Some factors influencing wheat production in West Punjab : cross-sectional analysis of West Punjab watercourses wheat production data

This paper is prepared for staff use. and is not for publication. The views expressed are those of the author and not necessarily SWP20 those of the Bank. IITERNATIONAL 3AK FOR RFZONSTRUCTION AND DEVEjOPMT Economics-Department Working Paper No. 20 SOE FACTORS INFLUMiCING 'rMAT PRODUCTOIO IN WEST PUNJAB Cross-Sec;ional Analysis of West Punjab Watercourses 'rneat Production Data (l964/65) This paper reports on a partial analysis of the performance of wheat production in 20 West Punjab watercourses during the rabi season 196h/65 with respect to selected inputs and soil factors in order to estimate the responses of wheat production to quan- titative changes in these factors. It is part of an exploration of the possibilities of deriving agricultural supply models for developing countries. The data were processed by the Bank's Statistical Services Division. Interpretation of the results is only tentative due to the deficiencies in the data base and the author's lack of familiarity with the specifics of West Punjab agriculture. June 19, 1968 InveL!ment Planning Division Prepared by: Bernard Oury Pesearch AsAistant: HIIung M. Kim  TABLE OF CONTEDN1 PaP,e I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . .1 II. The Data Base . . . . . . . . . . . . . . . . . . . . . . . . 2 III. Methodology .. . . . . . . . . . . . . . . . . . . . 6 IV. Estimated Equations ...... .. .. . .. .. . .. .. . 10 V. Response to changes in the Factors Investigated . . . . . . . 17 VI. Conclusions . . . . . . . . . . . . . . . , . . . . . . . . . 23 Tables Table 1: Identification of the TWatercourses. . . . . . . . . . . 3 Table 2: Acreage, Yield per Acre and Total Production of Wheat, by Watercourses (1964j/65) ............... . 11 Table 3: Data Pertaining to the Factors Investigated, by Watercourses (1964/65) . . . . . . . . . . . . . . 12 Table 4: Model I, Exponential or Semi-Log. . . . . . . . . . . . . 14 Table 5: Model II, Log-Linear. . . . . . . . . . . . . . . . . . . 15 Table 6: Matrix of Intereorrelation Coefficients Among the Factors Investigated. . . . . . . . . . . . . . 16 Table 7: Response of Production, Acreage, and Yield of Wheat to the Individual Factors Investigated. . . . . . . . . . 19  -1- I. INTRODUCTION 1. This analysis atte=pts to appraise tile performance of wheat production in twenty (20) West Punjab watercourses during the rabi season 1964&/65, with respect to selected factors such as labor, the number of draft animals, the pH of the soil, the depth of t;ie water- table, salinity, and irrigation, in order to derive some e7aluation of changes in wheat production that could be expected to accompany quantitative changes in these factors. It is part of an exploration of the possibilities of deriving agricultural supply models for developing countrieo. 2. A multiple regression analysis of wheat watercourse cross- sectional data is performed so as to estimate the direction (plus or miaus) and the magnitude of the changes in wheat acreage, yield, and production on the wtatercourses which could be associated with changes in the factors citea Above. At the same time, this application of regression enalysis to cross-sectional data provides an indication of its usefulness for purposes of agricultural project appraisal. 3. The cross-sectional approach allows us to analyze factors generally left outside the scope of time series analysis such as soil differentials as shown in this application. Cross-sectional analysis does not raise some of the complex problems of statistical estimation encouwAtered in time series analysis, notably the problem of serial correlation among residuals. Cross-sectional data may conveniently * -btained from sample surveys, which in developing countries are often the only meass of producing t.e requisite data base.  -2- II. THE DATA BASE 4. The data analyzed were collecterH from International Bank for Reconstruction and Development (IBRD), Program for the Develo?rnent of I1 Irrigation and Agriculture in West Pakistan, Comprehensive Report, Volume 10, Annexure 14, 1 95l , 'a e.ared by Sir AlexanK'er C-ibb and ')artners, International Land Development Consultants, and liunting Technical Services Limited. They are individual farm data aggregated into 20 watercourses, each related to barrage or feeder canal command. The 20 watercourses are about evenly spread geographically over West Punjab from Bahawalpur in the south to Sargodha and Gujranwala in the northeast. 7ney are located in the major cotton and wheat producing areas of West Punjab and identified in Table 1. The areas commanded by the water- courses range from 266 acres to 360 acres, with an average of 310 to 320 acres per watercourse, according to official or survey estimates. 5. Wheat production appears to have been practically nil in watercourses 4 and 16. Sone discrepancies were encountered between two different series of wheat acreage estimates jiven in the Report and a compromise series involving some guesswork had to be dorived. Adequate information unfortunately was znavailable for several watercourses for other vL..iables of critical importance,such as fertilizer (given only for wa ercourses 7., 9, 10, 12, 14, and 18),thus preventing a more complete analysis; irrigation actually applied to wheat was unavailable for watercourses 5, 6, 16, 19, and 20,thus weakening the validity of the 1/ Later in this paper referred to as the Report.  -3- Table 1: Identification of the Watercourses Water- P 1/ course NP Caral Distributary Outlet No. No. 01 P Lower Chenab Shahkot 42795 R. 02 p Lower Chenab Lakhuana Minor 11450 T.R. 1 Mohalwala 03 NP Upper Chenab Kailaske 35280 R. 04 P Upper Chenab Nurpur 346CO R. 05 P Central Bari Doab Niazbeg 131780 R. 06 p Central Bari Doab W'an 46065 R. 07 N? Dipalpur Dipalpur Minor 6522 L. 08 NP Dipalpur Probindabad 29295 L.- Nurpur 09 Lcwer Chenab Gojra 16561 L. 10 2 Lower Chenab Khixtra 103331 L. 11 P Sidhnai Hittar 43795 R. 12 P Sidhnai Venoi 79660 L. 13 ',NP Bahawal Sultan 11745 R. 14 2/ NP Abbasia Gilani 5964 R. 3966 R. 15 NP Rangpur Ganesh Minor 10714 L. Jagatpur 16 NP Rangpur Bukhari Minor 2U102 L. 17 P E. Sadiqia Jarwah 1211 R. 18 P E. Sadiqia Gajjyani 94850 R. 19 P Lower Jhelum Deowal Dist. 12975 R. 20 P Low-r Jhelum Chaukera 52713 R. 1/ P = Perennial irrigation, NP = lon-perennial irrigation. !/ Watercourse No. 114 is made up of two small watercourses. Source: IBRD, Programme for the DeveloDment of Irrigation and Agri- culture in West Pakistan, Comprehensive Report, Vol. 10, Annexu-ze 14, 195, p.6.  analysis performed in this loaer, t the estimaaces derived from this investigation still am)ear sensible. Information on salinity was un- available for watercourses 3, 14, and 1S. In view of these limitations, wiatercourses 3, 4, 14, and 1S have been excluded from tne analysis which was finally performed on a samrle restricted to 16 iratercourses out of a total of 20. In the absence of aDpropriate data, weather conditions had to be assumad to be the same over all watercourses during the rabi season 1964/65. The same goes for economic variables such as prices received by farmers for wheat and substitute crops wrhich were also assumed to be the same for all watercourses. o. In all, the acrea-e gro-n to wheat covered about 2487 acres (exclusive if 338 acres in w;atercourses 3 and 14) or, on the average, about 155 acres per watercourse. 'feld differences between watercourses were rather large, ranging from about 5 maunds to over 15 maunds per acre. However, half of the ,*atercourses had yields of around 15 maunds per acre, and the average yield rorked out to 13.5 maunds. 7. The watercourse data analyzed are derived from a survey of 564 individual farms. The individual farm data would of course have provided a better base but unfortunately they wsere not available. For instance, they would have made it possible to broaden the scope of the analysis, and to consider a larger num.ber of variables conceivably influencing wheat production in 7est Pun,iab. They would have made pos- sible the preparation of stratified samples in order to go deeper into  the analysis of the relationships of wheat production and its components (acreage, yield) to the various factors influencing them. I  -6- III. METHIODOLOGY 8. Three models have been used to perform the analysis and to ascertain the consistency of t-he findings. They are all centered around the identity: (1) Q N x Y where Q is the total production of wheat, N is the area grown to wheat, and Y is the average yield per acre of wheat. [.;. e.e: i l. ...i; :'CTZ)-_- euat i.on (1 ) bec-5* -ez (2) loge Q = loge N + 1