Sowing Harvesting Threshing & Storage Total 1. Dwarf 12.1 9.3 5.1 5.0 31.5 2. Desi 7.8 §;§_ 3;§_ 3;3_ 21.1 3. Added labor (2 from 1) 4.3 2.8 1.6 1.7 10.4 4. (3 as % of 2) 55.13 43.08 45.71 51.52 49.29 The man-days actually employed for both varieties are illustrated in Table 7.7. Evident are notable economies with increased output which indicate that the average cost per maund of processed dwarf wheat becomes less for barani smallholders. This is quite possible because of the specific way in which tasks are performed. Namely, harvesting is done by hand, cutting the wheat near the base of the plant with a small sickle. The relatively larger number of tillers on dwarf wheat places more plant stems in a unit area, compared to desi wheat, which means that the harvestor can grab more stems with little extra effort. _Before wheat can be threshed, a glgrali/ is prepared which requires the same amount of work, whether for five maunds of wheat or one hundred (if less than five maunds of wheat are harvested, the grain 155 usually threshed by beating the stalks against a log or rock). TAN! bundles of wheat are placed on the glare and are trampled by bLfllock teams. Winnowing is largely a function of the velocity of erld. A "windy" day makes the task of separating the chaff from the grfrin much easier and quicker. After the two are separated, farmers \ . . lg(A clara is a threshing floor prepared by driving bullock teams A" Clrcles on a particular spot in the field until the round is packed arwj. None of the farmers interviewed has a permanent cement) threshing (1c"3" because, as the farmers say. "they crack under the strain of the 1111(3Ck teams and the quick changes in climate from cold to hot." 149 measure the grain with a wggi_and empty the EQQiLE contents into a bag or basket. The grain is then taken into the home where it will be milled by the women when flour is needed. Straw is saved and stacked into piles. Large piles of straw are bound tightly and covered with a mud "ceiling" to protect against wind and rain. As in the other tasks, the additional labor required is less than proportional to the amount of grain available with the new varieties relative to the traditional. 3. Bullock Team Fields are plowed by bullock pairs. For every man-day of labor used in field preparation and sowing there is a pair of bullocks. Thus, a bullock team is employed, on the average, 12.1 days and 7.8 days for dwarf and desi wheat field preparation and sowing. A pair of bullocks works approximately one extra-day to prepare a clara. Bullocks are rarely rented. They are usually loaned in exchange for labor. Hence, the average price per day for a pair of bullocks has been estimated from separate pieces of information collected during the survey and is shown in Table 7.8. 4. Seed and Fertilizer The amounts and prices of seed, nitrogen and phosphate have been discussed in Section I. It should be added, however, that dwarf Wheat seed is now readily available and its price has converged to appr‘oximate the price of wheat grain sold for food in Hazara. On the average, wheat seed sells for around 50 paisa per seer. However, mSt seed is saved from one crop to another and very little is sold comnercially in village bazaars. As mentioned, the government attempts to standardize the price 150 of a pound of nitrogen at 0.50 rupees and a pound of phosphate at 0.40 rupees regardless of source. At these prices, the government is subsidizing the price of fertilizer by about 25 percent. TABLE 7.8. COSTS OF BULLOCK TEAM IN HAZARA DISTRICT, 1970 Bullock Team Item Unit Amount Market Value Rupees 500 Useful Life Years 8 Use Per Year Days 70 Salvage Value Rupees 250 Area Cultivated Per Annum Acres 6 Annual Costs: Feed/care/shelter Rupees 165 Depreciation Rupees 30 Interest Rupees 25 Death Loss, Rupees 10 Miscellaneous Rupees _331 Total: Rupees 250 Annual Credit: Manure Rupees 75 Net Cost: Per Year Rupees 175 Per Day Rupees 2.50 — .Igtel Revenue and Product Pricing Total revenue is the summation of all goods produced times 'their respective prices. In Pakistan, both wheat and straw have (Economic value, and are, in essence joint products. Since I"lenipak-65 and the popular desi varieties look (and many say taste) alike, they are equally priced at Rs 20.00 per maund in Hazara (Whereas desi wheats receive up to 50 paisa more per maund over dwarf wheat in the Punjab.) Dwarf wheat straw is short compared to 151 desi wheat. Hence, it is assumed that for every maund of desi wheat there is a maund and a half of desi straw produced. And for each maund of dwarf wheat a "grain : straw" ratio of 1:1 is assumed. Again, like other factors in Hazara, straw is generally not marketed. It is kept for feeding farm livestock. Farmers do say, however, that on the average wheat straw is worth approximately five rupees per maund if sold; either dwarf or desi straw, but the latter is preferred. This price is used in the computations below. 152 Budget Analysis of Costs and Returnslé/ Now that we have examined the major components used in partial budgeting analysis and have priced inputs and outputs according to their average values for 1969/70, the next consideration is to analyze the costs and returns associated with dwarf and desi wheat production. Such an analysis can help determine whether or not dwarf wheats are relatively more profitable than desi wheats and from that, infer whether or not the relative income position of barani smallholders has improved with the production of dwarf wheats. Table 7.9 gives a break-down for total costs and total returns on a per acre basis for dwarf and desi wheats. Cost items are divided into the costs for (i) family-farm inputs and (ii) purchased inputs. Assuming that the figures are broadly true for individual barani farms, the added "family-farm" inputs for dwarf wheats cost 15.95 rupees per acre and the added "purchased" inputs for dwarf wheats cost 32.57 rupees.. Together, the incremental expenses brought about from the "new package" of inputs cost the farmer about 48.52 rupees more than the "old package" per acre. However, the additional income from dwarf wheat grain and straw is 178.50 rupees over the return from an acre of desi wheat. Thus, 48.52 rupees of added expenditures lDrought 178.50 rupees of added income, a ratio of about 3.7 to 1. By deducting the "family-farm" inputs, since their opportunity costs 3'13 relatively low in the short run, the ratio of added income to added expenditures is even greater, a ratio of about 5.5 to 1. Either “Buy. the comparisons indicate that dwarf wheats, grown on a per acre K " lé/The analysis follows the same technique used by: Witt, Lawrence, PCOSt of Production And All That," mimeo, The Ford Foundation, Islamabad, ak‘Istan, August 8, 1969, pp. 14. 153 basis, are at least three times more profitable than desi wheats, using 1969/70 average figures. This implies that barani smallholders can afford significant amounts of added expenditure if yields can be increased substantially. TABLE 7.9 INCREMENTAL COSTS AND RETURNS FOR DWARF AND DESI WHEAT IN HAZARA DISTRICT, 1959-70* (Rupees/Acre) ITEM (1) ”heat Type INCREMENTAL DWARF (21 DESI (31 CHANGE (4) (n = 71) (n = 35) (2) - (3) FAMILY-FARM INPUTS: Land 0 Rs. 70.00/acre 70.00 70.00 0.00 Labor 0 Rs. 0.50/day 15.75 10.55 5.20 Bullocks 0 Rs. 2.50/day 32.75 22.00 10.75 PURCHASED INPUTS: Seed 0 Rs. 0.50/seer 18.25 16.10 2.15 Nitrogen 0 Rs. 0.56/lb. 22.79 3.36 19.43 Phosphate 0 Rs. 0.46/lb. 14.81 3.82 10.99 sub-total 55.85 23.28 32.57 TOTAL COSTS (rupees) 174.35 125.83 48.52 RETURNS: Grain 0 Rs. 20.00/md. 330.00 170.00 160.00 Straw 0 Rs. 5.00/md. 82.50 64.00 18.50 TOTAL RETURNS (rupees) 412.50 234.00 178.50 INCOME ADDED BY ONE RUPEE 0F EXPENDITURE FOR: PURCHASED INPUTS About 5.47 TOTAL INPUTS About 3.68 * Average input levels are shown in Table 7.6. 154 Understandably, the ratio can worsen by (i) higher input prices, (ii) lower product prices, (iii) reduced dwarf wheat yields or increased desi yields and, (iv) in general, reduced productivity of dwarf wheat resources. Analysis of changes in the ratio from anyone of these (or in different combinations) is possible with partial budgeting analysis. But because of the number of possible combinations and difficulty of demonstrating them all, only two examples are given. First, if we assume that the government subsidy on fertilizer is removed, so that nitrogen costs 70 paisa per pound and phosphate 60 paisa (and no change occurs in the quantity applied) then we would have the following calculations: Dwarf Wheat Desi Wheat Incremental Change Family-Farm Inputs 118.50 102.55 15.95 Purchased Inputs 63.96 25.28 3§;§§_ Total Costs 182.46 127.83 54.63 Total Returns 412.50 234.00 178.50 With the higher prices for both types of fertilizer, the ratio of added returns to added costs drops from 5.47 to 4.60 for income added by one rupee of expenditure for "purchased inputs." The ratio is lowered from 3.68 to 3.25 for income added by one rupee of ex- penditure for "total inputs." Second, if we assume that dwarf wheat yields are lowered by twenty percent (and, concomitantly, the straw yield drops) and we have the 1969/70 average input and price levels, then the ratio is lowered from 3.68 to 2.33 for income added by one rupee 0f expen- diture for "total inputs". The income added by one rupee of 155 expenditure for "purchased inputs" is still very favorable. Other examples can be developed, but with more reasonable assumptions it is evident that dwarf wheats are relatively more profitable than desi wheats. The favorable ratios of returns to costs indicate that barani smallholders can be made better off by adopting the "new package" of inputs. Also, the comparisons suggest (along with the conclusions of Chapter V) that if the ratio continues to be as favorable, then it behooves barani smallholders to undertake the additional expenditures and to produce the new varieites. From these conclusions, we can infer that the relative income position has improved for the farmers who have used the new, highgyielding varieties of wheat. CHAPTER VIII SUMMARY AND RECOMMENDATIONS Summary Background West Pakistan's agricultural sector faces two problems of con- stant urgency: (l) maintaining a rate of growth in food-grain produc- tion that at least equals or exceeds the rate of growth in population and (2) reducing the uneven spread in the distribution of benefits within the rural populace so that the gap between the "haves" and the "have pots" does not widen, hence creating the dismal prospect of social discontent and political unrest among the rural masses. For a country with limited resources, solutions to the two problems are not easy to come by. For example, if Government proceeds with guidelines specifically directed towards increased crop produc- tion (at least cost), then it must work to push the economic factors whose multiplier effects would be the greatest, and perhaps faster to adjust, in terms of attaining the desired result. In usual cases, this calls for programs with incentives to stimulate the major producers of the agricultural economy -- the large commercial farmers. However, such programs of change benefit large farmers in the first instance and widen the distribution of benefits in the second, a result which apparently conflicts with the other problem of lessening the disparity in incomes. Certainly such measures to increase production have created the spectre of only improving the conditions of the large commercial farmers and leaving the smallholder relatively worse off. 156 157 Whether or not barani smallholder farmers have benefitted from Pakistan's agricultural strategy to diffuse technologically improved varieties of wheat is thus of concern to policy makers. For if these relatively poorer farmers (whose crops are largely dependent upon rainfall) cannot benefit from the development strategy that moves large farmers, then other measures and programs may be required. But if the same strategy increases agricultural crop production and bene- fits both large and small growers then its expansion and reproduction is called for. The objectives of this study were to: (1) determine the extent to which barani smallholders use dwarf wheats, (2) identify the factors associated with dwarf wheat diffusion and adoption among smallholders, (3) examine the phenomena of off-farm migration and non-farm employ- ment which apparently influence farm management decisions and the income position of barani smallholders, (4) analyze the economics of wheat production on barani land, and (5) make practical recommendations for improving the well-being of barani smallholders. The importance of this study is further indicated by the fact that: approximately 30 percent or 3.8 million acres of West Pakistan's wheat is sown on barani land each year; wheat yields are invariably lower on barani acreage (by as much as 50 percent) compared to irrigated acreage in West Pakistan; over 6 million people live on barani farms and very little is known about the socio-economic conditions and constraints within which barani smallholders perform.» West Pakistan's principal barani or rainfed area spans the Northern territory between 33° to 359 North Latitude and 71° to 74° East Longitude. This area is affected by the summer monsoons and 158 enters into the Himalayan foothills of West Pakistan. For purposes of analysis "the barani" can be divided into three geographically distinct areas: (1) the Potwar Plateau and Pabbi Areas, (2) the Himalayan Foothills and Valley Areas and, (3) the Peshawar Vale and Kohat Range. This study focused on barani smallholders who live in Hazara District, an area which is largely representative of the Himalayan Foothills and Valley Areas. Hazara receives relatively higher rainfall than other barani areas, between 30 to 60 inches annually. In addition, the District faces problems associated with a dense and rapidly expand- ing population on a small base of comparatively unproductive cultivable acreage. The majority of the farmers grow primarily wheat in winter and maize in summer. Very little of the grain produced is sold commer- cially. Data for the study were collected during two field surveys in 1970 which coincided with the post harvest periods for wheat and maize: May/June and October/November, respectively. A special sampling technique was developed in order to have a widespread sample of respondents. It involved approaching and interviewing farmers who were in the field or on the trail in the vicinity of villages which were chosen at random by the interviewing team. The latter consisted of the author and an interpreter who was raised in similar barani surroundings. Special care was taken by the interviewing team to encourage reliable answers from the respondents. Altogether, 226 farmers were interviewed. The average size of holding cultivated by the interviewees is about 5 acres; 97 percent of this cultivated acreage is barani. Also, about 82 percent of the respondents own their land, 4 percent both own and rent land, and the rest are tenants. 159 Dwarf Wheat Diffusion and Adoption (1) In the survey locations of Hazara, it was found that there has been rapid adoption of new dwarf wheats (particularly Mexipak-65, the common dwarf) and chemical fertilizer. In the sample of 226 res— pondents, two smallholder farmers (0.9%) were sowing dwarf wheats in 1966/67. By 1967/68, a total of 30 farmers (13%) were trying dwarfs. In 1968/69, 75 farmers (33%) grew dwarf wheats and by 1969/70, the year covered by the survey, 150 smallholders (66%) were dwarf wheat adopters. The test of dwarf wheat endurance came when farmers were asked if they were going to plant dwarf wheats during the next winter season (1970/71). The response showed that 200 out of 226 (86%) anticipated growing the new variety, some for the first time. The pattern of fertilizer adoption was quite similar to that described for dwarf wheats, also starting with initial usage in 1966/67. With regard to the intensity of adoption, the findings show that barani smallholders moved cautiously at first by experimenting with handfuls of seed. More specifically, the findings indicate that small- holders will rarely sow more than 30 percent of their acreage with a new variety during the initial year of experimentation. As they become more familiar with the innovation, the decision to adopt or reject completely comes into their consideration. (2) Respondents in Hazara District adopted dwarf wheats because they: (i) gave better yields over desi varieties, (ii) fit the cropping pattern of the farmer and grew in a faster time than desi varieties, (iii) were subject to experimentation on a small scale by individual farmers, (iv) were not complex in relation to desi wheat with respect to the manner in which they were planted and cared for, and (v) had other 160 positive features like beards which protected grain against birds. In a few cases, Vbxipak-65 even tasted better than desi wheats. One complaint, however, was that dwarf wheat gave less fodder. But in all cases of adoption, barani smallholders were willing to substitute less fodder for a higher grain yield. With regard to yields, data were presented which showed that for each and every year, from 1967/68 to 1969/70, dwarf wheats out-yielded desi wheats by a consistently wide margin, nearly 100 percent on the three year average. On the average, dwarf wheat growers can expect around 16.5 maunds per acre with dwarfs and 8.5 maunds per acre with desi wheats. However, to date, dwarf wheats show a wide variability in yield, from a possible low of 6.5 to a high of 26.6 maunds per acre. 0n the other hand, desi growers experience a range in Output from 4 to 13 maunds per acre. Evident is greater risk and uncertainty associated with the new varieties. With better farm management prac- tices and increased experience with the new varieties it may be possible to reduce the variability in dwarf wheat yield in the coming year. However, complicating the process of adjustment to dwarf wheats is the finding that farmers have substituted much more than a new variety of wheat for an old. Namely, dwarf wheat growers apply more man-hours of labor, higher seed rates, and chemical fertilizer to their crop than desi growers. In essence, they have adopted a completely "new package." (3) There were two principal communication channels which informed barani smallholders of the new varieties of wheat: (i) mass media channels and (ii) interpersonal channels, including the sub- categories. The most effective mass media channel was the radio. Nearly 24 percent of the respondents who knew of "Mexipak" indicated 161 that they first heard about dwarfs over the radio. Overall, 39 percent of the respondents owned radios. Seven percent who did not own radios indicated that they were first made aware of dwarf wheats over this media. Less than one percent first learned of dwarf wheats from the written media. There were essentially three types of interpersonal channels which informed the barani smallholder of dwarf wheats: (i) interpersonal localite, (ii) interpersonal cosmopolite and (iii) the demonstration plots. While the radio was evidently important, the interpersonal localite contacts still had the largest impact of informing smallholders of new seed; 35 percent of those who knew about dwarf wheats first heard of them from neighbors. Another 16 percent of the sample first saw dwarf wheats in demonstration plots and asked the farmers growing them about the crop. These two findings show that interpersonal exchanges between farmers and the dwarf wheat demonstration plot (which showed striking differences next to desi wheat) were effective transmitters of information. Another 25 percent of the respondents heard about dwarfs initially from interpersonal cosmopolite sources such as the government extension agents and the distributors of farm supplies. It should be noted that Field Assistants (lowest level extension agents of the Agriculture Department) were very instrumental in diffusing dwarf wheat varieties. Aside from personally informing farmers of dwarf wheat potential, they were responsible for the installation of many of the demonstration plots on farmers' fields which, in turn, were catalysts in dwarf wheat diffusion. (4) Three types of independent variables were correlated with 162 awareness and innovativeness: economic, communication, and intervening variables (of migrants and their remittances). The results from the zero-order correlations show that: (a) With regard to awareness: (i) 0f the economic variables, farmers who owned most of their land and had comparatively larger families were significantly earlier than others in their community to become aware of dwarf wheats. A very weak relation tends to indicate that farmers with medium size farms (4 to 5 acres) score relatively higher in terms of awareness than farmers with large farms. In addition, it is quite probable that "wealth" (the combination of farm area and cash income) in a smallholder environ- ment is not a major determinant of early awareness of innovations. (ii) 0f the communication variables, both interpersonal cos- mopolite and mass media channels were significantly important in creating awareness sooner than other channels among smallholders. Also signi- ficant is the finding that the last farmers to become aware of dwarf wheats were all informed of the innovation by interpersonal localite channels. (iii) The data do not lend themselves to testing whether or not off-farm family migrants were important sources of first information in dwarf wheats; one can only surmise that migration leads to increased awareness of the developments in the world outside the village community. (iv) Taken together, the zero-order correlations indicate that smallholders who become aware of innovations before others are generally those who own most of their land, have large families, frequently listen to the radio and have more interpersonal contacts with change agents and sales representatives of farm inputs. 163 (b) With regard to innovativeness: (i) Judged by the zero-order correlations, income per capita per farm, mass media channels, demonstration plots, migrants and migrant remittances seem to have little influence on smallholder's innovative behavior (with dwarfs). (ii) Of the significant economic variables, smallholders who own five acres or more, have large families and relatively higher incomes are usually more innovative than others. (iii) Of the significant communication variables, smallholders who score high on awareness and interpersonal cosmopolite contacts are more likely to adopt innovations before others in their community. Two types of independent variables were also correlated with innovativeness by way of two regression models in order to identify whether communication or economic variables were more important in explaining innovativeness among smallholders. The results show that both communication and economic variables are significantly important and necessary in explaining innovativeness. Furthermore, it does not seem to be a useful argument to see which set of variables is most important since both conmunication and economic variables are highly interdependent and sequentially different from one another. Migration and Non-Farm Employment Migrant exodus apparently influences the farm production/sub- sistence environment of the barani smallholders in Hazara. In order to understand the nature of migration, non-farm employment and their possible influence on barani smallholders, a series of questions on these factors were included in the survey schedules. The preliminary conclusions are that: 164 (l) The "extended family" appeared to be one of the most important (2) (3) (4) (5) factors perpetuating the pattern of migration and non-farm employment. Migrants are generally young men with some schooling and they are usually from small farms with "medium size" extended families. Mostly adult males leave the village family. Married men will even stay away from their family for a year or more while they work in other places. Male family members go to such far-off places as Karachi (1,000 miles from the village area) to find various forms of employment. Jobs taken by barani smallholders are varied but the general job is low-paying, manual and not related to agriculture. There is a large cash flow from the places of non-farm employment to the needy villagers in the mountains; on the average, about 62 rupees per month is sent back to the village per migrant. Though off-farm remittances are an important component of family income for barani smallholders, the zero-order correlation analysis (discussed above) indicates that migrant remittances do not have much bearing on the farmers' willingness to use dwarf wheats. On the other hand, migrant remittances can provide for some degree of "risk insurance" against the failure of innovation. Among other things, off-farm migration serves to reduce the dispar- ities within Pakistan conditioned by the differential rates of natural increase in population and between the income earned by the many different groups which serve the country. The Economics of Barani Wheat Production Production relationships between inputs and outputs for both 165 dwarf and desi varieties of wheat were specified with the Cobb-Douglas production function. The objectives in fitting the function to the raw data were to depict the economic impact of dwarf wheats at the farm level, to evaluate the performance of the sample of barani farms in producing either wheat, and to estimate optimal levels for the use of seed and fertilizer. Limitations with the use of the Cobb-Douglas function and the estimates which were obtained were noted throughout the analysis. The noticeable economic impact with dwarf wheats is manifested in: (i) the development of a new production function and (ii) the changes in the production factor proportions. The two sets of production elasticities consistently revealed the following results: (i) an important factor in explaining interfarm differences in wheat output is land; (ii) the production elasticities for labor were negative and not signi- ficantly different from zero; and (iii) fertilizer appears to contribute relatively little at the margin to output for both types of wheat. The signs of the fertilizer elasticities were positive but only the elasticities for fertilizer applied to dwarf wheats were significantly different from zero. The results also came up with a negative production elasticity for "seed rate" applied for desi wheats and a positive elasticity for seed rate applied for dwarf wheats. The marginal productivities which were estimated with the Cobb- Douglas function indicated that: (i) the rate of return to land was relatively higher than those of other inputs; (ii) the MVPs for labor used in producing desi and dwarf wheats were zero, showing that there is surplus labor on barani smallholdings; (iii) it is not possible to say 166 conclusively what the seed rate should be for either dwarf or desi varieties sown under barani conditions; and (iv) nothing conclusively can be recommended with regard to fertilizer applied to desi varieties. 0n the other hand, a comparison of the MVP's for fertilizer applied to dwarf wheat with their MFC's indicate that optimal use of N at present prices should be between 20 to 30 pounds per acre and that between 30 to 40 pounds of P205 would come close to an optimal use of phosphates. By all indications, the performance of dwarf wheats on barani land can be improved substantially by a proper utilization of the inputs studied. However, the variability in the sample data suggest that more research and analysis is needed to understand what readjustments are necessary in the factors of production presently utilized. Cost and returns to dwarf and desi wheat production were examined with partial budgeting analysis. Problems with placing a price on the inputs of land and labor were discussed. Altogether, values were placed on land, human labor, bullock labor, seed, and fertilizer (both N and P205). Returns were estimated for the joint products of wheat and straw combined. Using the actual level of inputs employed and the output obtained (for 1969/70) for the sample of 71 dwarf wheat growers and 35 desi wheat growers, the costs and returns indicate that, on the average, dwarf wheats are at least three times more profitable than desi wheats. In addition, partial budgeting analysis indicated that the relative profitability with dwarf wheats compared to desi wheats was virtually unchanged when the price of fertilizer was increased by 25 percent in one case and when dwarf wheat (and straw) yield was lowered by 20 percent in another case, all other things held constant. Overall, partial budgeting analysis showed the potential for barani smallholders to 167 finance agricultural improvements themselves. With moderate-sized investments in dwarf wheats and fertilizer, farm income appreciably doubled over former levels with only traditional wheat and complementary inputs. Finally, one can infer from the above that as smallholder income and barani agricultural productivity increase, the prospects for main- taining an ever-growing population improve. Looked at another way, the "need" to migrate is essentially balanced out by the ability to feed more people on the fann. General Conclusions Concerning the Fourth Five Year Period Analysis of the survey results leads to the conclusion that new strategies and programs need not be a necessary component of the Fourth Five Year Period; that programs similar to the Food Self-Sufficiency Program should be designed to develop and diffuse new, high-yielding varieties of seed and complementary inputs to barani areas in order to make a substantial impression on the livelihood of barani small- holders. The survey has shown that smallholder barani farmers are respon- sive to economic incentives and communication stimuli. They will make rapid adjustments in resource allocation with high-yielding varieties of seed provided they become aware of the potential improvements with the new seeds and receive equal opportunities in terms of acquisition of the genetically-improved varieties. Concomitantly, the possibility of successfully diffusing an innovation so that it may eventually be adopted depends very much on the characteristics of the innovation and the risk and uncertainty bearing on the subsistence level of the farm producer. To be adopted, the innovation must show relatively higher profitability, 168 an ability to fit into present cropping patterns, and other positive features mentioned above. Also, the findings indicate that the lack of credit does not appear to be a limiting constraint on the ability of farmers with small holdings in Hazara district to adopt new, high- yielding varieties of seed, since they apparently have sources of income from local employment and migrant remittances. It was also found in this study that any program designed to diffuse agricultural innovations must first concentrate on creating awareness among smallholders in general. Specifically, the findings show that barani smallholders who own most of their land and have comparatively larger incomes and families can be informed most effectively of agricultural innovations by the radio and the interpersonal cosmo- polite channels of communication. Barani smallholders who rent propor- tionately more of their land, have small families and relatively little cash earnings depend almost entirely on interpersonal localite channels for information on innovations. The larger (in terms of numbers) intermediate category of farmers can be informed of innovations by both mass media and interpersonal types of communication. In addition, it behooves the extension staff to select those smallholder farmers, who exemplify the "best" farm management practices, to serve as demonstrators of further inndvations. It can be concluded that the impact of the new varieties has been to increase output, to employ more resources and to influence cropping patterns. And as we have seen, the relative income position of barani smallholders has been able to improve in the process. However, due to the small size of farm operations, the quantifiable differences may not show as major repercussions on the total economy of West Pakistan. 169 Recommendations The following discussion should be considered in light of the survey findings of this study. Furthermore, the recommendations presented are made with the fundamental desire of indicating ways to sustain the current production advances of the Green Revolution, to broaden the base of agriculture technology and cropping alternatives available to the smallholder, to build viable extension institutions to service smallholders, and to improve the transmission of information from both domestic and foreign research to the farmers and extension personnel. The following recommendations apply to situations found in Hazara District, and, as such, they are directed towards the Provincial Government of the N.W.F.P. They may also be applicable to parts of the Himalayan Foothills in the Punjab and Azad Kashmir where similar agro-climatic farming conditions prevail. For More Immediate Consideration (l) The most immediate action should be to develop a full-scale crop improvement program with_particular attentionggiven to developing and diffusing new crop varieties suitable to barani conditions. This entails intensified research in two interrelated areas: plant breeding_ and farm management. Plant breeding research should strive to develop newer and "better" varieties adaptable to barani conditions. New varieties need not be higher-yielding to be better. For example, although genetic structures of a variety do not change, disease organisms do. The pathogens that attack wheat, particularly the rusts, continuously mutate into new types with different degrees of aggressiveness against the wheat plant. 170 New varieties which are popularly grown may become susceptible to new pathogens. If Mexipak-65, which is the common dwarf variety sown by smallholders, becomes susceptible to new rusts, smallholder wheat production will be vulnerable to suffer drastic losses that may bring serious repercussions on smallholders and the total farm economy. The only defense against such a scourge is a research and distribution program which insures that alternate varieties are constantly being introduced. Farm management research is desperately required to identify the types of farming, cr0p and livestock combinations, crop varieties, levels of capital, and labor input which are needed to use land and labor resources fully and efficiently to the best advantage of the smallholders. In addition, farm management research is needed to analyze ways to im- prove the external environment of tenure, credit, marketing and price conditions which are required to accelerate agricultural progress and general economic development. If effective, farm management research can provide a feedback linkage from farm to policy-making agencies. A clear knowledge of averages, variations, and relationships within agriculture can lead to sound hypotheses concerning the effects that alternative policies may have on smallholders. Farm management research can also be useful for: (i) advising farmers engaged in agricultural production, (ii) teaching those who are to serve agriculture in operating capacities, and (iii) in- forming those formulating public agricultural policy. (2) Plant breeding research should be carried out on existing research stations with experimentation through a "farm testing program" in cooperation with the extension service. Pakistan has long-established 171 and successful plant research institutions that can give valuable support to crop improvement campaigns for wheat, maize, sorghum, millets, and rice. These institutions have also had strong ties with the Interna- tional Maize and Wheat Improvement Center (in Mexico) and the Interna- tional Rice Research Institute (in the Philippines). Such ties should be continued with periodic exchanges of breeding materials, crop specialists and students. This requires the availability of modest funds to import small quantities of materials, critical equipment, and skilled specialists. Together, both domestic and foreign institutions should devote more attention to develOping varieties which are attractive to barani farmers. The major characteristics of favorable varieties are: high yields, short growing periods, good taste, and fodder supply to feed farm livestock. Research should also focus on developing varieties which are drought, disease, and insect resistant. Perhaps, most important of all, research should concentrate on ways to reduce the extreme variability in yields with the new varieties. To accomplish this requires a thorough testing program of new varieties under different soil and climatic conditions. The research stations at Tarnab, Pir Sabak, and Mansehra provide ideal agro-climatic conditions suitable for developing varieties for "the barani". In particular, Tarnab and Pir Sabak facilities are adequate to test for drought resistant varieties of the major food-grain crops. The Mansehra station is well—located to serve Hazara District. It is not a new experiment station which is needed, though; it is a specific set of agricultural research outputs that provide new possible breakthroughs to protect and improve what has been attained. Research experiments carried out on farmers' fields in a "farm 172 testing program" should include closely supervised instructions to innovative smallholders on the use of new varieties. To do this, it will be necessary to have detailed plans for each test plot indicating varieties, treatments, etc. Farm test experiments can be developed and packaged by research station personnel. The selection and obser- vation of farm test sites can also be handled by the Department of Agriculture extension staff. Such farm tests, which could range from a quarter of an acre to a full acre for experimentation,can also serve as demonstration plots for the neighboring farmers to see. (3) Farm management research should be developed and implemented as a complement to plant breeding research. Since risk and uncertainty bear on subsistence farm production decisions and ultimately affect the intensity of adoption of new varieties, certain steps are required to assure a greater rapidity and extent of adoption of new technology. First, farm management research should be conducted through well-executed advanced preparations and carefully constructed interview schedules in order to gain useful information from village respondents. Second, farm management research should focus on determination of (2) optimal input levels and (b) optimal combinations of farm products. Knowledge of optimal input levels is more important in helping governments to allocate the scarce fertilizer supply, for example, among areas and even among cr0ps so that farmers can have prior knowledge on fertilizer avail- ability. Knowledge of optimal farm combinations is useful to the farmers as they become commercially oriented and influenced by changing price relationships for purchased inputs and marketed outputs. Third, farm management research should focus on the formulation of budgets which 173 show the economic effects of various combinations of changes in practices and enterprise combinations. The information can be used in advising farmers of the net returns they should expect by following different techniques of production. Farm management research does not need a large budget, but does require a few experienced and exceptionally talented researchers working together as a team. Institutions such as the Board of Economic Enquiry in Peshawar, the Pakistan Academy for Rural Development, and the agri- cultural and economic faculties at the University of Peshawar have individuals capable of conducting small farm management research projects. Their cooperation and funding could be engaged and coordinated through a "Farm Management Section“ in the Planning and Development Department in Peshawar. Farm management research need not require large and cumbersome samples of farmers. Such exigencies, in early stages of development, encounter problems with data compilation and scarcity of skilled inter- viewers. It might best be started on a small scale with initial surveys limited to 30 to 100 well-selected farms which have conditions conducive to effective comparisons on the points raised above. Emphasis should be placed on working out practical research methods and simple field schedules and procedures. Furthermore, the initial studies should be made to spot problems for more intensive study, to provide basic data for farm budgeting and to provide a learning experience for conducting more intensive studies. Each year at least one specific problem should be studied. At this juncture, the most important research can be found in the barani areas of Kohat, Potwar and Pabbi areas. The economic problems in these areas are also serious. At this time farm-level 174 field research is needed both to orient crop research and agro-economic research and to provide guidelines for government action in these areas. (4) Channels of communication between research outlets and farmers should be strengthened so that smallholders are kept informed of innovations available to them. If new varieties and production alternatives are specifically designed for the difficult conditions and precise cropping patterns of the barani, then the mass media and interpersonal channels of communication will need to remain dynamic. Since interpersonal con- tacts are especially needed to reach the smallest farmers, the extension agents will require additional training and experience in the use and limitations of the new technology adapted for the barani smallholders. Better demonstration plots, not necessarily more, will need to be installed in order to enhance the farmers educational needs. In addition, radio programming will assume greater responsibility which implies that its broadcasts will have to be timely and relevant with respect to the particular information needs of the farmer. In particular, there is critical need to improve the knowledge of Field Assistants who will be called upon by farmers who desire information on new varieties. Moreover, to alleviate the problem of excessive distribution of extension resources, the use of "farm test experiments" is recommended for involving Field Assistants in practical activities. Such an approach would be educational for the Field Assistants and the farmers alike. For Later Consideration Government attention should be given to enhancing the supply and extension of profitable factors of production pertaining to 175 innovations. More specifically, the success of any crop improvement program depends upon the availability of production inputs at sites conveniently located for the farmers well in advance of the appropriate time of utilization. If supplies are not available well in advance, the uncertainty affecting farmers' decisions may be such that the continued use of innovations is thwarted. An allied recommendation is to develop pure seed supplies. Seed multiplication should be encouraged and expanded among farmers in "the barani". Seed certification would be an important concomitant to seed multiplication. It necessitates records on "proof of origin," inspection of the growing crop to check for purity of stand, seed germination, and purity tests after harvest. Plant breeding research should give attention to developing diversified crop production and intensive land use techniques. Studies, for example, on the use of fertilizer with various levels of residual moisture should be undertaken to find profitable levels of its use with new crops. Multiple cropping should be studied to utilize "scarce" land more intensively and efficiently. Finally, it is recommended that an agronomist or a "subject matter specialist" position be established in each District to handle several related activities: (i) to work with Field Assistants (FA's) insetting up farm test experiments,(ii) to train FA's to upgrade their performance and to keep them abreast of the latest research output available for testing, (iii) to serve as a liaison between farmers, Field Assistants, and research stations, (iv) to attend annual research reporting and extension planning meetings, and (v) to experiment 176 with multiple cropping techniques on farmers' fields. A subject matter specialist should hold at least an M.S. degree in agronomy and have some field experience and practical expertise with farmers' problems. He should be administratively under the Extra Assistant Director of Agriculture and at least on the same status level with Agricultural Assistants (who attend to the Administration of FA's and the necessary paper work). [_, BIBLIOGRAPHY 10. 11. 12. BIBLIOGRAPHY Abidi, S.A. Anwar. "The Development of New Wheat Varieties in Pakistan." Conference on Food Production Increase in West Pakistan: PrOBlems and Effects, Pakistan Academy fer Rural Development, Peshawar, April 22-23, 1970. Aziz, Sartaj. "Problems and Prospects of the Green Revolution." The Planning Commission, Government of Pakistan, mimeo, May, 1970. L Bradford, Lawrence, and Glenn L. Johnson. Farm Management Analysis.~/ New York: Wiley, 1953. Brown, Lester R. Seeds of Change: The Green Revolution and Develop- a rent in the 1970's. New York, Praeger Publishers, 1970. Burrows, W.C., et. al., eds. International Conference on Mechanized Doyland Farming. Proceedings sponsoredfiby’Deere and 00., under the auspices of the Food and Agriculture Organiza- tion of the United Nations, Moline, Illinois. Copyright 1970 by Deere and Co. Bussink, H.C.F. "Population Estimates by Province and Districts in West Pakistan, 1970-1980." Lahore, West Pakistan: Harvard Advisory Group, May, 1970, mimeo. Cancian, Frank. "Stratification and Risk-Taking: A Theory Tested on Agricultural Innovation." American Sociological Review, XXXII, 6, December, 1967. Carter, H.O. and H.O. Hartley. "A Variance Formula for Marginal Productivity Estimates Using the Cobb-Douglas Function." Econometrica, April, 1958. Crosson, Pierre R. Agricultural Development and Productivity; Lessons from the Chilean Experience. Published for Resources for tfie Future, Inc. by the Johns Hopkins Press, Baltimore, 1970. Dalrymple, Dana G. "Technological Changes in Agriculture, Effects and Implications for the Developing Nations." U.S. Depart- ment of Agriculture, Foreign Agricultural Service, April, 1969. Davy, Dorcey F. "Improving the Training of Field Assistants in the Agricultural Training Institutes of West Pakistan." Islamabad: The Ford Foundation, March, 1967, mimeo. Davy, Dorcey F. "The Changing Role and Function of Government Extension Personnel." Islamabad, W. Pakistan: The Ford Foundation, March 5, 1970, mimeo. 177 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 178 Department of Agriculture, Peshawar Region. Report on Cropping Patterns for Peshawar Region. Published by the Director of AgricuTture,TPeshawar, W. Pakistan, 1969. Dichter, David. The North West Frontier of West Pakistan: A Study in RegTEhETIGeography. 0xf0rd,7EngTEnd?T ClarendOn Press, 1967. Eckert, Jerry Bruce. "The Impact of Dwarf Wheats on Resource Productivity in West Pakistan's Punjab." Unpublished Ph.D. Dissertation, Michigan State University, 1970. l ’Ti "The Beginnings of Change in West Pakistan's Agriculture.“ The Ford Foundation, Lahore, June, 1970. Elkinton, Charles M. and Aziz Syed. Pakistan Agriculture: Resources, .. Progress and Prospects. Lahore, Pakistan: Agriculture DiviSion, U.S. Agency for International Development, 1967, revised edition. Ezekiel, Mordecai and Karl A. Fox. Methods of Correlation and Regression Analypjs: Linear and Curvilinear. New York: JOhn WHTey anthons, Inc., third edition, 1965. Falcon, Walter P. "The Green Revolution: Generations of Prob- lems." Paper presented at the Summer Meeting of the American Agricultural Economics Association, Columbia, Missouri, August 9-12, 1970. Festinger, Leon. "Behavioral Support for Opinion Change." Public Opinion Quarterly, XXVIII, 1964. Gotsch, C.H. and W.P. Falcon. Agricultural Price Policy and the Development of West Pakistan. V01. I: "Final Report of the Orgafiiiation‘TOr Sodial and Technical Innovation." Cambridge, Mass.: February, 1970. Government of Pakistan. 1960 Pakistan Census of Agriculture, Vol. II Karachi, 1963. Ministry of Agriculture and Works. Farm Mechani- zatiBn in West Pakistan: Report of the Farm Mechanization 00mmittee. Islamabad, March, 1970. Ministry of Agriculture and Works. Agricultural Statistics of Pakistan., Islamabad, w. Pakistan, 1970. Government of Pakistan. Planning Commission. The Fourth Five-Year Plang‘l970-1975. Islamabad, July, 1970. Government of the Punjab. Bureau of Statistics. Statistical Handbook of West Pakistan, June, 1970. Planning and Development Department, Lahore, December, 1970. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 179 Government of West Pakistan. Season and Crop_Report. Lahore, 1965-66 to 1969-70. Planning and Development Department (Project Wing). Implementation Plan for the West Pakistan Food Self-Sufficieppy Programme. Printed by the Superintendent, Government Printing, Lahore, West Pakistan, August, 1967. Planning and Development Department. Bureau of Stat- istics. Rabi 1968-69 Crop Acreage of West Pakistan: By Districts and Tehsils. Lahore, 1969. [flu Griliches, Zvi. "Hybrid Corn: An Exploration in the Economics of Technological Change." Econometrica, October, 1957. Havens, Eugene A. and Everett M. Rogers. "Adoption of Hybrid Corn: Profitability and Interaction Effect." Rural Sociology, XXVI, 4, December, 1961. Heady, Earl 0. Economics of Agricultural Production and Resource Use, New Jersey: Prentice-Hall, Inc., 1952. Herdt, Robert W. and John W. Mellor. "The Contrasting Response to Nitrogen: India and the United States." Journal of Farm Economics, XLVI, 1, February, 1964. Hufbauer, G.C. "Cereal Consumption, Production and Prices in West Pakistan." The Pakistan Development Review, Summer, 1968. Huke, R.E. and J. Duncan. "Spatial Aspects of HYV Diffusion." The Philippines: International Rice Research Institute, Summer, 1970, mimeo. Hussain, Sayed Mushtaq. "Price Incentives for the Production of High-Yielding Mexican Varieties of Wheat." The Pakistan Development Review, X, Winter, 1970. International Maize and Wheat Improvement Center. CIMMYT Report 1967-68: 0n Progressing Toward Increasing Yields of Maize and”Wheat. Mexico 6, DIF. Mexico. Izuno, Takumi. "Maize and Millets Research in the NWFP of Pakistan -- Past Problems and Future Prospects." Islamabad, W. Pakistan: The Ford Foundation, February, 1971, mimeo. Johnson, A.A. Indian Agriculture into the 1970's -- Components of Modernization. New Delhi, India: The Ford Foundation, June, l970. Johnson, S.R. and Gordon C. Rausser. "Effects of Misspecifications of Log Linear Functions When Sample Values are Zero or Negative." American Journal of Agricultural Economics, LIII, 1, February, 1970. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 180 President's Science Advisory Committee. The World Food Problem. The White House, II, May, 1967. Qureshi, S. A. and I. Narvaez. "Second Annual Technical Report: Accelerated Wheat Improvement Program: 1966-67." Distributed by the Ford Foundation, Islamabad, West Pakistan, July, 1967. Qureshi, Sarfraz Khan. "Rainfall, Acreage and Wheat Production in West Pakistan: A Statistical Analysis." Pakistan Development Review, III, 4, Winter, 1963. Rana Tractors and Equipment, Ltd. "Report on Dryland Wheat Production with Massey-Ferguson System." Lahore, 1969, mimeo. Rizvi, S. M. A., et al. Consolidation of Holdings: A Study of the Process of Consolidation of Agricultural HoldingsinSelected Villages in Peshawar District. Peshawar, West Pakistan: Pakistan Academy for RUral Developemnt, January, 1965. Rogers, Everett M. Diffusion of Innovations. New York: Free Press of Glencoe, 1962. Rogers, Everett M. in association with Lynne Svenning. Modernization AmonggPeasants: The Impact of Communication. Neinork: *Holt, Rinehard and Winston, Inc., 1969. Rogers, Everett M. and F. Floyd Shoemaker. Communication of Inno- vations: A Cross-Cultural Approach. New York: Free Press of Glencoe, 1970. Ruble, William L., Donald Kiel, and Mary E. Fafter. "Calculation of Least Squares (Regression) Problems on the Least Squares Routine." Agricultural Experiment Station, Michigan State University, November, 1969. Sharif, C. M. Farmer's Attitudes Toward Self-Help, Pakistan Academy fOr Rural Development, PeShEWar, W. Pakistan, 1965. Sharif, C. M., ed. Food Production Increase in West Pakistan: Problems and Effects. Pakistan Atademy for RuralDeVElopment, PeShawar, JUne, 1970. Shaw, Robert d’A. "Jobs and Agricultural Development." Overseas Development Council, Monograph Number Three, Washington, D. C., 1970. Stern, Joseph J. and Walter P. Falcon. Growth and Development in Pakistan: 1955-1969. Harvard University, Center for Inter- national Affairs,Dccasiona1 Paper Number 23, April, 1970. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 181 U.S. Department of Agriculture in cooperation with U.S. Agency for International Development, Watershed Management Team. “Resource Management for the Rainfed Region of West Pakistan: Final Draft Report. Volume II: General BaCkground and Analysis of*Specific Problem Areas, May, 1968 - February, 1969. Water and Power Development Authority. Water and Soils Irrigation Division. Surface Water Hydrology Project. Annual Reports: 1960-1969. Lahore, West Pakistan. Wharton, Clifton R., Jr. "Risk, Uncertainty and the Subsistence Farmer: Technological Innovation and Resistance to Change in the Context of Survival." A paper presented at the Joint Session of the American Economic Assocation for Comparative Economics, Chicago, December 28, 1968. "The Green Revolution: Cornucopia or Pandora's Box?" Foreign Affairs, Council on Foreign Relations, Inc., New York, April, 1969. Wilcox, Wayne. "Politics of Distributive Policy in Pakistan." Paper presented at Conference of Economic Growth and Distri- butive Justice-- Pakistan, Rochester, New York, July 29-31, 1970. Willet, Joseph W. "The Impact of New Grain Varieties in Asia." U.S. Department of Agriculture, Economic Research Service, Foreign Regional Analysis Division, July, 1969. Witt, Lawrence W. "Cost of Production and All That." Islamabad, W. Pakistan: The Ford Foundation, August 25, 1969, mimeo. Witt, Lawrence W. and Ghulam Qadir. "Prospective Wheat Surpluses in West Pakistan." In C. M. Sharif, et al., ed., Food Production Increase in West Pakistan: Problems and Effects. Peshawar, W. Pakistan: Pakistan Academy for—Roral Development, June, 1970. Wright, Bill G. "Critical Requirements of New Dwarf Wheat for Maximum Production." Paper presented at FAD/Rockefeller Foundation International Seminar on Wheat Improvement and Production, Ayub Agricultural Research Institute, Lyallpur, Pakistan, March 26, 1968. Yotopoulos, Pan A., Laurence J. Lau and Kutter Somel. "Labor Intensity and Relative Efficiency in Indian Agriculture." Food Research Studies in Agricultural Economics, Trade and Development. IX, 1, 1970, Stanford UniverSity. APPENDICES APPENDIX A GLOSSARY OF PAKISTANI TERMS AND MEASURES Units of Measure 1. Crooping Intensity = the ratio of "net area sown" plus "area sown more than once" to "net area sown" times 100. Kanal l/8th of an acre. Maund standard unit of measure for agricultural produce = 82.286 pounds. Approximately 27 maunds = one ton. Merla = l/ZOth of a kanal. Rupee = name for Pakistani currency. One rupee = about 21 U.S. cents at the official exchange rate. Seer = l/40th of a maund = 2.057 pounds. E291 = a wooden or metal bowl used for measuring grain on the threshing floor (in Hazara District). One wodi holds approxi- mately 5-1/2 seers of wheat grain. Urdu Terminology 1. 2. Borooi - land which has only natural precipitation as a source of moisture; used synonymously with "rainfed." Boojor - land which is pebbly, undeveloped and generally uncultivated. Burra-sahib - burra means "big" and sahib means "Mister." Chowkidar - night watchman. Qloro - a threshing floor prepared by driving bullocks on a particular spot in the field until the ground is packed down. Desi - anything which is "local" or traditional. Kero - method of sowing seed by dribbling it by hand directly 182 8. 9. 10. 11. 12. 13. 14. 15. 183 into a furrow. Moiro - cultivated land near the household which generally receives farm yard manure. Mexipak - the term which is commonly used by farmers for identifying the new dwarf varieties of wheat. Patwaris - village revenue collectors, usually elder citizens. Eoro_- method of sowing seed by dropping it into a funnel attached to the traditional plow. EEEQQD.’ Muslim practice of female seclusion whereby women are concealed from non-family members. Shalwar/kameez - Shalwar means pants and kameez means shirt. Combined, the term refers to traditional dress. Ihéflfl - literal translation is "police station." However, it refers to an area of police jurisdiction as well. Zamindar - literally means "landowner" but generally refers to the farmers with wealth and large land holdings with tenants. APPENDIX 8 DEFINITIONS OF "ZERO-ORDER" AND “MULTIPLE" CORRELATIONAL ANALYSIS Correlational measures were developed by social scientists to measure the strength or degree of relation between socio- economic characteristics and population behavior. (1) Zero Order Correlational Analysis refers to the process of measuring the closeness of the relation between Ego variables. The standard measure is called the coefficient of correlation designated by the statistical symbol "r" which is generally estimated by the "product—moment" method for the coefficient of linear correlation. Estimation of "r" is useful in telling us whether variables are related to each other and the strength of the relationship. For ex- ample, if we find a perfect and positive correlation, the value of r will equal +1. For a perfect and negative correlation, the value of r will equal -1. If there is no relationship at all, the value of r will equal 0. Thus, r can have any value between +1 and -1. When two variables are correlated with each other and significantly different from zero, then the independent variable is said to "explain" variance in the other. The two variables are called independent and dependent, according to which one is viewed as depending on the other. In the following analysis it should be noted that the ordinary coefficient of correlation assumes a linear relation between two variables. Also, it cannot help us decide whether the Only a brief review can be given here. More detail is found in Ezekiel, Mordecai and Karl A. Fox, Methods of Correlation and Regression Analysis: Linear and Curvilinear, John Wiley & Sons, Inc., New York, Third Edition, 1965. 184 185 relation is one of cause and effect. (2) Multiple Correlational Analysis refers to the statistical method of measuring the joint relation between one dependent variable and two ore more independent variables. The multiple correlation coefficient among several variables is always positive and at least as large as the largest zero-order correlation coefficient. As such, it is almost always labelled "R". A regression equation describes the relation between the dependent variable and all of the indepen- dent variables. It has the following form: Y = a + b1 x1 + b2 x2 + --- where Y represents the estimated value of the dependent variable from the regression equation, "a" is a constant term, the b's are the par- tial regression coefficients, and the x's are the independent variables. As we increase the number of x's, visualization of relations between variables becomes difficult. Determining which independent variables should be added or eliminated also becomes a difficult problem. 186 APPENDIX C FACSIMILE OF FIELD SURVEY INTERVIEW SCHEDULE* Resource Use On Barani Farmland of Hazara Districtg 1970 - CONFIDENTIAL - Date of Interview: 1.0. No: Farmer's Name: Age: Village: Literate: Yes/No Part I. -- General Information Farm Characteristics 1. How many total kanals of land do you own? rent? 2. How many total kanals of land do you cultivate in a year? 3. Are any of the kanals that you cultivate irrigated? Yes/No If yeg, how many? What do you grow on the irrigated land? 4. How many fragments do you cultivate? Production 5. How many kanals of wheat did you plant this rabi (1969/70)? 6. How many maunds of wheat were harvested altogether? 7. How many kanals were planted with (a) Mexipak? (b) desi? 8. If both_grown, did you thresh the mexipak and desi separately? Yes/No 9. If mexipak_planted and threshed separateLy, (a) how many total maunds of mexipak did you harvest this year? (b) was mexipak planted on your best land? Yes/No If yes, was this land all maira? Yes/No ,9 This interview schedule incorporates the principle questions asked during the first and second rounds of interviews which are used in this study. Superfluous questions have been eliminated. 10. 11. 187 If desioplanted and threshed separateLy, (a) how many total maunds of desi did you harvest this year? (b) was desi planted on your best land? Yes/No If yes, was this land all maira? Yes/No What other crops did you grow this rabi? kanals used? Part II. -- Economic and Communication Factors If Mexipak was_growng ask the following: 12. l3. 14. 15. 16. 17. 18. 19. 20. 21. When did you first hear about mexipak (month and year)? Did you first hear about mexipak from the radio, a neighbor, a government worker, a store clerk, someone in your family or other? (a) Do you own a radio? Yes/No (b) Do you listen to any agricultural programs? Yes/No When did you first plant mexipak (month and year)? Why? How many kanals of mexipak did you plant: (a) the first year? (b) the following year(s)? How many maunds of mexipak did you harvest: (a) the first year? (6) the following year(s)? Will you grow mexipak next year? Yes/No Why didn't you grow mexipak sooner? How many seers of mexipak seed did you use this year? Where did you get the mexipak seed from? How much did you pay for your mexipak seed? 188 If desi was grown (and no mexipak), then ask the following: 22. Have you heard about the mexipak wheats? Yes/No 23. When did you first hear about mexipak (month and year)? 24. Did you first hear about mexipak from the radio, a neighbor, a government worker, a store clerk, someone in your family or other? (a) Do you own a radio? Yes/No (b) Do you listen to any agricultural programs? Yes/No 25. Will you grow mexipak next year? Yes/No Ij‘yeg, why? 0n how many kanals? Where will you get the seed from? Part III -- Resource Use Bullocks: 26. In preparing your wheat fields, did you use bullocks? Yes/No 27. How many bullocks do you own? 28. How many pairs of bullocks were used in preparing your wheat fields? 29. How many times do you plow each wheat field including the time for sowing? 30. How many days does it take to plow all your wheat fields one time? 31. How many hours per day do you usually spend in the fields plowing? Field Preparation and Threshiog: 32. Do you plow your fields more often for mexipak than for desi wheat? Yes/No If yes, how much more time is used? days hours 33. Did you use bullocks to prepare the clara and to thresh? Yes/No If yesz how many? How long did they work? 189 Labor: Family and Purchased 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. How many people helped you harvest your wheat this rabi? How many days did it take? How many hours did it take each day? Did you hire people to help you harvest your wheat? Yes/No If yes, how many? What did you pay each? How many relations (family members) helped you? How much is a laborer paid in rupees for one day's work in this village? (a) Does that include lunch? Yes/No (b) Does that include a share of the harvest? Yes/No, How much? How many people helped you winnow your wheat this rabi? How many days did it take to winnow your wheat? How many hours did you work each day? Did you hire people to help you winnow your wheat? Yes/No If yes, how many? What did you pay each? How many relations (family members) helped you? Fertilizer: 45. Did you use chemical fertilizer on your wheat this year (1969/70)? Yes/No If yes, then ask: (a) what kind of chemical fertilizer did you use? (b) where did you buy it? (c) how many bags did you buy? (d) how much did you pay for each bag? (e) what was the weight of each bag? (f) how much did you use on your mexipak? desi? (9) how many kanals of mexipak did you use it on? 46. 47. 48. 49. SO. 51. 52. 53. 54. 55. 190 How long have you been using chemical fertilizer (in years)? Part III -- Miscellaneous How much does a maund of flour (ata) cost on the local market? Did you sell any wheat last year? Yes/No Will you sell any wheat this year? Yes/No How much does a maund of mexipak desi flour cost? Have you taken any loans this year? Yes/No Part IV -- Family and Employment How many people are presently eating in your household? (a) adult (over age 12) male female (b) children (under 12) male female (c) total family total total Do you and/or any of your family member -— now in your house -- work in the village area? Yes/No IILye§, (a) who? (b) what type of work? (c) how much is earned? Are any of your family members away from the village at this time? Yes/No lf‘yee, how many family members are away? For each member of your family that is away, would you please give me the following information: Relation to interviewee? Are they married? Where are they now? Family here or in place What do they do? of work? How long have they been away? ______ How often does he visit What is their age? village? 56. 57. 58. 191 Years of schooling? How much cash do you Do they speak: Urdu/English? receive? Do they write: Urdu/English? What else do you receive? Why did he leave the farm? ____ Have you worked away from the village? Yes/No IILyeo, then ask: Where? Doing what? For how long? When did you return? Why did you come back? How do you and your family members know when there is a job available outside the village? What kind of work do you/they usually look for? MICHIGAN STATE UNIV. LIBRARIES I“llWINill)”I"NWIN“)WIN”WIVIHHHI 31293000968515