1 ]_ PROJECTIONS UF PRODUCT SUPPLY- AND FACTUR DEMAND ‘ _ .5: i_ if, UNDER STRUCTURAL CHANGE F0 ‘‘‘‘‘ KOREAN AGRICULTURE A: SYSTEMS SIMULATICN ARRRCACN DE SSCTIa’IIon foI the Degree (If Ph D ' ' MICHIGAN STATE UNIVERSITY ' ‘ ’ IEUNC IIAN LEI; , 1974 ‘51?7¥%52i59125241%2 LIBR/A I? V Michigan Hug“; University E}; 0 ABSTRACT ‘37 0\ ' PRQJEcrIONs 0F PRODUCE SUPPLY AND FACTOR DEMAND @. UNDER STRUCTURAL CHANGE FUR KOREAN AGRICULTURE: A SYS'I‘EMS SIMULATION APPROACH By Jeung Han Lee The primary purpose of this study has been to build a model of part of Korea‘s agricultural production system to be used as a com- ponent of the NBU/KASS model. Since the acreage response component has already been built, we have concentrated on modeling yield responses and factor demand of various crops in different regions. The basic emphasis of this study is on the yield effect of structural changes growing out of public policies, programs and projects designed to influence technology, institutions and people. It is recognized in this study that the major sources of productivity growth and development are structural changes. One important byproduct of this study has been to show empirically lbw different disciplinary theories and techniques can be combined to model a complex system more precisely and accurately. Useful neoclassical economics (undified or ummdified) , develop- trent and growth theories are incorporated in this model, along with concepts, theories and descriptive information from other disciplines. The systems simulation approach has proven a useful technique in IE'U' JemgHanlee integrating these diverse inputs into a yield determination component that can be incorporated into the larger KASS model for use in solving practical problems in the camplex multidisciplinary system, which is Korean agriculture. Economic development in agriculture is a complex process. An equally complex set of policy instruments is required to affect transformation of traditional agriculture. Thus, the model dealing with this complex system must be complex enough to measure important possible repercussions of complex policies, programs and projects. We have tried to meet comprehezsiveness, consistency and balance, clarity, workability criterion in a sector model for planning purposes. We specified a Cobb-Douglas type production function for every crop in each region under consideration with two basic kinds of vari- ables: conventional inputs and structural change variables. The latter shift the yield function as well as the factor demand function. There are three different structural change variables. The first involves biological technology and 111mm change (biological research and exten- sion of its results). The second involves land and water development. And the third is the variable exclusively related to perennial crop production such as tree crop age cohort and residual effect of the conventional inputs used in the past. The first two structural change variables are generated mainly by the public sector. The rate of land improvement has been modeled by a high-order differential equation as a function of public irwestment, among others. The same is true for biological research and dissemination of its results. We have also recognized the existence of indigenous innovation among the leading farmers and by the agribusiness sector. JemgHanlee In order to estimate input usage for conventional production factors under the assumption of optimizing behavior, we have derived a factor delend function from the production function. In doing this, we have used several behavioral constraints. First, we have imposed a capital budget constraint modeled as a stepped supply function for credit. Second, various elasticities of factor denand have been adjusted, based on the direction, duration and magnitude of prices of both products and factors. The model allows adjustment to take place as a result of regional specialization, long-term profitability and for other reasons. Once the relevant marginal rate of return to capital, as deter— mined by the supply and demand relationship, was known, it was a mechanical process to project input usage and hence output. This permitted us to use accounting equations to compute the relevant aggregate variables . After testing the model, through a series of sensitivity analyses, to determine whether it worked properly, we specified several policy operiments with variables. Then we made carputer runs for each level for each policy variable and several different combinations of policy Variable levels. First of all, we identified quantitatively the sources of pro- ductivity growth for each crop in each region in more detail and pre- cision than any study has thus far achieved. The major conclusions drawn from the policy experiment computer run can be smmarized as follows: First, important complementary relationship exist among the so-called conventional inputs, between these JeungHanLee inputs and structural change variables, and between technological change and variables governing farmer incentives. The major determin- ants of conventional input usage, especially fertilizer, seem to be: (1) varietal change and (2) land and water development. Second, it appears that biological technology is a critical and leading deter- minant of yield growth. The second important structural change variable in productivity growth was found to be irrigation. Another important structural change variable defined in this model was found to be age composition change for tree crops. Several values are important in the development of Korean agriculture. The simulated results of this study cannot be fully evaluated in terms of these performance variables Imless the model presented in this study is linked with other components of the KASS model. For this reason, we have tried to evaluate alternative policies mainly in terms of food self-sufficiency and, in doing so, have assumed that the producer prices, areas allocated to each crop and consumption needs projected by the initial version of the KASS model correctly represent the future. Recognizing that biological technology involving varietal change is a crucial factor determining yield increases, we made several alternative assumptions about possible biological research accomplishments on the part of the Korean agriculture in order to project the simulated consequences of these alternatives. In connection with this policy experiment, we have concluded that Korea is, at best, able to achieve her food self-sufficiency development goal in late 19703. In the case of the worst biological research assumption, Korea was not able to attain this goal even by 1990. JeungHanlee The degree of food self-sufficiency depends substantially on the commitment of resource to improve biological technology. All sets of conclusions reached here should be interpreted with reservations. This is so partially because various levels of inter— actions with other sectors or subsectors of Korean economy are not fully taken into consideration, partially because the model presented in this study needs some further refinements, and partially because the model's data base is rather weak. Needless to say, projections based on the model components developed herein and intended for use in evaluating public policies, projects and programs will be much improved when this componelt is linked with the rest of the KASS model. Limitations of the present model and further study needs for improving it were preselted. Needed additional study has to do with: (1) data improverent, (2) refinement of some model structures, and (3) linkage with other components of the KASS model. Nevertheless, the version of the model presented here seem to represent the real world situation reasonably well; that is, the model seems to be capable of projecting yield levels and related conventional factor demand and projecting the consequences of various policy alter— natives in terms of relevant criterion variables. With further refine- ment the model can be useful in evaluating policy alternatives for Korean agricultural development . PROJECTIONS OF PRODUCT SUPPLY AND FACTOR DEMND UNDER STRUCTURAL CRANE FOR KOREAN AGRICULTURE: A SYSTEOB SIMULATION APPROACH By Jeung Han lee A DISSEHATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of mm OF PHILOSOPHY Department of Agricultural Economics 1974 To my father, who laid the foundation, but died without seeing the completion of this work. ii ‘- n.— in Pit iiess Gratitude is given to Professor Glenn Johnson for the creative guidance he gave througl'out the author' s entire graduate program at Michigan State University. The author also wishes to express his appreciation to the merbers of his Guidance and Thesis Committees, Drs. Ttomas J. Manetsch, thesis supervisor, George E. Rossmiller, Robert D. Stevens, William J. Haley, Michael H. Abkin and Victor E. Smith, for their assistance during his graduate program. The author is grateful to the Agricultural Development Council, Incorporated, for providing the fellowship that made the author‘s graduate program possible. Profound appreciation is exteided to Professors Herman M. Southworth and A. B. Lewis, former associates of the Agricultural Development Council, and Earl 0. Heady, who encouraged advanced graduate studies and arranged the fellowship for graduate programs at Iowa and Michigan State Universities. The author also wishes to thanks to Drs. Tom Carroll, field leader, IVSU KASS project, and Dong Hi Kim, Director, NAERI, and his staff for collecting data and other assistance. Special appreciation is expressed to Sandy Clark, who edited the entire manuscript, To Enid Maitland, Judy Pardee, Ellen Vander Lugt and Patti Stiffler for typing the early draft and assisting with other secretarial matters, to Bert Pulaski. for administrative services, and t0 Claudia Winer, Christopher Wolf, Keith Olson, James Williams and many Others for assisting with the computer program. iii Finally, a nurber of colleagues in the Department of Agricultural Ecommics contributed assistance and encouragement during the period of the author's graduate program. The author is appreciative of these acts of kindness and frieidship. iv up"IVE... ... '1 O . . Q I TABLE OF CONTENTS PREFACE..... ...................... II HI PART I BACKGROUND AND PURPOSE OF THE STUDY THEORIES IN AGRICUETURAL SECTOR.ANALXSIS AND PIANNING ..................... Introduction ................. Aggregate Production FUnction Studies ..... mare Economdczkbdels oEIXDKSerent ..... Development Strategies ............ Sector Analysis Models ............ Sector Planninglfixkfls ............ NMDHUJACRICUUDJVQ.SECTOR.STUDY GONE» DIEEL PURPOSES, OBJECTIVES AND SCOPE OF THE STUDY Purposes of the Study ............. Scope of the Study .............. Objectives of the Study ............ PARTI MKHEQEIICAL STRUCTURE OFIIIEL PUBLIC INVESIPEhflF-IAND.AND WATER DEVEDOPMENT Introduction ................. PUBLIC INVEsnm--BIOLOGICAL RESEARCH AND EXJENSTION .................... Biological Researdh Subsector ......... Theory of Innovation Difosion ........ ltuimmatical.hbdel of Social Diffusion of Innovation ................. PRODUCTION FUNCTION AND PRODUCT SUPPLY HKDECEKDI .................... Agricultural Supply Studies in.Literature . . Yield Projection.Mbde1 for Annual Crop . V 66 66 88 88 94 106 129 129 147 Product Supply P. Perem’al Crop . Parmeter Estimn' D FACIUIIDIANDPRUE FactorDenande IaborDaTand Pro, BASIC RESUIID m Imm TALIINTNN AN Intmductim . . Ibdel Verificatii Sensitivity AnalI Land and Water Di Imletmtatim . WWI o. L! We limo Abdel . was Chapter Page Product Supply Projection Pbdel for Perennial Crop .................. 151 Parameter Estimation ............... 156 VII FACTOR DEVIAND WON .............. 167 Factor Demand Projection lbdel .......... 173 Labor Derand Projection Model .......... 198 PART III BASIC RESULTS AND SENSITIVITY TESTS VIII I‘DDEL VALlllATICN AND SENSITIVITY TESTS ....... 203 Introduction ................... 203 lbdel Verification ................ 205 Sensitivity Analysis ............... 214 land and Water Development Project Implementation .................. 225 D( POLICY EXPERIMENTS AND THEIR mum's ........ 236 PAKI‘ IV POLICY IMPLICATIONS AND (INCLUSIONS X POLICY IMPLICATIONS AND CDITULUSIOI‘B ......... 257 Pbdel Assumption EValuation ........... 279 Further Study Needs ............... 281 XI SW ....................... 284 Maj or Conclusions ................ 290 APPENDIX ........................... 296 BIBLIOGRAPHY ......................... 326 EI'IU 3.1 5,1 5.2 I I! II P. ll‘. 1‘: h; SImmryOf Policy (Imp IIIRelative to Altem Hypofiletical PlanedEI ofRate of Increase in AdjuSted by Proportim 1y be Used Nation Value of Disc: ATTENDANT WWW COefficieI: Farmers or Data to t We . we'lmgatim ijec Table 3.1 5.1 5.2 5.3 6.1 8.1 8.2 8.3 8.4 8.5 8.6 8.7 LIST OF TABLIB Summary of Policy Components of Alternatives II and III Relative to Alternative I ............. 63 Hypothetical Planned Expected Research Results in Terms of Rate of Increase in Yield at Experiment Station Adjusted by Proportion of Land Where Results Could Advantageously be Used ................ Emotion Value of Discomting Factor, DF (Hypothetical) 124 92 Productivity Gain Due to Exteision of Spontaneous Techrological Change (Hypothetical) .......... 127 Productivity Coefficients of Rice Irrigation ..... 162 Parameters or Data to be Varied in Sensitivity Tests . 216 Structural Elasticity of Accumulated Area of large- Scale Irrigation Project in Region 1 with Respect to Unit Cost in Selected Years ............ 217 Accumulated Expected Yield Increase Due to Research and Exteision, 675, Rice in Region 1, by Different Values of AEFA, in Selected Years ........... 217 Accumulated Expected Yield Increase Due to Research and Extension, (ZS, Rice in Region 1, by Different Values of AMEA, in Selected Years ........... 218 Accmlulated Expected Yield Increase Dre to Research and Extension, 623, Rice in Region 1 by Different Values of DFC, in Selected Regime .......... 219 Accumlated Expected Yield Increase Due to Research and Extension, 028, Rice in Region 1 by Different Values of mm, in Selected Years .............. 219 Expected Yield Level of Rice (National Average), by Different Initial Conditions of Rice Yield, in Selected Years (Ton/Ha) ................ 220 vii Tile II Expected Yield level 01 Differmt Values of PC! (Tm/Ha) ....... 5.9 Impact of Varying Value Variables, in 1985 . . Ill Impact of Varying Value Variables, 1111985 . . ml Imam of Varying Valu Rama Variables in l 3.1 11M of Varymg' Value ALDSI on SQleCtEd RESIX “3 Total haunted land Different Ratios Of Act 1985 (1.000 ha.) . . , m “’11 Aemulated Land M. by 1985 (1, 3'1 Total Des' “first Ratios Of A3 1'5 Total Ac . Cumlated Different AC 53% GIMME h) . 0 . . Table Page 8.8 Expected Yield Level of Rice (National Average), by Different Values of FCKPX, in Selected Years (Ton/Ha) ....................... 221 8.9 Impact of Varying Value of YLDPA on Selected Response Variables, in 1985 .................. 221 8.10 Impact of Varying Value of ASCs on Selected Response Variables, in 1985 .................. 222 8.11 Impact of Varying Values of FLBPDs on Selected 223 Response Variables in 1985 .............. 8.12 Impact of Varying Values of Productivity Coefficients ALPS, on Selected Response Variables in 1985 ..... 224 8.13 Total Accumlated land Areas Improved by Projects, by Different Ratios of Actual to Desired Investment by 1985 (1,000 ha.) ................... 230 8.14 Total Accumulated Land Area Under Inplementation by Projects, by Different Ratios of Actual to Desired Investment, by 1985 (1,000 ha.) ............ 230 8.15 Total Accumlated Desired Irwestment for Projects, by Differait Ratios of Actual to Desired Investment by 1985 (l, 000, 000 Won) ................. 231 8.16 Total Accumlated Actual Investment for Projects, by Different Actual to Desired Investment, by 1985 (1,000,000 Won) .................... 231 8.17 Total National Project Investment by Different Ratios of Actual to Desired Investment, by 1985 (1,000,000 Won) .................... 232 8.18 Average Actual Unit Costs (1,000 Won/ha) of Improved Land and Relative Efficiency, by Different Ratios of Actual to Desired Investments, by 1985, Computed from Tables 8.12 and 8.15 ................. 234 9.1 levels of Policy Variables .............. 237 9.2 Nbdel Response to Change in land and Water Development Investment in 1985 .................. 240 9.3 l’bdel Response to Change in Biological Research Outcome in 1985 ................... 242 file 91 9.5 9.6 9.7 3.8 3.9 ill 3.2 Ill lbdelReponsetOM lbdelRespmseton lbdelkecpasetodm lbdelRespometon Smplyinl985. . .. loc‘elResponseton Ratein1985 ..... MXlelResDonsetoCharg VETiAbleinBBS .. Scurces of Yield Prod; for Rice in ' 1. Set Specified in Table iomfomemRate °‘ °°EinRegionL1 setSpe‘ifiedinTable for Fmits in R5 . 1 HB’Pothetical Planied E El; oidflf Rate of I] Table 9.4 9.5 9.6 9.7 9.8 9.9 10.1 10.2 10.3 10.4 10.5 Page Model Response to Change in Extension Budget in 1985 . . 243 Nbdel Reaponse to Change in Product Price in 1985 . . . 243 Nbdel Response to Change in Factor Prices in 1985 . . . 244 Nbdel Response to Change in Government Credit Supply in 1985 ..................... 246 Model Response to Change in Government Interest Rate in 1985 ...................... 246 lvbdel Response to Change in the level of All Policy Variables in 1985 ................... 249 Sources of Yield Productivity Growth Rate (in Percent) for Rice in Region 1. Based on Medium Policy Level Set Specified in Table 9.1 ............... 260 Som:ces of Growth Rate of Fertilizer Use (in Percent) for Rice in Region 1, Based on Medium Policy level Set Specified in Table 9.1 ............... 262 Sources of Yield Productivity Growth Rate (in Percent) for Fruits in Region 1, Based on Medium Policy level Set Specified in Table 9.1 ............... 263 Hypothetical Planned Expected Research Results, in Terms of the Rate of Increase in Experiment Stations Yield, Adjusted by Proportion of Crop that could Advantageously Use Results (Biological Research Results are Assumed to be Forthcoming Earlier with the Sane Annunt Over a lS-Year Period of Total Accumlated Increase in Yield as that in Table 5.1 ......... 269 Hypothetical Planned Expected Research Results for Sweet and White Potatoes, in Terms of the Rate of Increase in quaeriment Stations Yield, Adjusted by the Proportion of Crop Area that could Advantageously Use Result ....................... 275 U 2.1 mumnowm. Sectorlbdel. . . ,, 2'2 WWW ' and minim 31' lbdified mama] fh WWW (he to fan 3'1 Amdified Version of 1 4'1 1%th beneen u: W or entermg in Meet. ., . . . 31 Six Ste” in a Problem- :2 2:121 0f 1.l'ldiViClual a LIST OF FIGURE Figure Page 2.1 Emotional Flow Chart of Korean Agricultural Sector Madel ...................... 41 2.2 Provincial and cropping region boundaries of Korea . . . 43 2.3 l'bdified functional flow chart of Korean agricultural sector model due to farm resource allocation canponent addition ........................ 47 3.1 A modified version of Korean agricultural sector model . 58 4.1 Relationship between unit cost and accumulated land improved or entering improvement process for each project ......................... 79 5.1 Six steps in a problem-solving process ......... 95 5.2 A model of individual adoption process—-the adoption 101 tree .......................... 5.3 A causal diagram of the modernization component (applies to crop modernization and introduction of mechanization) 105 5.4 Flow chart of innovation diffusion process (hypothetical) 107 5.5 Weight given to research outcane to compute diffusion parameter (DDF) (hypothetical) ............. 113 5.6 Weight given to extension effort to compute diffusion parameter (DDF) (hypothetical) ............. 113 6.1 Hypothetical supply function derived from resource fixity theory ...................... 136 5.2 Cultivated land classification by technology for Indian agriculture ................... 144 5-3 Production function shift among subfmctions due to addition of a new production factor .......... 163 7-1 Stepped capital supply function (hypothetical) ..... 184 x g. ;—- - ... L A.) (‘3‘ . ._.- ..r. - 24‘- 3.2 Indexmtmcted ton factog,basedondirec ofpncechanges . .. Iso-qumts ...... Relatimship between 1' invesmmt and efficie PTOJect iuplmentatim Wage national rice ' tree P01icy levels on mTable 9.1 and that 1 altenutive II . . , . T991 grain productim ferent POIiCV levels 1 “Mellon 7.2 7.3 8.1 9.1 9.2 9.3 9.4 9.5 10.1 10.2 10.3 Page Index coretructed to modify demand elasticities for factor, based on direction, magnitude and duration of price changes .................... 190 Iso-quants ....................... 197 Relationship between ratio of actual to desired investment and efficiency affecting delay time for project implementation ................. 229 Average national rice yield level projected under three policy levels on research outcomes, specified in Table 9.1 and that projected by KASS under policy 242 alternative II ..................... Total grain production projection based on three different policy levels, specified in Table 9.1, projected by KASS under policy alternative IIdenoted by * and consmlption needs denoted by 0 in 1971, 1975, 1980 and 1985 ........................ 248 Projection of yields of rice, barley, tobacco and other grains, based on the likely policy level set . . . 251 Projection of yields of wheat, pulses and potatoes based on the likely policy level set .......... 252 Projection of yields of silk, industrial crops and forage crops, based on the likely policy level set . . . 253 Projection of yields of vegetables, fruits and grasses, based on the likely policy level set ...... 254 Aggregate demand for fertilizer and total factor measured in terms of service and expenditure, based on the mediun policy level set specified in Table 9.1 . 266 Total grain production projection, based on medium policy level set, specified in Table 9.1, and that based on assumption specified in Table 10.4 (only difference between two runs is different assumptions between Tables 5.1 and 10.4) and KASS projection of consunption needs ................... Total grain production, based on worst case of research outcome where last three research outcomes for each crop in each region are not realized and fertilizer demand based on above assumption, and KASS projection of consumption needs, and total grain supply ........ 272 xi Flue lll Total grain PI'OdUCCi-G on reserch outmes, Table 10.5 wtere sleet assented (bibled or to pmjection ...... Figure 10.4 Total grain production projections, based on worst case on research outcomes, and on new experiment design in Table 10.5 where sweet and/or white potato yield is assumed doubled or tripled and KASS consumption need projection ....................... Page 276 Basically, this transf lnagriculture. '1'an it is not primarily a j rather a problem of de Mt take, fonts that e l“ agriculmre [Schultz The raport presented he 5953151101161 for Korea. 1:01 My 0“ the Pmdllction si Elwin this study is a 5 William 1‘bdel oonst: slim“ ““519 agri Wmmmfl We ee - . ulenes, Mile of Kor “P initial VeIsiOn of re PREFACE Basically, this transformation is dependent on investing in agriculture. Thus it is an investment problem. But it is not primarily a problem of capital supply. It is rather a problem of determining the form this investment must take, fonms that will make it profitable to invest in agriculture [Schultz (5.2, p. 4)]. The report presented here is a case study of agricultural develop- ment planning, based on a comprehensive and consistent agricultural sector analysis model for Korea. For the study to be manageable, the focus is pn‘marily on the production side of agricultural developmmt. The model presented in this study is a subsector model of the Korean Agricultural Sector Simulation Model constructed and being improved jointly by a Michigan State University agricultural sector simulation team and National Agricultural Economics Research Institute, Ministry of Agriculture and Fisheries, Republic of Korea. The initial version of the Korean Agricultural Sector Study (KASS) model is reported by Rossmiller, et a1. [R.7] in 1972. Since then, there have been several ongoing studies to refine the KASS model. This study is one of those attempts. Since the author is oriented toward produc- tion ecoromics, the attempt is to improve the production component of the KASS model. The first refinement attempt for the production side was to build in a linear programming component for the KASS simulation model. The Purpose was to guide resource allocation among various economic antivities within and among regions, as will be reviewed in the 1 mission of the KASS model Wt in Glapter II. The KASS teen W‘ poem in the Korean agn‘ Eiil. [F.S, I972]. Public 1' etiology, institution or ht leion possibility frontier, tecmbimtion of these. I pile sector plays an import meet in generating charg Me. For this reason, a o 13tlirlk8L1‘B Where 8 is production elasticity of capital. By differentiating the above equation totally with respect to time and dividing by Q(t) , we have: (t) _A(t) I°5. IU—Fugbn— Hun-Irimeuu ....Aa‘u —...::r.a‘.-.I. .105... )3. LG..> ...-.a-.. gar—av..- 4 in—v In-acoahflu cur-l lav-L..-'.v \ cub-4...— '—.nwl~c ..u.. 'lVaF-Iu a... ‘..-—‘.dIL-.". .bvl-v-v‘. long—Ii I .u.....~.. a...u~l\ .<. .... . .iw- r. b..- I: i all-JFvnchA- IInJhu.‘ n Filiill 2:33....«23; ...»... 4:. p: .. ..u :32 .......‘..>-. _.sw;.v/. .n.._:; VI ...—...~ —!u:Av-.¢\7.. hi ...-r\ nu....~ .q- fawn—al-Au «- ...»...1 .v—.-. / mIK xi . e hi i .2..~:.~ . vanaevra .~2an .w :ozaEuodc. 3325 Each .93. _: 003m 2 not-ammo”: 3:33:05; flotsam—59“: “mm—CLQOZZOU _ :ofiaumcnnuo; woo ul 3003:.— nooaarcocax— ’ Go nuovtm @ ON a. ‘ . mflzvoEEoo . _ .l I l I I~ I I I J _ . _ :OZLZZZCAXV ocmFCoC :_oE_m=_.o< from o:~ fil I I II I I. cant: vent—on . moi; _ radon 15:2... . .< act; _ _ .IIIIIIIL _ 6:3 on: _ 2255.3; . col moo _ _ a :< _ u.:_::_ojr.. I. III“ _ wouJOmom * ~N27L~flz WQOH 4 ~N1::f. A I In I I I I I. III I 3:... 72:12:: 7.0: .....ch¢: :l mac-....onch: 42. . 950.45 71:35:: . 3:9}. cofiozoota ~a..:::o_..m< Enacoz _ Tao“; isolator; mEouE. pea nicer. i .3 50.3 5 1.6.: ...;m z .. 4. ..QNZCuoL :50. new 10.... >2. 3:932 5:350 0.. .3 E ~01..th ocano monu— firemaom 30.3 it Lno>-:m—:\~:3L .530u1 ocm note—Nu ninauiv; «Douche annex ouaioutmsq. 303 Se .2_\:._2; D0503: NZQNUIuol poo—um 03a.) u<_ZM.H_mU .II I I I I I _ ._ . ......nnm... . Em...» acidity.“ pl I I III I L oucaomao “a. WEEHMOHAmg MISS model is shown minimal sector intc Iimludes tm mrtiwest I‘egimiinclude the fc him: and Region 3 1 mild cropping is relat There are nine fu 518m 2.1. Hmever, ’Ee not completed by th. m. (Rm. Thus the Hillielized Projections itIOped by Omittees 1 i1 Policy variables are 42 the KASS model is shown in Figure 2.1. The KASS divides the Korean agricultural sector into three regions, as shown in Figure 2.2. Region 1 includes two northwest provinces where single crop paddy is dominant. Region 2 includes the four southern provinces where double-cropping is dominant, and Region 3 is made up of other agricultm‘al regions where upland cropping is relatively important in production. There are nine fmetional subsectors in the KASS model, as seen in Fing'e 2.1. However, three of then (shown with a dashed outline) were not completed by the time the sector report was published [Rossmiller et al. (R. 7)]. Thus the remaining six components were used to make corputerized projections for each alternative policy set projection developed by committees for the three informal subsectors. Note that all policy variables except family planning programs, directly or indirectly conceived to affect the resource base and other farmer decision variables. These policy variables induce technological, institutional and human change. Despite the crucial fmetional linkage between policy variables and technological, institutional and human changes, the KASS model links these two sets of variables informally. For these informalized subsectors, KASS used what Johnson [J .12] calls traditional projection. BlaCk and Bonnen [B.8] and many others have used these techniques, WhiCh are reviewed in a later chapter. Let us look more closely at each component. Interested readers are urged to refer to the original report. Yield levels of 19 agri- Cultural commodities or commodity groups are projected for 1975, 1980 and 1985 by a committee for each of three regions and for each of 43 East Sea “r .4! .. . .,r_r ".(r «f.- Uri/(Ir? 2 Yellow Sea Eégié Single Crop Paddy - Double Crop Paddy \.r a, Upland .- - . --' - Provincial Boundaries ‘-—-—u KASS Cropping Region Boundaries Figure 2.2. Provincial and cropping region boundaries of Korea. me policy alternative iips with famner decisi man mi _Im_c basis. The resotn'ce allo min to each of 12 cm olmd development and ilocated to each crop 0: it Price relationshi] The samurai I ideal Pmlctiom level iii) by Hllltiplying yie] iiis Wait. The f It“ (“Ch as tree fin iii (MDT), WhiCh was 44 three policy alternatives. Behavioral or formal functional relation- ships with farmer decision variables were considered by the committee on an ad hag basis. The resource allocation component deals mainly with land allo— cation to each of 12 crops or crop groups, taking into consideration new land development and land disappearance due to urbanization. Land allocated to each crop or crop group was also projected informally, taking price relationships into consideration. The agricultural production component first determines the physical production level of the agricultural commodity or commodity group by multiplying yield and acreage. There are two dynamic aspects in this component. The first is simulation of perennial crop pro- duction (such as tree fruit) by using a subroutine of distributed delay (DELUI‘), which was developed by Abkin [A.l]. The second is a seasonal labor requirement profile so that mechanization level can be determined. The crop accounting and farm consumption component determines gross and net revenue and farm consumption. The difference between total production and on—farm consumption and losses at various market stages is called marketed surplus. An interesting feature of this component is change in on—farm inventories of farm products. Such irwentory changes in turn influence market supply over the season so seasonal price level can be determined by interaction with urban demand. The regional and national accounting component computes aggregate performance variables such as value added. The national input-output component, which is represented by 22x2nntrix (one for sector) was not actual] {use of the project. 1 m and growth rate V imbecaze an irput to Nation the size of w oration oonponmt. om mamas that income a As indicated in F; mOperational by the t- ”ileted. Instead, the if! than major grains 5 Wine by aCOOmting s so muld be Stabilize Ejorml’rices are tr The 130de and Bible directly r Elata 45 a 2 x 2 matrix (one for the farm sector and another for the nonfarm sector) was not actually incorporated into the model for the first phase of the project. Instead, projections of nonfarm gross national income and growth rate were made using informal methods. This projec- tion became an input to the urban demand component along with an urban population the size of which was generated from the population and migration component. One interesting aspect of the urban demand can- ponent was that income elasticities are time-varying parameters. As indicated in Figure 2.1, the price adjustment component was not operational by the time the first phase of the project was completed. Instead, the price levels of agricultural commodities ’ other than major grains such as rice were generated by an iterative procedure by accounting supply and demand interaction so that price levels would be stabilized at a reasonable level. Note that the major grain prices are treated as policy variables. The population and migration component generates many important variables directly related to the farm sector, such as population size by rural and urban, by regions, by sex and age, etc. However, this conpmmt is independent from the other components of the model in the sense that there is no interaction between them. We have briefly reviewed the initial version of the KASS model, Which was incomplete as a sector study model. Yield projection, resource allocation, price adjustment, and national input-output model Moments were treated as exogenous; interaction between farm and mnfanm sectors and among agricultural regions were not taken into consideration; the model is indadequate or is missing several impor- tant components such as modern input supply, food processing and dmbution, meme di nomant in agricultur Fortunately, the afthe KASS project is . nibdldjng in addition he and capable of pl; it here are several film a linear progr M10“: the price ad Ht“Went to deal i amnion, the mum me population and 1m distribution, income distribution, employment, etc., which are ectreiely important in agricultural development processes. Fortunately, the KASS is a continuous study. The second phase of the KASS project is conceitrating on refining the initial version and building in additional components so that the model is more real- istic and capable of planning objectives more precisely. At the present time, there are several ongoing projects dealing with this matter, including a linear programming component to deal with farm resource allocation, the price adjustment component, the government grain manage- ment component to deal in part with stabilization of seasonal price fluctuation, the national input-output model component, and refinement of the population and migration model component. The model presented in this study is also an attempt at refining the initial version of the KASS model, focusing mainly on the functional relationship among crop yields, factors used, and public investment. The model preSented here is primarily designed to supply the farm resource allocation component, a recursive linear programming model, with the necessary parameters over the planning horizon (1971 to 1985) . Included are yields, factor demand, objective function coefficient canponents other than prices, and land resource constraints. Thus, we will briefly examine the structure of the farm resource allocation component for better understanding of this effort. The discussion is based on papers by De Haei and Lee [D.7], De Haen [D6], and Lee [L.10]. The overall flow chart of the KASS model after the farm resource allocation component model is introduced is shown in Figure 2.3. Let US examine the changes. First of all, two carponents, the resource [i 4. _ t:3...f.?.-~n. 2.3.1....- «ut‘ >......:.:...; >... .223 a... LBJ-.4 ......IC.AU~ . ~G:QN o uaquflvné 3...: .... 2311.:~ INNO2CVIHV . — A , :9: ...—IE: .Iu..._.. .. . a < 2 .35: .7. 2.1.33- 2.1:. . . LbN..~ZLo~.+ IT-CLpp: 1....- t.L.;_r _ —1.::.Il~ 7.1;»! ....—~ » . < “ECO-uflhfioha NU-an "iii-f. ——C-Av— arr-I :3.- 1 >3. . IncuthLA» “Up-Crush? hnufilu —flur.nvla~.vwn anuflvhnv >Ad- L‘U>Ia.fl—:\.—.—.«L.~ has... 4 ’flao ‘.-\\-—-L.— Op..°~°bnb 1!..- lazinv—IU “...—G. . ...... ., .. .I. i I: i I i ~ 13%?! U$>1hi:.. -.var-Av.aa.¢ ‘..:.‘ ‘Ih.-.. - llada- 4.u...k.\ \ .h-u\-uunn.q ~ IV.I-fll' 'Vov-lw I ~§~ ‘ N\-I- I ..uvnlnvoiulfln I‘flvl-uo ‘L-—--.l¢.un‘< ...-ur‘ 1 .(.-“Il\|l|.l. . . . ny~.- 1-.V Vi. rt . asauwlwtukk ‘ |-.anv ol‘n 'II\I\I\$|1\~\A~1V I: .83.“on unmanned“. 8383.2... 8.58mi Show 8 can apnea Houoom HmHfiHdUH§m fimohox mo uHEo Bod Hucofiuofifi oowmaoae .m.~. ogwfim .23.. nonunion 5263 no 883 3653 an; 3&5?an . J no 55883 38:2: 3. sagging Eton—SHE: 38:3. ":1. 3:3 E 388:5 33......» 3.23 ® awn}; . SN swatch 8. inning—osmium .o. cuuono .. 3 E54330»; mco:uofio..a nona— en: :2 $6.50; a cocfisaou 8.2.52 .282 c0521.}. new cosflsaoa Each nwmm .2 uuzazono.) F “En meiotic; coonU 3. mom—a; 3 {on .0 :3 l ..u «.15. 3.— nEuum .350 2. noun .2 is: ... dam nacho 32:22: .N. son—um .N :ocaficanuo yam :— . ouWMOmW concur—anew}. .mmrHaOE avail «30.3.3 mounts-v.54 * do 393.3.“ ON 0. . mflzuoEEoU _ 1 u I u. ._I I I J _ 7 . _ :9:nE.2:o o In». «scarce: Eoc__m:€< III :Cm * vow a :mo..; o:a:.wv . 04C; _ ran—1m 32:3: . .4. .729 r - I n u n u ‘ ,oofimmflmeoomm , 183$“ u:2c:04u< a , 1:227. was .833.“ a: 35......“ L858“ Each _ “8383.2 51 20:92ch Sgowmm am “one.“ :35“: van 3535594 >5 2555 ...:cflig . “adenp cocoa—vogue “orangulut. Enacoz Typo—z ...:5 Jr: manic maniaunb 2.59:. ucn ntonxu 3330-:35 1‘ rerun .51.. / J4< panacea; :33 new menu >2. 3:932. "mcoiu a: .5 uvumuohe «Ease—u Lon: _d:0uoom 30...... >5 uao>lcaE\.:ot; 393 3.. ashram; .5393 one auto—«u sin—«.77.: .0503: gadolno; .lI I l I I I. wanna oann> . 355,132.... . mouth finisher—x do... can. one.» 0.2.: u u W‘I con—HST... v 2.. I < “I 31:.» .he—:23: ..(JKHMLZU I I II I I .L oucnvmaC - "a: MENLJ. Ou1© fixation and agrimlt 3.3mm versim of fizmresome allocati0 Idesigned to deal Hail Iborand others dealim ties). At the same tim mmm and 1 thhenndel, which the aemrmic adjusmmt c Hue basic inputs Flick, price levels of j free available. With t' Elbe categorized as £02 1. Inputs to the coupon ent: a) Areas allc for each I 48 allocation and agricultural production components in Figure 2.1 of the initial version of the KASS model, are now merged together as the farm resource allocation component in Figure 2.3. This nan component is designed to deal mainly with allocation of farm resources (land, labor and others dealing with different agricultural production activi— ties). At the same time, the model deals with the level and type of farm mechanization and feed grain imports. A more important aspect of the model, which the initial version of the KASS nodel ignored, is an economic adjustment of regional production patterns. The basic inputs to the new linear programming component are yields, price levels of products and inputs, and total land and labor force available. With these inputs, the main outputs of this component can be categorized as follows: 1. Inputs to the crop accounting and farm consumption component: a) Areas allocated to each commodity or commodity group for each region in each year. b) Production of each livestock oomrodity group for each region in each year c) Total value added 2. Input to other components: a) Capital requirement for farm mechanization b) Feed grain import c) Farm input requirement d) Others 3. Inputs to the next iteration of linear programming model itself: a) Capital stocks (farm machinery, perennials, large animals, etc.) b) smdm c) Others In stumry, the molar set of poliC bias or alternative (16 relation model. Ihe is translates these pt :5meer projections bution; hence, supp] Infomal form, t 2.1 maxim = subject to ‘ and but) stands for th actor of objective fUI 1‘3th) for a mtrh izsctor of constraint CE xenon coefficients are 2-2 Wt) = Pyj (t) be!” yj Stands for jth, Inth 49 b) Shadow prices for some intermediate products c) Others In summary, the new programming component does not have its own particular set of policy variables. Instead, it receives policy vari— ables or alternative develoth strategies directly from the existing simulation model. The programmu’ng model then describes how the farm firm translates these policy varialbes, which are revealed in a set of informal projections mentioned earlier, into forms of agriculteral production; hence, supply response through resource allocation. In formal form, the programming model can be stated as follows: 2.1 max H(t) = v(t)*5<(t) subject to g (t)*f((t) 5 I3(t) and 3((t) _>_ 0 Where H(t) stands for the value of the objective function, {/(t) for a vector of objective function coefficients, )2 for a vector of acti- vities, Q(t) for a matrix of input—output coefficients, and 13(t) for a vector of constraint capacities, in time period t. The objective function coefficients are computed basically: 2.2 {/(t) = Pyj(t)*Yj(t) — Pzi(t)*Zij (t) th Where Pyj stands for j output price, Pzi for ith input price, Yj for jth yield, and zij for 1th input for jth outputs, in time period t. As already implied, all elements in \7(t), Q(t), and I3(t) are exogenously determined, except some of I3(t) concerning the flexibility constraints and some of Wt). The kind and ty1 asntially the sane as KISS nodal and appearir aceptions: first, aft mdity group is divi 5600M change is to add PW- The third is ' Jo different processes. a”Mother with a pack.- chasm Mugment. Rice, being the n ifias Wtion, for 1%”?me tille Ed me DamsPlanter. ' WSW (1) With 50 The kind and type of agricultural production activities are essentially the same as those defined in the intial version of the KASS model and appearing in the footnote of Figure 2.1, with some exceptions: first, after combining other grains with pulses, this commodity group is divided into summer and winter grains. The second change is to add a new activity of forage production from pasture. The third is to divide each crop production activity into no different processes—~one with traditional production metl'ods, and another with a package of modern machine inputs, except for rice and pasture managemmt. Rice, being the most important crop in terms of production as well as-consumption, four processes are defined: (1) traditional methods, (2) power tiller, (3) rice transplanter, and (4) power tiller and rice transplanter. The grass production from pasture also has two processes: (1) with fertilizer and (2) without fertilizer. The livestock production activities are the same as those defined in the simulation model, except that a new activity of Korean cattle raising is introduced. The rest of the activities in the programming model are machinery investment, perennial investment, feed import, etc. There are a variety of constraints. The constraints directly related to the present study are for land and labor. The former is divided into three categories: summer paddy, summer upland and winter land. The latter has two categories: labor that peaks during Jme and peaking during October. The rest of the constraints are conven- tional, such as traditional flexibility constraints (Day, Singh, Mldaha, andAhan [13.3, 0.4, M.27 and A.6]). We have briefly allocation cmponait m usiderably improved ‘ odel in my respects r«lesion of the KASS orx hes not contain any ec his, the programming moral and hunan Chan vial farmer decisim ililient POint, 51 We have briefly discussed the main features of the farm resource allocation component model. The introduction of this component has considerably improved the resource allocation mechanism of the KASS model in many respects. However, the basic criticism of the initial version of the KASS model is still applicable: its programming model does not contain any economic development and growth theory. In other words, the programming model explains neither how technical, insti— tutional and human changes take place, nor how public investment affects crucial farmer decision variables. In fact, Falcon [F.l] makes an excellent point, "Agricultural production is typified by the wide range of input substitutions that are technically possible-- not fixed coefficients. . . (and) one of the primary objectives of an agricultural development program is to change the input-output coefficient." All but a few input coefficients are assumed fixed, and the possibility of factor substitution is very restricted in the farm resource allocation model of KASS. A programming model can be constructed to simulate the impact of various levels of public policies, programs and projects on the perfor- mance of the agricultural sector. A good example is found in the Mexico model {GB}. The programming model has many strengths, such as power to hmdle interdependency of economic development (consistency criteria) in addition to handling resource allocation, but it has weaknesses, too. The profit minimization assumption is often criticized [MA] . In addi~ tion, especially with multilevel, multiperiodic or recursive programing, CCmputer capacity is often restrictive for modeling of a large system With nonlinear relationshipS - In emery, the (1)1ovstn1ctm‘al cha dnge would affect th lo» the farmer decisio Face, product SUpply. action reSponse more : W119 and projects. 52 In summary, the major questions this study tries to ask are: (1) how structural change would take place, (2) how this structural change would affect the resource base and/or its productivity, and (3) how the farmer decision is related to this change in resource use, and hence, product supply. Mare specifically, we intend to explain pro- duction response more systematically, depending on public policies, programs and projects. - PURP(BEB, Now that the be: M and its farm resc header is eqllipped hit Purpose of the or. One Of the Host field PrOjection coup 1° to generate twhmlq afPlhlic Hides, pro; CHAPTER III PURPOSES, OBJECTIVES AND SCOPE OF THE STUDY Purposes of the Study Now that the basic aspects of Korean Agricultural Sector Study model and its farm resource allocation component have beei discussed, the reader is equipped with the minimum knowledge for understanding the basic purpose of the model presented in this study. One of the most important purposes of this study is to build in a yield projection component for the KASS model. A more crucial purpose . is to generate technological, institutional and human changes by means of public policies, programs and projects, and to link these changes to farmer decision variables. The impact of public investment is not directly revealed in the change in yield. Instead, public investment induces changes in the number, quality and quantity of inputs, which we call here teclmological, institutional and human changes. Hereafter these changes will be called structural change in accordance with Learn and Cochrane [L.3]. Changes affect resource uses, as well. Our first task is to explain how technological, institutional and human changes occur through public investment. Second, we need to exPlain how these changes affect resource use, which is factor demand. A Change in factor use changes the output level along a given production function, whereas the change agency discussed above shifts among sub— Pl‘ChilJction fmctions. As implied above and discussed in a later chapter, 53 it is theoretically p£ supply (yield) project alinear programing f Hsrver, the approach Imet. This study dev ofshich is another ba Each researcher fist his own teChnique ME real world and TygWeakness of his 0w: Imgth 0f “he—I discj WESS of our own. 1 PM “fools should be mm r9alistic. The Pullose of a W a Particular di. :r‘accurately HDdel a : 2W PM With 1 somant Mme of thj semi disciplinary appr ‘ dim in as well a 54 it is theoretically perfectly possible for factor demand and product supply (yield) projections under the structural change to be made from a linear programming framework by a series of linear approfimtions. waever, the approach is avoided here for reasons to be pointed out later. This study develops an alternative approach, the development of which is another basic purpose of this study. Each researcher in every specialized disciplinary seems to believe that his own techniques are dominantly realistic and capable of explain- ing the real world and criticizes other approaches while tending to hide the weakness of his own approach. We should be able to respect the strength of other disciplines and techniques while recognizing the weakness of our own. In other words, any technique and theory from other schools should be used whenever and wherever they are appropriate and more realistic. The purpose of a sector analysis should not be a simple applica— tion of a particular disciplinary theory or technique, but to reasonably and accurately model a sector's behavior to provide sufficient informa- tion for planning with respect to the problem of the sector. Thus, an important purpose of this study is to illustrate how two or more dif- ferent disciplinary approaches can be incorporated or merged together by feeding in as well as feeding back. From the above discussion, it is easy to identify the system we want to model. The major inputs to the system can be classified as follows: 1. Public investment by central and local government in the form of finance or subsidies. 2. Price polic 3. Credit poli The major syste 1. Yield level 2. Input level 3. labor demm 4. Available 12 These Variables location Wt of all be interested in El Ellerated indirect Wes' Seasoml run I“ land allocation is llllmt IIDdel. The Se lbtlnr Category 55 2. Price policies for products as well as inputs. 3. credit policies with respect to amount and rate of interest. The major system outputs under consideration for each region are: 1. Yield levels by agricultural commodities under consideration. 2. Input levels and variable costs by commodities. 3. Labor demand by commodities and season. 4. Available land by categories. These variables will be fed directly into the farm resource allocation component of the KASS model. The development economist might be interested in other outputs too. Off-farm income and employ- ment generated indirectly or directly by public investment are good examples. Seasonal rural employment level can easily be estimated, once land allocation is determined by the farm resource allocation component model. The same is true for some aspects of income distribution. Another category of variables, which relates output to input variables is called state variables. The various types of improved land, land that has adopted a new technology, etc., belong to this class of variables. The output variables described above are also a type of state variable in this particular case. The final outputs of total system in the context of the Korean agricultural sector are called performance variables as mentioned earlier. The system we want to model can be expressed as follows: 3.1 %l = Q(t) + gnome) Where 2 stands for a vector of state variables, {—1 for a vector of limits, and A and I} are, respectively, a matrix of parameter set. Now let us be: hescope of the stud; lemblic policies, F classified as follows: 1. Public lmve a. Biologi. b. Extmsiz C. Tidelanc e. Uplandd l~ Large-3c 8- Small-so h. Paddy lar 1. Paddy 131 1" UPland cc 1" Upland 1r 2' Fume Polici. a' Price p01: 56 Now let us be more specific about the kind of policy inputs, so the scope of the study and its objectives can be understood more clearly. The public policies, programs and projects mder consideration are classified as follows: 1. Public Investment Programs a. Biological research b. Extension c. Tideland development e. Upland development f. Large-scale paddy irrigation Small-scale paddy irrigation Paddy land consolidation 1. Paddy land drainage j. Upland consolidation k. Upland irrigation 2. Public Policies a. Price policies i. Product price policy ii. Input price policy b. Credit Programs 1. Credit available to the farm sector ii. Interest rate policy There are many other public policies, programs and projects that affect agricultural development and rural development or welfare. lbsher pm. 25] , among others, advocates the integrated development Strategy. Transportation, marketing facilities, electrification and L — other infrasmictlme provide producer witl program for the agri niagrimlmral proc mt midered in this size, partly because 1 mt clearly lawn, and llgmmther KASS caipo leap-imltural input blpiblic programs can let us again sm basic idea by a means a llwltural Sector Sim .3181th presented : lit side of the figure In this system, it hm in the figure. lltleother for public s . . 3 Walls. As seen in 'r 57 other infrastructure would be equally important variables that can provide producer with incentives. The same thing is true for public programs for the agribusiness sectors, such as modern input production and agricultural processing. The public programs in these sectors are not considered in this study, partly to keep the model at a manageable size, partly because production functions of these infrastructures are not clearly known, and partly because there is a possibility of design- ing another KASS component to deal with some of these public programs. The agricultural input supply function is still assured horizontal, but public programs can influence the price level by various measures. let us again sumorize what this study intends to do and its basic idea by a means of flow chart. A modified version of Korean Agricultural Sector Simulation model after the addition of the com— ponent model presented in this study is shown in Figure 3.1. The right side of the figure is exactly the same as that shown in Figure 2.1. In this system, there are two decision boxes represented by a diamond in the figure. One is for allocation of public investment, and the other for public policies such as price subsidies and credit programs. As seen in the figure, it is believed that the public invest- ment induces technological change, referred to here as structural change. Note that structural change is not induced by change in the relative price, but is generated by the public investment. Also, the public investment does not directly affect production level or resource use, but affects them indirectly through a change in input quality or quantity, material or immaterial, or in incentive. land and water development, biological research and innovation diffusion represent ...H__ COWuUw HOHAH - gun Na . EMUNE WOUUwh muUfilflm ”wig Manon—“INN 7 HF-anug .- NWT a 58 .Hoooe Menomm Hogufioofluwé Gmmuox m o Sowmumg oofiwfiooe < magnum noun ofim mumuwoua .woflofloo 039E monocmo l. 0 .H .m mflawwm condos? U995 m®>fiH c538 mopm co .3on medom :33qu was cowumdsmom .m . A i .1 wfiugooné Howe: Homo: no.6 50.383 panama wouflowmm .— $35834 chofimz Hows; 4'1 Hmcowwmm u: so 5qu IL F... US 3ng 8.30m nohm . posses 98ch Houomm meHu—Am Uwvfiumz mmug Basswood anon Hill ooummmmm H8933 cowuom moan . — Mange in inputs mmumts are r lutimovation diffu investment and outer: Variom land ( developnmt, and can the farm resource all mats dismissed abo sector emponmt. m defammicm product admbchct supply an]: Note that publi mum-my can d 94191wa be influezo M Resom'ce use is a We. are resource 1 dmfmction are deter “are 3.1, since Mfume. Both “531m 3.1 supply Wmtput coefficien lJECtiVe function of 11m far, the “7'11 nmtion mat ' First of all, no 59 the change in inputs in quality or quantity in Figure 3.1. The first two components are represented as a function of public investment alone, but innovation diffusion is represented as a function of public iJNesunent and outcomes of biological research, among others. Various land categories are direct outcomes of land and water development, and can be fed directly into the resource constraint of the farm resource allocation component. Hereafter the three subcom- ponents discussed above will be referred to as the public irwestment sector component. There is another subcanponent, which will be called the farm micro production component. This includes the factor danand and product supply subcamponents of Figure 3.1. Note that public policies can now have a role in agricultural production-~they can directly affect resource use. Thus, product supply can be influenced by public policies that influence resource use. Resource use is also influenced by structural changes discussed above. Once resource use and structural changes that shift subproduc- tion function are determined, product supply can be computed as shown in Figure 3.1, since product supply is an exclusive consequence of resource use. Both subcomponents of factor demand and product supply in Figure 3.1 supply the farm resource allocation component with input-output coefficients, as well as the physical components of the objective function of the farm resource allocation component. Sega of the Study Thus far, the discussion has indicated what will be done, now We will mention what will not be dealt with in this study. First of all, no particular mention will be given to income — (Estdbutionand och u'quality ofrural fismdelnmmgeable dealswlth same of tl Projectims of mingmtil project variables, such as to iedegzeeofattaim sufiysfmfldbe extent mmallocatim cc alsobelinkedwith th himbletocapture d formulating policy: thimflatimum‘ limits comments at Mommas sim Mitis almost impos: Wiscotypeof Whopproduc' “Rims. Evenif Memo ‘wmyeouldbede 60 distribution and other measurements of economic and rural development or quality of rural life. These aspects are ruled out simply to keep the model manageable, and particularly because the KASS model already deals with same of the variables. Projections of product supply and factor damnd have little meaning until projections are used to estimate higher—order performance variables, such as total production, value added, etc. , for examining the degree of attainment of development goals. For this alone, this study should be extended to feed in projections made into the fann resource allocation carponent model. The combined projections should also be linked with the existing KASS simulation model. This is desirable to capture the dynamic interactions among components and for evaluating policy alternatives in terms of performance variables. The KASS simulation model is a product of multidisciplinary team work. Since its components are disaggregated, it seems that aggregation of componmts requires similar mutual incorporation. All this implies that it is almost impossible to put all components together right after one new component is developed, especially in view of the inflexible time constraint faced by the author. As mentioned earlier, one of the objectives in building the farm resource allocation component was to deal with farm mechanization. Farm machinery is one type of farm input. The machinery service demand for individual crop production is of exactly the same nature as demand for Other inputs. EVen if machinery investment could be determined by the Programing model, allocationof the service from the existing stock 0f machinery could be determined more logically and precisely by the imidalversion of t] ismimcomistenq thdelp: ofmdflneryservice Intum, this , nuima'yimitumier abstiume. merean htfamnedzanizatir [twidesapositiye am WW1: inasmcided no projq Wofindividlal htlrlaborsaved d; hmufortradit Mbmdmnizatimp 61 initial version of the model presmted in this study. However, there is some inconsistency of allocation of variables among components. The subcompormt model presented in this study does not include projection of machinery service demand for individual crops. In turn, this creates one more restriction. That is, the machinery input under consideration in the KASS model is purely labor mbstitute. There are many economic and agronomic studies indicating that farm mechanization has a yield-increasing effect, but no study provides a positive answer.1 In other words, labor demand cannot be determined independently of demand for machinery, and vice versa. Thus, it was decided to project the labor demand for the so-called traditional processes of individual crop prediction defined in the programm‘ng model, then the labor saved due to mechanization will be attracted from the labor demand for traditional process in order to project the labor demarxi for the mechanization process. Thus, a few equations written in FORTRAN computer language would have saved at least one-third of the activities defined in the programming model, which is a big advantage, especially as a omputer capacity is a restrictive factor. 0n the other hand, the KASS model contains five livestock, one fishery, and residual production activities, in addition to 12 crops. The programming model carponent deals with crop and livestock acti- vities, not with the last two activities. It would be more logical if the model presented in this study generated input-output coefficients h lFormer Director of Crop Experimmt Station, Dr. Young (1121 0mg. highlighted the doctrine of heavy fertilization and deep lelng for several years, but was unsuccessful. forallactivities d4 ofpolicyvariables. Immediately smdyhas already tur fl‘etimeallowed. 'l‘h shipbasbeen postpcn Wearenowreat toil. Wewilldevel mmeoclassicalecorm modes. 'Ihenainpu sometime: “fitya‘dqtmtity, 1 kisimvariables, an: Wotmdelwithm Wmdevelommtpo Mun-1976). Ac Wis for agricul 62 for all activities defined in the KASS model over time as functions of policy variables. Reference to the field of activities excluded here seems extremely scarce. Furthermore, the scope of the present study has already turned out to be considerably large in the light of the time allowed. Thus, the model for livestock input-output relation- ship has been postponed until a later date. Objectivg of the; Study We are now ready to specify the objectives of this study in detail. We will develop a systems simulation component model based on neoclassical economics, including development as well as growth theories. The main purposes of this model are: (l) to link public decisions with change in the agricultural resource base in terms of quality and quantity, (2) to relate these structural changes to farmer decision variables, and (3) to supply the farm resource allocation component model with necessary parameters subject to the dynamics of d'Ianging development policies, as seen earlier. All these subcampon— ents constitute the production side of the KASS model. As noticed earlier, this production side will be linked with the rest of tlre model. The crucial variables affecting the performance of the agricultural sector are public policies, programs and projects, represented by diamonds in Figure 3.1. The KASS team initially examined three sets of policy alternatives in terms of these policy variables. These sets are summarized in Table 3.1. Policy alternative set 1 corresponds to the third five—year Plan (1972-1976). According to Rossmiller, et al. [R.7], the major POIicy goals for agriculture include the following: Table 3.1. Sum: and I PolicyCom Research and guid land and water de Labor sinstitutes Food prices Imort policies Import policies Infrasmicum'e Fully plamfing pm Saree: Rossmille: 1. harassing t1 amiusis on a partiudarly ‘ 2. Increasing ' the farme 3. Inpmvmg the smears and is major policy ' W as follows: 63 Table 3.1. Sinner-y of Policy Canponents of Alternatives II and III Relative to Alternative I. Policy Cmponent Emphasis or Position Relative to Alternative I Alternative Alternative II III Research and guidance programs Mare Same Land and water development Same less labor substitutes As Needed As Needed Food prices Higher lower Import policies Same Open Import policies Same Open Infrastructure Nbre less Family planning program More More Source: Rossmiller, et a1. [R.7, p. 65]. 1. Increasing the production of agricultural products, with emphasis on attaining full self-sufficiency in food grains, particularly rice, by 1976. 2. Increasing incomes for farmers, with emphasis on narrowing the farm-nonfarm income gap. 3. Improving the quality of rural life, with alphasis on infra- structure and public service developmait. 'I‘ne major policy instrtments for attaining these policy goals are stated as follows: 1. Establishing an expanded agricultural production base. 2. Improving agricultural research and extension efforts. 3. Improving the market system. 4. Pheauraging the export of agricultural products. L. + The model pres« follwing policy qua 1. Met muld war the p] a; Alterna b. Alterna results c. Alterna- d. Alternai 2. What would 1 Policies be? 3. What would t amagem 4' WmWouldt for attainin 64 The model presented in this study will examine or deal with the following policy question more specifically: 1. What would be the impact on product supply and factor demand over the plarming horizon of 15 years (1971-1985) of: ac Alternative land and water development policy? b. Alternative biological research and diffusion of its results? c. Alternative product and input price policies? d. Alternative credit program? 2. What would the dynamic interaction of these individual policies be? 3. What would the impact of these policies, individually or in a package, on a set of performance variables be? 4. What would the optimal strategies of agricultural developumt for attaining developmental goals be? W PAKI‘ II WICAL STRUCI'URE OF I'DIEL PUBLIC] In this part, . mdelof public inve hflapter IV. Gap: mddiffusim for it: intreceives the out optochictim functic weddistinglfish hem Medal finctim. Finally, Chapter VII lewqmt variables < fimtim shifters. late than 350 x l. \ CHAPI'ER IV PUBLIC mVES'D’IENI--IAI\1D AND WATER DEVELOPMENT Introduction In this part, we present structural equations of the model. The model of public irwestments for land and water development is constructed in Chapter IV. Chapter V discusses models that use biological research and diffusion for its results. In Chapter VI, a production function that receives the output variables of the public investment subconponents as production fmction shifters is presented. The production functions used distinguish between annual crops and perennial crops. From this production function, a product supply projection model is derived. Finally, Chapter VII derives a factor demand equation that receives the output variables of the public investment subconponents as demand function shifters. More than 350 variables or parameters and about 300 equations or relationships are defined in this model. Presentation of every technical detail would confuse the reader and might obscure the essential feature of the model. For this reason, in addition to space limitations, only the basic essential structure will be given. Technically minded readers or those who are interested in technical details are urged to refer to the computer program written in FDRIRAN in Appendix A. 'Ihevariablenames appearinginthis sectionarethesameas 66 — dose in the m medifferent, howe in does not low t! more. On the c made available to us mooted in the RR analytical equation. lastly, it is .- mter pmgraln- '11 is recessary project (see balm). In addi mealso utilized: 1] MVP. Several fmc have appropriate. The center pn Wat submt i: othereseamh and e: 67 those in the computer program unless otherwise indicated. Subscripts are different, however. This method was designed to help the reader who does not low the FORTRAN language clearly mderstand the model structure. On the other hand, whenever a certain subroutine already made available to users is used, the corresponding call programming is presented in the FORTRAN language, along with the counterpart of an analytical equation. lastly, it is appropriate to describe the composition of the computer program. There are five subroutines constructed for making the necessary projections; PUBINV, SOCDLF, FDYLD, TEMP and IMPMFP (see below). In addition, two subroutines readily available to users are also utilized: DElDD, which is essentially a modified DELHI and HELLVF. Several functions are also used: TABILE, RANF and AMJD wherever appropriate. All this is written in FORTRAN. The computer program corresponding to the land and water devel- opment subcomponent is sham in subroutine PUBINV. That corresponding to the research and extension subcanponent is covered in subroutine SOCIDF. For technical reasons, the factor demand and product supply subccmponents are put together in subroutine FDYID. These three are the main subroutines essential for the model presented in this study. Powever, there are a mmber of variables that are endogenous in the total system of the KASS model at the present state of development or at least in the near future, but are treated as exogenous in this model. Examples are land areas used to produce each commodity, PTOduct prices and so on. At the same time, there are a mnber of Other variables used for more than one subccnponent, such as distributed — leprices, lone-m hemmed in stir meslmmin srbroul beexplainei in the Inthis diapt Btudevebpimt su inquantityas well . fomflat allows stn simlatedover time. Tue type of la dissunywas descri - typeofinvesflnmtha beefit—oost analysis ltismllhmn, how madysis cool. or analysis technim isthit for themdel \w— 1Inthis paper, impaled by ' tLP“Border to anal “Salvador. 68 lag prices, long-run profitability, etc. All variables of this sort are computed in subroutine TEMP. All necessary initial conditions are shown in subroutine DIPMFP. Other subroutines or functions will be explained in the text. In this chapter, we will construct a model of the land and water development subcanponent. The model first relates the change in quantity as well as quality of land with public investment, in a form that allows structural change in the resource base to be simulated over time. The type of land and water development under consideration in this study was described in Part I. The technique of analyzing this type of investment has traditionally been the ad E type of the benefit—cost analysis. There are a nmber of examples of this, approach. It is well krmn, however, that this approach is inadequate as a sector analysis tool. This author has demonstrated that the benfit— cost analysis technique can be much improved when a system simulation approach is incorporated [L. 111.1 In other words, net present worth, internal rate of returns and the benefit-cost ratio can be more realistically and accurately derived from the model presented in this study, though no attempt is made to do so here. In principle, this type of investment can be analyzed in a framemrk of programming model, as mentioned earlier. The trouble is that for the model to be more realistic and accurate, the matrix 1In this paper, the major components of the social benefit and cost are modeled by difference equations with exogeneous policy vari- ables in order to analyze the impact of establishing a credit union for El Salvador. size nust increase abject matter. 4.1 ILflJt) = i=1,2,3 “me Eik represent developtmt project : mach year by kth 1 M. 113C, which is c Several factors. figim time lag or Mich land eneters We defined as fol 69 size must increase. We seem to have digressed somehow from the main subject matter. At any rate, to seize precisely the indirect effects—— induced and stemmeduas well as the direct effects of an investment proj ect,2 the model should be able to mathematically trace out over time all important interacting variables. For this reason, a more comprehensive and consistent model is constructed for analyzing an Investment project. Let us discuss how to compute the accumulated land improved by means of public investments. That is: t 4.1 Tij(t) = TLik(°) + J; DSCi(t) dt i=l,2,3 k=l,2,. . .,8 th Where TLik represents accumulated improved land of k land and water development project in region i, and IBC:ik represents land area improved in each year by kth project in region i. land area improved in each year, DSC, which is often called the delay output, is determined by several factors. Implementation certainly involves time lag. With a given time lag or delay, DSC for each year is determined by the rate at which land eneters implarentation process. In a simple case DSC can be defined as follows: 4.2 DSCik(t) = Eik(t- T) =. 2Gittinger, in his book [G. l] , recommends that the secondary ! effects such as "induced" and "steamed" can be disregarded without a significant loss for analyzing an agricultural project. The result based on this methodological suggestion brings about a lower Priority of agrimflthral investment as contrasted to what the real Situation would be, resulting in a biased information for policy-makers. ihezeEikequals firprojectkand betwmstarting iject. dealingwithan placeimmyp hgmwmledng Thedelay isasmnedtobe ra smhas theErland: somtheprocesm mdeledbyadistrfl equation is: 70 Where Eik equals the rate at which land enters implementation process for project k and in region i, and T equals the necessary time lag between starting and completing project implementation. For an indi- vidual project, this formation would be all right. Since we are dealing with an aggregate model, one sort of project can be taking place in many places at the same time. This implies that the time lag on completing a project varies among different locations. The delay output rate such as DSC of the project implementation is assured to be randomly distributed with a specific density function such as the Erland family of probability density function. In other words, the process of the aggregate project implementation can be modeled by a distributed lag model. The appropriate differential equation is: D‘k degtg [D‘ k‘1 dk'lYgQ 4.3 -— + K — t ..... +K [‘3] %Q+Y(t) = X(t). Where: D = average expected time lag, K = order of differential equation, Y = output such as DSC discussed above, and X = input such as E discussed above. We parameters govern the shape of output distribution with a certain input signal: D and K. With a given average expected delay (D), the ShaPe of distribution is determined by K. When K = 1, the shape is exactly the same as emonential distribution. That is: ihich is a first- dwpe We infinity, the s isa discrete My. There are a 11.3 to met differ adPark M6), and muical method us tieCmplter librar: me three reasms ft mites the time f1 timof differential We average e396 directly computes la 71 4.4 D 9g? + Y(t) = no Which is a first—order differential equation. As k increases, the shape approaches that of the normal distribution, but when k approaches infinity, the standard error of distribution approaches zero, which is a discrete delay. Discrete delay is a special case of a distributed dealy. There are a variety of numerical solution methods for Equation 4.3 to meet different needs using this type of formulation [Manetsch and Park (M. 6), and Llenellyn (L.l9)']. The particular computerized numerical method used here is called DELLVF subroutine, developed by the Computer Library on Agricultural Systems Simulation [C.7]. There are three reasons for this selection. This subroutine automatically computes the time fraction necessary to secure stability of time solu- tion of differential equations, called 1131‘, deals with a case of time— varying average expected delay of the project implementation and directly computes land area under process of project implementation. The calling statement to this subroutine in FORTRAN is: 4.5 CALL DELLVF [E(I,K), DSC(I,K), RINT (l,I,K), STRGP, PLRP(I,K) DEL(I,K), DELDP(I,K), DI', KDEL(I,K)] Wnere : E - rate of land entering implementation process, DSC = rate of land leaving implementation process, RINI‘ = intermediate rate under processing for each stage of KDEL order, STRGP = sum of RINI‘ that is total amount being processed, PlRP = loss DEL = curn DEIDP = laggi m = time IKEL = order Specific pun: be eel—aimed later. Total land a- called storage, can equation: 4.6 STRGPikl he Cmputer program SOhltion Mind is: 72 PLRP = losses during processing, DEL = current time delay, DELDP = lagged time delay, DT = time augment for computation, and DKEL order of differential equation. Specific purpose for choosing this particular subroutine will be explained later. Total land area imder the project implementation which is often called storage, can be computed with the simple first—order differential equation: t .. 4.6 S'I'RGPik(t) = STRGPflJo) +L [Eik(t) - IBCik(t)] dt The computer program corresponding to this equation by Euler's numerical solution method is: 4.7 STRGPikgt+D = STRGij(t) + 131* [Eik(t) - IBCik(t)] However, we did not use this formulation here since the subroutine DELLVF computes this storage directly. But we did check the convergency of this storage by both computing methods (Equations 4.5 and 4.7). Once the area of improved land by each category of land and water development in each year, DSC, is determined, the change in each land class can be easily computed using the first-order differential equation. Before presenting the model, we specify several Imderlying assmptions: 1. Agricultural land can be classified based on a variety of criteria or purpose. The simple scheme adopted serves the purpose of this suidy. For each region: 73 A. Paddy land, TPTANDi(t) l. Permanently irrigated paddy, PITIi(t) Semi-permanently irrigated paddy, PITZi(t) Temporary irrigated paddy PITBi(t) Rainfed paddy, PIT4i(t) Consolidated paddy, CSLP.1(t) Drained paddy, DRDPi (t) B. Upland, TUIANDi(t) 1. Consolidated upland, CSULi(t) 2. Unconsolidated upland, UCSULi(t) sashes/ON 3. Irrigated upland, ULIGi(t) Each irrigation type of paddy could be classified by consol- idation type and further by drainage type. The number of paddy type would then be 16 which would make the model unnecessarily complicated. Unconsolidated and mdrained paddy are missing in this classification because they can be computed directly by subtracting improved ones. from total paddy, which is computed as the sum of the four irriga- tion types. Likewise, we need one type of mimproved upland to compute total upland; unconsolidated upland has been chosen arbitrarily. It is assumed that a certain amount of agricultural land in each region and year will be transferred to urban uses (high- way, industrial, urban residential sites, etc.) and termed T'Ri (t). It is also assumed that the fraction of each category of land classified above that transfers to urban uses is promorti are term WIR. wh ment for . Each lam indepemd the large perform: addition and drair large-sea ' The small only the ‘ Paddy 00m type of i i1'I‘igated Pmportiq - Still 8110: paddy imt< —f—_————— 74 proportional to the era of each category and these fraction are termed PLRi, i = 1,9. At the same time, the parameter WTRi, where i = 1,9, is designated to control land retire— ment for each category if necessary. 3. Each land and water development project can be implemented independently. To simplify the model, it is assumed that the large-scale irrigation project would be multipurpose, performing land consolidation and drainage, if desired, in addition to irrigation. It is also assumed that consolidation and drainage are proportional to the unimproved land in a large—scale irrigation project. 4. The small-scale irrigation project defined here augments only the area under semi—perfectly irrigated paddy. 5. Paddy consolidation or drainage can be performed on any type of irrigated paddy. The proportion of each type of irrigated paddy consolidated or drained is assumed to be proportional to area of each type of irrigated paddy land. 6. Still another type of irrigation that transforms the rainfed paddy into the temporary irrigated paddy is not considered in this study, since this may not require any form of public investment. 7. For irrigation and consolidation, a certain fraction of land is required for inserting some structure such as an irrigation ditch, path, etc. That is, area available for cultivation is reduced due to these land improvements. This fraction is represented by PADSCi. associati. —:———, 75 8. Tideland development augments paddy that is perfectly irrigated, consolidated and drained, and upland development auguments only unirrigated and unconsolidated upland. With these assumptions, the time path of each land class defined above can be described by means of the first-order differential equation. That is, remembering that k index goes 1 to 8 and corresponds to: r mien; Tideland development Upland development large-scale irrigation project Small—scale irrigation project Paddy consolidation project Paddy drainage project Upland consolidation project CDNO‘WJ—‘WNH Upland irrigation project 1. Permanently irrigated paddy, namely the paddy in "irrigation associations" as reported in the official publications, PITli(t): t 4.6 PITli(t) = PITli(o) + ’0 [(DSC13(t) * (1.0 - PADSCl) + DSCil(t) — PADSCB * PITPli(t) * DSCiS(t) - WTRl * PLRli(t) * TRi(t)] dt 2. Semi-permmently irrigated paddy, which is called irrigated paddy in the official publications, PIT21(t): 4.7 PIT21(t) = 13min») + I: [(msciam * (1.0 - PAmscz) — PTZli * DSCi3(t) - PADSC3 * PITP2i(t) * DSC15(t) - WI'RZ * PLRZi(t) * TRi(t)] dt 3. Tamera: 4.8 PH 4. Rain-fed 4.9 PITA 5- (bnsolida 4.10 cm 6' Draimed p, 4-11 DRD] 7‘ COUsolidat 4.12 (SUI. 8' UnconSolid 4.13 [Jon] —:————” 76 3. Temporary irrigated paddy, PIT3i(t): _ t 4.8 PIT3i(t) - P1T3i(o) - f0 [P1311 * IBCi3(t) + I’l‘32i * IBCi4(t)+PADSC3 * PITP3i(t) * IBCiS(t) + W * HRBi(t) * TRi(t)] dt 4. Rain-fed paddy, PIT4i(t) _ t 4.9 PIT4i(t) — PI'I‘4i(o) - ’b [PT41 * DSCiB(t) + P1?42i 7" DSCi4(t) + PADSC3 * PITP4i(t) * IBCiS(t) + m4 * Pl.R4i(t) * Talon] dt 5. Consolidated paddy: CSIPi(t): 4.10 cosmic) = (3121(0) + 10‘: [Dscfim * (1.0 - PADSC3) + msci3(t) * (1.0 - PADSCl) =1: (1.0 - min» + 0501105) - WI'RS * PLRSi(t) * T&i(t)] dt 6. Drained paddy, DRDPi(t): 4.11 DRDPi(t) = DRDPi(o) + [0‘3 [lBCi6(t) + DSCi3(t) * (1.0 - RDPi(t) * (1.0 - PADSCl) + Dscfl(t) - wrms * 113610;) * mica] dt 7. Consolidated upland, CSULi(t): —- t - 4.12 CSULi(t) — CSULi(o) + f0 [DSCi7(t) * (1.0 PADSC4) - WI'R7 * PLR7i(t) * TRi(t)] dt 8. Unconsolidated upland, UCSULi(t): 4.13 ucsuria) = ucsurim) rot [DSCi7(t) - Dscizu) +er8 * mice) * mice] dt 9. Irrigat‘ 4.14 U] 77 9. Irrigated upland, ULIGi(t) 4.14 music) = music) + rot [IBCis(t) * (1.0 — PADSCS) - er9 * mi * mien dt Where: PAIBCj, j = 1,5 = land losses due to project, WI‘R. j = 1,9 = control variable to restrict the conversion J of a certain type of land into urban uses, TRi(t) = total land transfer to urban uses in each region which is an exogenous variable to the model, HSCik(t) = the rate of land implemented in each year for each of land and water development projects, as we defined before. Other variables are defined as follows. The purpose or reason for having these variables are explained briefly above. PIT‘Pij (t) is the proportion of each irrigation type paddy to total paddy, that is: 4.15 PITPij (t) = PITij (t) /TPLANDi(t) Where PITij (t) , j = 1,4, is paddy area by irrigation type defined above, and TPIANDi(t) is total paddy in each region. PLRij (t), j = 1,9, is proportion of each land category defined in Equations 4.6 — 4.14 to total in each region, and total land TIAND, is defined as the sum of Paddy and upland in each region. In principle, the less perfectly irrigated paddy can be trans- formed into any type of more perfectly irrigated paddy. For example, it is not necessary for the temporary irrigated paddy to be transformed into a semi—perfectly irrigated paddy first in order to be transformed into a perfectly irrigated paddy. A large-scale irrigation project usually transforms some of each type of less perfectly irrigated paddy into perfectly ir and P142i specify gated paddies tr: emple, P132i is irrigation we 2, 4.16 H21 4.17 n32 In the above filtering the ham given. This is th. Plblic bud89t to w mi”Mable is t} by ”*0 factors: tc N lflit Cost Of ea Let Us diSCUS 78 into perfectly irrigated paddy. Parameters, PI'Zli, PI‘3liPl‘32i, PI'4li and 1’1'42i specify the proportion of each of the less perfectly irri— gathed paddies transformed into a more perfectly irrigated paddy. For emmple, F'I'32i is the proportion of irrigation type 3 transformed into irrigation type 2, so that the following relationships hold: 4.16 171‘21i + PI'31i + 171‘41i = 1.0 4.17 P'I‘32i + P'I'42i = 1.0 In the above discussion, we did not specify the rate of land entering the improvement process, Eik(t)’ and assumed this variable was given. This is the variable government has the power to allocate the public budget to various segments of policies, projects and programs. 'Ihis variable is the true policy variable and is exclusively determined by two factors: total budget allocated to each project in each region and unit cost of each project. let us discuss the unit costs necessary to implement each project. Can we assume without loss that this unit cost is constant over time, no matter how much land is transformed? There are reasons that unit cost would be an increasing function of time as low-cost projects will be implemented first. It is assumed here that the unit cost curve is an increasing function of total land improvement for each land and water development project, as shown in Figure 4.1. Here the independent variable could be either improved accmulated area or accumulated area entering the improvement process. For the purpose of computing the required budget, we concluded this independent variable to be a simple average of unit costs computed both ways. That is, the unit cost for each project, APE 4.18 APSC]. Utli 1; Cost: (F‘unction Value) ll Figure 4.1. APSI. = . 1k 11111 lam APSZik = unit 79 each project, APSCik(t) is now: 4.18 APscikm = [APSlik(t) +APSZik(t)]/2 8 :> L) g .5 J.) E Accumulated land improved or entering improvement process Figure 4. 1. Relationship between unit cost and accumulated land improved or entering improvement process for each project. Where: APSlik = unit cost computed from Figure 4.1, based on accumlated land entering the improvement process, APSZik = unit cost computed from Figure 4.1 based on accumulated land inproved. The accumulated land entering the improvement process can be computed: t 4.19 Sik(t) = Sik(°) + 1;) Eik(t) dt Where: S ik(t) = accumulated land entering improvement process, %m=m There is an function of time. decline over time rate would mama; each other. Hem rate of APPA. m S'muld be read; 4.20 APscik The exact 11a Matted land is no 1311th is welI TABLE function is FDMN is: 4'22 “’32 (1 119m: VAL“ (1.x) . 31M 80 Eik(t) = the rate land entering improvement process in each year. There is another reason that unit cost would be an increasing function of time. Costs related to the use of heavy equipment may decline over time. On the other hand, it is quite certain that the wage rate would increase more rapidly, so the two factors may not compensate each other. Hence, total unit costs are assumed to be increased by a rate of APPA. Thus, Equation 4.18 actually used unit cost function should be read: * 4.20 APSCik(t) = [Ammo +APSZik(t)]/2.0 * eAPPA t The exact mathematical relationship between unit cost and accu- mulated land is not pursued here since interpolation or extrapolation by computer is well developed [See (L. 19)]. For present purposes, TABLIE function is chosen, and calling statement to this function in FORTRAN is: 4.21 APSl (1,K) = TABLIE [VALCI‘ (l,K), SMAIL, D1 (1,10, KCOST ('I,K), S(I,K)] 4.22 APSZ (1,10 = TABLIE [VALCI‘ (1,10, SMALL, D1 (1,10, KGOST (1,10, TL (I,K)] Where: VALCI.‘ (1,10 = an array of unit costs or function values derived from Figure 4.1 corresponding to each segment of independent variable, SMALL . = the smallest value of independent variable defined, the origin value, Dl (LK) = interval of independent variable augument, KGBT (1,10 S (LR) and We distingu desired investmem or actual budget, 5311’ gestation peg his, in turn, reg At any rate, rate of land enter Mded by Wblic rate of land enter 4'23 Eik(t) lhere APSC is Wen tatim budgEt for E 4'24 TCPDik( I“here: mm = nor m& = tot Equ A“ iIIelicit a WW1)? dish-i1), Wetion of land I hemmm my emShifm‘g the ——7————_——’ 1 81 KCDST (I,K) = the number of segments of independent variable divided, S (I,K) and TL (I,K) are Sik(t) and 'ILik(t) in FORTRAN state— ment, respectively. We distinguish three types of public budget: intended invesUnent, desired investment or implementation budget, and realized investment or actual budget. Project completion is often delayed beyond a neces- sary gestation period because the desired investment is not realized. This, in turn, requires more investment. At any rate, the intended investment is assumed to determine the rate of land entering improvement processes in each year, is exogenously decided by public resource administrators, and termed here Bik (t). The rate of land entering project implementation, Eik(t) is: 4.23 Eik(t) = Bik(t)/APSCik(t) Where APSC is average unit cost. Required investment or the implemen- tation budget for each year, TCPDik(t) , can be computed as follows: 4.24 'I‘CPDik(t) = [ASPCik(t)/DELPPik] * STRGPik(t) Where: DELPPik = normal average expected time delay to implement project, S'I'RGPik = total land under implementation process computed in Equation 4.6. An implicit assumption is made here that budget requirement is uniformally distributed over the time period between initiation and completion of land improvement. The main purpose of the public investment is to play an important role in shifting the production function among subfunctions, hence, product supply 31 defined in severe Because of equation are defi inEqmtims 4.6 - 1. Rate of 4.25 St 2. Rate of 4.26 SC 3. Rate of 4.27 SC 4. Rate of: 4.28 SC] "" F 82 product supply and factor demand functions. This shifting can be defined in several ways, depending on the production function defined. Because of the way prediction flmction and derived projection equation are defined in Chapters VI and VII, we transform the variables in Equations 4.6 - 4.14 into the rate of change in various land classes. 1. Rate of change in perfectly irrigated paddy: 4.25 SCRil(t) = [PITPli(t) - PITPli(t—l)]/PI'I'Pli(t—l) 2. Rate of change in semi-perfectly irrigated paddy: 4.26 SCRi2(t) = [PI’IPZi(t) - PITP2i(t—l)]/PITP2i(t-1) 3. Rate of change in temporary irrigated paddy: 4.27 SCRi3(t) = [PITP3i(t) - PITP3i(t-l)]/PITP3i(t—l) 4. - Rate of change in consolidated paddy: 4.28 SCRi4(t) = [RCPi(t) - RCPi(t-l)]/RCPi(t-l) 5. Rate of change in drained paddy: 4.29 SCRiS(t) = [RDPi(t) - RDPi(t-1)]/RDPi(t-l) 6. Rate of change in consolidated upland: 4.30 SCRi6(t) = RCUi(t) — RCUi(t—l) 7. Rate of change in irrigated upland: 4.31 SCRi7(t) = RIUi(t) - RIUi(t-l) Note that 1: dated, undrained not been transfor me mt supposed defined in Equati‘ the 13nd and wate: Also note t1 143“ the previous plane Consolidatj 1970, ,0 initial \ Motion fUImtic We of Cobb~ iIIduped land to t of the Zero Value Where : PI'I'Pi j,(t) j = 1, 3= proportion of the first three types of irrigation to total paddy in each region, respectively RCPi (t) and RDPi (t) = proportion of consolidated and drained paddy to total paddy, respectively RCUi (t) and RIUi (t) = proportion of consolidated and irrigated upland to total upland, respectively. Note that paddy irrigation type 4, which is rain—fed, unconsoli- dated, undrained paddy, and unconsolidated, mirrigated upland have not been transformed into the rate of change because these variables are not supposed to shift the production function. These variables, defined in Equations 4.25 - 4.31, are termed as structural changes in the land and water development subcomponent. Also note that the last two equations are computed differe1t1y than the previous ones for technical reasons. These two projects, upland consolidation and irrigation, are assumed nonexistemt prior to 1970, so initial values of RCU and RIU are zero. On the other hand, the production function adapted in this study, as presented in Chapter V1, is a type of Cobb-Douglas production function, having the ratios of improved land to total land as independent variables. The logarithm of the zero value is not defined. This problem can be solved in some Way. The truly difficult problem is that the rate of change in ratios for this particular case turns out very large, such as 100 or 200 percent, at the beginning of the planning horizon. Then the production response is overestimated with a constant production elasticity. Thus, for these two variables, and variables of SLD Rfl and SLDR.12 changes in the re hep: wills 84 the ratios of developed new land to appropriate total land discussed in the preceding page, the time—varying elasticities are applied as we will see in Chapter VI. Among land and water development projects, tideland development (k = l) and upland development (k = 2) are designated to augument paddy and upland, respectively. Can the productivity of this new land he assumed to be constant or the same as that of old, existing land without losing generality? It is quite certain that, while other land and water development projects act to shift the production function among subfunctions upward and rightward, these two projects in fact act to shift the production function among subfunctions downward and leftward. This is because the productivity of new land is generally low, so the more new land a region has in cultivation, the more low average pro- ductivity will be realized in the region. To more accurately deal with productivity growth over time, we assume that new land productivity will grow as follows: the first year's productivity will be 30 percent of existing land productivity; second-year productivity will be 35 perce’lt; 45 in the third—year productivity; fourth-year productivity will be 60 percent; fifth-year 80 percent; and sixth-year 100 percent. This productivity growth rate is termed WGj. To compute a weighted average productivity of the new land, we first compute total land developed during the past five years, including the current year. For example, for tideland development: 4 . 4.32 smim = .E DSCil(t-J) J-O due hem We ale relati: dflltax Whom 85 Then we compute the sum of weighted productivity as follows: 4 4.33 weenie) = z wcj *Dscfloz-J) j=o Where: WP'I'LDi (t) = sume of weighted productivity of tideland, WGJ. = weight given to land developed in each year. Finally, the weighted average productivity of new land, WAPi1 (t) , is computed: 4.34 WAPfl(t) = Wl‘LDi(t)/SU1‘YRi(t) As we computed structural change variables in Equations 4.25 — 4.31, we also transform total new land into the rate of change. First, the relative quantity of new land to total paddy for tideland and develop- ment and total upland for upland development is computed, respectively: 4.35 RI'Di(t) = SDIYRi(t)/TPLANDi(t) 4.36 RUDi(t) = SDUYRi(t)/TUIANDi(t) The rate of change of new land is 4.37 SLDRil(t) = [RTDi(t) — RI'Di(t—l)] 4.38 SlDRiZ(t) = [RUDi(t) - RIJD(t-l)] Thus, it is equivalent to saying that region-wide average productiVity will decline if new land is added by: 4.39 PRNLfl(t) = WAPil * sti1 (t) Acro them to 0th 86 4.40 PRNLiZ(t) = WAPi2 * SLDRi2(t) A crop-specific model will appear in Chapters VI and VII. In summary, this model component, together with the model in the next chapter, has been designed for product supply and factor demand projection. Therefore, all output variables for the model components described in this chapter can be said to be intermediate output variables. The variables that will be transferred directly to other subroutines are: TlANDi(t) = total agricultural land SCRik(t) = rate of change in the proportion of each kind of improved land to total paddy or upland, respectively APSCik(t) = average project unit costs land area improved in each year DSCik(t) WAPil(t) = weighted average productivity of new land SLDRil(t) = rate of change in relative area of new land As seen above, the major output variables of the public subsector CGDponent modeled in this chapter are: (1) total agricultural land, paddy, upland or a combination, and (2) accumulated land improved in terms of irrigation, consolidation and drainage. In other words, the model componeit described in this chapter is capable of simulating behavior of these variables over time, based on alternative public policies, especially in terms of public investments in various land and water development projects, and in terms of alternative land use DOIicy in relation to agricultm'al land disappearance. This model Component can also simulate consequences of alternative patterns in allocating actual investment in relation to desired investment as will be discussed in Chapter VIII. | {mega-.7 87 In summary, this model component is relatively simple in terms of model structure and data requirements. Fewer restrictive assump- ticns are required. There are four major parameters in this public subsector model that will probably affect the behavior of the output variables: Unit costs, average expected time required to complete each project, fraction of land required to insert sane land-improve— ment structure, and land disappearance due to urbanization. These variables or parameters (except the last one) are technical ones, so the engineer can provide additional information on the related data in nature. In short, data improvement and support for collecting data are needed for model improvement. CHAPTER V PUBLIC INVES'D’JENI‘uBIOlDGICAL RESEARCH AND EXTENSION This is the second chapter on the public investment sector. first, we will develop a primitive model of the biological research subsector, consisting primarily of a table of possible research outcomes from indicated research investments. Much more attention will be given to the third section, where we construct a social diffusion model of research outcomes. This mathematical social diffusion model is based on some useful decision—making theories and earlier work done by the systems simulation team at Michigan State University. These will be reviaved in the second section. This subsector model is essentially independent of the land and water development subsector model with some minor exceptions, although both subsectors compete for public investment funds. Interaction with output variables from both model canponents and subsectors will be presented in the next two chapters. Biological Research Subsector It seems that there exists a functional relationship between research outcome and research investment. However, it also seems that research productivity is not well known. Furthermore, research outcome seems to have a large probability or confidence range. The research outcome appears to involve high uncertainty or risk. The PrOductivity and confidence range for its probability distribution seems t0 depend on many other things, such as accumulated knowledge, 88 89 coordination among specialized disciplines, development of knowlege in other continents or comtries, etc. What is emphasized here is that the research outcome is not only a function of the current domestic public investment, but also a function of past investment, domestic or abroad. Evenson [E. 3] formulates a production function as a function of these two types of investment in addition to other variables. Hayami and Ruttan [H.9, part 4] discuss the extent to which biological tech- nology can be transferred among countries. I‘lbseman [M. 21] discusses building biological research systems. At the same time, Fishel [F.6] presents several articles by different authors dealing with econanics of biological research. Despite much work on the economics of biological research, the common conclusion seems to indicate that social returns to public investment in research are high. The impression is that they have studied only successful cases. The analytical framework for coming to this conclusion seems to have been primarily the cost-benefit analysis technique. Examples are found in Griliches [G.8]and Peterson [P.S] and Schultz [8.4] summarizing this part of the study, done mainly by Chicago school people. We do not follow this analytical framework. The reason is very simple: first of all, we intend to model how economic variables dynamically interact with each other to capture all possible direct and indirect effects of research activity. Secondly, we intend to project yields rather than examine the internal rate of return. Biological research in the Korean agricultural setting can be Classified into three basic categories: 90 1. Basic and environmental research 2. Breeding 3 . Culture All this research aims at a set of canton goals. These goals can be put into two groups: desirable or ”good“ outputs, and undesirable or "bad" outputs. l . Desirable outputs: a. b. e. f. Yield increase Uncertainty reduction Quality improvement Early maturity Savings in production factor requirements Others 2. Undesirable outputs: a. b. c. d. e. Increase in material costs Increase in labor requirements Increase in uncertainty Increase in credit needs Others Various combinations of desirable and undesirable outputs can be brought about by research activities. What would the shape of production function of public investment be in terms of the research outcomes listed above? What would the productivity coefficient of public investment be? One way to estimate this productivity coefficient is regression analysis, using time series data where yield or other variables are 9l dependent variables and public investment or expeiditure is an inde- pendent variable. The independent variable may be lagged or accumulated public investment in agricultural research. There are several diffi- culties with this approach. First, past experience is not always repeated in the future. Second, the production function shifter, scientific findings abroad and training are important. Training abroad is not necessarily financed by the Korean government. All this implies that at least some of the public investment in other countries must be counted as independent variables in the production function. A further complication is that agricultural research is a sort of joint product enterprise. A more systematic modeling of agricultural research systems might well become a good topic of another Ph.D. dissertation. To make the model presented in this study manageable, we adopt a pragmatic approach. That is, the first assumption is that the research outcome is a sort of package having certain combined levels of attributes in terms of the outcomes listed above. The second assumption, which is more cmcial is that a planned agricultural research outcome would be realized. A hypothetical set of planned research outcomes for each crop or crop group appears in Table 5. l. The figures indicate the average productivity gain. These figures do not directly represent the pro- ductivity gain at the experiment station level. Suppose the produce tivity gain of a research outcome is 30 percent above that prevailing at the time. Also, suppose that this particular technology can be disseminated to 50 percent of the area in the region with a gain in no.0 nN.o nN.o on.o nN.0 nn.0 on.0 0¢.0 nN.0 nm.0 o¢.0 nm.o on.o kuoy nwma ewoa nH.0 no.0 no.0 nH.o no.0 0H.o 0H.o mwma 0H.o 0H.0 0H.0 nH.o Nme 0H.o nH.o meH 0H.o no.0 no.0 nH.o no.0 nH.o na.o omoa no.0 mnma nH.o 0H.0 no.0 wmma nH.o no.0 no.0 nH.0 no.0 nH.o nH.o nH.o mnma m“ 0H.0 nH.0 nH.o nH.0 nnma nnma no.0 no.0 no.0 nH.o no.0 0H.0 no.0 0H.0 no.0 quad mnma Numa no.0 no.0 0H.0 no.0 no.0 no.0 no.0 no.0 no.0 no.0 no.0 0H.o HsmH mono , Hench oHnmu canoe mmmno Immqu xawm owmuom ooomooe cumuom lowo> momasm undue umnuo umonz noanmm moflm “mow ammo 8 enhancement/E 380 338m mamas 33 mo Eafiocoum .3 Suwanee. commum madden mm 3me as 386$ no 8mm no meme 5 333m nonmemmm 83an Emma Wodehouse in Emma L 93 productivity. Then the average regional productivity gain would be 15 percent (0.3 x 0.5 = 0.15). In other words, the figures in the table are interpreted as computed by multiplying productivity gain at experi- ment station by proportion of land where results could advantageously be used. The research outcome for a crop can come about more than once during the planning horizon (1971-1985) . In the computer program in Appendix A, RYINCRijk stands for the productivity increase at the experiment station, RYDISSijk for dis- seminable area in proportion, and RYDIFFijk for the average produc- tivity gain, where i = l, 3 for regions, j = 1, 13 for crops and k = 1, 5 maximum for the number of research outcomes. The variable I]3EXYRJ.___.Ik stands for the year in which kth research outcome for jth crop in ith region is materialized and ready to be disseminated. As pointed out earlier, agricultural research is a highly risky enterprise. In other words, it is highly uncertain as to when a dis- seminable research outcome will take place at what level, with what attributes and with what effects on total accumulated productivity gain during the planning horizon. As Johnson, et al. [J .15] correctly point out, there are numerous possible decision-making rules for a highly risky enterprise. We can hypothesize the consequences of alter- native courses of action or assmiptions. That is, the disseminable research outcomes with certain levels of productivity gains at given POints of time postulated in Table 5.1 will be treated as a starting Point. What will happen if actual research outcomes are different from this situation in terms of tinting, productivity gain at each Point of time when the research outcome is materialized, or accmlulated 94 productivity gains achieved during the planning horizon? We will come back to this issue in Part III, whei we discuss sensitivity analysis and policy experiments . rIheory of Innovation Diffusion The research outcomes defined in Table 5.1 are interpreted as at emeriment stations, not at farms. They must be communicated to individual farms through diffusion channels, such as the extension service. Diffusion of innovation does not take place entirely spontaneously. Hence, we need to model the media that channel infor— mation about this innovation to materialize the potential productivity gain. Diffusion of an innovation does not take place instantaneously. The condition of the experiment station in terms of agricultural resource base is likely above that of the average individual farm. These are some reasons why the actual productivity gain at individual farm levels may be considerably less than that at experiment stations. With a given research outcome having given attributes, the rate of diffusion and hence, actual productivity gain, will depend on the magnitude of the stimulant if other conditions remain unchanged. In the next section, we will present a social diffusion model to describe the interaction among research outcomes , stimulant, diffusion rate and change in actual productivity gain at the fam level. First, however, we review some useful theories on decision-meking and diffusion, in economics as well as sociology. According to Johnson [J .10], a problem-solving—oriented decision-making process can be Shown in a diagram as in Figure 5.1. There are six steps or sub processes involved in making a decision. This diagram shows that “WW. .UNM \WN ..NWN» .Mm\\% 95 Normative Problem Non-normative Conce ts of H . _ . Good :nd Ba Def1n1t1on ‘ ’ Concepts as to right goals and acts i " l \\ v \ Observation Analysis De cision-making Action I R esponsibility Bearing Figure 5.1. Six steps in a problem—solving process. (Source: Adapted from A Study of Managerial Proce_sses of Midwestern Farmers, Johnson, G. L., Halter, A. H., Jensen, H. R., Thomas, D. W. Iowa State University Press, Ames, Iowa, 1961. See also "The Role of the University in Economic Development," J. S. MzLean Visiting Professor Lecture, Department of Agricultural Economics, University of Guelph, Publication No. AE 70/2, March 23, 1970.) ’1‘!“ 96 problem-solving—oriented decision—making is an iterative process, each subprocess interacting with normative as well as non-normative concepts or information, and there is interaction among the subprocesses. As Bradford and Johnson [B.l4, p. 15] point out, ". . .stability in farming is abnomel--change and the need to study and adjust to change are normal. In farming, as elsewhere, partial ignorance is universal; the need to learn and adjust is the main problem of farm management. " In other words, the need for manath and decision-making would largely disappear [V.3, p. 12—13] in a static world where there is re change and resulting uncertainty. In a problem-solving—oriented decision-making process, the obser- vation phase plays an important role with a given problem definition. In fact, the extension institution is largely designed to help farmers gather the necessary information. The extension worker also plays an important role in helping farmers formulate value of problem definitions [J . 7]. The knowledge accumulated by observing available information is critical in making decisions for adopting an innovation or a new tech- nology. Johnson and his associates [J .13] classify the knowledge situation as follows: 1. Certainty; positive as well as negatiw-z Inactive situation learning situation; voluntary as well as involuntary Forced action situation; positive as well as negative .U‘ZI-‘LON Subjective risk situation; positive as well as negative. 97 Inactive, learning and forced action situations together are often called uncertainty. That is, they are classified as modified version of Knight's knowledge classification of certainty, uncertainty and risk. Whether a farmer adopts an innovation or not is exclusively determined by the knowledge he gathers, other things being equal. This knowlege situation can, of course, be altered by observing available information. Johnson and his associates [J .13] discuss the type of information source in detail. The idea of communicative and rxoncommmmicative sources is directly adopted in formulating the diffusion model preseited in this chapter. Before making a firm decisim, the message receiver needs to analyze the available information. In this analysis phase, it seems that the economist tends to emphasize the economic variable exclusively as the subject matter of analysis, whereas the sociologist claims that sociological factors are more or at least equally important, depending on Specific situations. Griliches [6.6] and Schultz [8.2, p. 164] claim, respectively, that profitability or an economic variable is a strong explanatory variable or major determinant for adopting a nev technology or irmovation. Schultz adds, “it is not necessary to appeal to differences in personality, education, and social environmen ." The same sort of idea is expressed by Griliches: ". . .in the long run, and cross—sectionally, [sociological] variables teid to cancel thetselves out." Both scholars are criticized by a sociologist, Rogers [R. 2, p. 143—144] , for their extrere position. At the same time, Mellor [M. ] emphasizes uncertainty or risk consideration as being efinally important in the adoption behavior of peasants. Some examples 98 of sociological variables considered important in explaining adoption or rejection of a 116/7 idea, innovation or technology are found in Bowde'i [B.lZ], Pbulik and Lokhande [M26], Feaster [F.2], Fliegel, et al. [F.8], Chattapadhyay and Pareek [C.2], Havens [H.6], etc. The sociologist does not completely exclude economic variables in his model, however. A good example is found in Fliegel, et al. [F.8]. Johnson, et al. [J .13] identify different kinds of information used by fanner decision—making. Rogers [R.S, p. 137-160] discusses the type of perceived inno— vation attributes that affect the rate of adoption. The main points he makes can be summarized as follows: 1. Relative advantage, in terms of monetary as well as non— monetary matters. 2. Compatibility, in terms of values and needs, as well as idea or technology previously introduced. 3. Complexity 4. Triability 5. Observability. As far as seed technology for a major crap is concerned, it is not hard to believe that the new technology would be disseminated rather rapidly, since all criteria advanced by Rogers are likely to be ful- filled. In other words, economic variables might be said to be the main variables affecting diffusion rate in the long run in this instance. In this sense, Griliches and Schultz are right since both are primarily Concerned with seed technology, although there could be an emeptional Case for a subsistence crop [see, Rogers (R.5, p. 142—l49]. 99 On the other hand, Rogers [R.S, p. 183-185] presents adoption cateogries as follows: 1. Innovators (first 2.5 percent of adopters) Early adopters (next 13. 5 percent of adopters) Early majority (next 34 percent of adopters) late majority (next 34 percent of adopters) assets Laggards (last 16 percent of adopters) This characterizes the shape of adoption rate distribution curve. This curve, based on the figures given above, is a bell-shaped normal distribution, and the emulated frequency distribution is S-shaped. In reality, however, it seems that the level of the perceived attributes of an innovation determines the specific shape of the adopter distribution curve. That is, the magnitude of the perceived attributes seems to contribute greatly to the determination of the mean, standard deviation and skewness of the distribution, or parameters of Erlang family of probability distribution, K and D in Equation 4.3. Nevertheless, Rogers [R.S, p. 179] concludes that "It has generally been found that adopter distributions follow a bell-shaped curve over time and approach normality." Is this conclusion true regardless of changes in aspiration, value system, level of perceived attribute of an innovation, degree of fulfillment of needs, level of stimulant, etc.? Is there an alternative form of adopter distribution, adapted to a more specific environmental condition? It would seem that a bell—shaped normality curve prevails in a case where diffusion takes place more or less spontaneously, there are many alternative means to satisfy a need, the society is more or less in a stationary situation, and there is not finch stimulant in adoption, even if the perceived attribute is real. 100 As a matter of fact, Perry and his associates [R4] in a diffu- sion research report say that "A generally accepted belief is that adoption of new farming practices follows as S or growth curve," and concludes by saying that, "Perhaps in American society, at least, inno- vation is rapidly becoming the norm and the diffusion curve will soon more nearly approximate a J—curve than an S—curve.’ What is being said here is that the diffusion curve can be a J shape if certain con— ditions are met even in a less developed country; productivity gain is high with little uncertainty, the relative price is sufficiently in favor of adoption, an adequate amount of information is supplied, and so on. This argument does not deny the usefulness of adopter categoriza- tion as advanced by Rogers. One may find some innovators and venture- some or progressive farmers in any society. The progressive farmer adopts innovations first. Once he is successful, the diffusion process speeds up rapidly. Rogers [R.S, p. 185—187] generalizes socioeconomic characteristics of the innovator. Nevertheless, there is a report [Malone (M. 1)] that finds no difference in adoption rate of a package program in India between social status classes. As implied in Figure 5.1, decision-making in adopting a new tech- nology involves an iterative process. According to Campbell [0.1] , the traditional model of the individualadoption process currently used is a five-stage model: (1) awareness, (2) interest, (3) evaluation, (4) trial, and (5) adoption. After criticizing this traditioml model, he advances an alternative model called the adoption tree. This model is illustrated in Figure 5.2. lOl Trial: Rejection ——-—* Reevaluation Adoption Trial jection —-+Reevaluation -——v Rejection Evaluation Figure 5.2. A model of individual adoption process-—The Adoption Tree. According to this model, the analysis or evaluation leads to two alternative decisions: one to trial and the other to rejection. The trial also leads to alternative decisions: to adoption which means the perceived attribute of the innovation proved advantageously or the inro— vation fits the individual farm situation or practice, or rejection, which implies that the innovation is not advantageous to the specific farm situation, practice or both. As far as an innovation turns out profitable to at least some farmers, the re-evaluation process will be repeated several times until the knowledge situation becomes certain, while adjusting individual farm practices. If an innovation is profit— able to some farmers and there is a need that can be satisfied by adopting it, the innovation eventually will be adopted by the wlole population, regardless of the price level. A good example is the case of hybrid corn in the United States. The trial stage can be viewed as an adoption in a broad sense. Sane find the innovation viable, but not all. Rejection after trial may be called dropout. This dropout is not a consequence of a decrease in the relative price of the crop. Kislev, et al. [K.5] might call this drop out a process of an innovation cycle. They define an innovation as either a new product or production method that appreciably affects the supply of an existing 102 product. As an innovation is adopted, the industry supply increases, by definition. Then the price level will drop, so the first adopters will be driven out of the production of the new product or use of the no» method according to Kislev, et al. They call this process an "innovation cycle." The proposition that what is good for the individual fimr is also good for the industry does not often hold for a competitive industry like agriculture. As Cochrane [G. , p. 949] points out, despite the industry's depression, the individual farm continues to adopt the new technology. Suppose that, due to a new technology, industry supply has ecpanded until the price level has dropped, say, by a third. Suppose also that the new technology produces a 30—percent higher yield with negligible cost increase. Should the first adopters stop using the new technology? It is quite possible that they can restrict area allocated to this particular crop. But they will never stop using the mac technology as long as they produce the same crop. Cochrane [C. p. 96] puts the matter this way: "To stay even with the world these average farmers are forced to adopt the new technology. The average farmer is on a treadmill with respect to technological advance." he continues "In the quest for increased returns , or the minimization of losses, which the average farmer hopes to achieve through the adoption of some new technology, he runs faster and faster on the Treadmill. But by running faster he does not reach the goal of increased returns; the treadmill simply turns over faster." What Kislev, et al. talk about sounds like a production cycle, not an innovation cycle. The production cycle can be observed for 103 a certain class of products in any country (see recent articles such as Talpaz [T.l], Meadows [M.8], and Hub and Lee [H.28].) It seems that we are now equipped with absolute minimum amount of diffusion theory to be ready to model a social diffusion model of innovation. Before describing the mathematical diffusion model advanced in this study, let us quickly look at how a system scientist or economist using a system science approach could deal with the social diffusion process. The literature revieved here is exclusively the work of Manetsch and his associates. The social diffusion process is often discussed under the heading of modernization program in their works. There seem to be a variety of modernization models. The modernization model of Brazilian textile industry by Manetsch, Ramos and Lenchner [M.5] seems to assure that promotion is not required. In this seise, the model can be said to be an equilibrium model, since the supply is always equal to demand. However, the modernization model of the beef herd management sector in Brazil by lehker and Manetsch [L.13]is different in nature. This model allows both necessity of promotion and assumes a lagged response. This is a disequilibrium dynamic model, since supply exceeds demand, and a dynamic model since the response does not take place instantaneously. The modernization model of cotton production in Brazil by Manetsch, Ramos, andIenchnerlM.3] seems to be the first study of a social diffusion process linked directly with a research sector and modeled with the systems simulation approach. Pbdernized land that now uses a new technology is modeled by a higher-order differential equa- tion such as Equation 4.3 and stimulated by extension effort, together 104 with other variables. Mademized land is supposed to update the yield level, which is a function of public research expenditure accumlated, without additional extension effort. Them, the technical assistance from extension workers, credit requirements and total modernization costs are carputed. A more interesting feature of the model is an atteipt to compute some sort of cost-benefit ratio of the modernization program; total costs being research expenditures plus other modernization costs required, and benefit being defined as the public revenue increase due to a high productivity. The nature of this cost—benefit analysis is viewed from the standpoint of the public sector alone. In fact, the existence of a public sector is justified by its externaltiy. In general, a public expenditure can rarely be justified by this nature of the cost-benefit analysis, without counting external or indirect effects. Otherwise, it would be quite possible for a profit-motivated private firm to enter the market. A social diffusion prototype model appears in Phnetsch and Park [M.6, Ch. 15 ], constructed after several years experience. A repre- sentative application of this model is found in the Nigerian agricul- tural sector simulation model by Manetsch, et al. [M.4]. Since this model is more realistic and a later version by Manetsch and his associates, and since the model presented in this study is a departure from this model, we will review it more thoroughly here. The causal flow chart of the Nigerian modernization cmponent is reproduced here as Figure 5.3. As seen in this flow chart, there are three major processes in the diffusion of a new technology. The dissemination process is 105 .Dfimmooo mo «6303.6 chooses." msounm Goa .3an . . . . . u- .uon. we can geumneoumonu echo ou $.3ng ufifiomfioo oowuwmemhooou mnu mo eWMMWanqumeuoqm .m .m 95mg 30:09:00 m2< on. tetanus-=00 3:33:03. 0H. San—Eda «anaconda—noon... Sconuauuooa Juana—none. 359:” >£>flu=o0um ”Simon—5:005. Aucomucuuou :03 use moan—boa omauo>< >$>309uounm -mnEanooE. . uncomzonuh . >3>Co=v0un~ nuooofi o o «coo Enumenm unanno>0 uoflunncuooOZ .uouam own? finch. 30:09:00 ofiasuoxm . . somuoswcuovofi cu. voumoaom 33733 comncuiw Anfiuuuoue :owmnoaxo . Eek: smooch“ numommnsm confine.“ «to: E 3.3 \ mo :80 33:58 59—: unouonm nominee—hove: Soar: mo 5 need . >3£nm1fi>< O :oSscaD «533:0an 3mm ..oCuoo4 comncooxw . I, '- mogul ocmd negative: 0 \Momgwnom o o 0 $822 rainstorm 1.833.; camera TEN Mazda: ACSSGUZ—iw I. _. J ~ g. .F >::£.:CCLL 106 divided into two categories: (1) adoption due to promotion by the ertension service through direct cumunication, and (2) dissemination through the roncomnmicative source, which is termed the diffusion process in the original report. The third process is the dropout or rejection process. The adoption through commmication source is defined as a function of profitability, including subsidies, extension effort and input availability. In actuality, they rule out the input availability constraint in implementing the simulation run. Instead, they assume that the input required for adopting a new technology will be supplied. The dropout rate is defined also as a function of profitability without subsidies, and the ratio of the actual extension service to the required one. The diffusion rate is now determined by the amount of modernized land and profitability. Then, as usual, various require- ments, modern productivity, etc. , are computed. In summary, wewill extendandslightlymodifythemodel of social diffusion advanced by Manetsch and his associates in such a way as to adapt it to the present study and reflect some of the realities and theories discussed above. Mathematical Pbdel of Social Diffusion of Innovagi_on_ First, we discuss a model of social diffusion of an innovation. Then we discuss computation of the average yield increase due to inno- vation dissemination, taking into account interaction with the resource base. The overall flow chart of an innovation dissemination process hypothesized here is shown in Figure 5.4. Remember that we have a . Aamoeuofioobc mooooum scam—fig Gauge—EM mo ounce 30.3 Jim cobweb mmH 107 108 maidmum of five research outcomes during the planning horizon, which will be indexed k; for each of 13 crops or crop groups, which will be indexed j; and for each of three regions, which will be indexed i. What the flow chart in Figure 5.4 shows is the diffusion process of a single research outcome or innovation, which canes about at a given point of time for a crop and for a region. Will the diffusion of an innovatim take place instantaneously? Do all farmers adopt an innovation at the same time? Does the exter- sion institutim try to disseminate new knowledge to all farmers right after the new knowledge is materialized? In practice, wheiever a new teclrmology comes out, the extensim agency will try to introduce it first to a small number of farmers with favorable socioeconomic char- acteristics. This is due to a limitation of manpower, budget, seed multiplication capacity, or even krowledge of the technolog7 itself. This first group of farmers may be called innovators. Even for innovators, some gestation period is required in order to decide whether to adopt or reject the new idea, or whether more observation is needed. This gestation period will differ, depending on character- istics of different farmers , including their knowledge situations. This means that perhaps the adoption of a new idea by the innovator class can be crudely modeled by a distributed delay model rather than a discrete delay model. As shown in Chapter IV, a distributed delay process can be modeled by a higher-order differential equation that has tie same property as the Erlang family of probability density finctions. As indicated in Chapter IV, there are a variety of DELAY Subroutines. But here the DEIDI‘ subroutine is chosen for solution 109 stability of the delay output, and to use the time increment 131‘ = 1.0 for the efficieicy consideration. However, the DELDI‘ subroutine is found to incorrectly compute the delay output for this particular model.1 Thus wehave slightly modified that subroutine and called it DELDD. DELDD appears in Appendix A, together with other subroutines. The call statement of this subroutine is: 5.1 CALL DELDD [EDF(I,J,K), GA(I,J,K), RGA1(I,J,K,), DGA(I,J,K) IDIGA(I,J,K), DI‘, KGA, AR(I.J.K)] Where: EDFi'k = rate of land entering the delaghprocess for the J kth research outcome for the j crop and for the i region at a given time. GAijk = rate of land leaving delay process (mmodified) ARijk = modified GAijk’ delay output rate RGAijk = intermediate rate in delay process DGAijk = time length of delay lJII‘GAi.k = fraction of time increment, Dl‘, needed to make the J output stable Dr = time increment KGA = order of differential equation. Once we know the input and output rates of the delay process, total land in the process of adoption can be easily carputed using a 1The DELDI‘ subroutine assunes a smooth and cmtimms input rate, which is termed here EDF, over time. However, in this particular model, the input rate changes rapidly in the beginning of diffusim process. This causes trouble with computing delay output correctly, termed here GA. What is modified is that a mechanism to correctly compute the out- Put rate is added inside the subroutine. This modified output rate is tamed AR, and is used to compute related state variables such as Imdem land, which is termed here AMP, as stated in the text. 110 simple integration foomula as done before. This variable is termed '13P, and is computed by summing the intermediate rates, RGA, after other parameters into consideration, instead of integrating the difference between input and output rates. 'Ihat is: 5.2 'ISPijk(t) = 1'12?1 wijkm“) * IDI‘CAijk(t) * DGAijk(t)/KGA The accumulated land that has a new technology, termed AMP, can be computed as follows: _ t Where: Alhjk is the delay output. (hoe the transition and modern land are known, the land remaining with a traditional technology, termed TDP, can be computed as follows: Where: TI'Pijk is total land. What lu‘nds of factors determine the rate of input, or land mtering the delay process (EDF) , the time length of delay (113A) , dropout rate (RRF) , and the order of the delay (KGA)? Would these variables be constant over time regardless of the level of extension promotion, perceived level of the research outcme, importance of the crop, degree of regional specialization, etc.? It is known alpirically that behavior of the delay output is not very seasitive to the order of the delay KGA, in the range between 5 and 10. Thus, the rest of the variables (input rate (EDF) , 1.61ng of the delay (DC'A) and dropout rate (RRF) are hypothesized here as scue fmction of : 111 1. Distributed lagged extension effort for a specific crop in each region, in terms of budget (BEXD) . The distributed lag value is used since it is believed there would be same carry- over effect. Long—run profitability of the specific crop (PROFTY). Change in the long-run profitability (PROFCH). Size of crop in a region in terms of area planted (SEC). Degree of regional specialization (RSP) . Level of specific research outcome (MDIFF) , which is swab)» defined as the expected rate of increase in yield when it is materialized, in earlier material in this chapter. What would the exact functional form between each of the depen- dent variables and die set of independent variables cited above be? The exact fmctional relationship is not well kmwn. However, it is not hard to conceive that: (l) the higher the level of the research outcome and the larger the input rate, the shorter the length of delay we can expect, and so on, (2) the depmdent variable may not be a linear function of individual independent variables, etc. For sim- PliCiLy, we derive one camon factor DDFijk(t)’ to link the depmdent Variables with independent variables, and hypothesize the follcm'ng relationship : 5.5 DDFijk(t) = mmmmijkm + RDEDFijk(t) [mg (t) * {PFEDFij (t) + PCEDFiJ. (t) + CSEDFij (t) + RSEDFiJ. (t) }] 112 Where: ROEDFi.k(t) = effect of research outcome on diffusion for kth J outcome for j crop and 1th region EEEIDFij (t) = effect of extension effort on diffusion PFEDFij (t) = effect of profitability on diffusion PCEDFij (t) = effect of profitability change on diffusion CSEDFij (t) = effect of crop size on diffusion RSEDFij (t) = effect of regional specialization on diffusion Note that (1) when the effect of the research outcome is zero, DDF is equal to zero (2) where no extension effort is given, research outcome is the only variable that affects DDF, and (3) exteision effort is designed to supplement research outcome and other variables supplemnt these two variables. Johnston and Soutl'morth [J .20] point out that "agricultural extension, for example, pays little return mless and until research has produced and tested profitable innovation to extend." This belief is directly adopted in this formulation. Precisely what are all these effects? It is sufficient to illus- trate for two variables only how weights are assigned to each variable, Since the necessary data for the other variables are given in the com- puter Program. let us see how the weight system is constructed for research outcome effect (ROEDF) and extension effort effect (EEEDF). AS seen in Figures 5.5 and 5.6, the weight given to the research out- comes and extension effort is, respectively, a function of research Outcome and extension effort. The weight, however, is not a linear fimction. Then, RDEDF, EEEDF and other variables are interpolated by me311$ of the TABLIE function. For example, for ROEDF: 113 ROEDF 2.0 /—\ 1.0 A A 0.05 0.10 0.15 0.2 Perceived level of research outcome Figure 5.5. Weight given to research outccne to compute diffusion parameter (DDF) (hypothetical). EEEDF 1.0 0.5 0.1 o‘.2 of3 o'.4 ofs Extension effort per 1,000 ha (Million Won) Figure 5.6. Weight given to extension effort to compute diffusion parameter (DDF) (hypothetical). 114 5.6 ROEDF(I,J,K) = TABLIE [VADF6, SMALL, DIFDF6, KDF6, RYDIFF (I,J.K)] Where: VADF6 = function value SMALL = smallest value of independent variables, in this example, zero. DIFDF6 = difference between adjacent elements, in this example, 0.05 KDF6 = number of intervals between elements, in this example, 4 RYDIFFijk = research outcome We were not very specific as to how the long-run profitability and change in profitability, crop size and regional specialization are computed. All these variables are computed in subroutine TEMP. The long-run profitability is defined as: 5.7 PROFTYij (t) = PDij (t) * YDij (t) — g PXDil(t) * lemma) Where: .. .. ' 1 , di u'ibuted PDIJ, PXDij’ YDij and FXDlJl are, respective y s 1: Product prices, factors prices, yields, and factor demand for the z faCtOI'. jth crop, and ith region, which are computed as follows: dPD..(t) 1 _. 5.8 m 7% 75.1— + PDiJ. (t) — PAVGiJ. (t) dPXD. (t) 2. _ 5.9 D * j; + mefla) — Pxig(t) dYD..(t) 1 = . t 5.10m*—d%— +YDij(t) YLDiJO E- svbn cate: 011d Mm land 1 item in Re, rElatj iIIPort W3 115 dFXD.. (t) * 1 9. = 5.11 D + + mm“) mm“) Where: PAVGij , PXfl, YLDij , and Exij 2 are respective unlagged current variables, and D is average eqaected delay on adjustment. We will come back to the cmputation of these variables later in connection with subroutine FDYLD. Crop size (SEC) is defined as proportion of total land area allo- cated to each crop, that is: 5.12 SZCiJ. (t) = Aij (t)/TIANDi(t) Where: A“ = Area allocated to each crop in each region, which will be 1*] determined by the farm resources allocation carponeit of the KASS model. Until this model is linked with the pro- corresponding to Policy Alternative II). TLAND = paddy plus upland in region. Alternatively, crop size could be defined as the ratio of area to cultivated total land. If land is cultivated only once in a year, both methods will come up with the same figure. If all regions cultivate land equally intensively, there would be no difficulty in comparing intensity among the regions. Where the latter definition is used, rice in Ragion 2, where rice is more important in absolute as well as relative terms than in other regions, has a ratio indicating it is not in31301‘t<'31'1t. The same conclusion is applicable in computing the regional Specialization index, RSP. This variable is computed: 5.13 RSPiJ. (t) = szcij (t)/[ASORj(t)/STAND(C)] 116 Where : 3 3 ASOR.(t) = z A..(t) andSTAND(t) = z morn) 3 i=1 13 i=1 1 As mentioned before, the distributed lagged extension budget is used for computing the weight needed to construct the parameter, DDF. That is: 5.14 D * $21353)— + BEXDij (t) = BEXIJij (t) Where: BEXIJ. . = the unlagged extension budget for each crop in each 1*] region per unit of land, and D is average expected delay to adjust. How does goverment allocate total extension budget over crops and regions? First, we assume that total extension budget of the central government (GBEX(t)) is growing at a constant rate; that is: 5.15 GBEX(t) = (1.0 + GGBEX * T) * GBEXI Where: (3th = the rate of growth of budget GBEXI = initial total extension budget Then, this total budget is hypothesized to be allocated over regions as follows: Emmi (t) 3W1 (t) 5.16 BEXiCC) = W +Wt7— *GBEXOZ) Where: BEXi = regional total extension budget r: "M . W 117 TA = total cultivated land in each region STA = arm of cultivated land over regional or national total cultivated land 'IIAND = total agricultural land in each region S'I'LAND = sum of total land over regions or total agricultural land BEXA and BEXB = parameters Then, to guide regional extension budget allocation to each crop, the following score system is used: 5.17 5001an13. (t) = 300121 * PFEDFij (t) + SCDRZ * mmij (t) scorn 7': csmmmijo) + $009.4 * RSEDFij (t) + PSCDREi J. (t) Where: PFEDF, PCEDF, CSEDF and RSEDF are the weights used to conpute the parameter DDF, as seen before. PS(I)REi j is a policy-determined score given to each crop to implenent a certain policy in support of export or attainment of necessary food grain production. SCI)R1,2,3,4 are the respective weights given to individual factors in the equation. Extension budget for individual crop per unit of land is hypothesized to be allocated as follows: SCOREi (t) * BEXi(t) 5.18 BEX'IJij (t) = SSCOREi t W Where: l3 SSCOREi(t) = jil SCOREiJ. (t) BEXi = regional total extension budget TAi = sum of total cultivated land in each region 118 Now that we have learned how the parameter DDFijk(t) is cmputed, let us examine the role this parameter plays in determining the length of delay (DGA), the dropout rate (RRF) and the input rate (EDF). The length of delay, DGA, is conputed once at the beginning of a year, when every nev research outcome is ready to be extended. Then this value is carried over until that research outcane is canpletely disseminated. This is done mainly because the DELDI‘ subroutine is not capable of the time-varying delay. 5.19 DC‘Aijk(t) = MAX [1.0, (MAM —- DDFijk(t))] Wnere DGAM is a parameter given that reflects a mandmm delay in year in the worst case. Note that as the parameter, DDF, increases, the length of delay is shortened but restricted to a minimum of one year. Also, note that the length of delay or expected average delay represented by DGA has a different notion from that often used by the sociologist. The expected average delay used by the sociologist in adopter distribution is the weighted average time between initiation and oonpletion of an innova- tion adoption for a whole population. But, the expected delay used, here, DGA, is changing with time and corresponds to the adoption delay of each year's input rate (EDF). It is possible to construct a model of what the sociologist talks about, however. The only modification from the one we constructed is to treat the input rate as an inpulse canposed of the whole population. That is, the whole population is assumed to be an adopter candidate fran the beginning, instead of assuring a fraction of the population are Candidates in each year. 119 Similarly, the dropout rate, RRF, is defined as: 5.20 RRFijk(t) = MAX [0.0, (RRFM - RRFA * DDFijk(t))] Where: RRFM = maximum dropout rate RRFA = parameter to conver DDF into a suitable magnitude As DDF increases, the dropout rate (RRF) decreases, but remains non-negative. Arnother type of restriction is given to this rate, too. Where the undem plus transitional land become more than 80 percent, the dropout rate (RRF) is equal to zero. In determining the input rate, EDF, we put what Manetsch and his associates call irnput rates due to promotion and due to diffusion together to form one category. Thus, we have only one delay process in our model instead of one for promotion and another for diffusion. If the expected average adoption time, here termed DGA, and adopter distribution parameter, KCA, are not different in both delay processes, we can consolidate then into one process without much loss. In addi- tion, we assnmne that both the modem land (AMP) and transitional land (TSP) have demonstration or diffusion effect, the latter having sanehow less effectiveness than the former. The input rate for each research outcome for each crop, region and point in time is defined as: 5.21 EDFijk(t) = AEEijk(t) * TDPijk(t) +AMl‘ijk(t) * AMPijk(t) * + DFC * AMI'ijk(t). TSPijk(t) Where : TDP. = traditional land ijk AMP =modernland ijk TSPijk = transition land DFC = a parameter 5.22 AEEijk(t) = AEEA * DDFijk(t) 5.23 AMIijk(t) = AMEA * DDFijk(t) Where AEFA and AM‘A are again parameters. Thus, the first term in Equation 5.21 is the input rate due to promotion, and therefore AEE is a fraction indicating what percent of the traditional land is going to enter the delay process per year. The next two terms in the same equation correspond to the irnput rate due to demonstration or diffusion from noncommmicative information sources. Thus, the variable (AME) is a kind of multiplier indicating how many farmers a modern farm or tran- sitional farm can try the new technology. Now two conditional restric- tions are imposed on the input rate (EDF). The first is a logical restriction: the input cannot exceed the traditional land: 5.24 EDFijk(t) = Min [EDFijk(t), TDPijk(t)] EDF in the right side of the equation is the input rate computed in Equation 5.21. The other restriction is that whenever modern land (AMP) plus transitional land (TSP) exceeds 80 percent of total land population, we let the input rate, EDF, equal the traditional land TDP. That is: 5.25 EDFijk(t) = TDPijk(-t) 121 The reason was explained earlier in the discussion of the restriction on the dropout rate, RRF. This is the essential overall picture of our social diffusion model. Before discussing how average productivity changes due to inno- vation diffusion, we add two more points. First, now that we explained how the dropout rate (RRF) is determined, we modify the definition of the modern land (AMP) given in Equation 5.3. The correct one actually used is: 5.26 AMPijk(t) = AMPijk(t) + rot ARijk(t) * [1.0 - mm (12)] (it The reader may wonder about the other delay subroutine, such as DELVF, which can directly count the loss rate, which is equivalent to the dropout rate (RRF) . The reasons are: (l) to demonstrate the usage of different delay subroutines, and (2) to make the dropout or loss rates functions of time. The second modification involves the definition of the land. The population can be defined in terms of nmber of farmers or acreage, and in absolute or relative terms. In this model any population is specified in terms of percent of land. In other words, the initial value of the traditional land (TDP) for any campaign of the research outcome adoption is set equal to 100, which is total land appearing in Equation 5.4, termed TI‘P. What the modem or transitional land (AMP or TSP) shows is what percentage of land allocated to a specific crop has adopted or is trying a certain new technoloy. There are two basic reasons for adopting this definition. First, it prevents additional complexity. Second, even if we use the absolute acreage, 122 we still have to convert it to relative tenms in order to compute average productivity gains. In connection with this definition, an additional assumption is necessary. The area allocated to a specific crop changes over time, either due to economic adjustment or the fact that some agricultural land transfers to nonagricultural usage. Is the more modem or traditional land likely to transfer to the other crop or nonagricultural usage? The expected change in area response is small, owing to assump— tionsmndeeitherintheinitialversionoftheKASSmndelorinthe farm resource allocation component, we can generally assume that the conversion of cropland will be the same for modern and traditional land. It seems appropriate to specify one more implicit model assump- tion. We assumed a maxim of five successive campaigns for a crop during the planning horizon. Would it be more realistic to assume that only the farmner who have adopted the first campaign are eligible for the secornd, only those who have adopted the second campaign are eligible for the third, and so on? This is not necessarily realistic. Furthermore, there is a time lag between successive research outcomes. Therefore, if we were to adopt this restrictive assunption, it might not correspond to reality. In addition, there is no apparent reason that a farmer cannot adopt the new technologies in a different order. This is one of the critical model assumptions. Now that we have specified the necessary model structure and assumptions about the social diffusion of an innovation in detail, we are ready to present a method for cemputing the expected average 123 productivity gain. Individual research outcomes having a set of per- ceived attributes (RYDIFF) have been defined as productivity gain in terms of yield at the experiment station, not at individual farms. Would the productivity gain at individual farmns adoptirng the new technology be the same as that at the enqnerimnent station? It is not hard to imagine that some farmers' resource base or technology, other than that to be disseminated, would be quite similar to the experiment station situation, but many would not. We may well asanmne that farms with a good resource base or equiped with better knowledge would adopt a new techrnology first. This argument then implies that as a new technology is disseminated over farms, the productivity gain on individual farms would decline. This conclusion is quite consistent with findings from the green revolution process by many researchers, such as Evenson [E.3], Wharton [W.3], etc. If we assume there is no difference in other technology or resource bases between the experiment station and individual farms, and that the transitional land will experience the same productivity gains as the modern land does, the expected average productivity gain due to an innovation diffusion in a region will be: emulated productivity gain due to successive innovation diffusion for a crop on a regional basis will be at a given point of time: k . . . = . . t 5 28 (:1?le (t) kil anleC ) 124 What the first equation says is that if the modern land (AMP) us the transitional land (TSP) turns out 100 percent, the produc- wity gain on a regional basis will be the same as MDIFF. Now let us assume that: (l) the resource base or other techno- gy of individual farmers at their disposal is not the same as that the eqnerimnent station (2) there is some difference in productivity in, by a factor DFC, between the modern and transition populations, 1 (3) the productivity gain diminishes as more fanners adopt a new :hnology. Then the expected average productivity gain due to dis- m‘nation of a new technology in a region will be: 5.29 (Eijk(t) = RYDIFFijk(t) * [1.0 - DFijk(t)] * [AMPijk(t) + DFC * TSPijk(t)] re DF is an average discounting factor and computed as a function the modern land (AMP) by using the TABLIE function. Function .168 are given in Table 5.2. Table 5.2. Function Value of Discounting Factor, DF (Hypothetical). Nbdern land (In Percent) 0 20 40 60 80 100 DF 0.03 0.03 0.035 0.042 0.51 0.063 The farmners ' resource base or other technology changes over Thus, to correctly estimate this average discounting factor the function shifter mnust be incorporated in one form or r. However, this type of interaction among technology level, 125 [put use rate, and the resource base will be considered specifically nd incorporated into the product supply and factor demand projection mponent models in later chapters. The accumulated productivity gain on a regional basis can be mputed easily at a given point of time as: 5 5.30 casijkm = kil Czijk(t) need to compute the rate of change in productivity gain due to the 7 technology dissemination. Let us define it as the total factor nductivity growth rate and label it YZ. 5.31 Yzij (t) = [CZSij(t) - (ESij(t-1)] [100 + CZSij(t-1)] ce the gross productivity gain, es, is defined in terms of per- tage, we must divide the successive difference by 100, plus the ' us year's gross productivity gain. Thus far, innovation diffusion and the resultant productivity due to the public investment have been considered in research. ‘he public institution the only one where an innovation or a new nnology is being generated? It is obvious that some farmers act or less as inrnovators in selecting seed, using production factors, pplying husbandry suitable to his specific farm location. Other ers imitate this progressive farmers. Therefore, indigenous nological change is made available by the leading local farmers selves. As a matter of fact, this has been the major source of nological changes prior to establishing the modern experiment .on. It also seems that this source of tectmological change is 126 important even at the present timne. It is also a well known that the agribusiness finn that supplies the fann sector with n inputs or processes farm products engages in research arnd .opment and disseminates findings to farmers. The next question is whether the extension worker is the only of bringing new information from the public research institutes rmners. The extension worker's job is to facilitate communication en the farmers and the public research institutes, as well as farmers themselves. This includes finding better practices or ndry at a farm or location, and introducing them at other farms nations. In sumnary, it is quite possible that some technological change place even slowly and is disseminated among farmers or agri- '33 firms without the help of public research and extension utes. It is also true that the extension worker can accelerate fusion of this type of indigenous and spontaneous technological The productivity gain due to this process is hypothesized as ion of the distributed lag extension budget with a positive apt, as shown in Table 5.3. Then the function value of the :ivity gain for each crop in each region is interpolated by ELIE function. Let us term this productivity gain YW. Then rroductivity change due to research and extension turns out: .32 YZDij (t) = Yzij (t) + Ywij (t) 127 1e 5.3. Productivity Gain Due to Extension of Spontaneous Technological Change (Hypothetical). Extension Budget, Million per 1,000 Ha. 0.0 0.04 0.08 0.12 0.16 [0.20 tivity gain 0.0003 0.007 0.01 0.012 0.0138 1 0.015 In sumnary, the total factor productivity growth rate (YZD) is only variable in this subsector model to be transferred to other psector models discussed in later chapters for the present purpose the study. The perceived levels of research outcomes (RYDIFF) and 5 derived total productivity growth rate (YZD) are crucially important .iables in our whole model. The variable RYDIFF is defined as the biological research result terms of the rate of increase in the yield level at experiment station. figures for INDIFF in Table 5.1 do not represent what will occur :nnditionally in the future or a target the public sector wants to Leve. Those figures simply represent one set of all possible arch outcomes. This does not necessarily mean that this set of arch outcomes will be likely to occur. It is simply a point of We for examining the consequence of some possible research omes on the overall performance of the Korean economy. In this e, the point of departure can be ectremely high or low as canpared .nat will happen in the future. We need to run an intensive policy experiment on this variable, 2 the research enterprise involves rather high uncertainty or This policy experiment is intended to provide public decision— 128 with: (1) some information on desirable biological research so they can design and build a research institute to secure e desired levels of research outcomes, and (2) some possible specific strategies to reach somne of the agricultural development 3. In this chapter we have modeled the complex and ill-structured en of biological research and dissemination of its results. The tionships described in this public subsector component are not nical, but mostly socioeconomic ones. These relationships have neen well studied thus far. The model presented here is based on art than science. This is one way to model a complex but ntructured system. The basis of constructing this kind of model (1) related theories and (2) experiences of experts in this field, a "if experienced observers believe that is the way things really that is the way they should be in the model, even if the para- s carmnot be measured" [Kresge (K. 8) ]. What is needed for improv- '3 component are: (l) in addition to further testing and ement of model structure, data gaps should be filled with more edgeable estimates fran researchers, extension workers and other red personnel, and (2) based on more reliable data and structure, additional computer runs should be made for the purpose of zing possible errors and further model validation. CHAPI‘ERVI PRODIIZTION FUNCTION AND PRODUCE SUPPLY PRNECI‘ION This chapter proposes a form of production function that has a mic long-run property and is used to project yields per land unit 13 crops or crop groups in each region. First, we review some of supply response studies and projection techniques used for product Ly in section one. The mathematical model is discussed in two ions. The second section discusses the projection model for 91 amps. Then, in section three a yield projection model for nnial crops, fruit and mulberry-silk is developed. Lastly, in on four, we discuss the sources of data and parameter estimation 61$. égricultural SuppIy Studiiin Iitfltgrg The study on agricultural supply of either aggregate study or dividual crop is a relatively rich part of agricultural economics ature. Historically, product supply studies have revolved 1y around hypotheses that either farmers are not price responsive ack profit motivation or that agricultural supply functions have icities near zero. We do not intend to intensively review studies, models or nations accounting for agriculture's failure to contract or to ioutput. In fact, Heady [H.ll, p. 675-676] and Hathaway 129 130 .5, Ch. 4], among others, summarize the hypothesis concerning this Ltter. On the other hand, Johnson [J.l6, Ch. 3] has reviewed some E the main explanations of agriculture's historic failure to con- :act or expand output, in order to advance a new hypothesis involving nvestment and disinvestment or the lack thereof as determinants of pply response. The product supply function can be derived from the static neo- assical economic theory, which is normative as it assunes maximizing havior of producers and consumers. Assuming diminishing returns to ale with some fixed assets, the marginal cost curve derived from nduction function having the above nature is interpreted as the art-run supply function of the firm. Based on this theory, implied :imation function of the supply has a form: S=f(Py) are S is supply and Py is product price. This form assumes first all that input prices are fixed, which is certainly not the case the real world. With input prices as the supply function shifter, estimation form is: S - f F , P ( y _) eri the estimation function oversimplifies the real world. That is, stands for the relevant input price. Even in this formula- firm has limited amounts of resources and usually produces more one product. With this consideration, the supply function is ified: ' 131 s = f(Py, 12x1, Pei) re P a. stands for the relevant price of alternative product. Thus far, we have implicitly assumed current prices would influ— 2 current supply by affecting yeilds. As Fox [F.9] discusses, general, the quantity of a crop ready for harvesting is determined acormdc factors that operated before planting time and during growth ges in which yield-influencing practices or materials may have been .ied, and by noneconomic factors such as weather. The logical impli- .on of this argument is that prices used as independent variable in rly estimates should include at least one year lagged ones with or out current ones. It seems that most supply studies prior to the 19503 were based be above lines of tlu’nking.l Nerlove vievs supply response as an Cation process in an uncertain world where lags in the adjustment asses exist. Because of uncertainty and sticky adjustment processes, ssunes farmers' adjustments do not take place instantaneously in ggregate, since farmer's expectations and lags differ. The basic t of the Nerlovian dynamic system can be surmarized as follows Nerlove (N.6, p. 53—63)]: _ 7% St ' St—l — Y[St ' St-1] St stands for the current or actual output, 3: for long-rm ibrium or desired output, and St—l for one year lagged actual 1For a review of supply function studies, see Nerlove [N.6, ] and Knight [K.6]. 132 t. The desired output can be expressed as a function of farmer's * ectation on future price (Pt) such as: * * St = a + B Pt , this expected price can be expressed: 71’ _ P P-k Pt ‘ Pt—1 ‘ M t-l ' t—l]' re P:_l and Pt-l are, respectively, one year lagged expected and a1 price. 7 and a are called the respective elasticity or coef- .ent of adjustment, depending on whether output or price is expressed ogarithmic or absolute terms. These three equations are derived from Nerlove's basic system of mics. Let us interpret this system. First of all, it is a differ- al equation system with a numerical solution. 'mat is, the first :ion, for example, can be written: D ‘ dS * t _ dt +St"st D stands for the expected average years of adjustment of the 1 supply to the desired one. Euler's numerical solution to above ion is: * St = st—l + DI'ID [St - St—l] is exactly the same as the first equation of the Nerlovian system noting that Dr is a time increment, and DIS/D has the same meaning Secondly, in this system, the actual output is adjusted or 133 pond to the desired or long-run equilibrium. What would this mean? e seen to be few theoretical problem in this formulation as it tains little thoery. Nerlove's work has stimulated much subse- t research on supply or factor demand estimates. However, this s not necessarily mean that this formulation explains the fanmer's supply response very well. We will now examine what is needed more realistic supply prediction. First, a personal comment. Nerlove and all his subsequeit owers assume, almst without making specific comment or criticism, i the repercussion of a disturbance over farmers' reaction is :esented by a simple exponential function, as modeled by the first- :r differential equation. Do they do this for simplicity or because ’believe their assumption is realistic? What are the empirical ications for the innovation adoption process modeled with S or J ed response curve? Would supply response studies be improved by ling with a higher-order differential equation rather than with first-order differential equation? The second comment has to do with forces determining the "adjust- coefficient.” Nerlove [N.7] makes it clear that the estimated fluent coefficient is unstable over time. What forces change this icient? Does this involve invesmemt and disinvestment behavior 1 on direction, duration and magnitutde of price change? At around the same time, Nerlove advanced the pioneering techni- f estimating supply responses, there was some theoretical advance .derstanding farmers' response. This has to do with the asset y theory involving investment and disinvestment due to Johnson 134 his associates [J.4, J.5, J.13, J.l6, B.l3, E.l]. This theory the theory concerning the knowledge situation are the major cam- ts used to modify the neoclassical economics. The modification extension of static tl'eory in this direction intmds to explain the banana of chronic disequilibrium in the U.S. agriculture more cisely. According to Johnson [J.l6, p. 24], "At least three different lines of reasoning have same importmce as explanations of the teldency of American agriculture to expand but not contract production. The first and most important of these deals with technical advance. The second deals with improvenmts in the human agent. The third has to do with the role that various economic adjustments, mainly specialization, play in increasing productivity." ng felt these explanations to be inadequate, he [J .16, p. 26] ludes that "A substantial gain in explanatory power is achieved “he”: the neoclassical analysis is modified: (1) to recognize explicitly that acquisition costs may be less than, or equal to, or less than zero; (2) to recognize imperfect knowledge (as D. Gale Johnson did) of the techrology, education, and other ckmmges . . The key concept of the resource fixity theory is the distinction en acquisition price and salvage value of resource, which the neo- ical economic theory fails to make consistently. Then the defini- of the resource fixity is as follows: "An asset will be defined, very simply and curdely, as fixed ('if it ain't worth varying'). Nbre elegantly stated, an asset will be defined as fixed so long as its marginal value productivity in its present use neither justifies acquisition of more of it or its disposition" [Johnson and Hardin (J .17)]. me definition can be stated in mathematical form: oosza>MVP>sz_:_0 135 an stands for the acquisition price of an asset, sz for its vage value, and MVP for the marginal value productivity of its ent use. Compare this definition with the corresponding implica- of the conventional definition of the fixed resource in the run, which can be defined: m=an>MIP>PXS=O What, then, would implication of this theory be in terms of yield, 1 organization and production response, and aggregate supply? Since 'papers [J.16, Ch. 3 and Appendix, E.l, J.4, J.5, H.5, Ch. 4, and Chs. 6‘ and 7] deal with this subject matter in detail, we briefly oduce here the basic idea as in the context of the present study. Dse a firm producing a product where one factor is variable, such ertilizer, and the other is fixed. Further suppose that the pre- farm organization is represented by P1, Q1 in Figure 6.1. Assume the product price has increased to P2. In this situation, more 1e variable input will be used. How about the fixed input, the1? : are two possibilities: (l) the MVP of the fixed input may still as compared to its acquisition cost, and (2) the MVP is suffi- ly increased to justify the purchase of an additional unit of the Let us assume the first case. Then supply quantity will be ased slightly, such as toQZ, since one resource is fixed so nroduction is subject to diminishing returns. Now, suppose the :t price has increased to P3. Then, without doubt, the use of ariable input will be increased. For the fixed asset, one of 136 a two alternatives discussed above will come up again. But let us me that this much price increase now justifies the additional of the fixed input. Hence, the production function shifts to the er subfunction. Product Price c P3 r b P2 . P4 ’ d Pl - a Q1 Q2 Q4 Q3 Quantity of Supply a 6.1. Hypothetical supply function derived from resource fixity theory. In the first case of price increase, only one input was variable, 3 in the second case, both inputs are variable and supply quantity h increased, such as to Q3. Thus, the segment of the supply on is more or less flatter than that of the first price increase, m in Figure 6.1. bw suppose the price has dropped to P4. What might happem to red resource? It is certain that the MVP Of this asset is accord- tropped. Had its salvage value beer equal to its acquisition 137 'ce, the farm would not face a capital loss by liquidating same of '5 resource so that its MVP would be equal to its market price. ever, its MVP may be greater than its salvage value but less than acquisition cost. If this is the case, the previously acquired full unt of this resource will be still used on the farm. Thus, the 1y quantity is now slightly decreased such as to Q4. Note that e consequence of the definition of the asset fixity yields a partially eversible and kinked (at Point b) or discontineous supply function. F important conclusion is that the price elasticity of the aggregate pply is different, depending on the direction, duration, magnitude 1 recent history of price movement. Smith [8.13] is the first research to consider two prices, :uisition and salvage, for farm-produced inputs in a linear program- g model in order to examine the nature of farm organization. This hodology can also be used to determine the optimum investment and investment for any durable inputs. For example, lee [L.5] applies 'Lfferent wage rate for disposing of own family labor from that hiring labor. Mare significant application of this resource fixity theory can btmd in the supply study. It seems that Halvorson [11.3] is the t researcher to study the impact of the direction of price change upply elasticity estimation (milk), although Johnson [J.4] and m and Johnson [B.13] have tested this possibility with time series :ross—sectional data, respectively. Later, Barker [B.l] tested supply theory hypothesis that the elasticity of expansion under lg prices exceeds the elasticity of contraction under falling 138 ces (for milk). Recently, TWeete1 and Quance [‘I‘.5] confirmed the type of hypothesis for aggregate agricultural srpply. Quance [Q.l] ' es capital gains and losses by means of the resource fixity ry. The resource fixity theory does not imply assymetrical responses ice increases and descreases; instead it implies adjustment to a target. Empirical evideice that the response is assymetrical t contradicted by the theory. In most applications of this theory in supply parameter estima— , it seems that they pay attention exclusively to the elasticity germs of the direction of price change. That is, it appears that 'e is no empirical study to test the hypothesis that the elasticity lso differmt, depending on the duration and magnitude of price ge as implied by the theory. In other words, they have studied segment of a study curve such as bc and cd in Figure 6.1, but not ants ab and bc where a kink occurs. This ldnk may not be obser- 2 in the aggregate data, whereas Figure 6.1 is drawn on the basis firm. This is, perhaps, an important reason that they have not .ed this kink in deriving supply estimates for the aggregate data. .ct, the supply function of each individual firm may be kinked fferelt levels of supply quantity, or the price expectation may fferent from each other. Many problems are involved in the analysis of agricultural supply. [H.131 and Nerlove [N.7] discuss these difficulties or problem ' The common factors can be sumarized as those concerning: mplexity of production system, (2) technological change, (3) ation, (4) fixed or quasi—fixed production factors and (5) ainty. 139 learn and Cochrane [L.3] conclude that "regression analysis of e—series data is an imperfect tool for supply analysis where struc- al changes have occurred during the time period analyzed. . .Regres- n analysis has rarely provided satisfactory supply estimates in the t. . ." Staniforth and Diesslin [8.15] conclude that ”The major single limitation is that it (regression analysis) cannot be used for prediction in light of new variables and structures. Regression models. . .reflect historic relationships and at best describe present relationships. . .prediction is more important than the record of the past." we have noticed, technological or structural change which is one thant supply function shifter among subfunctions as defined by .n and Cochrane [L.3] plays an important role in economic develop- : or adjustment process in agriculture. Dean and Heady [D.5] hypothesize that technological change cts supply elasticity itself. How should we handle this important able in modeling the supply system? Many seem to consider this able in supply estimation as they do in aggregate production func- estimates. However, none of then can be considered adequate. m and Cromarty [B.ll] put it this way: "For the technology and weather variables the data used are inadequate but are better than complete emission or oversimplification which accompanies the use of a linear time trend." Johnson [J .8] suggests that "A principal problem encomtered in synthesizing macro supply estimates from micro data has to do with pre— dicting which inputs or resources are changed and which are not changed. . .Changes in inputs to be considered include, of course, those necessary in introducing new technologies, and securing the benefits of regional, sector and farm by farm Specialization and diversification.“ 140 InhereIt limitations of regression models in predicting the supply antity seem to have caused research workers to turn to budgeting, agrammdng, production function or related techniques. It is not ssible to review here all relevant supply studies based on these tech— :1ues, however. 2 In fact, the production function is the foundation supply. Once the production function is specified, it is easy to idle supply problems and factor demand. Indeed, Wipe and Bawdem .4] derive firm—level supply equations from empirically estimated )duction functions to predict firm output and estimate supply [sticities After making a comparison with the actual data, they Lclude that derived supply equations are not erpirically reliable. lt is, they find that predictions ranged from slight underestimates extreIe overestimates of actual output, with the latter being most valent. Derived supply elasticities were generally found higher .1 those obtained by direct regression analysis. This conclusion a not seem new in literature, however. This author [L.4] has ) discussed these properties of derived supply function. The degree lisagreement between actual response and response derived from the tuction fimction would become even greater when the Cobb-Douglas , of production function is used as by Wipe and Bawden. What are the possible causes of this general overestimate and isistency of estimates or predictions they find? A critical factor 2’For basic methodologies, see Heady and Dillon [H.181 , and ' and Theater [H. 19] for production fimction approach; Heady and er [H. 16] and Day [D.3] for programming approach; and Johnson and Mighell and Black [M. 16] for budget approach. 141 is that the farmer is not a straightforward profit maximizer. He has a much more complex set of variables to consider than that dictated by the static micro economic theory. He has a set of restric— tions in making production decisions. That is, he has limited amomt of capital, takes uncertainty into account, does not respond to change instantly, etc. In addition, the production parameters may not reIain stable over time. We must consider all relevant decision environments of farmers if we are to make a realistic prediction, especially when adopt a production function approach. We want to close this subsection by quoting Heady [11.13]: "The efficiency of either (positive and normative) thus depends on whether the relevant variables are included and accurately measured in the empirical model and low well they correspond with the real world condi- tions as they will exist during the period for which predictions are to be ma ." Before moving into tie next subsection, it seems worthwhile to riefly discuss literatures concerning supply studies in the LDCs. It spears that the marketed surplus model of the subsistence crop zveloped by Krishna [K.9] is the first supply model unique to We. at is new is that the price elasticity of supply in this model is directly estimated, since the standard technique is difficult to ply due to a lack of data. On the other hand, after criticizing the weakness of Krishna's 1el in terme of assumptions made, Belmman [B.5] advances a new lel of the marketed surplus of the subsistence crop, which is entially an extension of Krishna's model. According to Behrman, "Same economists contemd that the supply response of the agricultural sector in less-developed comtries is quite simmilar to the response in countries with high per capita income. Others argue that this response is perverse in the sense that increased prices result in smaller quantities. E;- 142 "A third group maintains that institutional constraints are so limiting that no significant reaponse to economic incentives is likely to be observed." His own position is that, 'To some degree, these different opinions have resulted from the failure to distinguish explicitly between the supply of a single crop and the supply of all crops, betweex total production and the marketed surplus, and between short-run and long-run responses." He them estimates acreage response elasticity of the subsistence crop to price change. In fact, what we need to know in the context of development is not acreage response, but the supply of total crop production and yields. We know acreage responses to price change reasonably well, but what is less well known is whether or not total production and yield respond to economic opportmities. Schultz [3.2, p. 37] hypothesizes that, 'There are comparatively few significant inefficiencies in the allocation of the factors of production in traditional agriculture." Nevertheless, they are poor, not because of malallocation of resources but because of the low-level equilibrium trap. In fact, Schultz‘ s hypothesis is intended to derive a proportion that the reorganization of the existing resources at the farmers‘ disposal may not serve to increase total production. Now, let us turn to methodologies they used to make production Projection. One of the conclusions described above was that neither classical supply studies nor orthodondcal production function studies are adequate for making projection of product yield. This may be eC{Uivalent to saying that there is no established metlodology that Can be used for varying situations. Perhaps for this reason, they formerly used quite diverse varieties of methods. For example 143 Blakeslee, et al. [3.9] used extrapolation of the past trend of yield in projecting world food production, demand and trade. The same technique was used by Desaeyere, et al. [D.9] in maldng a long-term projection of supply and demand for Belgium agriculture. As everybody knows, the trend does not have much meaning. Rather, this technique disguises some important economic relationships. Black and Bonnen [B.8] base their supply projection on same judgment as to what the impact would be if technology made available by experimental stations is adopted. Johnson [J .12] claims this methodology works out well and names this type of projection as the traditional pro— jection in the sense that it is not computerized. Bormen [B.lO] reviews several projection studies and tests the accuracy against real world data. One of his conclusions is that the projection is underestimated, because they do not consider some aspects of economic adjustment, such as regional specialization. Tom [TA] discusses steps in economic projection. After adding several difficulties with production projection, such as the fact that annual crop production in Northern Nigeria varies up to 40 percent, which is attributed mainly to climatic factors and irresponsiveness of subsistence farming, he introduces the methodology of yield pro- jection used for Nigeria. According to him, the factors considered are improved husbandry and equipment, improved seed, seed dressing, and insecticide in storage. He does not indicate what adjustments the farmer makes to adopt technology, possibly due to space limitation. However, it is hoped that this plan is not the type of plan Dandekar [13.2] criticizes. 144 In addition, Tom briefly introduces the National Council of Applied Economic Research's projection methodology used for India, one of the few comprehensive studies of agricultural supply in a developing country, which will be reviewed here in detail. The National Council of Applied Economic Research [N. 2] presents projections of demand for and supply of agricultural commodities for India (1960—61 and 1975-76) . The basic method used in projecting production or potential supply can be best explained by a diagram, as shown in Figure 6.2. First, they project areas with or without irrigation facilities and then improved seed. With fertilizer With improved seed ’ Irrigated Without fertilizer Area Without improved Total area see Under "" With fertilizer Cultivation With improved seed [$19th Without fertilizer Without improved seed Figure 6.2. Cultivated land classification by technology for Indian agriculture. Source: National Council of Applied Economic Research [N.2, p. 159]. Then it is assumed that land with improved seed is entitled to be cultivated with fertilizer. Basically there are six different lands for each crop. This type of land classification is based on the availabilities of the various inputs considered. Next, they determine yield response for each of the various land categories from 145 various sources of data. Once areas and corresponding yield rates are known, it is very simple to compute total production or average yield. There are some interesting features in this study. First, they take into account the effects of what they call intangible factors, which include community development programs, credit facilities and marketing, agrarian reforms and improved agricultural practices. How- ever, it is not very clear how these intangible factors affect produc- tion. The other point is that they claim that the production level obtained represents the potential, and the actual production will be adjusted by whether or not pesticide is used. Cownies, et al. [0.9] presents a slightly different model for projecting production for West Pakistan. In their scheme for land classification, the with-or—without-fertilizer category is drOpped from that done by the National Council of Applied Economic Research, or from Figure 6.1. Instead, they assume that yield in each of the four land categories is: Yi(t) = Yi(o) + ni Mt) - £(o)] Where Yi (o) and {(0) are, respectively, the yield and fertilizer application rate in the base year, whereas Yi(t) and at (t) are, respectively, the corresponding current rate, and 11:.L is fertilizer response coefficient. Both models base their projection on several highly subjective assumptions . No economic basis is given in allocating resources among crops. The yield level is assumed to be a linear function of only 146 fertilizer application. Variable input levels are treated as exogenous. Farmers are assumed to have no ability to adjust to changes in econanic variables. A similar model is used by Lele and Mellor [L.lS], not for making projection, but for identifying the source of productivity change in food grain for India. They use historical input and output levels determined by farmers. The national Council of Applied Economics Research model does not seem to be worked out well. The result of the model projection shows that India would attain self—sufficiency in food grains by 1964, and then would have over production. But nobody insists that India has achieved food self—sufficiency at the present time. Either pro- jection of demand or production or both must be considered unrealized, even in approximation. This kind of projection or planning model seems to invite a criticism, such as that given by Dandekar [D.2]. let us quote again what he says: "A plan is a plan in the true sense of the term only when it is a proposal for action on the part of the one who makes it. The reasons our plans in agricultural development have not been plans in the true sense is that they have not been essentially plans for state action. It is obvious that making the farmer's resources allocation or related decisions is not a job that can be done directly by a state. In summary, there seems to be no established methodology that we can directly apply to projecting yields over time. Thus, in the fOIIOWing section we propose computerized methodology that allows various economic factors to play a role in determining yield levels. 147 Yield Projection Model for Annual Crop A change in production level and hence supply must be inter- preted as a consequence of either a change in the rate of the so-called conventional input used, a change in the teclmology level or a change in random disturbance factor such as weather. The supply Emotion most commonly used is expressed as a function of the product price or com— bined with factor prices, with or without other variables. This is incorrect formulation. The price change influences product supply indirectly through forcing a change in input use. Johnson [J .5] points to the heart of the weakness of this prag- matic formulation by saying: "One of the more serious difficulties faced by the camodity supply analyst is the lack of data on the amounts of different resources used in the production of each farm product. Lack of such data requires researchers to deal with price-output rather than with production function relationships, thus limiting their alternative approaches.” We reject here the price—output relationship approach. Instead, we adopt what might be called an economic system approach capable of modeling the effect of change in input used as well as change in tech- nology. The model must also be capable of dealing with dynamic aspects of agricultural economic systems. The production function used here is not a sort of aggregate production function disembodied technological change. The term "technological change" used in this sort of aggregate production function is, according to Schultz [8.2, p. 137], "not an analytical concept for eXplaining economic growth." This is so because, according to him, "1‘0 use it for this purpose is a confession of ignorance, because it 148 is only a name for a set of unexplained residuals." A number of authors agree that to measure technological change in that way is a 'ineasurement of our ignorance." One of the main purposes of studying an aggregate production function is to provide a basis for prescribing policy alternatives designed to affect the growth rate. According to Denison [D.8], this is equivalent to a "menu of choices available to increase the growth rate." The menus provided by the aggregate production function studies, including the Denison work itself, do not seem very precise inasmuch as the policy—maker can widely select an appropriate set of policy measures without confusion or ambiguity when applied to the agricultural sector. To identify more precise sources of the growth or yield function shifters in order to provide a more precise menu of policy choice and to provide a framework of structural analysis of an economic system, we propose an aggregate production function:3 a. . Q(t) _ 1.] dlYfi@)~A um, buk¢)( ij 13 ii: (t) m B t) which is essentially a Cobb-Douglas-type production function, and where: Yi‘ = production per unit of land yield of jth crop in ith 3 region Xi' SL = 2th the so-called conventional production factor for J producing jth crop in i region = kth the so—called conventional production factor or what we call technological or structural change variable in i region Aij = aij 2 and bijk are appropriate production parameters. 3Notation used for variables or parameters here are not the same as those used in computer programs just for the simplification. However, We will specify the notation used in computer program at appropriate Places for major variables. 149 The yield function given in Equation 6.1 does not seem much different from aggregate production function or micro firm production function often used. First, this yield fimction has the property that the source of the growth or technological change are disaggregated in detail. Second, this yield function has a property of a long—run context in the sense that all production factors are identified. Third, all independent variables (Xijk and Zik) are endogeneously determined, based on various economic variables or public policies, projects or programs. To simplify the model, we adopt Taylor's expansion series of the production function in Equation 6.1 [see, Allen (A.8, p. 457)]. Ignoring the higher—order terms, we have: BY. . (t-l) 6.2 Yij (t) = Yij (e1) +i [xij1(t) - Xij1(t-l) * a‘fijfif-D aYi. (t-D + z (z. (t) - z. (t-l)]* ‘Z‘JT‘I‘, k 1k 1k 3 .1 t- ' By rearranging, we have: 63 Y 1+ 1 *Xfli'£(t) 2b * . . . = . . t— . . 1] (t) l i alJ R,( )) ijfi. t- + k 13k iikc) * mgr—m 1 Yij‘t‘l) 0 Where: X (t) = ijz (t) — xij Q(t—l), and 21km = 21km - 3nd?”- Xijz Note that the intercept term Ai j in Equation 6.1 has disappeared in the manipulation process. In a case of the aggregate production of 150 disembodied technological change, the intercept term (Aij) plays an important role. That is, the term is assumed to be a time-varying variable, and the rate of change in intercept appears in a derived equation, such as Equation 6.3, and is known as total factor produc- tivity growth rate, which is a residual and criticized as a "measure- ment of ignorance" by Schultz and others. We assumed the intercept term (Aij) to be time invarient, since we deal with a sort of embodied technological change. That is, we intard to include all possible sources of productiVity change, and assume any change to be embodied in one of the production factors considered in the production function. Also, note that we have treated the production elasticities or productivity coefficients as if they were time invarient. In fact, the productivity coefficients ofthe function shifters, bijk’ are assumed to be time invarient at the model's present stage, with some exceptions, whereas a. .1 is time varying. But this variable is very 1 slowly changing over time. For this reason, and to avoid extreme complications, we assumed aij 2 to be time invarient only in mathema— tical manipulation to get Equation 6.3. Muation 6. 3 is the final equation used for making yield pro- jections with a minor modification to be explained later. Once the Optimrn input rates of production factors (X132) , which will be discussed in detail in Chapter VII, and the rate of change in Structural variables are known, the yield level (Yij) is ready to be projected. The structural variables considered here, 211(k) /Zik(t-l) , «J. n0 .l “P X 151 have already been modeled in the previous chapter, and are listed here again under different notations. They are: 1. Rate of change in proportion of perfectly irrigated paddy, SCR. (t) 11 2. Rate of change in proportion of quasi—perfectly irrigated paddy, SCRi2 (t) 3. Rate of change in proportion of temporary irrigated paddy, SCR. (t) 13 Rate of change in proportion of consolidated paddy, SCRi4(t) Rate of change in proportion of drained paddy, SCRiS(t) Rate of change in proportion of consolidated upland SCRi6(t) Rate of change in proportion of irrigated upland, SCRi7(t) Rate of change in proportion of new paddy land, SHRfl(t) Rate of change in proportion of new upland, SLDRiz(t) 10. Rate of change in what is called total factor productivity, due to research and extension, YZDij (t) . sesame Note that, because of the metl'od of camputation for SCRi6' SCRi7, SLDRfl, SLDRiz, the corresponding terms in Equation 6.3 must be read as [bijk/Zik(t—l)] * [Zik(t) - Zik(t-l)] and the corresponding coeffi- cients appeared in the computer program must be interpreted as [bijk/Zik(t—l)] rather than bijk for these variables. Product Supply ProjecM W Among 13 crops or crop groups covered in this study, at least two are perennial crops: fruit and mulberry, which are numbered j = 5 and j = 11, respectively. The perennial has a distinctive property in the production process. This necessitates different treatment of Perennial and annual crops. Grass, some varieties of forage, and some industrial crops are also saxdperermial and in this sense are 152 similar to perennials, while in other respects, they are similar to annual crops. For simplicity, we treat these as annual crops. First, we review some useful theories in methodology or model applied to perennial crops. Based on the theories reviewed, we propose a model for perennial crop precinct supply projection. The study of production or supply of perennial crops can be classified into tm categories: (1) the conventional supply response study, and (2) projection of production or supply for agricultural development planning purposes. These categories cannot be independent, hmever, since the projection must be based on what farmers would respond to possible change in economic opportunity in future development. Fernando [F.3] has made a projection of potential supply of tree crops for Ceylon. He indicated that "the state of supply analysis is relatively unsatisfactory compared to demand analysis" and adds that, "There are few really operationally useful elasticities of supply." He essentially assunes that "the area under tree crops has beccme stabilized." In this model, there is no room for economic and other social factors to affect the number of area planted, uprooted or cultivated. He also indicates that "the projection of the average yield per acre presents many more problems than the projection of area." He adds that "average yields are affected by several factors simultaneously and the effect of each on output can be isolated and measm'ed only with extreme difficulty." His principal projection methodology is to first separate areas with traditional tree varieties from those with high-yeild varieties, and small holdings from plantations and then to multiply each of these areas by fixed yield levels, respectively 153 to get total production or supply. (hoe again, there are no economic or other social variables involved in determination of the yield level. Fernando's study is by no means the only one that uses an account- ing approach for projecting perennials or annual crop. We have seen already many examples. The essential point of this nonstructural or accounting model is, as already implied, an assmption that the farmer implicitly does not respond to a better economic opportunity, especially in terms of yield. However, there are many indications that the per- ennial crop farmer does take advantage of the better economic opportunity in the real world, although there are serious debates over the measurement of the producer responsiveness. Bateman [B.3] reviews and sumarizes the supply response studies on tree crops in developing countries in context of the market period (harvest), the short run as well as the long run. After discussing some shortcomings of the models he reviewed, he advances four alter- native models. One common characteristic he introduces is a sort of unique farmers' expectation on prices of own product or alternative crops. In other words, famers' expectations are formulated by Nerlovian expected price. That is, the supply is expressed as a function of the distributed lag prices. In this respect, his approach is ahead for examples of those by Stern [$.16], Wah [W.l], and many others, for example, who define supply as a function of a discrete lagged price, with or without other set of variables as independent variables. However, even Bateman assures, as usual, that the yield level is fixed or given, with the excuse that ”The expected yield pattern of most tree crops is relatively stable, and changes in potential yield occurs slowly." 154 Wah [W.l] sees one important realism correctly. That is, the change in composition of matured tree crop cohort affects the average yield. According to Wharton [W.Z], the Marshallian-Comot distinction between the short run and the long run is desired, especially for perennial crops. He points out that, even in the short run, output is still a function of variable inputs that Batman, for example, recognizes but rules out. More important in perermial crop produc— tion, as he sees it, is that there exists a residual effect of vari- able inputs used. The same thing my be true for armual crops; however, the carryover effect is more significant for perennial crops and should not be neglected. The theoretical consequence of this realism of the carryover effect, according to Wharton, is that "this fact suggests that stimu— lants (variable input uses) are useful for upward response to price but would perversely affect any subsequent downward price movement." We have discussed some important facts that any useful model should take into consideration. First, a change in the composition of tree crop age cohorts is important in explaining changes in average yield. Second, the carryover or residual effect is also important in explaining supply response. The same is true for large animals such as dairy and beef cattle. Third, farmers' decisions on resource use for tree crops are not based on current price levels nor discrete single lag price such as p(t-T) where r = 1,. . .,T, but on a distri- buted lag price. The same thing is more or less true for other crops or livestock. It is needless to say that variable input uses and Structural changes defined in Chapters IV and V as important to Perennial crops as to annual crops. 155 A production function for perermial crops that directly or indir- ectly reflects the theories reviewed above follows (subscripts are omitted to shorten notation): 9. I W£(t)*a£(t-T) 6.4 Y=A* II II X12, (t-T)* i=1 r=o k n 23‘ (t) *Vc(t) x = 2 so—called conventional production factor Zk = so-called noncomrentional production factor and what we call structural change variable 1 <: H age cohort composition 5" m , b, c = appropriate production coefficients 2 II R a weight given to coefficients of current and past input I level, and Z W(t-r) = 1.0 W T = maximum time range affecting on the current output or carryover period. Equation 6.4 can be written: W *a£(t) Y, i=1 1‘ bk (t)* n zk X2“ afi ('3 = :11 X2 MT) T (t-r) T: Where : and 22¢) is here interpreted as a distributed lag input and W20 + W21 = 1.0. Once again, an economic interpretation of this formulation ‘__ LI‘his age cohort ccmposition can be easily computed from the perennial production subcomponent of the KASS model and the farm resource allocation component. 156 is that the past input rate influences the current output level with a weight of W(r). The past input level can be, without generality, represented by a distributed lag value of the input. That is: d? .JQ - = 6.6 T * dt + XMt) XMt) lbw much past input level will influence current output depends on the magnitude of Wo or W1. Parameter Estimation In this chapter, we used a shorthand notation for each variable quite different from that used in the computer program. For conven- ience while discussiong parameter estimation problem, we try to introduce the variable names used in the computer program. For con- venience, parameter estimation problems are discussed for both the variables shown in this chapter and the next as to how to estimate production parameters aijl’ bijk’ cij’ etc., in Equations 6.1 - 6.5. Before discussing this problem, we will briefly discuss the validity of the Cobb-Douglas function. A problem stems from the unitary elasticity of substitution of the Cobb-Douglas function among production factors. In reality, the elasticity of substitution may or may not be unitary, or may vary. Substitution between labor and capital inputs is usually discussed. The estimation method of the constant elasticity 0f substitution for two factor case is given by Arrow, et al. [A.9], and Brown [B.l6], and for a many-factor case given also by ‘Brown [3.16]. Variable elasticity substitution is discussed by Lu and Fletcher [L.ZO] . 157 Even if labor is not one of the production factors included in Equations 6.1 and 6.4, it is suggested that the elasticity of substi- tution between labor and capital for Korean agricultural production be estimated in order to infer the validity of using the Cobb-Douglas function. Hayami. and Ruttan [H.9] followed this procedure to justify ' using the Cobb-Douglas function. On the other hand, Lu [L.21] first checks whether the elasticity is constant or variable. In the case of constant elasticity, he tests whether the elasticity is unitary to make sure the Cobb-Douglas function is appropriate. While testing the validity of the assumption of variable or constant elasticity of substitution, he assumes there are only no inputs, labor and capital. By contrast, he includes many prediction factors whm fitting Cobb— Douglas production functions. Lu's approach seems more or less logical. To follow Lu's pro- cedure, we need to add just one more independent variable, capital input over labor input, to the conventional estimation equation of the constant elasticity of substitution: log (V/L) = a+b logW+c logK/L Where V stands for value added, L for labor input, K for capital input and a, b and c for appropriate parameters to be estimated. According to Lu, when c = O, the function above reduces to the CFS function. When c = O and b = 1, it reduces to the Cobb-Douglas function. When C = 0 and b = 0, it reduces to the fixed coefficient production function. It canbe shown thatwhenb =0 andch, theaboveequa- tion also reduces to the Cobb—Douglas form. 158 However, here we adopt a simple procedure used by Hayami and Ruttan. The derived equation follows: 6.7 log (V/L) = 0.133 + 1.084 log w, R2 = 0.984 Where V stands for the average value added per farm from farming, L for weighted average labor input per farm by age and sex, and W for weighted average of farm wage rates by age and sex. All data cane from yearbooks of Agricultural and Forestry Statistics. The regres- sion coefficient or elasticity of substitution is significnatly different from zero at a high level. The null hypothesis that the elasticity of substitution is not different from unitary is rejected- at 5 percent, but accepted at 2 percent of probability. As a result of this statistical test, we do not insist that projections based on Equation 6.1 will always turn out unitary elas- ticity of substitution among capital items thanself or between labor and capital. The behavior of the elasticity of substitution will be affected not only by relative prices among inputs, but also by the nature of technological changes generated by public policies, proj ects and programs. The insistence of Evenson [E.3] that the elasticity of substitution between biological and mechanical inputs is relatively low, and of Srivartava and Heady [5.14] that the elasticity is changing should be interpreted as consequences of particular public policies. Let us now discuss the problem of estimating production coeff- icients, first for aijJL in the production function in Equation 6.1. Any of several types of data could be used: time series data, cross— sectional data and experimental data; however, none of these sources 159 was readily available and they are expensive to collect. Even if one of then is available, it is not very easy to obtain appropriate signs of the regression coefficients for variables considered. There are a umber of examples showing that one or more of them do not have the signs predicted by the theory. For example, Alcantara and Prata's . IA. 7] production elasticity derived from total cost function has a negative sign for machinery input. Should the farmer be interpreted as behaving irrationally, since less or no machinery use will bring more production, according to his result? What is wrong? It should be interpreted that his survey design, resource classification or specification was inappropriate. This means that it is difficult to secure even the correct sign, not to mention correct magnitudes. Because of this difficulty, we use an indirect method of esti- mating the production coefficients based on the assumption that farmers behave rationally under the perfect competition. Assume the following production function for simplicity: 6.8 Y = aXb The first-order condition for profit madmization is: 6.9 dy/dx = Px/Py or 6.10 dy/dx = baxb'l = bY/X = Px/Py when we rearrange it, we have: 6.11 b = [mm/[Pym 160 Where Px stands for input price, Py for production price, X for input level, and Y for output level. Of course, the left side of Equation 6.11 is the appropriate production coefficient or elasticity, and the right side is called the factor share. In other words, this factor share, which can easily be collected even in the least developed countries [the terminology adopted from (0.1)], is used as a proxy of the production coefficient in this study. This approach to estimating this coefficient is not new in economic literature. Many studies on aggregate production function use this approach, including that of $010 [8.11]. Ray and Heady [R.Z and R.3] use the same approach in formulating a simulation model for U.S. agriculture. Tweeten and Quance [T.5] use this concept to estimate aggregate supply elasticity, and Quance [Q.l] uses it for estimating the marginal value product to examine capital gains and losses, based asset fixity theory. Next, we must ask whether the factor share or productivity coef- ficient is stable over time. Tyner and Tweeten [T.6] find that the productivity coefficient changes over time and diverges from the equilibrium position, and suggest an adjustment model to correct the factor share estimates. This. model is based on the Nerlovian distri— buted lag adjustment model. We approach this problem slightly differently. That is, the productivity coefficient, ALP, is estimated as follows: 6.12 ALPijg (t) = mama) * miji(t)/[PDij (t) * YDij(t)] Where PXD, FXD, PD and YD, respectively, are distributed lag variables 161 of factor price, PX, factor used, FX, product price, PAVG, and yield, YLD. In an ideal situation, it would be very nice if all parameters However, this 'I'nus, what were estimated simultaneously from one data source. would be very difficult, if not impossible, in practice. we have done is to construct the production function after separate estimation of each of the individual partial production elasticities; that is a pragmatic synthetic approach similar to that used by Ray and Heady [R.Z, R.3] in their simulation model. The first attempt is to obtain production elasticities of rice irrigation. Ruttan [R. 10] hypothesizes that rice yield differences among regions or locations are due to differences in irrigation levels. since other technology would be similar among regions.2 Based on this hypothesis, the following production function is fitted, using time series provincial data from 1962 to 1972: 6.13 Y = f (zi, Di) Where Y stands for average yield in each province in each year, Zi for areas urnder various irrigation types, and Di for some dummy variables including trend. The actual model adopted in this study, among many alternative specificationns is a linear function in log, both for dependent and independent variables. The result is shown in Table 6.1. A similar analysis was done by Lee [L.lZ] for Taiwan agriculture. Can it be said that the coefficients of Zi in the first low in Table 6.1 “—n 2This hypothesis appears in several articles, with or without CO~author, for example [H.27]. W‘— 162 represent only the effect of irrigation? An improvement in or addition of irrigation certainly induces us to use more conventional inputs. The above coefficients represent the joint effect of all these factor uses. This joint effect is illustrated in Figure 6.3. Table 6.1. Productivity Coefficients of Rice Irrigation. Independent Variablesa Constant Z1 22 Z3 $231 Trend R Coefficients 1.571 0.200 0.159 0.107 —0.361 0.464 0.435 t-Value 2.33 3.27 2.28 1.01 1.82 1.22 0.02 0.01 0.02 0.30 0.05 0.20 Probability 821 = perfectly irrigated paddy/ total paddy Z2 = quasi-perfectly irrigated paddy/ total paddy Z3 = temporary irrigated paddy/ total paddy Source: Year Books of Agriculture and Forestry Statistics, MAF. 1962—72. D.1e to a change in the other production factor use, the production function has shifted among subfunctions upward and rightward. An empirical illustration of this nature is found in Herdt and Fallor [11.23], Where they derive rice—fertilizer production functions in the U.S. and India and contrast them to show the nature of production function shift Suppose P'P' is parallel to pp, Then the optimum input due to differences in factor use. which is a price ratio line in Figure 6.3. and output level with addition of a 11.617 factor are, respectively, f end a, whereas those without a new factor are e and c, respectively. AS clearly seen, the effect of the new factor on the production level 163 Output , Y i (P Y = f(Xl|X2,. . . wheels1 = 0 Input, x1, (xllx2,. . .,Xn) Figure 6.3. Production function shift among subfunctions due to addition of a new production factor. is certainly not ad, but ab, annd bd is an effect of an increase in the so—called conventional input use from e to f. What we observe from the real world is not ab, but ad. There- fore, to measure the correct net effect of adding a new factor we must first measure the difference in conventional resource use with and without a new factor, ef, and compute an increase in production due to the increase in factor uses, bd. Then the net effect is Obtained by subtracting bd plus ce from af. In mathematical team, this can be accomplished: 164 Z. ..-— 7': .. * w'c * 1; Where : SCEYYY = rate of increase in yield due to technological change in terms of land and water development. YLDPA = coefficients observed from the real world, such as coefficients in Table 6.1 ALLPA - parameter to allocate improved land SCR = rate of change in improved land ALP = productivity coefficients of conventional inputs ASC = elasticity of factor use with respect to change in improved lannd Thus, the first term in Equation 6.14 corresponds to ad in Figure 6.1 and the second term to bd. Implicitly we have stated a method of computing the net effect of adding a new factor in terms of improved land. waever, the same principle is used for computing others, such as variety change, age cohort change, etc. How about productivity coefficients of the other strucmral change terms in the production function in Equation 6.1. Basically, they can be obtained by statistical analysis of farm management data, or we can observe the difference in productivity between improved and unim- proved land from cross-sectional data. One interesting method of obtaining productivity coefficients of irrigation for upland is given by Parvin [P.2]. He contrasts the yield in dry years to that in wet years, using data from eqneriment stations where the techno- logy used is likely stable. At any rate, the estimation technique Seems available for varying situations. However, the necessary data 165 are not available or readily collected. Thus we temporarily use what might be called "gestimate" data based on this author's eqnerience and results from various case studies for structural change terms other than paddy irrigation, total productivity and past input use for perennials. The same type of data is used for ASC, which is elasticiqr of factor use with respect to structural change. In summary, one of the basic structural relationships in this chapter is the production function in Equation 6.1 and therefore the derived projection Equation 6.3, including those for perennial crops. The projection equation (Equation 6.3) can be derived from any form of production function, including exponential functions such as Equation 6.1. There are two reasons for having the Cobb—Douglas production function explicitly in this study: First, with this form of production function, mathematical manipulation is much easier, for example, in deriving factor demand function. Second, this form of production function can reveal certain interaction effects among inde— pendent variables in the production function; however, it was found later when computer outputs were analysed that Equation 6.3 was unable to handle interaction effects among structural change variables. The level of structural change variable is determined indepen- dently from that of the so-called conventional input. In other words, interaction effects among and between so-called conventional inputs and structural change variables are considered separately in Chapter VII. The interaction effect among structural change variables are important. To allow this interaction in projection, there seem to be at least three options: (1) to use the other explicit form of the ‘I in III" |.\\ 4.. 166 production functions, (2) to use the production function in Equation 6.1 directly for projection purpose, and (3) to insert some mechanism to reflect interaction in Equation 6.3. This problem needs to be discussed relative to availability of required data for estimating productivity coefficients, as well as for parameters in factor demand annction that will be disucssed in the next chapter. While collecting data for the purpose of refining the model presented in this study, more appropriate forms of production functions or yield projection equations should be sougnt. CHAPTER VII FACTOR DEMAND PROJECTION Production levels and, hence, supply responses for agricultural products are the consequences of resource use. Thus, demand for pro- duction factor plays a very important role in explaining changes in production and, in turn, growth rates. In the previous chapter, we did not explain low demand for production factor is determined; this is the main subject of this chapter. It is logical for us to first discuss specifically what kinds of production factors we are considering. Production factor can be classified in many ways, depending on the objective of study. Heady [11.11, p. 299-300] gives some expamples and basis for classification. Jolmnson [J .4, and J .16] classifies productive resources in order to study the nature of resource fixity and production response. Our primary purposes in this study are: (l) to explain production response, and (2) to supply the farm resource allocation component model with variable costs for each crop in each region and input-output coeffi— cient requirements. Inasmuch as each production factor classified appears in the production function together with the other classified factors, classification must be based on a useful theory. A basic theory is given by Johnson [J . 2]. According to him, resources or production factors that perfectly complement or supplement each other should not appear in a production function together as independent variables. Otherwise, multicollinearity problems arise. 167 168 For our purposes, and based on a theory suggested by Johnson, the production factor is classified as follows: 1. Fertilizer 2 Pesticides and insecticides 3. Other materials 4 Labor in spring labor peak season 5 labor in fall labor peak season Note that: (1) land is not included, because we are dealing with production per unit of land——yield. (2) What are often called fixed resources, such as buildings, are not explicitly included. Since we are dealing with the service of flow, not stock itself, maintenance, depreciation, and other costs associated with these fixed resources 1 are included in other materials. (3) The other type of fixed resource, farm machinery, is also ruled out, as explained in Chapter III. (4) Still other types of fixed resources, fruit and mulberry trees, are also ruled out since the farm resource allocation and perennial crop production subcomponents deal with that matter. As ecplained in Chapter III, the last two items, two types of labor, do not appear directly in the production function, but demand proj ec- tions for labor are made. How do we go about making proj ections of these factor demands? Despite "While the problems of agriculture are directly those of commodity supply and price, basically they are problems of resource demand and supply" [H.l9, p. 2], it scene that "one of the neglected m; 1This cost item includes costs associated with seeds, buildings, farm implements, materials, etc. 169 areas in agricultural demand analysis has been the demand by farmers for inputs produced by nonfarmers" [C.lO] . Some examples of factor demand studies are found in Griliches [8.9 and 6.10], Cromarty [C.lO], Heady and Yeh [H.ZO], Heady and 'Raeeten [H.IO, p. 154-374], Hayami [H.8], Sung, et al. [$.18], etc. Worthwhile to mention in connection with these studies are: (l) Griliches [G. 10] and Sung, et al. [8.8] use a Nerlovian distributed lag model. (2) Cromarty [C.lO] introduces farm income and interest rate in demand function as independent variables, where Heady and Yeh [H. 20] include income trend in addition to these variables, as a proxy of capital budget. (3) the demand structure connecting dir- ection, magnitude, and duration of price change does not seen well studied, despite Boyne and Johnson [5.13] clearly finding some evidence. (4) The model by Sung, et al. is constructed primarily for making demand projection for fertilizer. Inasmuch as they try to include some teclrmological changes as function shifters, higher credibility should be given to this effort. However, they mriss a very important variable in explaining a change in total demnd for fertilizer. Due to genetic characteristics of crops, some require more fertilizer than others. For example, vegetables, fruits, potatoes, etc. , belong to the former type, and demand on these commodities and, hence, produc- tion are expected to grow more rapidly. Without considering this adjustment, the projection for fertilizer demand would be strongly underbiased. There are also some other evidences for underestimation. They do not consider possible demand shifts such as upland irrigation, drainage, consolidation, Etc- 170 On the other hand, there is another type of factor demand study, which is derived from production function, based on profit meidmization and other static assumptions. The example of deriving factor demand from a production function is give: in Heady and Tweeten [H.19], Heady and Dillon [H.l8], Lee [1..4], etc. A good example of a factor demand study of this nature based on cross-sectional data is give1 in Ruttan [R.9 ], which we will soon examine camparatively and more inter- sively. Similar demand functions can also be derived from a linear programming model through parametric programing techniques [Lee (L.5)] and of budget procedures. Ruttan‘s study cited above projects demand for irrigated land. He begins by saying: "Economists have long been concerned by the fact that projection of resource use have been made on the basis of the quantity of inputs "required" to support some projected level of final output. . .Ihe "requirement is ordinarily determined by applying a factor or coeff- icient to a projected level of output. . .The difficulty with such a projection is that they cannot encompass the tremendous capacity of the economy to adjust to changes in due availability and cost of resources. Requirement projections implicitly assume that the projected amount of the input will be used regardless of the costs of supplying it"[R.4, p. 1]. After criticizing the requirement approach, he advances a metlod for determining the economic demand for irrigation. First, he derives a regional agricultural production function to get the productivity of irrigated land. The production function is specified in terms of not only the irrigated land, but also other relevant production factors customarily used. Once the productivity coefficients and costs involved in irrigation are known, he is ready to determine the optimum use of irrigated land. Then he projects demnd for irrigated land to 1980. 171 He makes two sets of projections, one based on what he calls the demand model and the other based on What he calls the equilibrium model. Both models stem fran what he calls the productivity model, and the only difference seem to be in assumptions. In the demand model, he assures a regional production growth rate, whereas in the equilibrium model, without assumptions as to production growth rates, the optimum input of irrigated land is projected with the usual assumption of profit maximization. Now we are ready to propose an alternative factor demand pro— jection model. The regression analysis approach has several disad— vantages, as noted in previous chapters, but provides many insights to be considered in a useful projection model. In short, again a pragmatic approach based on the structure of the agricultural produc- tion system itself and some basic relevant findings are incorporated into the proposed model. First, we make an assurption of profit maid.- mization. If this assumption and those made in Chapter VI are accepted as a first approximation, projection of factor demand is a mechanical matter based on a simple optimization technique. Whether or not the Korean farmer responds to the economic opportunity and to what degree is an important question. This question along would be a good topic for a Ph. D. dissertation. Thus, we leave this question without inten- sive examination, but call attention to relevant studies such as these by Hub and Lee [H.28], lee [L.4], Seol [3.6], Kim [K.4], and Ferris and Suh [F.4]. The next question to be ecemined is Whether or not the profit Hmdmfization assumption is appropriate. There are endless examples 172 of economic models that adapt this assumption for application in either the LDCs or others, although this assumption has occasionally bee1 challeiged. Is this assumption totally inappropriate? If so, in what respect? It seems that the bad thing is not the assumption itself, but that the researcher often seems unable to consider other behavior of producers. That is, producers, regardless of whether the firm is large or small, and whether or not it operates in an LDC, maximize more than net money return.2 If a researcher fails to recognize this fact, he must be criticized. Is the farmer in the Midwest in the U.S. a straight forward profit maximizer? Instead of answering this question, let us ask if the giant American corporation such as (M or Ford a straightforward profit maidmizer? How do you interpret the fact if they do (or used to) hesitate to hire a Negro worker, other things being equal? If so, should they be called nonprofit-motivated firms? What we want to emphasize is that there are many sets of norma- tive behavior constraints in making production decisions. Researchers ought to not only criticize profit maximization assumptions, but also identify other values that may well have a trade—off relationship with monetary values. Rossmiller, et al. [R.7], for example, identify value constellations for Korean agriculture at the macro level, as seen in Chapter III. Otherwise, we have to reconstruct a body of a new economic theory, after destroying established standard economic ——__ 2For a more detailed study of managerial processes of farmer decision, see Johnson, et a1. [J.13]. 173 theory. Fortunately, the existing economic theory is not so bad that it should be destroyed. Then what are the most appropriate constraints a useful micro model has to consider in order to be more realistic? As a matter of fact, all of the relevant constraints may not be studied, mainly due to a lack of appropriate measurement of variables, such as psychological factors, due to difficulty of conceptualization, etc. Constraints or modividations presented in this study have speci- fically involved resource constraints and behavioral constraints. Apart from constraints of both types defined in the farm resource allocation component model, fixity of capital budget, and uncertainty and resource fixity, which are the modified neoclassical economic theory due to Johnson [J .16, Ch. 3], will be specifically considered in this model. Factor Demand Projection Pbdel With this intorduction, let us constant a demand projection model. The production function presented in Chapter VI is represented here without regional subscript and time index for the convenience in Equation 7. l: . 5L b 'k 7.1 YJ. = AJ. mg}, msz with this production fimction, the resultant profit function is: 7.2 H = ZIP .Y. - ZZP X4 - F - ZRiKi - A[£ZPij£,+F - Kl ym J X’L-J ‘Kz‘Ks’K4+K5] "U1(K1'r(1) ‘ U20(2 ‘ I22) ‘ “30‘s ' 1t3) ' ”40% ‘ ‘4) U5(K5 ' I21) 174 Where: Py = product price Px = input price F = fixed costs (if any) K1 = farmer's own capital used for farming K2 = credit from government-supported institutions K3 = credit from private loans with low interest rates K4 = credit from private loans with high interest rates Ki = respective total available amount of capital budget A’Ui = respective lagrangian multiplier K5 = famer's own capital disposed for nonfarm uses 1% = appropriate interest rates paid or received . The meaning of Equation 7.2 will be self-evident to the reader familiar with the elementary matheratical ecoromics. That is, the profit real— ized is defined as the difference betweer total revenue, ZPyjYJ. , and total costs, ZZPXRXj z + F + ZRiKi, but (1) total expenditurestZng Xj J, + F, should not exceed the capital budget made available, K1 + K2 K3 + K4 - K5, (2) farmer's own capital used for farming should not exceed that available, (3) credit from government source or others should not exceed what is made available, etc. and, (4) farmer's own Capital salvaged should not exceed what he has. The first-order conditions for profit maximization are: OLY. 3H — _J_ _ =0 7.3——P. —Px.£ 1ng 2)ij yJ OLij 3H _ _ ~ 7411--Rw+)\ 1.11:0 175 7.5 -—=-R2+A-u2_>_0 3H __ 7.6 mg—'R3+A-u3_>—O 7'7 Tf‘Rafl‘Uno 7'8 W=‘R5‘A‘“530 H 7.9 ——= . - _ _ _ _ ZZPx£XJ£+F k1 k2 k3 k4+k5—0 7.1oin—= 42—30 7.llai=k2- >0 7.123i=k3-E§30 7.133i=k4-§30 7.143—”-=k5-§30 The factor demand function of Xj g is defined as the solution of ElNations 7.3 - 7.14 simultaneously. That is, the optimum input rates and borrowing or disposal rates are determined by the system of equations above. But how do we solve this system of equations simultaneously either to get the optimum rates or to derive factor demand function? Heady and Dillon [H.181 and Heady and TWeeten [H.191 discuss the solution methods for the optimum input rate generally or the uncon- Strained Cobb-Douglas production function. lee [L.4] discusses a method of deriving it from the constrained production function in quad— ratic form. Iau and YotOpoulas [LB and Zarelbka [Z-l, Ch- 8] provide 176 a method of driving the factor demand function from the unconstrained Cobb-Douglas production function. At any rate, there is without a doubt no analytical solution for the above system of equation. Then how do we go about deriving factor demand function subject to constraints imposed? We are going to give an indirect solution method here, without losing any generality. Let us first write Equation 7.3 in explicit form by substituting Yj in Equation 7.1, assuming that there are three production factors. Then we have: a. -1 a. a. b. ~ an = P}, 31 32 J3 1k _ - = . . —l a. b..— an _ I— 831 a32 33 3k _ _ = 7.16 3X]; — Pyj l—Ajaj2 le ij X33 112k sz Asz2 0 a. a. a. —l b. 311 __ 31 32 33 3k _ -}.Px=0 where the lagrangian multiplier is differentiated among inputs, just for exposition, but this differentiation will soon be withdrawn. At an equilibrium, all the variable inputs slould be used so that the least cost combination is secured. The least cost combination between le and ij is secured when the following condition holds: (1 + Al) le/Py. BY. BY. 3X.2 a.1X.2 7.18 —J—/—J—=—J-=J—xi—=ZI‘+TTH27# which can be simplified as follows: -1 _ (l + )1.)le (1 + )2)sz 7 19 X —-——-—-—-——" a'l-lajzle 177 which is called the iso—quants equation [see Heady and Dillon (H.18)]. The iso—quants equation between le and X33 can be similarly derived as follows: -1 (1 + n1)px1 [(1 + n3)1>x3] _1 Pyj Pyj J an E”5'3le 7.20 X33 = Now let us substitute Equations 7.10 and 7.20 for X. production function of Equation 7.1 to obtain: a. a. a. b. = 31 -l -l 32 -1 .‘1 J3 1k 7.21 Yj ijjl [V1V2 ajl ajZle] [VlV3 ajl aj3le] HZk where: (l + )‘iwxi Py. 7.22 Vi = J Noting that the production function in Equation 7.21 is now a function of le alone, it can be rewritten as follows: -aj 3 -aj 2-aj 3 ajz aj 3 ajz ajl ajz aj3 a. +a. - _ 32 J3 7.23 Y3 — Ale V2 V3 a. 4a. +a. b. 31 32 33 3k Km "4. The Optimum input rate of le is determined when the marginal value product is equal to the marginal factor cost, or the marginal PhYsical product is equal to the price ratio; that is: 178 BY. a. +a. -a. -a. ...L= 12 J3 12 J3 7.24 2)le [ajl + ajZ +aj3lAjV1 V2 V3 a.1"aj2"aj3 a.2aj2 a ajs x ajl'i’ajz'iajB‘l J J 1'3 J'1 11251.1( = (l + )\1)Px1 Pyj Solve Equation 7.24 in terms of le in order to derive the factor demand function: — a. +a. -1 -a. -a. -a. -a. a. a. = J2 J3 J2 33 32 J3 .33 J3 1 LSjAjvl 1 V2 V3 a31 ajZ aj3 b—lI-S- Hzk 3:! J 7.25 Xj Wher S.=a. +3. +3.. e J 11 12 J3 Now let us assume that the net marginal returns to capital expenditure on each factor, AIL: are equal to each other. In fact, this is essentially one of the efficiency criteria for resource use. Equation 7.25 can then be written: —s. s-l ' a. -a. -1 J 1% 7.26 . = . .a. J .Px 1+ Ha. sz le [AJSJ 31 PyJ 1 ( A) [21“ 2 J _1_ b. l—S. 3 J “2.. Which is the final derivation of factor demand function of le. The factor demand function can be written more generally as follows: 179 -S. s.-l -l a. -a. .2 . = A.s. a. J . J + ma: 3% 32 7 7 XJn [ JJ Jn Pyth (1*) L32 sz 1 b. "1* HZkaJlSj where n is a dummy to indicate which input among IL inputs we are taking under consideration. Before proceeding to the next step, let us digress to examine some important economic relationships from our final factor denand function. Let us define the elasticities of denand for factor le with respect to own price, cross prices and end product price, respectively, as follows: =_J'l.___=J' ' 7.28 51,1 an le .1133— 1 3X. P -a. 7.29 81,2 =31?Ll . —x-3=1—_§3 X2 .11 3 3X. P -a. 7.30 e1,3=§§l]—‘~)—(-x§=1:§—3- x3 31 j ax P . o1 - l 7.31 €1,y = "_‘3P '.' ' X0 _ 1_S. YJ J1 J Observe that the following relationship lnolds: 3 7.32 E el.i = -el.y i=1 The above relationship insists for example, that when all factor Prices and product price increase by the same proportion, there is no 180 change in optimum irnput level. Would this be true regardless of the capital budget level? Clearly not. Then what is wrong? The net marginal returns to capital expenditure, )1, is a function of all kinds of prices and budget constraints, in addition to technical coefficients. In order to derive the correct and true relationship and go ahead to the next step toward solving the system of equations in Equations 7. 3 - 7.14, we insure that total expenditures will not exceed total available capital budget; that is, we make the relationship in Equation 7.9 hold. What we need to do is: (1) substitute the factor demand function for all inputs for each product in Equation 7.9, and (2) solve the resultant equation in terms of the net marginal returns to capital expenditure, A, and (3) substitute the resultant equation for A into the individual factor demand function. Then the final equation will be true factor demand function constrained by capital, and we will be ready to derive more realistic homogenity or other relevant consistency conditions. Unfortunately, there is no analytical solution for this particular type of production function. This author [L.4] has shown that the fimdamental relationships in commdity demand advanced by Frisch [F. 12] fold equally in factor demand with a set of production functions in a form of quadratic function. The relevant relationships are Engel aggregation and homogenity condition. The former says that the sum of budget elasticities weighted with the budget proportion P max]. £122 PM. Xj 2 is unity, and the latter says the sum of demand elasticities With respect to own and cross factor prices is equal to budget elasticity in the absolute value for a commodity. That is, the homogenity condition in Equation 7.32 no longer exists in a case of 181 a constrained profit function. Instead, when all factor prices and budget increase by the same proportion, there is no change in factor demand. This slould be interpreted carefully. This relationship holds only for marginal changes in the neighborhood of the previous optimum input levels. Let us now go back to the main subject. To insure that the relationships in Equations 7.3 - 7.14 hold, Equation 7.27 is expanded by the Taylor series, as in Chapter VI, to obtain: :3“) 8.. 7.33 x.. = [1+T_ 1117—5 +;;_1.1& 1J2 L1 Sij zl‘Sij P . (t) {7. b.. x11 1 1 t 1 2k ——(——D—+ —§—2D +2 I‘d”— P .2 t‘ l-Sij Yi t" k " ij zlkc) Zith-IS .kxijnlfit 1) Where y = l + A, and i for region, 3 on crops and IL for factor. This is a final projection equation of factor demand for annual crops. For the perennial crops, the appropriate terms should be added. But y, is still an unknown variable. Now in order to make Equation 7.9 hold and to determine 7, let us substitute all factor demand functions into Equation 7.9, and after including particular terms for perennials and rearrarngenant, we have 182 7.34JZZIPXl xi£ Xij£(t)=zz M(t)X £(t'l) P ..(t) + 2: 2: alijz $316271: . qu(t) Xij£(t- -1) - z z 2 EX—igéifip ox (t-l) a ijJL Pxiz t- Pidé ijIL I3 . (t) - X 2 G3,. x12. 13% mm}; t- Pxil(t) Xij£3 PXi Q(t) X. j£(t) in Equation 7. 34, solve the resultant equation in terms of Yi(t) to obtain: 13 ..(t) . _ I— 1442210.: 1 1+2): 311.21, t_ ijz. J yij II 7.36 Yi(t) P . (c) I3 . (t) x12 _ x11 ’ 22052131 P_.£(t—1) £203in mezt-IS ik(t) 222015in —-—-(——I) + 183 7. 36 (cont.) Xi(t'1)Ki(t) - _ - Y-.(t-l) 220.413. 2? ma) xijza 1) i which is the final projection equation for the gross marginal returns to capital expenditure for variable inputs. Let us make sure we understand the meaning of y; that is: P ..aY.. = = M 7.36 Yi (l + Xi) axijy, / Pxil where A is a sort of net marginal rate of internal return to capital. With this mderstanding, let us play a game to insure that the relationships in Equations 7.4 — 7.8 and 7.10 — 7.14 hold, since we have already made Equation 7.3 and 7.9 hold. And let us postulate the following stepped capital supply function. First, substitute the farmer's own capital, K1, into Equation 7.36, and solve it. The possible outcomes are: 1. yi>R1 2. Ai=R1 3. AiR2 2. Ai=R2 3. Ai_Riup to K=K1 + K2 + K3 + K4, where the latter two are again credit from private noninstitutional sources with different rates of interest. Every reader will realize that the restirctions in Equations 7.10 - 7.14 are clearly fulfilled by this type of game. How about the restric- tions in Equations 7.4 - 7.8? Ri's are some sort of market prices of capital and Ui are some sort of shadow prices for the constraints imposed. [Ri + Ui] can be interpreted as the opportunity cost of each source of capital. Again, by the game played above, the internal rate of return, A, in each region is greater than or equal to the oppor« tunity cost, Ri + Ui‘ In other words, whenever A is less than R1 + Ui’ that source of credit is not used in the corresponding region. Now we have proved that our solution satisfies all conditions imposed on our profit function. However, this does not necessarily prove that our solution is stable. But thanks to the facts that: (1) individual productivity coefficient is designated to be greater than zero and (2) the sum of these productivity coefficients is also less than unity, for each individual crop, as inferred from Chapter VI, the so-called second—order conditions autamatically hold for this particular system. Once the optimum rate of internal return (A) for each region is determined, we are ready to project factor demand with Equation 7.27, in turn, to project yield levels with Equation 6.4. i 186 When all these dexand fmctions are substituted into the pro- duction function, the resulting function is called the supply function. However, we do not try to present the resultant equation here. Instead, let us compute some relevant variables involved. The reader may wonder how to compute the farmer's own capital used for farming and nonfarming, and capital borrowed from government sources, for example, whex the marginal value product curves are NMPVl and NMVPB, respectively. First, we compute the total expenditures, which are not weighted by area planted to each crop in each region such that: 7.37 svci(t) = ZZPXiz(t) inj 2(1:) where PXiIL stands for input price and FXi j 2 for factor input levels used in the computer program, respectively, for P x'il and Xij 2' If this total expenditure, SVC, is less than the farmer's own capital available (FOK), that is: 738 SVCi(t) < FOKi(t) then we know that the farmer's own capital used for farming is SVCi, which is K1* in Figure 7.1, and that for used nonfarming, K1 - Kl*, is FOKi(t) - SVCi(t). Likewise, when: 7.39 SVCi(t) < [FDKi(t) + GLi(t)] where GL is the government loan in caiputer program, K2, then actually borrowed credit from government source, K2* - K1 = SVCi(t) - FOKi(t) and so on. let us try to understand the meaning of capital budget constraint T «.4 187 used here, thereby, total expenditure, SVC. The concepts of these variables are somevhat different from camnn sense. Since we wnat to first determine the intensity of factor use per land unit, total e4 a) 4.0 " , .0 a ' I f g ' KASS Yield 9.0 , , E 3.0 1’ 2.01 L L 1 l I J j I 71 73 75 77 79 81 83 85 Year Figure 9. 1. Average national rice yield level projected under three policy levels on research outcomes, specified in Table 9.1 and that projected by KASS under policy alternative 11. 243 relatively low, however, extension is certainly one of the effective measurements of attaining development goals. Table 9.4. Pbdel Response to Charnge in Extension Budget in 1985. Response Variable Unit Extension Budget Change Relative to Basic Run 0.75 1.00 1.50 Average Rice Yield Ton/ha 5. 798 5.879 5.993 Total Rice Production 1000 Ton 7,301 7,405 7,547 Table 9. 5 shows the national average yield response to the product price level. The price elasticity turns out to be small. In fact, product price levels alone may not be a very effective means to affect the physical production level. That is, price policy should be interpreted as a complementary factor with technological change and also determined from the viewpoint of the terms of trade or welfare consideration for the farm sector. As seen in Table 9. 5, total value added changes proportionally with the price level change. Table 9.5. l’bdel Response to Change in Product Price in 1985. Response Variable Unit Product Price Change Relative to KASS Alternative 11 Price Set ‘ 0.8 \ 0.9 \ 1.0 Average Rice Yield Ton/ha 5.858 5.879 5.907 Total Rice Production 1000 Ton 7,379 7,405 7,439 Total Value Added Billion Won 2,308 2,660 3,001 244 How about response to changes in factor prices? As shown in Table 9.6, and as expected, production and, hence, total value added decreases and total cost increases as the factor price level increases. Again, the structural elasticity seens very low and policy for factor price should be aslo determined by considering not only production, but also welfare of the rural sector, more specifically income redistribution, together with product price policy. Table 9. 6. Mndel Response to Change in Factor Prices in 1985. Response Variable Unit Factor Price Change1 Relative to 1970 Level low Unchanged High Average Rice Yield Ton/ha 5.881 5.879 5.844 Total Rice Production 1000 Ton 7,407 7,405 7,360 Total Material Costs Billion Won 304 320 322 lFor 'hign” factor price, prices of fertilizer, pesticides and other materials are 1.5, 1.5 and 1.2, respectively, as compared to those in 1970, and for "low" factor price, those are, 0.8, 0.8 and 0.9, respectively. The standard textbook on micro economics teaches us that produc- tion response is zero with respect to proportional changes in product and factor prices. That is, when both prices change by the same pro— portion, there would be no change in factor use and, hence, in production. Ch the other hand, according to Tables 9.6 end 9.6, the supply elasticity with respect to product price change seems higher than that with respect to factor price change. There are some reasons for the difference. Production is certainly geared to producer incentive. This incentive can be measured by economic 245 returns to producer‘s owned fixed resources, land and labor. When the share of the rest of production factors is relatively small, the level of producer' 3 incentive, value added or profit is much more sensi- tive to product price change than to a change in factor prices. There— fore, it does not matter whether the product price is increased by, say, 10 percent or prices of factors supplied by the nonfarm sector are decreased by 10 percent. The level of producer's incentive is considered in adjusting factor demand elasticities (see Equations 7.40 to 7.42). That is, the profit level, among others, is supposed to shift the elasticity function in Figure 7.2. This profit level is also assumed to accelerate the adoption of new technology (see Equation 5.5). At the same time, farmer-owned capital is defined here to be proportional to the gross revenue. Hence, changes in product price will affect factor use more than changes in factor prices. A lower elasticity of rice yield with respect to factor price change stems from other model assumptions, too. The most important reason is that a lower value of demand elasticity for factor is assigned since it is found to be extremely low [see Lee (L.4) ]. Second the productivity coefficient of the so-called conventional input is positively related to changes in factor price (see Equation 6.11). This means that when factor price increases: (1) demand for this production factor will decrease, but (2) production elasticity tends to increase since factor demand elas- ticities with respect to factor prices are far lees than unit. Thus, the effect of factor price changes on production response will some-mat cancel out each other. Tables 9.7 and 9.8 present model response to government credit Policy, first for the amournt of credit supply and second for interest rate. 246 Table 9.7. Model Response to Change in Government Credit Supply in 1985. Response Variable Unit Government Credit Supply Relative to Basic Run 0.5 1.0 1.5 Average Rice Yield Ton/Ha 5.879 5.879 5.880 Total Rice Production 1000 Ton 7,405 7,405 7,405 Total Capital Cost Billion Won 56 54 54 Table 9.8. Midel Response to Change in Government Interest Rate in 1985. Response Variable Unit Government Interest Rate Relative to Initial level 0.5 1.0 1.5 Average Rice Yield Ton/Ha - 5.894 5.879 5.861 Total Rice Production 1000 Ton 7,424 7,405 7,381 Total Capital Costs Billion Won 42 54 6O The government loan rate does not affect physical production. This seemns to sten from enough credit being available from the noninstitu— tional private money market. But the farmer is asked to pay a higher interest rate and this fact is reflected in total capital costs in Table 9.7. On the other hand, a change in government interest rate affects physical production slightly and total capital costs greatly. Thus, both policy variables also influence income distribution. Thus far, we have assumed that only one policy variable is variable, other policy variables being kept at the medium level, as Specified in Table 9.1. let us now examine what happens when all 247 policy variables change in directions favorable or unfavorable to the farm sector. That is, all the policy variables are assumed to change from medium levels to lower or higher level at the same time. let us look at what happens to total gain production. Grain as defined here includes rice, barley, wheat, other gains, pulses and potatoes (whose yield level is already specified in terms of grain equivalent). The result is shown in Figure 9.2. For comparison, total gain production projected by the initial version of the KASS under policy alternative II is also presented, denoted by * in 1971, 1975, 1980 and 1985. Again as seen in the figure, the present projection, regardless of the policy level, is higher than the KASS projections, especially after 1975. The possible reason for this was already discussed. Total consumption needs (to be taken into account of the market losses or production deflators) are also shown in Figure 9.2. In order to achieve food self-sufficiency earlier, more development effort has to take place earlier, since with the higher policy level, it is only possible to achieve the food self—sufficiency after 1979; with the medium level, after 1981; and with the lower level it is not possible even by 1985. We will come back to this problem later. Going back to Figure 9.2, note what happens after 1975. Remembering that all policy levels except the biological research outcomes change. Total grain production is more or less depressed for 2 or 3 years right after 1975 with the lower policy levels. Total grain production is accelerated from 1975 with the higher policy levels. In any case, grain production is growing smoothly. This is so because we assured that all policy levels were constant over time, that the 248 14 r- Medium 13 . Total Grain Production (Million 12 . Ton) Lower KASS Projection ll. of total consump- tion needs of gains 0 KASS pro] ection of total grain production t a n a a L I 4 1971 73 75 77 79 81 83 85 Figure 9. 2. Total grain production projection based on three different policy levels, specified in Table 9.1, projected by KASS under policy alternative I denoted by 7" and consumption needs denoted by 0 in 1971, 1975, 1980 and 1985. biological research outcomes materialzied contirnuously, and we did not consider a main factor causing production to fluctuate over time (weather). let us find out what happens to other response output variables due to a change in the level of all policy variables. Partial results are shown in Table 9. 9. Notice that all output variables specified are responding well to the policy input levels. As noticed earlier, total rice production here is high as compared to the KASS projection 249 which is 5,451 for 1985. The KASS projection of total material costs is 141, which is lower thann that under any policy level considered here. Total value added projected by the KASS is 1,166, which is higher than that under the lower policy level, but less than that for the other levels. However, it is worthwhile to note that the KASS total value added includes all agricultural commodities. Thus, we should say that all proj ections made here are considerably higher than those made by the KASS under policy alternative II. This fact stems directly from the higher yield levels of major crops, which, in turn originate largely from superior biological research outcomes. Material costs also turn out higher in order to support higher production levels. The labor demand projection by the KASS is available, but we hesitate to compare these results since we did not consider mechanization in this model. Table 9.9. lVbdel Response to Change in the level of All Policy Variables in 1985. Response Variables Unit Policy Level lower mdium [Higher [Likely Total Rice Production Million Ton 6,564 7,336 8,119 7,389 Total Material Costs Billion Won 174 187 207 231 Total Capital Costs Billion Won 41 26 13 37 Total Value Added Billion Won 1,019 1,361 1, 716 1,296 Total Labor Demand Million Hrs 5,356 5,085 4,812 5,113 250 lastly, we want to present the national average yield for individual crops, projected on the likely policy level set. These results appear in Figures 9.3 through 9.6. As compared to those projected by the KASS under policy alternative II, yield levels of rice, wheat, fruits, vege— tables and potatoes projected by this study are higher, those of barley and tobacco are approndmately the same, and those of other grains, pulses and industrial crops, which are minor crops, are lower. As indicated earlier, the sources of the difference seen to stem largely from: (1) difference in initial yield levels, (2) assumed biological research outcomes, and (3) some other model assumptions, such as change in age cohort of perennial crops, past input use effect on yield level for perennial crops, etc. At any rate, compare the gowth rates of yield with perceived accumlated rate of increase in yields due to the biological research in Table 5.1. There is a close correspondence between them. Thus, we can conclude that the biological research is the crucial factor affecting physical productivity while other policy variables are complementary with research outcomes or measures for attaining other development goals, such as income level, income redistribution, reduction in uncertainty, labor requirenent, costs, etc. This conclusion does not imply that the other policy measures are unimportant. What we mean is that biological research and diffusion of results are jointly more important in achieving physical productivity gowth. In summary, the results of policy experiments in this chapter are consequences of assumptions (many of which are based on poor data) of the model presented in this study. In almost all chapters, we have 6.0 5.0 E \ § ,6- 4.0 H a) -.-n >-‘ o §é° o E 3.0 Ti .3 U Q Figure 9. 3. 251 / Tobacco ’ .— ....—_-——— ‘ ...—fi/d ’l/Other Grain ‘ . ‘ . l l 4L 1971 73 75 77 79 81 83 85 Projection of yields of rice, barley, tobacco and other grains, based on the likely policy level set. 252 6.0 ’- Potatoes / 5.0 - / /// 33 4.0 - 8 4..) 'o‘ r--l 3 >0 3 0 Wheat q) I g ’/ o if / ,9, 2.0 - ‘a‘ Z 1.0 - ”MI/Pulses ” “M 1971 73 75 77 79 81 83 85 Figure 9.4. Projection of yields of wheat, pulses and potatoes, based on the likely policy level set. 253 1.1r Industrial Crops / / o.9_ l/J / 0.8. / -35 rage E \ 5 0 3E 0.7, / o 52"“ Forage Crops / $03 0.6. .30 it? .'> <1 0.5. Hg / / go / / cow 2 ’/ 0.3 - Silk /,,—~/ 0.2,, 1971— 73 75 77 79 81 83 85 Figure 9.5. Projection of yields of silk, industrial crops and forage crops, based on the likely policy level set (scale in right—hand side for forage, and that on left for others). National Average Yield of Forage ton/ha L VEgetables 15 14- 12 l/’ Fruits ll- / 9 - ‘_'///////l 1” Grasses ... u". l l I I + 1971 73 75 77 79 81 83 85 Figure 9.6. Projection of yields of vegetables, fruits and grasses, based on the likely policy level set. 255 discussed weakness of the present study, especially, in terms of data. This author does not intend to project the future values of the relevant variables accurately in this study but, instead, to make a model that can be used to project these variables when better data become available. Thus, the results presented here should be inter- preted as tentative. The purpose of the computer experiments in this chapter is to aid further model refinenent. The possible areas to be refined conceptually have been discussed in each mathematical model chapter, and the need for better data has been discussed repeatedly at various appropriate points. PARTIV POLICY IMPLICATIOI‘B AND CONCLUSIONS CHAPTER X POLICY IMPLICATIONS AND CONCLUSIQIS After constructing a mathematical model in Part II, based on the theory presented in Part I, we tested the model in Chapter VIII to determine whether it works properly. We also tested the model to determine whether it responded properly to changes in policy inputs, and whether it can help identify a set of policy variables and their levels that can contribute to attaining the development goals in Clnapter IX. It is in order that we draw policy implications and evaluate the model. Part IV contains two concluding chapters. In the first chapter, we seek policy implications and conclusions based on the simulated results of the model. At the same time, we want to discuss again the weaknesses of our model and further stLrly needs to make the model more realistic for policy making. Finally, Chapter XI sunmarizes the overall study, including conclusions. What kinds of conclusions can we draw from the simulated results of the model in terms of policy recommendations? In discussing these policy implications, we must keep several points in mind. First, the object of the study was to project yields for specific crops under consideration as a result of public policies, projects and programs designed to directly or indirectly affect the production level. A Change in the yield of a crop or a group of crops is likely to affect 257 258 the pattern of resource allocation. Changes in yields and land alloca— tion will certainly induce change in the producer price level and structure. This, in turn, induces change in the yield level and allocation of land, labor and other production factors. This reper- cussion takes place unless the demand elasticities for different commodities are identical and the yield level and factor demand change proportionately among crops. However, these conditions can hardly hold. In other words, we are not able to fully evaluate the policy alternatives from the model presented here Lmless it is linked with the resource allocation and camodities demand component of the overall KASS model. Second, the policy input signals in this model are kept as simple as possible. Thus, the model provides limited information. In order for the model to supply more useful information for policy making, interaction with public decision makers is needed to determine (their) interest and the general direction of policy variables, programs and projects. Third, a simple mechanism has been adapted for certain behavioral relationships. Take the farmers' own capital for investment for example. As indicated earlier, farmers' consumption, saving and invesment decisions are more or less jointly determined. The mechanism for this joint determimtion is not clearly known. Since our simple mechanism does not fully reflect this complicated process, errors in making pro- jections can be expected. Fourth, several important variables that may affect production rates are omitted. Transportation, electrification, etc. , are examples, 259 In addition, in the process of economic transformation, farm size is likely to increase, which will likely affect productivity. These relationships are not included in the present model. Lastly, the data base for themodel is poor. That is, inadequate checks have been made on how well the data used here represent the real world situation in many instances. This may cause a strong bias in drawing policy implications. With these reservations, let us now return to our main subject. We will discuss policy implications exclusively in terms of production, partly because the present model is designed for the production system, and partly because other development goals can be better evaluated when the model is merged with other model components. To evaluate policy alternatives, we must first identify those policy variables that will affect the system. We have specified one set of conventional production factors, and, at the same time, another set of structmral change variables in our production function for each crop in each region. How much is each of these variables conceived to be contributing to the growth of productivity? We have three different conventional inputs and ten structural change variables for each of 13 crops in each of three regions. Thus, it becomes too complex to discuss all these variables for all crops at the same time in order to draw policy implications. Therefore, we select two crops in Region 1, for illustrative purposes--an annual crop, rice , and a perennial crop, tree fruit. Structural variables are grouped appropriately, as shown in the following tables. Table 10.1 shows the sources of rice yield productivity growth 260 Table 10.1. Sources of Yield Productivity Growth Rate (in Percent) for Rice in Region 1. Based on mdium Policy Level Set Specified in Table 9.1. Year Conventional ‘Research and water and Land New Land Total Input Use Extension Development Development Change 1971 0.11 1.18 1.18 -0.03 2.44 1972 0.34 2.30 0.84 -0.01 3.47 1973 0.75 5.69 0.56 —0.06 6.94 1974 0.37 2.12 0.46 -0.02 2.93 1975 0.35 1.78 0.42 —0.03 2.52 1976 0.45 2.87 0.40 —0.25 3.47 1977 0.26 2.12 0.38 -0.22 2.54 1978 0.33 3.01 0.37 -0.15 3.56 1979 0.79 7.54 0.36 -0.08 8.61 1980 0.27 2.43 0.35 -0.03 3.02 1981 0.31 2.75 0.34 -0.00 3.40 1982 0.23 1.94 0.33 0.01 2.51 1983 0.32 2.73 0.32 0.01 3.38 1984 0.80 6.68 0.32 0.02 7.82 1985 0.24 1.89 0.32 0.02 2.49 Ibtal 5.92 47.03 6.95 —0.082 58.98 Average 0.39 3.14 0.46 -0.05 3.93 from 1971 to 1985. This projection is based on.what we call the medium policy level set, defined in Table 9.1. For the 15-year period, rice productivity increases by 6 percent due to increased use of con: venticnal inputs, 47 percent due to reseaICh and extension, 7 percent due largely to water development, and decreases by 0.8 percent due to 261 bringing marginal land into production for a net total of about a 60 percent increase. The simple average annual growth rate is 0.4, 3, 0.5 and —0.0005 percent, respectively, in the order of the factors listed above, and the annual total average growth is about 4 percent. As seen in Table 9.1, the medium policy level set assumes the factor price level remains unchanged at the 1970 level. Thus, the sources of conventional input use change are other than changes in factor prices. As shown in Table 10.2, in the case of fertilizer for rice in Region 1, the sources of factor use change are product price change by 0.45 percent, research and extension by 7 percent, and land and water development by 0.3 percent annually. Average annual total growth rate is about 8 percent, and total fertilizer use increases by 117 percent from 1971 to 1985. The effect of research and extension and land and water develop- ment are computed such that the effect of change in the conventional input use induced by these structural changes are subtracted from the gross effect. On the other hand, the effect of research and extension includes not only research outcomes made available by the public sector, but also innovations by leading farmers. Table 10.3 presents the same sources of productivity growth for fruits in Region 1. But two additional sources of the growth are considered: past conventional input use and age cohort changes. Total and annual average growth rates for 1971 to 1985 are 62 percent and 4 percent, respectively. The order of importance of the sources is age cohort change, research and extension, current inputs use, past input use, and land and water development. As expected, addition of IIEicrginal land causes the average yield to decrease. 262 Table 10.2. Sources of Growth Rate of Fertilizer Use (in Percent) for Rice in Region 1, Based on Medium Policy Level Set Specified in Table 9.1. Year duct Own Cross Budget Research Land and Total Price Price Price Change Extensions water Change Change Change Development 1971 0.0 0.0 0.0 0.89 2.67 0.50 4.06 1972 0.16 0.0 0.0 0.0 5.18 0.56 5.90 1973 0.63 0.0 0.0 0.0 12.73 0.49 13.85 1974 1.44 0.0 0.0 0.0 4.72 0.41 6.57 1975 1.90 0.0 0.0 0.0 3.93 0.36 6.19 1976 2.03 0.0 0.0 0.0 6.33 0.32 8.68 1977 0.38 0.0 0.0 0.0 4.67 0.30 5.35 1978 0.16 0.0 0.0 0.0 6.62 0.28 7.06 1979 0.07 0.0 0.0 0.0 16.62 0.26 16.95 1980 0.03 0.0 0.0 0.0 5.37 0.24 5.64 1981 0.01 0.0 0.0 0.0 6.11 0.23 6.35 1982 0.01 0.0 0.0 0.0 4.32 0.22 4.55 1983 0.0 0.0 0.0 0.0 6.08 0.21 6.29 1984 0.0 0.0 0.0 0.0 14.93 0.20 15.13 1985 0.0 0.0 0.0 0.0 4.23 0.19 4.42 Tbtal 6.82 0.0 0.0 0.89 104.51 4.77 116.99 Average 0.45 0.0 0.0 0.06 6.97 0.32 7.80 263 0H.¢ No.0- H0.0 sm.H 0N.H se.0 H0.0 «mammsa N0.~0 em.0- 0H.0 sm.0N NH.0~ s0.0 HH.0 H0009 maum No.0 H0.0 00.0 00.H 0s.0 0e.0 000H 00.0 N0.0 no.0 00.0 00.H 00.0 s0.0 e00H 0a.m N0.0 H0.0 0H.N 00.H N0.0 00.0 mama 0e.q No.0 H0.0 0N.H 00.H «0.0 00.0 N00H 00 m 00.0 H0.0 0s.0 sw.H 0s.0 00.0 HO0H 00mm No.0 H0.0 00.0 s0.H 00.0 H0.0 000s 00.0 0.0 H0.0 m~.0 s0.H «0.0 0~.H mama 00.0 s0.0- H0.0 HH.N 00.H 00.0 0w.0 mama 00.0 0N.0- H0.0 0H.H 00.H 00.0 s0.0 seas 00 0 0N.0- H0.0 00.0 00.H Na.0 Ne.0 0N0H «sum 0H.0- no.0 N0.0 00.H 00.0 ms.0 mama N0.0 No.0 H0.0 w0.H 00.H 00.0 ms.0 sums 00.0 HH.0 H0.0 00.H 00.H 00.0 as.0 ms0a 00.0 0N.0 H0.0 0s.0 00.H s0.0- 00.0 mama 0H 0 00.0- 00.0 0s.0 00.0 00.0 0N.0- Hams owqmou pager/on c9583 owcmoo ow: owmmoo “8839/3 EH3 Ea fiasco madam 8: smash H309 pg 062 new swumz noummmom ow< ummm 3038280 Home 4.0 03a. 3 Bahama 8.... H83 radon gems . . no nomad .H gwwom 5 mug pom Sachem ad 3mm £380 figfiuooooum 3on mo 80.58 m 0H 3an |||| 264 A change in age composition of tree crop will certainly affect average yield. However, it is doubtful that fruit yields actually increase by 1.7 percent armually for this reason. A possible source of bias is clenge in age composition over time. We used some rough tentative data because the correct figure will be generated when this model is linked with the other components. The other source of bias is the response elasticity. This is applicable for all structural change variables, however. What conclusions can be drawn from the analysis of the two cases indicated above? One may easily conclude that the highest—payoff input is tectmological change made possible by biological research and dis- semination of its results. Cochrane [0.5, p. 88-90] observes that two basic factors contributed to an increase in total farm output in the United States: farm technological advance and an increase in the size of the total fixed plant-land. He notes that the former was a minor cause and the latter a major cause during the nineteenth century, whereas the former was the major cause and the latter minor in this century. In another paper, he [C.6, p. 46] claims that "the engine of modern farm production is farm technological advance." No one would deny validity of this conclusion, even for develop- ing countries, if the productive land frontier has been exhausted. However, we need to understand the mechanism of the engine of the growth. Johnston [J .22] refers to the new technology known as the "Green Revolution" in the Southeast Asian countries as the "seed— fertilizer revolution." The Korean experience shows that a 30¥percent increase in rice yield at the experiment station is associated with about a loo-percent 265 increase in fertilizer application. In other words, the new high—yield varieties are bred so as to be highly fertilizer-responsive. The basic philosophy Lmderlying this direction in crop breeding is to accelerate the supplementation of scare land with fertilizer that can be more easily augmented. Auer and Heady [A. 10] estimate that from 1939 to 1961 the sources of growth of corn production in the U.S. are com hybridization, 35.5 percent; fertilizer, 31.4; regional specialization, 17.9; and other, 15.2. In other words, the source of productivity growth can be more than seed and fertilizer in the emerging countries. Wharton [W.3] and many others discuss complementary inputs associated with the "Green Revolution." Barker [8.2], for example, after remarking that "many writers appear to believe that the technological change begins and ends with the introduction of the new rice varieties," insists that "a highly complementary package of inputs is associated with the new rice varieties. This package includes irrigation and water control, fertilizer, methods to control weeds, diseases, and pests. Use of these inputs allows the new varieties to express their yield potential. Without these inputs the grower carrot expect a good yield." He also notes that "the greatest production gains from the new technology have occurred in the best irrigated areas." Figure 10.1 illustrates how total fertilizer and material costs will increase under the process of economic and technological trans- formation. Both total factor and fertilizer demands have doubled between 1971 and 1985. Without this package of complementary inputs, it is certain that the new seed technology would not have exhibited its potential . Figure 10. l. 18 L 16 " l4 ' / Total Factor Demand 12 ’ 10" / 1971 73 75 77 79 81 83 85 Fertilizer Demand Aggregate demand for fertilizer and total factor measured in terms of service and expenditure, based on the medium policy level set specified in Table 9.1. 267 Krishna [1K.10] points out that "As a part of development policy, agricultural policy has generally been used negatively-~to keep bread and raw materials cheap for a growing industrial sector, and to maximize and transfer to the city for investment the profits of trade in agricultural commodities," and addsthat "If the circumstances of a country permit this critical minimum rate of agricultural growth to be realized while the terms of trade of agriculture are depressed against it in the traditional way, there would be no need for a positive agricultural price policy. But the evidence shows that in many developing countries the minimum rate of agricultural grwoth consistent with rapid and sustained general growth can be quite high; and that a negative price policy cannot be followed witl'out risking failmce to achieve or sustain the desired growth." Krishna also feels that input price subsidization is not a complete substitute for product price guarantees, and that both are needed as complementary instruments of policy for different reasons. He states that ". . .product price guarantees are needed in addition to input subsidies because it is rot a matter of in- difference whether the profitability of a crop is increased by raising the price of the crop or by lowering the prices of input 5," and adds that "Tlms if a support program does accelerate output growth it turns out to be a very profitable investment for the food consumers of a society." As seen earlier, in the process of economic transformation, a number of economic and technological influences are forcing farms to become more capital-intensive. The resulting demand and the arrange- ments under which capital is made available to agriculture are deter- mining factors influencing the structure of the farm sector, as Brake [B.15] suggests. 268 In summary, not only are technological inputs completentary, but factors governing farmers' incentives, including credit and credit costs, are also complementary to varietal improvements. The important question is, however, which is the most critically limiting factor, or which ones can or cannot be supplied easily at reasonable prices in the present Korean agricultural setting. It is not hard to ideitify this crucial variable from the results of the present study--biologica1 research and dissemination of its results. Let us momentarily assume that this variable can be easily controlled by public institutions. We have seen that it is impossible for Korea to achieve food self—sufficiency until the 19803, in light of projections by the KASS and this model. We do not seek an optimal strategy for achieving this goal as early as possible, since that involves a host of economic and technical variables, some of which are outside of the preset model. However, we do try to grasp some implications about attaining this goal by manipulating the policy variables in the model. The first experiment was to determine what would happen if the technological breakthrough takes place much earlier than anticipated (Table 10.4). It is assumed here that biological research results for good grains are forthcoming earlier with the same amount of total accumulated research outcomes over a lS-year period as that in Table 5.1. This would imply a big push in biological research in the mid-19703. The result of this new experiment on is compared with that based on the medium policy level set in Figure 10.2 in terms of total grain production, where the KASS consumption projection is denoted by *. 269 Table 10.4. Hypothetical Plarmed Expected Research Results, in Terms of the Rate of Increase in Experiment Stations Yield, Adjusted by Proportion of Crop that Could Advantageously Use Results (biological Research Results are Assumed to be Forthcoming Earlier with the same Amount Over a lS—Year Period of Total Accumulated Increase in Yield as that in Table 5.1. Year Rice Barley Wheat Other Pulses L Potatoes Grains 1971 0.10 0.05 0.05 0.05 0.05 0.05 1974 0.15 0.15 0.20 1976 0.20 0.20 0.20 1977 0.15 0.15 0.15 1978 0.05 0.10 1979 0.05 1980 0.10 1981 0.10 0.05 1982 0.05 0.05 0.05 1983 0.05 Total 0.50 0.35 0.40 0.35 0.40 0.55 The only difference is in the assumptions slown in Table 5.1 versus Table 10.4. Under the new assumption, food self—sufficiency can.be achieved in 1978-1979, whereas at the medium policy level set it was possible only after 1981. How can we achieve an earlier tedhnological breakthrough? Remember the underlying assumption.in computing the expected research outcomes at the experiment station.by the proportion of crop area that could advantageously use the results. we implicitly assumed 270 14 [ -/ ,1" fl) Medium 12 .. Higher Research W" 0 5 Outcome in Early...” [‘5 E9 Perl“ /* KASS Projection 8 / of Consmption :1 Needs 5" i’ 10 - 8 ~.-4 3 o 3: 8 - 33 H f3 3 6 - n l l I I l l l 1971 73 75 77 79 81 83 85 Year Figure 10.2. Total grain production projection, based on medium policy level set, specified in Table 9.1, and that based on assumption specified in Table 10.4 (only difference between two runs is different assumptions between Tables 5.1 and 10.4) and KASS projection of consumption needs. 271 that this area would be 50 percent in all cases. Thus, for example, with the same expected rate of yield increase at the experiment station, if the new technology has characteristics that would allow dissemination throughout the country, the expected rate of yield increase in Table 5.1 or Table 10.4 is doubled. Thus far, we have assumed that the planned research results would be realized. As indicated earlier, the biological research enterprise involves mnuch uncertainty or risk in terms of when and in what degree results will be realized. We have shown the conse- quences of realization of these research results optimistically. It scene logical to see what would happen if a planned research result is not realized. For this purpose, we assume that the last three research out- comes specified in Table 5.1 to be zero for each crop in each region. The result, based on the medium policy level set, in addition to the assumption made above, is slown in Figure 10.3, together with the KASS projection of supply and need for consumption food grains. Note that the nation would not be able to attain food-self-sufficiency during the planning horizon considered here. We do not intend to evaluate here overall performance of the economy if this pessimistic result to planned research outcome turns out true. But we do want to emphasize that there is no absolute and definite guarantee that this unfortunate phenomenon will not happen Simply because it is unfortunate. There is certainly some chance for this unforuntate phenomenon to occur. But we cannot state its likelil’ood in terms of a probability. 272 13 r' % KASS Projection of " Consumption of Grains ,f o 11 ' ,r”” ,’ Total Grain Production //° / ‘5 / 9 ,z’ll1::;7 I’,a” / KASS Projection of Total 0/ Grain Supply / ,,. Fertilizer Demand W 7 ' / // 5 . / J 4 4 A 4. 4L 1 fiL—o 1971 73 75 77 79 81 83 85 Figure 10.3. Total grain production, based on worst case of research outcome where last three research outcares for each crop in each region are not realized and fertilizer demand based on above assumption, and KASS projection of consumption needs, and total grain supply. 273 We can glean something from this experiment. That is, as shown in Figure 10.3, first rotice how closely total grain supply projection made by this model compares with that mnade by the KASS initial version. The major source of the difference between the two projections seems to originate from a difference in the initial condition. Second, demand by fertilizer would be growing slowly in contrast to the situation show in Figure 10.1. This is because the seed and fertilizer or other materials are complementary to each other. Thus, in this situation, more production of fertilizer or other materials has little meaning except for export purposes. What is the implication of these outcomes of the research enter- prise? Let us assune that early attainment of food self—sufficiency is the important goal of Korean agricultural development. Then, all one can say in light of the above analysis, is to mke a big push in research activity so that high levels of research outcome can be realized earlier and the uncertainty involved can be minimized. Thus far, we have talked about research outcomes as a package. We will now examine the consequences of the degree of the research outcomne for specific crops; as crops differ in terms of production consumption, and in the chance of attaining research outcomes. Rice is certainly the dominant crop in terms of production as well as consumption. We have already examined possible research outcomes for rice for a range in the accumulated rate of increase in rice productivity from 20 percent (which corresponds to the last policy run) to 50 percent (which corresponds to the "higher" policy input level in Table 9.1). W'— 274 For the small grains, including rice, the acctmflated rate of productivity increase by means of crop breeding would be at best 10 percent per decade, according to past experience. This is the primary reason that IRRI 667 is called a "Green Revolution" variety. However, there are other crops that are easier to breed for high yields——potatoes, forages and vegetables belong to this category. We observe that the leading farmer's record of the yield level of sweet potatoes, for example, is more than 56 ton per hectare, which is about 17 tons of grain equivalent. This figure is more than four times the national average yield of potatoes (sweet and white), about five times that of rice and six or seven times that of barley. On the other hand, in Japan experiment results show that the yield of sweet potato can be increased to 168 tons per hectare, which is about 52 tons of grain equivalent, a record about 40 times the national average potato yield in Korea. What are the implications of this? It implies that the national average potato yield can be greatly increased by breeding or improving cultural practices or both. Based on this, we conducted another experiment. It's design is shown in Table 10.5. Possibility I assures potato yield can be doubled and Possibility II trebled during the planning horizon. The result is shown in Figure 10.4, together with total grain production projection, based on the pessimistic prediction for research outcomes discussed before and the KASS projection of consumption needs for grains. For comparison purposes in this policy run, we have assumed that the research outcome for the other crop is the same as 275 Table 10.5. Hypothetical Planned Expected Research Results for Sweet and White Potatoes, in Terms of the Rate of Increase in Experiment Stations Yield, Adjusted by the Proportion of Crop Area that Could Advantageously Use Result. Year Possibility I Possibility II 1971 0.05 0.05 1974 0.30 0.60 1977 0.40 0.90 1980 0.15 0.30 1983 0.10 0.15 Total 1.0 2.0 that in the worst case. It is now clear that food self-sufficiency can be achieved by 1979 under Possibility II and by 1981 under Possibility I, whereas it is not achieved by 1985 under the pessimis- tic prediction for overall research outcomes. Thus, more emphasis on breeding or improving cultural practices for sweet and white potatoes or both is certainly one means of attaining food self- sufficiency goal of Korean agricultural development. The possibility of attaining food self-sufficiency by introducing more potatoes into production and consumption is examined elsewhere lee (L.6) 1. However, we have to be careful in making policy recom- mendations. In order for potato to become a larger part of the Korean diet, directly or indirectly, several preconditions must be met. 1The basic idea of this possibility is also in this author's unpublished comment on the KASS report. ' Figure 10.4. 276 14 - 13 , Possibility II / 12 - Possibility I 11' /< o / 10 ’ KASS Projection / Total Grain of Consumption . Production Based 9 . Need on Pessimistic Predictions O 8 I 7 r I; a g A A 4 L A 1971 73 75 77 79 81 83 85 Year Total grain production projections, based on worst case on research outcomes, and on new experiment design in Table 10.5 where sweet and/or white potatoe yield is assumed doubled or tripled and KASS consumption need projection. 277 A high—yield techrnology must be produced; new methods of cooking, processing and feeding livestock must be developed and adjusted to Korean taste; storage systems at the farm level as well as in the marketing process must be improved; edible oil sources must be developed and extended since edible oil is strongly comple- mentary with potatoes (less for sweet than white) for cooking or processing, etc. None of these preconditions seem hard to attain, however. It seems that if the public decision-maker pays as much attention to these problems as he did to introducing more wheat products into the Korean diet (which is essentially a short—run solution to the food problem and not to income or other develop- mental problem which, furthermore, involves the potential danger of making Korea a permanent food import country), we can greatly increase the degree of food self-sufficiency and attain other development goals as well in the near future. One possible bottleneck might be an adequate supply of edible oil. This author with others [K.l, L.8] also examined the possi- bility of increasing the edible oil supply fromn rape, which can be grown as a secornd crop with rice in the southwest provinces where the double crop ratio is comparatively low. Thus, rape oil production can be increased witlnout significant curtailment of other crops such as barley. Encouragement of more sweet and white potato production and consumpti --directly or indirectly—-muld accelerate food processing and livestock production, while reducing the need for imports of food 278 and feed grains and, hence, foreign ecdenge drains. In this process, income and employment opportunities would be generated or expanded. This author and others [L.7] examined the possibility of replacing feed grain with sweet potato silage by using a production function approach and concluded that this is an important possibility. The same idea was tested for a typical farm base using a linear program— ming framework [Lee (L. 5)], and it was concluded that under the present structure of prices and yield technology for sweet and white potatoes, it was not profitable to transform potatoes into meat. The possibility of transforming potatoes into meat and increasing the double crop ratio can be examined better when the component model presented in this study is linked with the farm resource allocation component of the KASS model. In summary of the policy experiments, we examined a mnber of policy alternatives for attaining food self-sufficiency as a develop— ment goal. We stressed the canplenentary relationships among mnaj or instrumental policy or economic variables. At the same time, we emphasized that the means of achieving this food self—sufficiency goal must be consistent with attaining otherdevelopment goals such as income level and employment opportunity. The key variable seems to be seed technology, which must not be a once-and—for—all change, but rather must be evolutionary in nature, as Barker [B.l] suggests. However, it also seems that the research enterprise producing, new seed technology involves great risk and uncertainty. The national planning and food demand—supply budget is based on optimisitc outcomes for this risky enter— prise will turn out to be the "tiger in the picture" in many instances. 279 This was true in the case of Hi-nong Il—ho (rice variety in mid-603) and is partially true of TERI-667. More seriously, commitment of policy based on this unduly opti- mistic predictionn has brought three major consequences to the nation's economy: (1) the wasting of scarce public budget, (2) the need for unexpected and large amounts of food grain imports and (3) price instability. We must have same provision for making a plan that involves less: risk and uncetainty. We need a measure of security for a case when expected outcoztes are unfortunately not realized. At any rate, we must establish a long-term policy with respect to food self-sufficiency that is consistent with attaining other development goals. This long-term policy must be able to provide the nation with a cheap balanced diet. The pattern and mainsprings of development must be sought in the land and people, and in the system of social and economic organization, as Kravis [K.7]suggests. The policy should not just imitate what the other countries are doing if their social and economic organization is different from ours. Mndel Assgption ENaluation let us now evaluate the model. First, we want to evaluate some of its basic assumptions and its utility for planning purposes. As the reader remembers in Chapter VII, we adapted the profit-mafimization assumption, which is often criticized as unrealistic. We admit the values sought in farming everywhere include more than money profit. When we neglect nomonetary considerations, prediction based on the straightforward profit maximization assumption may well turn out to be unrealistic. This is a case proven by Wipe and Bawden [WA]. 280 Supply of elasticities are different, depending on whether or not capital budgets are considered, whether or not uncertainty is considered annd whether or not provision is made for investment, disinvesement and resource fixity. The profit-maximization assumption is not really wrong. The error is omission of nonprofit considerations. Johnson [J .16] writes that "This is not to state that all farmers are maximizers at all times under all conditions of risk and uncertainty. However, they do enough maximization so that theoretical mmdmizing models are useful in analyzing that part of their behavior having to do with response to price and resource allocation." Second, in Chapter V, we failed to develop a theory relating biological research outcomes to public investment. Instead, we assumed a set of research outcomes with same possibility of materializing in specified magnitude of specified times, and then projected the con- sequences of each outcome. In a sense, this seems to have been the most urnsuccessful part of our work. Heady [H.lO] states that ”Research and education are not purely stochastic phenomena with chance occurence relative to their initiation and outcome. . .The probability of scien- tific discovery for a particular product, function, or service depends on the quantity and quality of resources allocated to it." However, as Jensen [J.l], among others, points out, "The economics of technological changes remains as one of the least developed areas in economics--both in theory and application." As repeatedly stated, biological research is risky and uncertain. Although we admit that the research outcomes are not purely a chance occurrence, we decided not to attempt to formulate a research production function, but, 281 instead, to frankly recognize our inability to formulate such a function with reasonable accuracy, instead, we decided to project the possible consequences of alternative research outcomes with a simulation component. Third, the reader may wonder why we need such complex models for innovation diffusion (Chapter V) or land and water development (Chapter IV). It is almost certain, for example, that a research outcome would eventually be adopted within, say five years. Is it not possible to model the diffusion process by a simple discrete difference equation with a five—year lag with due account of the discount factor? Such . a model is appropriate when its purpose is to make a long-term pro- jection, as did Black and Bornnen [B.8]. Our purpose has been to ask not only what will happen in, say, 1985, if certain public policy measures are adopted at the present time, but also to project the consequences year by year. In this situation, modeling by a simple difference equation is inadequate. As we noticed in the text, difference equations are a special case of the differential equations. Further Study Needs“ Before leaving this chapter, let us briefly discuss some futnn'e study needs. Throughout this report, especially at the end of each mathematical chapter and in the beginning of this chapter, we have indicated several areas for further study. let us summarize these study needs. First of all, about the model structure itself. We set up Several tentative behavioral relationships. Examples include the farm consmption—saving—invesment relationship, the noninstitutional 282 private money market structure, the real price structure including the interest rate, etc. Mare understanding of the behavior of these economic variables is needed. Second, as repeatedly indicated, the data base used in this study is tentative and weak. Even data on initial conditions fall in this category. For example, we took yield initial conditions for various crops by regions from publications that we krm to have some inaccuracies. In the case of factor demand, our knowledge of initial conditions by crops arnd regions is poor. Technical or other behavioral coefficients are even worse. However, we have used theories or method- ologies for estimating these coefficients from various data sources that are the best available to us at the present time. Constant attention to revision and use of better data is necessary. Standard econometrics or statistics will help us in estimating these parameters. Third, there are inaccuracies in the model specification——that is, we assumed that once land is irrigated or the new seed adopted, productivity is instantaneously increased. For example, in the year following installation of irrigation structures, the yield level may not be the same as that on land where the structure was installed, say, five years ago. For long-run projection purposes, this assumption may not be bad and approximates the real world situation. However, for year—to-year projection, some provisions may be needed to account for this assumption. Fourth, there are several other policy or environmental variables that might affect major output variables of this model. Eramples include improvement in transportation and market systems, rural 283 electrification or other infrastructures, and change in farm size and in migration patterns. Improvement in transportation, for example, will stimulate more regional specialization. This variable certainly has much to do with productivity growth, as illustrated by Johnson [J.3 ], among others. Nevertheless, the present version of the model fails to model this aspect accurately. lastly, model verification from historical data does not seem to be sufficient. However, this historical verification seems to fulfill a necessary condition. Due to constraints on data and time, we hesitated to undertake a broader attack on this task at the present time. However, this seems worthwhile to conduct to the extent that data are available. In summary, we have written about further needs for improving ‘ the model presented here, to be more realistic and to do more and detailed policy analysis. As indicated above, the version of the model presented here contains defects that indicate further study needs. Nevertheless, the version of the model presented here seems to represent the real world situation reasonably well; that is, the model seems to be capable of proj ecting yield levels and related conventional factor demand and projecting the consequences of various policy alternatives in terms of relevant criterion variables. With further refinement the model can be useful in evaluating policy alternatives for the Korean agricultural development . CHAPTER Ki SUMVIAIW In this chapter, we briefly summarize objectives, structure end results of the model, and end with a few general remarks. The primary purpose of this study has been to model part of the production system for Korean agriculture as a component of the KASS model. Since the acreage response system was already built, we have concentrated on modeling the yield response, and hence, factor demands of various crops in different regions. The basic emphasis of this study, however, has been on explaining how public policies, programs and projects concerned with technological, insti- tutional and human changes affect yield response. It is concluded in this study that the major sources of pro- ductivity growth and economic development are structural changes generated largely by public policies , programs and projects. Thus , the basic task has been to determine the kind and levels of policy variables that contribute to attaining agricultural yield goals for Korean agriculture. The specific objectives of this study were: 1. to proj ect: a. total agricultural land by paddy land and upland, for each agricultural region over time. b. improved agricultural land area by irrigation, consoli- dation and drainage types for each agricultural region over time. 284 285 c. yield levels by crops and regions over time d. conventional factor (fertilizer, pesticides and other materials) demand for each crop in each region over time e. labor demand for each crop in each region in major labor peak seasons over time. 2. to identify the source of yield growth, including biological research results and their dissemination, and 3. to evaluate public policies, projects and programs in terms of attaining development goals. One important intermech'ate purpose of this study has been to show empirically how different disciplinary theories and techniques can be used together to model a complex system more precisely and accurately. That is, after recognizing that one of the primary pur— poses of economic development policy is to alter input—output coeffi- cients in agricultural production, we have tried to internalize the production rate and, hence, factor demand, which is subject to various levels of the public policies and other economic opportunities, by using a systems simulation approach. The results of this model are to be fed into the agricultural resource allocation component of the KASS, which is a type of linear programing model that assumes a fixed input requirement in producing a given amount of output. Some neoclassical economic (modified or unmodified) development and growth theories are incorporated in this model. The systere simulation approach has proven useful in solving practical problems involving system complexity, lagged adjustment, feed back and forth, uncertainty, and situations where few time series data are available and for which the classical economic models are not very adequate. 286 Tyner and Tweeten [T.7] put this matter in this way: "Relationships between variables in agriculture and between agriculture and the nonagricultural sector are complex and dynamic and are not always suited to analysis by conventional optimizing quantitative tech- niques. Quantitative procedures are needed which can include time lags, nonlinearity, and secondary and tertiary effects over a reasonably long period of time. The simulation procedure meets these requirements and allows the recursive aspects of the agricultural processes to be mostly effectively portrayed." Economic development in agriculture is a complex process. Equally complex sets of policy instruments are required to affect transforma- tion of traditional agriculture. Thus, the model dealing with this complex system must be carplex enough to measure important possible repercussions of policy inputs. Therefore, we have tried to meet comprehensiveness, consistency and optimality criteria in a sector model for planning purposes. In structuring the model, we specified a Cobb-Douglas type production function for every crop in each region under consideration. We have two basic inputs: conventional inputs and structural change variables, which enact to shift the yield function as well as the factor demand function. There are three different structural change variables. The first involves biological technology and human change through extension of biological research. The second has to do with land and water development (three types of paddy land irrigation, land consolidation for paddy as well as for upland, paddy drainage, upland irrigation and consolidation, and upland and tideland develop- ment) . The third is the variable typically and exclusively related to perennial crop production such as tree crop age composition and residual effect of the conventional inputs used in the past. The 287 first mo structural change variables are generated mainly by the public sector. The rate of land improvement has been modeled by a high-order differential equation as a function of public investment, among others. The same is true for the biological research and dis— semination of its results, but we have recognized the edstence of indigenous innovation among leadirng farmers and by numbers of the agribusiness sector. All independent variables in our production function except the conventional inputs have been internally generated according to specified public policies. Instead of tryirng to specify a research production function for public investments, we specified a set of possible biological research results for each crop in each region over time. It is assumed that biological research results can be attained by the public sector within the limits of known scientific methods and knowledge through direct public investment. It also involves institutional reform. Since biological research involves biological processes, results are subject to uncertainty. Further there is doubt as to whether the traditional concepts of a production function applies to research programs; this was a basic reason for not trying to specify a research production function. Innstead, we constructed the model (Chapter V) so that consequences of alternative biological research results could be simulated to determine the impact of possible outcomes on various performance variables for the agricultural sector. To project input usage for conventional production factors, we have derived factor demand functions from camodity production functions with an assumption of profit maximization. In doing this, we 288 have considered several behavioral constraints. First, we have imposed a capital budget constraint with a stepped supply function for credit. Thus, government policies on credit and interest rate have explicitly become one type of policy variable. Second, factor demand elasticities have been adjusted, based on the direction, duration and magnitude of prices of both products and factors. It is also allowed that an economic adjustment in the sector can take place, based on changes in regional specialization, long— term profitability and others. In connection with this, there are two major policy variables: product prices and production factor prices. We have computed the marginal internal rate of return to capital with a given supply of capital. The demand function for capital is derived fromn the budget constraint equation, after each independent variable has been replaced with the relevant constrained factor demand function for all factors and crops in each region. We have secured all relevant first—order conditions for optimaltiy, and all relevant consistency relationships known as the Fresch scheme [F.12]. In computing the marginal rate of return to capital, since there is no analytical solution, we used an iterative numerical method. Since the supply function of capital is a stepped function, we determined whether or not a given supply of capital is fixed, by comparing the marginal rate of return to capital with appropriate interest rates . For farmer-owned capital, we assumed the farmer can dispose of part of it when the internal rate is less than that he can earn by using other than in farming. When the internal rate of return is higher than the off- farm opportunity cost, we determined whether or not credit could be borrowed at a higher interest rate. 289 Once the marginal rate of return to capital was known, it was a mechanical process to predict the optimal input rate, and hence the output rate, since we already had all relevant functions, parameters, market or policy variables, and structural variables previously gen— erated by the public investment. Notice again that the production response was exclusively the consequence of factor use and the behavior of function shifters. Then we were again ready to compute the relevant aggregate variables. In some cases, we have used the area allocated to each crop projected by the KASS initial version of policy alternative II. We assumed that labor inputs were not a limiting factor of produc— tion; on the other hand, structural change variables were assumed to be labor requirement function shifters. Many public investment programs defined above are designed to save labor. Aggregate agricultural labor use will be determined by the level of mechanization when this model component is linked with the farm resource allocation component of the KASS model. Production elasticities with respect to the conventional produc— tion factors have been estimated by the factor share technique, which is estimated by a method similar to that suggested by Tyner and 'meeten [T.6]. In other words, factor shares have been estimated from the distributed lag prices and input and output rates. The demand elas- ticities for the conventional inputs have been derived from the produc- tion function. Other production parameters have been estimated separ— ately, using various techniques and various sources of data. After testing the model to determine whether it works properly, through a series of sensitivity analyses, we designed policy experiments 290 involving different levels of public investment for land and water development, different price policies for products and production factors, different policies on interest rates, and several sets of possible research outcomes. We made computer runs for each level of each policy variable and several different combinations of policy variable levels. Results of our work are as follows: First of all, we have quanti- tatively and formally identified the sources of productivity growth for each crop in each region in more detail and precision than any study has achieved thus far--that is, we computed how much each of the struc- tural change variables and conventional inputs increase yields for each crop in each region over time. Maj or Conclusions The major conclusions drawn from the policy experiments can be summarized as follows: First, cmplementary relationships exist among the so—called con- ventional inputs, between these inputs and structural change variables, and between these technological inputs and variables governing farmer incentives. The major determinants of the conventional inputs, especially fertilizer, seem to be: (1) varietal change and (2) land and water development. Without these changes , there seems limited room for fertilizer to contribute to yield growth. Second, it appears that biological technology is certainly the most critical factor determining growth in yields. Without such advance, the contribution of the conventional inputs to growth is certainly less than with it. Hence, policies on product and production factor prices and on government credit and its interest rates can 29l play only a limited role in increasing crop yield when there is little advance in technological change. This seemns to stem from the fact that the Korean farmer presently uses nearly optimum levels of con- ventional inputs with a given technology. In this situation, a "positive" price policy [Krishna(K. 10)] may not encourage farmers to use more of the conventional inputs, although a "negative" price policy has a danger of reducing used conventional inputs. The second most important structural change variable to produc- tivity growth was found to be irrigation. However, the Korean irriga- tion system is fairly well developed. For example, the paddy area under the Irrigation Association (termed the perfectly irrigated paddy in this thesis) is about 40 percent in each region. Suppose that: (1) this paddy area increases to 60 percent by 1985 and (2) the produc- tion elasticity is 0.2. Then total yield productivity growth due to this land improvement project by 1985 would be 10 percent when the canplenentary input is adequately supplied. On the other hand, we should not evaluate public investments only in terms of yield productivity, since public investments achieve multiple purposes, such as reducing uncertainty in production, reducing labor requirement, etc. , in addition to yield increase. The other important structural change variable defined in this model is found to be age composition change for tree crops. In summary, it appears that there is an interaction effect between Varietal change and water or land improvement, and between land and water improvement. However, the mnodel presented in this study has failed to appropriately account for this interaction effect. 292 The number of values to be attained in developing Korean agri— culture is certainly mnore than one. Customarily, the degree of develop- ment is measured in terms of total production of food, income levels and distribution, and employment. This study cannot be fully evaluated in termns of these and other performance variables unless the mnodel presented in this study is linked with the rest of the KASS model. We have tentatively evaluated alternative policies in terms of physical production as one of the most important development values for Korean agriculture to be food self-sufficient. Bread is certainly not sufficient for modern living but without it nobody can live. In assessing the degree of food self-sufficiency, we have assumed that producer prices, area allocated to each crop, and the consumption needs projected by the initial version of the KASS model would correctly represent the future. Food grain is assumed to be composed of rice, barley, wheat, other grains, pulses and potatoes and production levels for each are measured in terms of grain equivalent. Since biological technology involving varietal change is a most crucial factor for productivity growth, we have made several alternative assumptions as to research outcomes and have then simulated to project the consequences. In connection with this policy experiment, we have concluded that Korea is not likely to achieve food-self—sufficiency before 1980. In the case of the worst biological research outcomes studied, Korea would not be able to attain this goal even by 1990. Assuming that early attaiment of food self—sufficiency is one of the most desirable goals, we have tried to identify some measures 293 to achieve it. For this purpose, we have made two additional policy experiment runs. The first one was based on some notion of ”big push" in biological research activities such that greater biological research output can be realized earlier; that experiment result has shown that the food self-sufficiency goal could be attained in late 19708. The second policy experiment was formulated on the basis of the following: (1) The leading and crucial source of yield productivity growth, biological research, involves a good deal of risk and uncertainty in terms of when and how much production is made available. (2) In the case of small grains such as rice, barley, wheat, etc. , productivity gains by means of breeding would be, at best, 10 percent per decade, according to past experience. (3) The potential productivity gain is different, depending on genetic nature of the specific crop, for example, yields of "potatoes, vegetables, forage crops and so on can easily be increased through improvements in varieties and cultural practices or both. Especially well known as a crop with a great potential to produce with a given resource is the potato. Potato yields can be doubled or even tripled by improving the variety and cultural practices. On the other hand, it is known that per capita consumption of potatoes in Korea is much lower than in Euroepan or some other developed countries. Potato consumption in Korea could be greatly increased if: (1) potatoes can be produced more cheaply through technological change in production, and (2) potatoes are processed in forms adjusted to Korean tastes. This involves improving cooking and processing methods. 294 Potatoes can be processed either at factory or through livestock. That is, potatoes can be transformed into meat or milk. Once potato yields are greatly increased, potatoes can become a good substitute for concentrate feeds. Along this line, we have made other alternative assumptions on possible biological research to investigate the consequences of impro- ving the variety and cultural practices for potatoes. We have designed two ecperiments. First, potato yield is doubled, and second, it is tripled through biological research during the planning horizon under consideration (1971-1985). For both experiments, we assumed biological research outcomes for the crops would be the lowest one we examined before. The experimental result shows that total grain equivalent needs projected by the KASS model can be produced domestically by 1978—79 under an assumption tripled potato yields, and by 1981—82 under the assumption of doubled potato yields. Improving cooking or processing methods for potatoes involves several problem. It requires an adequate supply of edible oil, which may be fulfilled by encouraging rape production using idle winter land in the southwest provinces. Also required are research and extension for improving cook methods and for using potatoes as feeds. On the other hand, the policy alternative encouraging more pro- duction and consumption of potatoes has other advantages, too: (1) the double cropping ratio can be increased, especially in the southern provinces, by producing either more white potatoes or rape. (2) The transformation of potatoes into meat or milk would generate more income and employment, and (3) depending on the degree of self-sufficiency 295 attained, imports of food and feed grains can be reduced and foreign erchange saved. The conclusions reached here should be interpreted with reser- vations. This is so partially because various levels of interactions with other sectors or subsectors of Korean economy are not fully taken into consideration as this model component has not been incor— porated into the total model end because the data base of the model is rather weak. Needless to say, the projections this model makes and its use in the evaluation of public policies, projects and programs will be improved when this component model is linked with the rest of the KASS model. In the earlier chapters, we discussed limitation of the present model and further study needs for improving it. This has to do with: (I) data improvement, (2) refinement of some model structures, and (3) linkage with other components of the KASS model. Nevertheless, the version of the model presented here seems to represent the real world situation reasonably well; that is, the model seems to be capable of projecting yield levels and related conventional factor demand and projecting the consequences of various policy alter— natives in terms of relevant criterion variables. With further refine— ment the mnodel can be useful in evaluating policy alternatives for Korean agricultural development. ZIPPEIMDIXLZX CC»4PLHHZR.I?R£X3RA¢4 PROGRAM MAIN (INPUT. OunPUTn MAIN 2 COMMON /CONYR/ 0?. 9? 7° .DTYoFINEDTaIALTplALTEX(3)oIDY:chSD: CONTR 2 1 1nROT4. IPRP(6).IPRPRD IRUN:ISENS:ITIHE.IYEAR. CONTR 3 2 JPER Ncufl, NCROP.NDTPOP.NTDPR2.NREGN. NRUk NT: CONTR 4 NYxflfionYEARS.NYRPR2.T VEARoVEARAG YEAR0.NCONAG CONTR 5 3COMMON IVAnc/ rxu"1(3.15. 3n. rx¢3.11.3). Fxn(3.13.3>.PxD(3.3). VAIc 2 1 IREXYR(3:1305) PD(3¢1$). YO(3.13). AcTcIspns). VAIc 3 2 GZSTH1(3.1JI. RANCS): UclsgSl. BEXDI3o13); VAIC c 3 AHP(3.13.9). FLBD(3u13o2).PITP1(3)n PITPZIS): VAIc 5 4 TSP(3-13.9)a PROFYY(3.13) PIT93(3)4 PIYP4(3): VAIC 6 5 TmPI3.13.9). CSLPtsn.. PIT13). PII2(3). VAIc 7 6 RGAII5.13:5). DRDP(3)o PIT3 (3). PIY4(3)o VAIC 8 7 RGA?(5.13.5). UcSULIJ). D1TP(3.4). D2YP(3.4), VAIC 9 ' 8 RGA3(5.13a5). RcP(3)o RDPI3). RCUIJ). VAIC 10 9 DGA(3:13:>). RIU (3 ) RTD( 3): RUDIJ). VAIC 11 1 RINYI5;3.U). SYRGPI3:8)o Dscc3.0)n 8(3.a). VAIC 12 1 2 cs ULI ULIGIJ). TL(3.B).. GLIPDIS) VAIC 13 COMMON IPfsr/ PAVG(3.13’- AI3.13). YAtsn. PXI3.3). PTSF 2 1 AS 0°(13). SZC(3o13)o RSP(3.15)a AGREV(3.13)o PTSF 3 2 PXDN1I3:3): PDH1I3c13). YLANDIJ): YZP‘3413). PTSF 4 3 AOTCH1(3.13).SCR(3.7). APSC(3.8’. HAP‘SoZ). Prsr 5 4 SLnR(5.2). SYLAND. STA. SAGREVIS) P757 6 COMMON IVPRY/ TPLANDCJ); TULANo(3). ICPA(3.6). SruCEYIsn13). vPRT 2 1 TCPn(3:5)o GTZS(3-13). GZS(3.13)o PIEVLFI3015)0 VPRY 3 2 GZ(3.13.5). GTZIS:13.5).IREV(3). ACEVVVISoIS); VPRY 4 6 TVCIJ). 60(3). TPCOSTIIIa YZDYYY(3:1J). VPRT 5 4 INFINIS): TPPtso13)- SIPPI13). SCEYVYIJ:13). VPRT b 5 , trowta). TGL(3). TPVL1(3n. SVNLAN‘S:13): VPRY 7 6 TPVL2(3). VLD(3.13). AYLDI13n. AFX(13:3): VPRT a 7 AYrLBI13)- AFLB(13.2). TFL9(3.13)o FLBI3.13.2). VPRT 9 a sraFv. IVC, sucns, scc. VPRT no 9 516057. STerN. SIF SYGL. vPRT 11 1 SanL1. STPVLZ. ero(3.13). ssrx0(3). van 12 2 ssr1(3>. FOK(3). GLI3) pVL1t5). VPRT 13 3 PVLZIS) van 14 T I 0.0 MAIN 7 CALL INPMFP ”MN 6 YEAR = 1970.0 MAIN 9 HT s 1.0 MAIN la 00 700 I=1.15 MAIN 11 CALL PUPINV MAIN 12 CALL TEMP MAIN 13 CALL SOCDIF MAIN 14 CALL FDVLD MAIN 15 T ‘ T ‘ 0; MAIN . 15 PRINT 94HA1N 17 FORMAT (. TIME = o.F1c.2n MAIN 15 YEAR = YEAR * DT nun; 19 0 CONTINU HAI‘J 20 MAIN 21 296 297 INPMFP 3 - 2 33333*}Uiné¥PP§Su1«3.1s.an. rxasnnsasn. CXE§J;§?‘3)'::$é:§3;SI. 3:}: 3 7 30131: D o a I . 1 Iaexvn(3.13.5)oPD< sense: 13). VAIC Rnn(3), UCISIJI. o 5 2 OZSTH1(31¥3’I _ PjYPZISM VAIC 3 ) FLBD(301302)0PITP ‘3’! 6 i ¢SEE§Iisiini PRO;TY§3.13)AF:;;I§?’I :iigiiif' 32}: 7 13 9) CSL (3 I , ' 5 5 32:;?§.13.5i. DRDPIS): PITSI31- P!;;‘g’;, Ifiig 9 ; neA2(5.13o5?: UcSULISl. Rg;:;;3.4n. ggu(g): ' VAIc 1. a RGA3<5.13.51. RcP<3IO ' RUDIS) VAIC 11 3 5) RIMS). TDI3): I 12 Z gifiiiéisiuni STRGPISaeh $3593?“ 2333, mg 1; ‘ 16(3). i l a 2 OSULI3)I UL . px, 3,, PTSF ) A(3 13) TAISIa- ' 3 1COMM°N [PTSP/ iéégfisif ' szc23.13n. $5P¢3.§:II :ggggii3§f’- gig; 4 ( ) PDH1(3;13, L‘N( ' ' 5 § :533:1?33133.5cn(3.7na ' Ap9c<3.an. :fizéggfg; :;2; s a a 2n STLAN°a - VPRT 2 oooaoaavaawiii’aiviaai mm; 552333;: 35:55::31331 W 3 1 Tcpn(3.8)o ' ' ' ncevvv:3.13). VPRT T1(3.13.5).TREV‘3’- n s 2 92‘3-13'5" G rzmvvvns.13n. V? T cans) TPCOSIIJ): a 3 Win am;- zsszmgzign: x22; 7 K IGLIsh VLL ' R 8 5 $5333. momma "LUMP "38932, 3m 9 ; ATFLB(13)1 AFLB(1302)0 g5:?é§:13)l £tCn ‘ ' VPRT 1‘ SYVC 0 ' WM 11 8 3:25;;- STNFiN‘ STFOK' - 512:; ‘3) VPRT 13 9 STPVLio STPVLZI SFXO‘3113)I :VL123ln. VPRT 13 3 ssrxnsn. F0K<3Ia GL‘3" VPRY 1‘ 3 PVLZI3) INPHFP 6 c INPHFP ; c INITIAL CONDITIONS :2:::: 9 pITP1 (1) I 0.24985 P rp 1° PITPI ‘2) : 0025734 INPHF 11 Elfpl (3) a 0.17837 INPHF: 12 PITPZ (1) = 0.41492 INPH 13 RITPZ c2) . 0.423‘3 {393:2 14 PITPZ (a; I g'gfiggé INPHFP 1s PITP3 (1 - - PITP3 (2) - 0.2 395 {figgfig {3 PITP3 <3) = 0.26798 INPHFP 1° P1794 (1) I 0.1?231 INPHFP 19 PITP4 (2) - 0-11528 - 2° . 79 INPMFP eITP4 c - 0'10 4 INPHFP 21 PITiI1) . 86o23°§3 INPHFP 22 Ple‘Z) I 199:06203 INPHFP 23 PIT1<31 ‘ 23:72267 [NPMFP 24 Pnnennn - 153.74996 INPMFP as PITZIZ) = 329.9491: iNPHFP as RITZ‘S) I 69.9761‘ INPHFP 27 8173(1) = 65-16955 ,NPHFP 2, 9173(2) : 138.20631 INPHFP 29 9173(3) : 23.81823 INPHFP 3. PIT4I1I . 35.93375 INPHFP 31 EIT4(2) - 66-94875 ,NPHFP 32 9174(3) . 11.61405 INPHFP 33 CSLPII) - 44.79634 INPHFP 3‘ CSLPIZ) - 96o27452 INPHFP 35 CSLP(3) - 16.93167 nnnp<1n - 273-3150‘ 12963 9(2) I..5._7039856 3239(3) . 103-é!“M UCSUL(1) - 217-9211 I~Pnrp UCSUL‘Z’ l ‘72,!126 INPNFp J6 UCSULI3) - 236-5707 I~PHFP 37 o1??¢1.1) - 0-4497! {~P"r as 011991.23 - 0o2‘530 INP“'P 3’ 0179(1.SI . 0o! [NP p 0179(1.o) - o-o ’ "PP ‘3 n:19<2.1) - 0-83525 ’NPHFP 4‘ n1rptz.2) . 0-9’270 NPHFD ‘2 9179(2.3) - on INPHFF ‘3 017P(2.4’ - 0'0 INPHFP ‘1 o 5 L . 1“ _ ’NPnrp 9s 5 01TP¢3aL’ . 0-0- IMP"? ‘ DZTP¢1.1) - 1.90733 INF” 6 DZTP(1.2) - 3.17264 1 PP ‘7 021P<1.3) . 3-9‘464 'NP r” ‘° uzrr:1.4) . 5.24989 ’NPHPP ‘9 0219(2.1) - 3.91554 1N "FD 5° uzrpcz.2) . 3.37910 ’ PNPP 5‘ 9219(2.3) - c.9530: IN "P” 5’ ozrpcz.4) .1;.3o313 INPHP” 5° nzrr<3.1) - 1.96859 NPHFP s Dzrr¢3.2) - 3.44709 INP"" 5‘ 0279(3.3) - 4.28502 INP"'” s 0279(3.4) a 5.76397 INP"' 5‘ sea (1) - 9-1 INPMFB 57 BC? (2) - 0.1 INPM’D 5. new (3) n 0.1 I~P”Fp 5’ an» c1) - v.73 I~""’» 6° an? (2) n 3.13 ‘Npfirp 6‘ RDP (3: - 9.!3 'NPH P °2 sea (1) . o. INF";p 6‘ ecu (2) - no '"p" a 6' ecu (3) I on leM;F 65 BIU (1) - o. 1"” Pp 6‘ RIU (2) - o. 'NpH a 6’ aIu (3) - o. _ ’"P F” 6° RYD (1) - 0.04193 'N "P“ 9 ETD ( ) - o. 3169 ’“PHP" 7° era (3) - o. ’N“:P' ’1 Run (1; n a. 5761 ’"” r, ’2 auo <2; - o. 3934 '"PSP’ 7’ auo (3) - o. 5341 ,anrp ' Do a I - 1.3 1"""r" 7’ chU“ a 0.0 ”PM? ’6 ULIG([) . 0.0 INPNFF 77 no 3 K . 1.6 1"""PP 7° 5(IOK) : 0.0 INPMFP 79 ILII.KJ - 0.6 - I~P"‘p 0' SIRGPII.K) . 3.3 [~P"‘P a; 7 . 1‘5 __ [NPMFD 3? scc1.Lo3> - 0:0 1~p"’p a; aINIIL.IaK) ' n.n INP"P: 8‘ 7 coNIINuE 1~p"’” g: e CONTINUE [anrp °6 STRGP(1:3) - 11-71EZ INPHF" 7 SIRGP<2.3) - 27.6703 {~P"F” “3 svnap(3.3> . 4.8116 I P"” 9° STRGP(1.4) . 11.6270 132"?” ’1 INNHF“ 93 1 "Pp 93 INPNFp 94 NPWPO 5 2599 - 87° I P P(2 47 9 13 § ‘ ~ HFP 3:53p1334’ ! 1-°1’ INPMrp 97 . 1.5 . INP 96 3?NT(L.1o3’ ' ‘-°"ng INP:;P 99 ’3) I 11506 P 81%;:tzg'3’ , 1,924g; {anrp :00 BinHLdo” ' 7'75‘77 NPHFP ”1 a NTILoZo‘> ' 9'991?7 INPMp, 1 giNT(L.3.4) - 0.678- ’NPNFP 03 a 00-4 , . 1,3 INPnrp 103 K . 1.8 , INF"; 105 no K) ' ”IN751"'K u no PER ' INP P 4 gggtsgen IN 1970 , HILLIO 1! 1000 N4 ”4ng 11:; c PAC 1 . 1"3_ P 0° 1° . a on INPnr. 10¢ 352??1f11 - iigasa {NPMrp :09 rxcx.1.2» - 3s§1° INPHFp 11¢ rx1!.1-3) * 6'“ NF" 9 111 Pub?!“ ' 1:22;; '13:”; 11.3 FX(‘02'2) I 42550 I "PP 118 3) I 0 NP r ;;:;:§:1; - 1:.zgo {figng 5:; a 1 _ ;;{::§:3, . 3:470 INP:;P 1:: 4.1) - Q-‘o INp ’ mum» - mpg» 1," 4 3) - 039‘ 1ND 9 9 Fxggtszz) . 24‘ 70 [Nng9 12¢ Fx‘x'5'3) I 62. INPHF‘; 123 2:“:vo - 9:33 mafia?" 3' FX‘IO6'2) ' ' INP P 4 3) ! §.590 '1‘} 2 §§{}13I1; ! 30.225 {fighrp 12: rch.7L§; - ;-gg§ ,Npggn 1;; 7 l 9 fl 0 $31813; . 193:; 13:29 :go rxlloao ' ' 9 ”In P G a 3) ' 10505 Mr; 1; :3: 39:1’ ' 17"‘ ffipfirp 13% X‘ ‘9'2) x 2.760 I ”MP; 183 F ‘ 9,3) l 45.62, [NPHPP 13. ;;1 :1a.1> ' 5-333 ’35:?” 13; 0 2) l 19 PF 13 F§§ 110:3) - 1-”°- {flgnrp 133 c 11.1) - 24.229 ,N "Pp 133 :x‘ 111.2) - ‘4.7oo KN:"P” 139 ”t 11.3: u4o.52 ,anrn :4. FX '12‘1) - 1 .63§ INpHFP 1‘; PX‘ '12‘2) . 1,939 leHPP :1 FX: .1213) I go360 leMF’ 113 5?“ Zoom - 953° 1.19.75; 3; 2) I Q- 1 FXC 'igz3) l a.15_ lzsnrp 1“ rxg }'1,1; . 11.839 INP:;’ l‘? txn‘l‘1.2, ' 3'319 ’NPHrh 3:8 ”0(11113" 5’04" ”MFP 359 FxnII-Zal’ ' 1"‘15 'NPHF: 1s: FxD(anI2’ ' 1'28» XNPHPp Is; "nu/2:3) " 3””! ”*9 F9 153 ;:D(Iu301’ ’ 13°72“ 3CX) r n: .3.2) 9 $.26- 7:1“ .3.“ 9 3.479 INPpr rxoc .¢.1) - 6-30 'NPHFp 15! 7*D‘ "'2’ 9 £3294 'NPHPP 159 rxnt "'3’ ' a '7‘ INPMp, 16¢ m" 'S'éi 2 iii: mm. m ‘ 0 Egg: 15,3) I 62.890 INPHp, 163 rxa: .6.1: .339 INPin 15; fth .a.?) I 1-959 ’NPHrp 164 FXD! 0643’ '; é'5 a ”PM” “3 rxot .7.1) 9 30-235 leHFP 166 rxn¢ .702’ ‘ "°°2 ’NPMPp 167 rxo¢ .7»3) 3 ’-2 2 NPnp 16 CXD( 5511’ 150225 INF P O FXD‘ ’5'?) i 0595 IN HF’ 169 UN 0553’ 9. 1'5” 1 PM} 17' VXO‘ .9,1) I 17.5‘_ INPHFP ‘7‘ FXD‘ I902, ' 20760 INPHFp 17a EXD( 19.3, ' 46062_ NPHFp 173 FXD‘ .10.1) - 5.339 ’NPnr, 1,‘ VXD‘ 010027 I 1.92! I~PHFP 17 rxot .10.3) - 1.590 lenr' 17! rch .11.1) ' 24-37 INF", 1 5 rxn: .11.2> . 4-700 INPHPF 7? rxbc .12.1) - 11-635 INF” p 179 FXD‘ .1202, ' 1.930 [NP r5 rag rxoc .12-3! ' 3-36° lenr’ 1 rxn: .us.1) - .266 .Np"Pn to; rxu¢ .L3.2) . -05 ,anrp 10; FxD .L3.3: - .15 ,NPHPn 13. fxnna¢ .1.1> ' 11.830 1" "Pp 13, rxonzt .1a2| ' 3.;10 ,NPNPp 156 rxnnzc .2.1’ ' 15.515 INPHPp la FXDHl‘ 0202) ' 1028 I Pnrp 1.: rxnn1: .2.su - 3.480 :fipfiPn 19 rxDH1¢ .301) ' $30720 ; p"'n 19' rng1< '3'2’ ' 1'23 1N""Pn 19% rxnn1( ,3.3) ' 3°47° 1 ”Mr. 19 rxnn1( ,4.1) ' 60‘“ 3"“"Pn 19: fxDH1‘ [402. F $.22 INpH . ‘9 ron1¢ o4a32 ' 0.97 13P"'“ ‘9; fXDMi‘ I512. I 80°06“ I PHF; 197 FxDHlt .5.3) ' 62-599 gzp"Pn 19: rxon1( :6-1' ' . 3° INSHFp 20 rxnn1¢ ,6.2p ' 1~°2° 1N,"Pp an: rxDH1( .o.s» ' 1-59° [NPHFP an; fxgn1( .7.2) a 4.062 Ianrp Po. rxDH1( 0703) ' . 7:202 [NPHFF 203 rXDH1( 03:1) ' 160225 INpgrp 20g rxDH1( o8a3! ' 1'5°5 INpS’P 906 rani‘ .9.1> - 17-5‘ t~pni“ 209 rxnn1( o9o25 ' 2.750 Ian,’ :1! “an“ fit“ ' “'52 ”PM: 2:! rng1( .10.1) 6 5-330. "Par. 21: rxnn1( .1o.2’ ! 1.920 :anrp 21‘ INPHFp 21s NPHFp 215 prnpp 2 7 .3()l 9o H (1,1053,‘ 1.5 INF E§3"1"'11'1’ , 2:';:° INP:F: 219 rxnn1(x.11.2) a .52 IMP"; 22' FXD"1“"2‘ ’ ' 11‘936 1 MVP 22 rxnntgfo:§.§) : 3:360 rgffiit 22: I I ;;g::.x,13,13 - 0.3;6 ’"PHFA 22' ’*“"1"'13‘2’ ' °'15 INPHFp 225 rxnn1¢1.13-3’a' °- 'NPnrp 22‘ EXD‘IOJ.’ I 10 INPHF’ 227 PXD‘I'a’ . 1.9 INPnr 222 fixn‘li:"-6::g° 'NPnr: 229 $3312» ' “~°'=' {ggnra 28¢ éntla3’ ' 2"39 1 HP» 231 9011.4) ' 3"‘° ,"“Hr6 282 60(1-5) ' ’°'°5 "’HPn 233 90"'°‘ ' °°’°° INPMrp 28¢ 'D“'7’ ' ‘1'5‘ luPurn 235 0:1.6) - 49-10 ,~p r. 23 an 1.21 - 1:7.27 :NPHr, as; I ' ' away I ‘06:“ 1:23." 233 12’ I 1160‘: [Np D 239 PD h ”PD 24. ['13, . . ’NPnrp 261 10 corlli. - 3.46533 :3:”’9 2‘2 ’3 1122 - 1.90m Mp” n 24, YD 103’ ' 2022’“. INPHPF 2“ :0 1:4) ' °°72351 IN "’9 21, VD 1:61 - 0,75956 13PM?“ 2‘, YD :,7) 6 9.75963 ’ ”Mr. 24° 70 1:5) ' 3!7°‘5° ,"”Hrp 21; YD 1.9) ' 1-‘19‘7 ,“Pfirn 250 m 1.1” ' gm [NPm-p as; 011.11: - .26662 ,“Fnr. as; Yn(1.12) - .69119 sz"'F as; vntx.1s> - 6.0 _ r pHPP 25‘ v (2.11 l 3.34.50 INPHI‘“ as, tn¢2.2: - 2 2573‘ [upnpp 256 !0(2.3; . 2.30999 ,anr. 257 yu‘z ‘) I 0.71.09 INPHFF as. yuczi5)" 5099755 IND”,° 259 YDt2o6) - 0.70746 INPMP“ 26° YD(207’ ' 14'3365‘ szNPp 26; YD‘ZOa, I 4633351 XNPMI‘D 262 YD‘2'91 ' 1'43290 ;~:"F» 263 va¢2.1o) - 25.6 I~6"’“ 26. YURI“) ' 0'2"" ”Wm-P 265 yn¢2.12) - 6.95661 xup"’9 266 YD‘2013) . 7’9 -. INPHPF 267 3313.3) - 2.15m mpg? 227. yn(3.4) . 1.23516 lunar: 27‘ (3 5) - 4.10837 ’NPnp, 273 $3.56. - 0.75951 War» 2;. x 6.71636 INF": 2, YD‘307’ p 5 ‘3 5) I 3.74721 MPH“; 76 Y” ‘ MW“: 377 13(12 {3.9) - 1.5”“ $313.10: - 25-0 ’NPnrp 28' 1013.11) - 9-27250 HF, 28 1013.121 - .7605? I~P”" ‘ YD(3.13) - 6.5_ Input, 28? c7: 11.!) - 9-‘ INPHrp 28' acre 12.51 - n-z lunar, 26‘ A076 (3.5) . a. 5 'NPnr, 205 acre 11.111 - 3.5 INF"; 29 A676 (2.11) - _. I~p ’ 2°; ‘cfc (3:11, ' 005 :NFH'P BAH¢12 - 1.230 1 Hr; as. B‘"‘2’ ' 1'157 "Put, 209 cancsi - 1.326 :anr’ 29' ucxs<11 - 0.0 "Par, 2,: "C‘s‘z’ - 0‘0 ’"Fnrp 29a ”C‘s‘s’ . 0’0 ’“Pnrp 292 11 x ' 1.3 - ’"PHP 29 I'LBM 010;: I £03}: ’NPHF: 2,; :10 ' .' . l Etgg: 0201) I 9.302 ’zgg‘.’ 229‘ r1201 .2.2) - 9-320 I", r. 9; rLaD‘ 0301, ' 96276 "‘PNF. 2’. r1931 .3.2) ' 9-2'¢ lup”'P 29. 5189‘ 0‘01, ' 9009. [anrp 30' rLen( 6‘02) . 0-152 x~ Hr. an; FLBD( . .1: - 9~690 I “Mr. 303 mm .m) - 9.245 [mu 3" new .661, ' 00°93 NPHPp so. rLaoc .6.2: - i 150 ’"Pnr, 305 7L3!" .761) I :25? ”“pr 30‘ FLON '7'3 : .3: 113:”. 3°; '8' ' Hr :tgg: .a.2> - .231 ’N’nr, 3:. “30‘ 0901’ ' 53554 ”Why, 3 9 71901 .152;)-, gogf: :3:”" 3:: rLao( .1 . . .2 3 1 "r, 313 > rLaot .1o.2) .19‘ ‘~.",' 3;, 'L9”‘ '11'1’ ' 0° 9 "Par; 31 .11.2) - 1.106 tun” ‘ :33: "3'23 2 3'13 {392 a; .1 6 ,0 gtgg: .13.1) . 3.010 x~.::’ §17 tLanc .13.:1 . .626 ’"Pnr: 3f: 11 ”mm {35.7%. 32. °° ‘2 ' "‘3 Mi: 3? xeexvnn§.:.;; : :3;: §3£"'- 3;: R a a 135:32 1.1.3; 9 i:;; :z::;: gs: YR 6010‘ ' HP 3 :gE:YR 1010’) W 19.2 :Np'fl': 3:: IBEX"? [0221) ' 1971 “hi”. 3:. laixvn 1.2.2) 6 1976 {NPnrp 32’ xaexva 1.2.3) 6 1970 tNPnr. 33. :BEXYR ‘02:” .1 1982 23:34:.” 33‘ xaEva 1.2.5) ! 1°99 rupnr’ 83; IBEXYR x.g.%; a 13;: 'angg §g= ‘BEXYMIJ I ' u“. a ‘ lasxvn BEXYRCIaBaJ) BEXYR‘IIBJ‘) BEXYR(X.81§) BEXYR‘Xp9ot) BExyR(I.9.2J xasxvntl.1o;1) leeva(!.1a,2) Iaeva(I.1o,3) xaEXYR(I.1o.4> lasvacI.1o.5) Ieeva(I.11;1) IBEXYR(I.11.2) IBEXYR11.11}3) Iasxvncx.11,4> IaExYR¢!.11.5) IaEva(I.12.1) IBEXYR(I.12,2> laexvn(l.12.3> IBEXYR(I.12.4) Ieeva¢1.12.5) lesxvng1.13,1) IBEXYR(1.13.2) Ieeva(l.13}3) IaEva¢I.13.4> IBEXVR11.13.5) (n m X U A a s VI \4 INC-I‘lflu OOOOOOODO 010.000.... BEXD(1.10) - v ‘I‘QJw-‘In‘ 'I '-Z.- 0211T11l) . 0170411) PITP111) - PIT1())/YPLA~D1X 01TP211) . PlY2(I)/TPLAND(l 9170311) - PXY3(X)ITPLAND(X 21TP411) - P)T4(1)/TPLAN011 ucsLP11) . TPLANDtx) - CSLP 0020011) . TPLA 0011) - URD P ) ) ) 3 1 ( 7ULAND(I) I CSUL(I) ‘ UCSDL( TLAND(I) I If (AHOD . P1711x)/TLAN0(I) - PIT21T)/TLA~011) I PIT5(l)/TLAND(XI I P1741l)/TLAVD1I) - CSLP(l)/TLAND(XI I DRDPCX)/TLAN011) - CSUL(I)ITLAVD(II - uc50L1x)/TLANu(T) - ULIo1l)/TLAN011) TPLAND(I) 0 YULAN UULIG!!! I TULAND(I> - 0L!G(l I I I D ) ) ) (I! ) 2350g°°0 5B7 500 5 00/ DATA PAnsc1 0A05c2LPAnscs.0A05c4.0105c5.0La: / . 3. 003. 0. 007.0. 005. 0.003. 21. DATA HTR1.HTR2.HT93:HTR¢.uTR§.HTR6.HTR?.HTR8.H7R9I901.DI T/ 550000 132000 132.5/ 3.50 2 50 1. B. 1050 10250 1610100/ IN 0 10 100 1000 10 000,0 006150 10u/ I PlTl(I) ‘ PlTZ‘l) o Pl?3(1) 0 P1711!) (701. 0) .NEu 0. 0) 60 To 1? INPHrP INPHFP INPHrP INPHFP INPHFP INPHFF INPHFP INPHFP INPHFP INPHFP INPHFP INPHFP INPHFP INPHFP INPHrP XNPHFP INPHrP INPHFP INPHFP INPHFP INPHFP INPHFP INPMru INPHFP 23 19 307 snrvn(x) - nsc<:.1) suuvn(1) . nsctl.2) HPTLD(I) . us<1)-nsC . H9(Lo1)nanP(x.L) HPULD(I) - HPULD(I) . H0(Lo1)-027P(1.L) sorvn). PROFTY(3.13).PXT?$(3)0 PIYP4(3)0 TDP(3013.5). CSLP(3)0 PXT1(3). P!Y2(3). RGA1(5a13n5’o DRDP(3). P!Y3(3)o PlTGtJ). RGA?(5:13I5). UCSULtJ). D:TP(3.4). 027P13o¢)a RGA3(5.1305)o RcP(3)o RDP(3). RCU(3): DGA(301305)a RIU(3): R?D(3)o RUDtS’n R!NY(5:3.8). SYRGP(3:5)1 DSC‘JJO)D 5130.): CSUL(3)a ULIG(3)0 TL(3.8). GLIRDtS) Cannon IPTSr/ PAVG(3o13la A(3.13). TA(3’. chs.s). OUNH COMMON IVPRY/ ‘IOWQUNH “NHOOVOH‘OINO‘ OV°“&GNO‘ ASORtlS): 510(3013)o RSP(3.131: AGFEV(3013): PXD“1(3.3’: PDHi(3o13’. TLAND(:’I 71013113): ACTCH1(3.13’DSCR(307)0 APSCI33015 "A'1302)0 5Lna(s.2)o STLANDa SYA SAGREV(3) O TPL‘ND(3)0 YULAND(3): YCPA‘306)3 SFUCEV13013)0 rcpn¢3.n). GYZS(3913). GZS(3.1;). PIEVLD 3:13): 02(3.13.5). 0T2(3o13.5!.TREV(3 ). AC HY v3.15). Yvets’. cc(s) TPCOST(3)o anvvv 3.13): TNFIN(3). YPP(3.13). SYPP(13). SCEVVV'Ja13)o Yrox¢3). L 3). TPVL1‘3): SYNLkN 3.13). TPVL2(3). YLD(3013)0 AYLDtiS). arxt1;.3». AYFLB(13)I A'L8(1302)0 T7LB(301J,I 'LB‘331302)0 STRFV. tvc, UCIS. ScCo SYCnST. S'NFIN. S'FOK. Q STPVL1. SYPVL20 srxo(3.i3). ssrxo¢3)p SSFX(3). F0K(3:. GL(3). PVL1(I)o va 2( 3) DIMENSION DEX EXIJ(3.13), sscoae(;). scons¢so13). PSCORE(‘.13).SROFCN(3013)061(3013.5)o EEEDF(301310 a. nyoxss; 195~o 5 / VADF1/ 0- 75¢ V‘D'Z/ 09 ,6: VADFJ/ e. 750 VADF4/ 00 560 VADF5/ 0 750 V‘D'b/ 100 9 VADV7/ 0. 63.1030 1'06: 5900: 3'97: §c750 0:970 "CEDF(3.15). CsEDFt3.13)a RSEDV(3013). VADVZ‘O). V‘D'3(5’D ROED'C301315)0 VAnr5(5). VAD'6(5)o VADF 7(6 I Aee¢3.13.5). Au?(3.13.5). RYDl'Ft! 13.9). KRF(3:13D:;: 0;:3013; 5). ”DY ‘(301305’0 PROFN1t3-13) 2450.1. 3.0.20 9'0.1o 900.3. 6-0.2. 3‘g.10 303.2. 3‘0.3; 6-a.i. 3'0;3n 300.1. 3.0.2. 3.9.3: 3-§ 2: 300.1. JcJoSo orniia 3.082. 6‘0920 6'Jo5: 3‘u.20 6'53 3: 1500. 0. 6.5.2: 3.0;;. 30613- 3‘000/ 1.250 ip401 1.‘ao 1.56 7 10020 1.170 1.22» 1.25] 1.13! .25/ 0.950 i010: 1.20; 1.230 1.25, 1.130 1.22: 1. 231 pm a N N s ... 10 . v - MS. a. ass. 0. 042. oo.o51. 0.063! SOCDII to Y9 SOCDIF SOCDIF OOVOUQQNWQUQ’N 51 6° DAYA DAYA DATA 1 DATA DATA DAYA DATA 313 vAuro/ o. fils. a. on! n.01.o.g12.mo.61§0. AAAAA/ SHALL.9HACH, BEXAABEXBAKG AI D 6.5. 0.9. / DIFDF1ADxrprzénxVDESADIFDFAAnxrnrsAolrnro.Derr7I H04 105. .D 20.0 KD'1AKD’2AKDVJIKDFEA ”0'50K0'60K0r7’ 5A5A4A‘A404A5/ SCOR1,3COR2.SQOR3 WCOR‘I‘ i" AD/ AEEAA AHYAAurc. noAfisnarn. MD ’A E’s, 000A 11. Do D01. DA 65’ PSCORE/ 39‘1 OBEXI. OGBEX/ :5191130 DA D3/ DA RYEAR I YEAR O 5.00 0001 ' I 1. 6 o acasx-rgoclexl DO 900 I 1A BEXAI) I (BEXAA7A¢!)/37A A BEXBITLANDII)/STLAND)O.BEX DscoRE(I) . 0A6 JI '13 PROFH1(IA J)1 I PROVYVIIA J) Pnorcu(z.J) - (Paorrv (IAJ) . naoru1(IAJ))/Pnorn1(xAJ) IEED7(I. J) I BLIEIVADElASHALLAU M7 “or BEXDIIAJ)’ EVEDFIIAJ) I Y:8LIE(VAD'2ASNALLA) 7072AK072:PROFYVIlAJ)) PCEDFIIAJ) I IABLIE‘VADCJASNAOHAI FD'SAKDFSAPROVCHIIAJ)) BSEDF(I.J) I YABLIE(VADF4.SHALL.I FDFA.KDFA,SZCIIA J)) BSEDFIIAJ) I YABLIE(VAD'5.SNALLA) EDFSAKDF5.RSPIIAJ)) 1% J, I SOOR1OP‘ED’1IIJ, O u- q:-r::ur(1 J’ ‘ SOORSICSEDf(IA J) O “"30‘ .RSEBEIIAJ) ¢ PSCOREIIAJ) 1SSCOREII) I SSOORE¢I) I 560“ “I DEXIJ(I.J) I (SCOREIIA J)/SSCDRE(I))I(aEX(I)/IA(I)) ‘25 (IA J) 3718(IAJII I D: 0700 1 Ir (IBEXYR(IAJAK)AE°AD) 00 to 700 _ Ir (KYEAR .LT. (xaaxvn(xAJAx) . 1.0)) an to 705 OYDIFF(I.J.K) A RVINCR1IAJAK)0RYDISSIIAJAKI ooenr¢xJ - 1AALIE(VADFAASHALL.nlranAxDrAAAVDxrr(:AJAK) ) DortlAJAK)K - RoenrtlAJAK> v No can?(xAJAK)-(Eeenr¢1AA)'('riDr(loJ) 1 IDPIIAJAK) D I .g a Cl. C L . X V II.- 1ED'IIAJAKI If IDOAIIAJAK I. NEADAD) "0 . rcsnr(1. J) A csgor(xAJ) - assortlAJAA) 1i6. o - AN'IIIJDK) - vsr¢x. 4.x AEEAIDDFIIAJAK) AHYAADDF(IAJ. AEE(IA JA K)'TDPCIAJAK) 0 AHTIIAJAKAAAHPtln JAN) 0 D'CIAMI(I, 4. KgAtsgtlAJA 50 DOAIIA JAK) I AHAX1I1ADAIDDAK I DD'IIAJAK))I IDTOAIIAJAK) ' ZADIDY'KUA/DDAIIAJAKI AD anrtlA JAK) I AHAX1IDADAIRR'" - RRrA-antIAJAK))) ED'IIA JAK) I AHIN1IEDFII_'JIKIIYDPIIAJAK I, If IIAHPIIAJAK’ O YSPIIAJAKII.LI. CD0 0) 60 YD 51 lDrIIAJAKI I ID'IIAJ A K) .RrIIOJlK I 'TZIIAJAK) I (ANVIIAJAK) A 73'(IAJA KDD‘RYDI'FIDDJDK) OYZSIIA J) l GIZSIIA J) I GIZIIAJ DFIIAJAK) I TABLIE¢VADF7A SHALL: DIE ,077. KDF7AAHPCIAJA K) .z‘IOJOK) I (1. D ° DEI IAJIKIIIRYDIF'I[AJAKAIIAI'IIAJAKI ‘ DEC. MI AK!) IZSIIAJ) I 025(IA OJ) ’ GZIIAJAK) 60 1 ISECIAJAK) I 5.3 1.50.3) 60 o no BALL DELDD (EDP!IAJAKAA°A(IAJ.K)ARDA1¢1AJAKA.DDA¢IAJAKAAIDTDA (IAJAKIADIAKQAAAKII. JAKII SDCDIF SOCDI' 314 no . 1.5 . ‘ socoxr 61 359:: 4.x: o YSP(! 4.x) . DGA(IoJoK)IKGA'ROA1(NeJaK)'lDTiA(!oJyK) igcngr o 1 ca 7 70 CALL DELDD (sprgx. u.K).94(l.J.K)oR8A2(1oJ.K).DIA(l5JoK)' IDTOA socoxr ex v.07. x9A.Ancx. 4.x )) ggggfr , 35';10J0K’ I 0.3 1. 5 71 TSPCIAJO’” ‘ TSP(X. Jo‘) ‘ Dc“X0J0K31K0A0R0L2(NDJOKT‘HITGI‘llJoK, SOCDU' 0 no 060 socolr 00 CALL DELDD (ED'(I 90K) uA¢l 4.x) RGA3tloJcK)IDdAIlsJoK).1970A socnlr (h KhDhK‘M-M‘H. JOK SOCDI' TSP( L“hm I 6. SOCD r DoaLn-1.socn r 01 TSP! oJlK) n TSP(I. JnK) ° DEL(I.JcK)IKGA'ROA3tflnJvKI'IDVOAIInJcK) 5060 r 30 T 6 0 5060 f 600 CONT”NUE _ _ socn r ANN “MU - “want.“ 0 DY'tum s RRHthHuMIchn 50ch r 700 CONT NUE SOCDIF VZ(1,J) I (BZSKI. J) - 015T"!!! J)>/¢160. o GZSYH1(I. J’) socnxr GZSYHitl J) c 0251:.Jsocoxr rzn .J) I V2 Z(X.J) 0 YAUL15(VADFBoSHALLaDlFDf1.KDF1.BEXD£1.4)) 5060 r C Di: RlithD LAO EXTENleN BUDGET 900D r "EXD laJ) l BEXD(!.J) t 010(06le(loJ) ' BEXBtloJ)O 5060 f 600 ON NU: SOCD r 900 .0" Nu: SOCD r RETURN SOCD r 3ND 5060 t 3333 8~""5 1 10n07u.IPRPIOI.IPRPHD.IRUNaISENSoITINEoIYE‘Ro 2 JPER vcuw.NCR0p.N01POp.MroPnzmHE N. NR Up,» 7. 3 NTIHEoNVEARS:NYRPR2pT VEAR.VEInAG.YEARo Moon‘s COHHON IV‘IC/ FXDH1(3.15.5). rx‘3.1503Il FXDISO’3;J’OPXD(303’I 1 IBErvfltsolza5)aRn(3o15). VD(3.13)o AcTctsais). 2 strn1(3 13).RA"(3I Uczscs). aexn(s.18). 3 ANPI30131’In VLBD(3:13.2).PIYP I3): PIYPztsio 4 75Pt3o13.5). PROFTYC3.1J).PITP c3). PITPQIS): 5 TDP(3o13.9)a CSLP(3). PITil SI. P17213)a 6 RGA1(5.13.5). unorcs). thacst. 9174(3). 7 R6A2(5.13.5). ucsuL(3). nxrp¢s.4). D2rpts.4). a R6A3(5.13.5). cPtx). RDPtsI. 0(3). 9 DGAtso13.9). qu(3 ). 1 (3). aunts). 1 RINTI’aSoUIo STRGP(398’A DSCISOOID SISoOIo 2 08 UL ), UL 16(3). TLIJ.BI: GLIRD(3) COMMON IPTSV/ PAVG(3.13). A(3.13). YA(3). PX¢5.3). 1 2 3 4 NH DIMENSION DIMENSION VOWOUNP 0‘0“.“NMOOV0ubufllb‘ “NHOO‘OULU 315 ”N rR/LD DY 07’. UYYoVINEDTnIALToIALTEX(J)0IDT.IPCSD. ASOR(13). szc(s.13). RSPt3;i3). Aanev13.13). pxpu1cs.3;. Pnn1(3.1sg. TLANO(3)o an(3.13>. Acrcu1(3.x31.scn(s.7). APSCIS.OI. HAPI302). LD't3o2)o STLA . SAGBEV(J) 3 ND COMMON IVPRT/ TPL‘NDISIa TULANDIJII TCPAISaOIo SFUCEYISI13IO TCPD(S.O)¢ 0T2$¢3¢13)n 615(3a13In PIEYLD(3013)0 OZ‘3.1305)0 GT1I301305) TREVISI. 6C5'YY‘3013)! TVC(3): CC‘SIc TP COSYISIa TIDVYVIBat3Io TNFIN(J). TPP(3.13)o STPPI13). SCEVVY‘5113)0 TFOVIS): TGLIJIn TP VL1(3). SVNLANI3113Io TPVL2I3Ia YLDISOISII ‘VLD‘13I. ‘FX(1§03II ATVLBIiJ). gita(13.2)a ;:L:é3013)o gL:(3a!302I0 ca STREV STcOsT. STNFINo SW STPVLi. STPVL2. SFXg(3-13Ia SSFXO‘S): SSFX‘S): f0K(SIo GLCSI.1I3I; SI PVLZI ALP(3;13.J>. §ALP(3.13)o PoR(3.13); Pan(3;a). Actcncs.13). §PPEFD¢3). Fncr05(3.i3).PPEFD(3.13.J). srperp(3)- uAP(S.13.J). SeAP(3). MD P(3.18.3). Av2(3.13.3). svero(3). VE'°(30£303)I'PE'D‘301303,l 83CEPDI3): VIURI3.1303).AACP(3I13n3IOEFDOP(3013933t APIut3.13.3).sPIPD(3). BAPAtS). rPcFD(3.13.3). BAP 8:3 0(3). BAPDIJIa erncP(s.1s.3). RAHTiI3). :AH72(3). RAntstS). rFrDS(ao1s.3). RAHv¢t3); APP1(3); APPZISI. scEFD(3.13.3). rnPP(3.13.3).rxat3.13.3). SACEFDIJI. ASC1I3-1307): SCEVLD(3.13).VLnPc13,13). AcEVLn(3.i3).AHc2(3,x3.7). GLIRA(3). “1270(8 13). GL IR(8I M-clt3.13.7)o VA61I9). VA52(7). VAE 3(5). AHEFD(3o1s-3). VAE4I’). ADPDEZ¢3c13IpADFDES(3:13)aP FDS(3:13.3)0 ADFDB4t3 IJIIEFOCPSI3013)IVLDPBC3113I0 BIE'DI301JoSIn FDC'USI3.13).PSC(30010 PSCPISoO’n FICEV(SI13o3I0 ScEYDp(3,13).VZDYDD(3.13).AcEYDb(3.i3).YLDPA(3.13.7). 87X 3: 3). ”X (301313). FXO(3.13-3)o ADFDEgtlo1303) ADFDEfiIS 13:3).ALLPA(301307)0 SBEAPPIBIo SFP 0FD£3I¢ ALLPBI3-1302II IEAPII301303ID PX‘1301303I0 SPXX(SIO SBEAPlISIIJoJIo386£P2I3c1303I0559XX(1)0 ‘AVLDIiS’c SHEAP3(3.13.3).SBEAP¢I3.13.3).AAFx(il.3). FXN1(301303)0 SBEAP5(3.13.3).38EAP¢¢3.15.3):YDHi(3.13)o “Aunt!“ SHEAP7(3. 13.3).saeara(3.13.3).sv1n(3). PVL191(3). rLarn1«3 13.2).PLBF02(3.13.2)onLIn2(31.”Amms rLarnsts.13.2).rLaPAi(3.13.2).uvc:3.i3). svcss). UNP‘OD 316 rLePA2(3 13.2) VLBPA313.1J.2).REV¢3.131. YFLePt3.13’o rLBYZDt3:1J.2)nFLBPB(3o12) PLBACE(3, 13.2).PLBPC(3.13.21. AAFLB(13.2). AATEL8(1J)o FLBSCE(3.13a2).PLBPDi(3.1JA7) Pnz<3.13.7).PLBPPP(3.13.2) LB DIMENSION FOKPAI3))GLPB(3) nVL051CJo13) DATA SHALE1,S"ALE2, SHALtJOSHALEAI .302: 0. 0 0.00 U. 0’ DATA DIFFE1. Dxrrsz Dxrr63.01rr54/ 0.05. 0.1025 10 0 DATA KE1 KEZ, K53 KE‘l BI 6: 4n 6/ DATA VAE1/ 0.3. a. g“. 0.165. 6.03. 0. o: o. 05. 01275 0.94: 055/ DATA VAE2/ o o. 0.02. 0.05. 0.13. 5.235. 0. 285. 03/ DATA VAES/ 0. Q. 0.5350 0.150 O. 265: 0.3 DATA VAE4/ 0.2. a. 035. 0. 09. 5o. 21. 1.325. c. 375. c. D‘TA AYZ/ 302-”, 3.1. 5' 3‘1 3.1.53.2.0301001‘3'2.fl. 3.200. 1 3-1.‘. 3.2. a. 3.1 o: 3'1.u: 3 2 3.0.5; 3.0.3: J‘Oc3g S'Jo3p 3'0. 5: 3t0r34 3'0.’- 3.013: 3 3.0.3} 3.0.0. 3‘0.3o 3'J.31 3'0 1 300.3} 300.2: 3.0.2; 3'J.2n 3'0. 5: 300.20 3'0.5. 300.2: 5 3.0.5; 300020 3.005. 3.1.2: 3'0 DATA VLDPA/ _ 2 ' .2. 0.00 0.010 0100 0.0: 3 .01: 3.00 12.000. U:0I U.'2a ‘ 3'010: 3'0 Dis: 900 9, 5 3. .159. 2"“000 3.00015 9‘00 00 3'00’070 6 36. .6; 3950. a 3.9.1, 900 01. 15'01'3 7 3. .02. 9.0.00 39.0000 3.0:00 12.";030 a 3. .15: 3'nn40 3.0-2a 3'0.i50 3'0.‘0 9 0.1. 3-n.o/ DATA ASC1/ 3-o.§5. ago. 0.01. 1 0.0. 0.0: 0001! 0.90 9.03.! 2 .0. 0.030 010. 0.9: 0,.30 3 .00 3.000! 010: 0.920 030' ‘ 9‘ In! 3'n'1l 36.000. 3.0.050 36.0.”: 5 3' .10 00°. 0001! 0.60 0.0: 6 0°15 00°. 9.00 0. 009‘ 0.030 7 on: 0'00 01030 0.0! 3.0.50 a .0. "0°20 000- 9.039! 3.003! 9 .0: 0902! 0005 0.00 00.21 1 .0. 30.0130 3.0.0: 0.95 0.52. 2 .0 ".00 0.020 3'0.02: 3 31.05. 3'n.o1. Moms. 3-o.o:. 3.0.15. ‘ 6.00”!) 3'fl.Oo .030: 0.00 0":l 5 9.0: “c 0 09°10 0.00 3.0.3: 6 3.0.51 3.0.20 b'Jy’. 3.0.2! 3.005: 7 3.6.1: 3.0.050 3.010, _ _ DATA ASC2] _ 300.03. 3600-0: 3'n5041 1 36.9100 3.0002! 35.000. 3'0.C50 0.00 2 9001; 0.0. 099; 0. 010 0.0: 3 30'91Q0 3.30050 00°; 0.920 0.0: ‘ 9.00 ".02. 01°: 53'0. 6. 0.0. 5 9.01, 0.0. 0.0: 0.91. 0.0. 6 3'9-01A 3.00020 3.0101: 3.0.02! 6.0.010 7 3.9.00 3.0.01. 6.010. 3.0.9! 00.: a 9.01. 00°! 01°; 0.01! “.0: 9 3.0.10 3.00‘51 9.391. 3.0.050 3.000! 1 3.n.02. 6'0.0/ DATA ASCJI - Stoofiao 0,0: 0.01- 1 90°: "0°! 0' 010 0090 30.0.0. 2 3-q.o¢. So-o.o. Son. 02. 36.0.9. 3-o.ns. 3 1.0. 0.01. o. . 0.1. 0.31. 1 0.3. 30.0.0. 3.9.05. 0.0. 0.02. FDYLD DATA ”MOOMOUbunu 9 DATA DATA H NHOCNOUIOUMH DATA DATA DATA 1 DATA R1 I 001 317 9.0- 0.00 0.02: 0.90 30.0..) 3'00Q0 12.0.02! 3°0'011 3'0. 02: 15.0..1I 3.9.0» 3.0.00 6.0.01: 3'0 1: 3.0.150 9.9.1! 6'".05. 5.00020 3.0.030 3.0.0/ AACPI 12.0000 3.0951 15.0091 3.0.5. 6.9.0: 12'9-0- 3.0050 15.00J0 3.0.5. 6.0.0: 1 '9-0: 3.0350 1,000Jl 30095: 6.0-BI YLDP8/1200.0. 502.0. 15-o.no 3-3.o. 6.0.0/ PXBI 3901. 5. 3901. 6: 59.1. OI APP1AA’PZI 306.15: 30001“ PSCI 3°C- 02:300-0103‘0002A3‘00J2. 3‘0.003.6¢0.0203'03003/ ALLP‘I 273.196 ALLPBI 3‘1oOn 0. U: 0 302’0001 1020 1‘00.00 0.2! 500.00 0011 10.6.0: 3.6 O. 3'0. 0. 3.0010 3.7150 6.0.1; 3.0.0! 3".2! M5.1.3'0 0/ PLBPA1I 239.0: 02939.0 017 FLBPAzl 3900.92.39-0. 01/ 7L8? A3I 39"0950159“0156I 'LBPB I 3900.55: 39.0 903/ FLBPC/ 12'o.a;3'05 “1015'0 003'J910600. 0012'0o003‘0o8115'0130 1' .a.0 I 'LBPD1/3.:20 250 '0050, '1:2°I “9050: :0.50. in: 'Oo50. 12.0.00 3'-0-40a 3.030: 3':°:200 9.0000 - 3.91020! 30.0000 39.0-00 3.02.750 .0075. V1.50: 05.750 ...SOJ «1.26: -c.50; 1200.6. 3--o.50. 3-F:o. 3O¢0.5°A 9.0100" _ 3t'1.500 -a.50. '2.50: '0-501 -o.50. ~1.25. -o.Jn. 1201.6: 3--6.Soo 3'500- 3.5.9, sac-3. no. 3.0.6. 3. .0: 35"1:5§o 3.365, 'LBP02/30'2o25: '0050. *11201 -@.50: :3 50: ”HM!- -D.50. 12.0.0! 3..00!°0 3.0300 5010.40: 9.0.0: _ 30'1AZOA 35'0.fla 390Joc- _ 30-2.75a ‘D.75. ‘1150u -g.,5: :0, 500 -1.20» -o.Su. 12-0.oa 3'-D.50A 3-350. at‘ioSOI 43.50.}. '2 .50! 'OQSQI 9.9500 v1.25. -o.30. 12-ooo. 3c-6.§oo 305: o. 2 9.0. 3:300. 35:¢3.00. ngog. ”c 1- 50a T'LBPI 3:2:1310601:7770303g213:3‘3.67733.93660I3'6i193. i5. 729. 3&2. 3070303. 50503011.OB9.109.16‘0502. UOI rOKPA rOKPX/ 3.15: 0.12. 00170 '0. 02/ GLXRB. OLPA/ D. 03: 0.0 05/ PGng1aPGL1R2: PGLIR306LXRRO DAB I .D- 3. D: a. 06. 300/ GLPB/ 20.: 30.0 ‘01, o:ANr(1) I I hJ' GLIR¢I) ' OLIRD 3V!R(X) - POLIRicGLlfltl) PVL1R1(I) - paLxR2.sLtn(1) PVL1R2(I) . Pchns.sL1R«I) _ GLIRA(I) - aana . (vaULlRR'o.9'GLlRR) FDKtI) - FOKPA(I)OSAGREV(l)-EXP¢V0KPX'T) 100 318 'EXP‘DLPAIT) I 1 3 tr c4.ea. 5 .BP. J .50. 1:: oo 70 ‘LP‘IOJIL, I FID‘I.JDL)'PXD(10L)/ (TDtan)'PD(loJ)) I0 T0 ”D ALP(I JIL’ I FXD(I JoL)‘P:D(IoL)/ (VD!) J)'PD(IaJ)D'Du§ C1O§UM 0F PRODUCTIVTTY COEFFIC CA ”L (I; J) I SAL P(l. J) 1 ALP(I.JAL) C RELATIVE CMANDEP IN PRODUCT PBIC ES “I J) - ORPUH§(|4J))/PDH1tloJ) E R( Io J) C RELATIVE CHANGE: IN FAC'O N! L) l XD(!.L) V P8D"1(loL))/PXDM1(I.L C RELATIVE CHANGE( IN TREE CROP AGEo MOM OSITIDN AND PAST INPUT USE (J 0106 106 108 107 5095 ,OR; J,E0.11) DDT DO 0T0 ACTCR‘I J) I AOTC(!aJ) ' ACTCHt(loJ EIUR(IaJ- L) I DFXDKInJoL) ~ FXDNITIoJoL))/FXDH1(!AJ0L) CONT IN U C ADJUSTED FACTOR OF FACTOR DEMAND ELASTICITV HRT CROPI SIZE ME C ADJUSTED FACTOR 0F FFAcTORD EM AW ELASTICiY (1.4) - TABL! E (vaez.SHALezoulrr$2.K62cszw ESL! J) - TAB BLlE (vAea. SHALES.DIFPE3.K63.RSP( I: C ADJUSTED FACTOR 0F FACTOR DEMAND ELASTIC iTY HRT PROFITADILtTY 5‘1, J) . 1AaLxe (VAEA.SMALEA.DIFFE4.KEA.PROETY(loJ)! ADF C ADJUSTED FACTOR 0F FAcTDR DEMAND ELASTICITV HRT FACTOR PthE 112 D0 W1 L . . m ADFDE1€!: JoL) 3i TABLIE (VAEI:8MALE10DIFFE10KEIARXDRLIn L)) co N IN us c ADJUSTED FACTOR DEMAND ELA§TTCITY an: own FAcToR PR I: 101 as Ernopcx JnL) - ADFDE1(l:joL)'(i.o . AnrDEatlsJ) o ADFDESLIAJ) . 4!! c ADJgSTED6 FACTOR DEMAND ELA§TICITT “RT OHM As HELL A8 CROSS ERXCE 1‘ 116L ,3 (I. J.L) - ALPtloJ L)/<1.6 ~ 3AL:(IA oJ)a-EPDOE(I JoL) W0 C FACTOR DEMAND ELASTICITY NRT PRODUCY AD7055(I,J, L) a TABLIE (VA61. SMAL51.DxFF51.KEi.PDR¢I J)) 1EFDPPUAJaL) I ADFDE5(IDJAL)'(ioD A ADFDE2(I J) ‘ ADFDESLIDJ) ‘ 054(loJ!)/(1. 0N ' 5ALP . VErD(I.J.L) 0 Scaro(loJoL) . ACEfOIIaJaL) . Plero(1.J) I,J,L) I FX(!.J.L) Fx(l.J.L) a rDPP¢X.J.L)'FXH1(l.JoL) C C COMPUTE INDIVIDUAL CROP VItLD LEVEL C 166 165 172 174 162 FXR‘IDJIL) I (fXIIaJAL) ‘ FXH1(IIIILII/FXH1II5JILI DO 166 H I 2 r stLA~(x.J)i'svuLAu(x.J) I (1.6 - dAP(IoH))ISLDRIIoMIIALLPetloJ-H) ) ru scsvnn¢z,4; . SCEVDD(I.J) . ScethloJ.L)IALP(l.J0L ‘cEvnp([,J) . Acevnn(t.J) o ALP(I.J.L)IAcern(I.JoL) YZDYDD(],J) I szvDD(X.J) o VEFD‘IIJIL).ALPIIIJDLI Fuc5v(1.J.L) - AL°(I.J.L)'FXR(I:J-L) SFUCEY‘IOJ) I SFUCEVIIaJ) I FUCEV(10J.L) DO 174 K ' 1.7 SCEYLD(I1JI I SCEvLD(l,U) o YLDP“IOJIK’.SCR‘IIKI..LLPAIIDJOK) CONTINUE _ 'IEYLDIIIJ) I 0.0 CONTINUE 175 176 176 no 177 J-1.:§ Ir IJ.I°5 we ICEYLDIIIJ) . i BIE YLDGIIJ) . o lo 70 7a ACEYLDII- J) .3? I 1 QIEVLD‘IDJ’ I R ‘ZDYYY‘IDJ) I SCEYYYIIIJ) I 8 ACEYVVCIAJ) I I 1r (SCBYVYII. J1. Ir IACBVYVII.JI VLDPC¢I J) u 1. YLDN1IIIJ) l V YLDII: J) I YLD 321 ‘c‘ocllI 60 70 175 I0 LDPBIIOJ’.‘CICRII0’) IEVLD‘I. J) I ALPII:J;LI"IURII.J0L) YIDIIBJ) - Iznm cevLD (I. J) - scsvnnIon) OEYLDIIcJIZ- :CEYDDIIoJI LI. 0. o) scevvaI,I) . o. o LY. 0.0) AcevaII.JI I o. Io hJ) . IIEVLDII. J) . ACEIvvcx.J) . :zn UY'YIIIJ) . SCEVVYII J) - squANII.JI M! J) cII. JIIYLDH1IIoJ) C E COMPUTE LABOR BEHIND 9V SEISONS 402 408 I01 no 401 N I 1.2 ELB'DIIIAJIN) I fLBFDZIIoJ,N) I ELB'DSII:J,N) I 'LBYZDIIoJ. N) I IF IJoEO. 9 .0! 00 Yo 03 'LBACEIIAJIN) I 0 'LBACEIIIJIN) 0 WI CONTINUE ‘ 'LOSCE‘IIJAII I 'LBSCEIII J12)?! 004 I 1 FLBSCEII Ja1) I FLB'4IIIIJINI0FXRIIIJa1I FLBP‘2(IIJI:,"XRIIIJ. 2) 'LOP‘S‘IIJIN I'VXRIIIJ. 3) FLBPBIII JON).YZDIIOJI . J.EO.11) no to 492 FLaPCIIoJoNIIAcYCRIIaJ) i. 5 § é o. o FLascEII-Jut) . FLBPDiII.JoK)IBCB(IoK)' A 1 L" WI 405 FLBSCEIIIJARI I FLBSCEIIOch’ o VLBPDZIIoJIKIISCRIIIKII 406 177 00 406 N l 1 2 rLaPPPII.J.N; - rLaroIII.J.~I . FLBFD:II.JIN) . rLernscl.J.uI . FLBIIoJ. N) I (1 ALL'AIII J KI flZDIIoJoNI I FLBACEIInJoNI) I ELBSCEIIAJIN) 0L 0 VLB'PRIII J. NIIIFLBOIIIJoN IrLaII.JI a (FL8(I.Ja1) . rLaIi.J.2)I.rrLaPII.JI g COH'UYE PRODUCTION COS? AND VOIAL REVENUE 182 151 183 100 c U5230 00‘ Io.fl avctx) I 0.0 o zoo J I a.1s UVCIIiJ) I {.0 I I 3 3cI’.JI - uGcII. J) . PX(I LIIFXIIanL) exxc anL) - see EAPIIIoJoLIIRAnRtI) GONT' as ‘N 3V VCI I I SVC(II YVOI ) I YVCIII o UVCIIN o UVCIIIJI'AII J) Nue 1 row IurnAsvnucruIe 322 rDVLn ° 'SJtB n c K)-PSGU , "4 uc;;§§,“:"c?sI" o PSCHhKIIDSCUJK) :33 a ”RED ox( )) rox(1) I svc¢l FDVLD 3; :§§§:}’ ;t§°',°,§‘, . augléI)?L(I) . schtv . roxcx) '“3t3 Kc!) . L1) rn 'VL1(I) - sv c( 70 . o o _ . r Y D I; i35t1§§§hfiav. ('3A«x§:1‘3Ltl,S.« APP:(I)IIPVL1(I) - (70K!!! rgvto ‘ roan ‘pVL2I:?‘§'§55«x) - 70K(I) - 0L(II - PvLItI) :33tg " ‘PVszééi'fifi)°E3un . ILsz-eLm . PVLIMHMPVLIHD . mm 1QCC!’ . uxna(x).PVL2(Il)/18. '°.Il:“ FD;Lg Hflicosnl) I TVOII) ° UCIs‘I’ ‘ :gytp trox(l) - rox¢:)/1§.OIY:(I) rova vath) I aL¢!)/£3.itYA .; !, FDYLD TPVL1II) - pVLicx)/13. lI" roan TPVLZtl) . PVLBQI)/13. a-TAt PDYLD SP xx u» . 920 roan TREV(;IJ"01 13 FDYLD D $39Ix.JJ - YLD(I.J)'4(I9JI J) F3¥t3 sew .J) - PAVOIIIJI'TP'III FDYLD YREVH) - T:EV¢I) 0 “Nb.” FDYLD 3mm - Wm ' H . W. - - am: “'33:; 39:53:35. « ° ”"°”’ mtg v o 166 °°?I~(;, a YREV(!) . TPgOSTCI) :gvtn rnan S conpure DISTRIBUTED LAOS ruan c FDVLD c ”‘3;"{:ngE E‘gLI 5(1, . or.¢oLG.(1) . oanott’)/oAa :33tg no 199 J- .1 ruan IEIIEIS I quxijJ. DYI(YLD(I-J) - YDCIJJII/DAB E8$t3 O 196 L . FDVLD DxDH1(j :;L) F];:?(3:LIH . DT,('X(“J,L) . rxotl JDL))/DAB 'gItg 196 rxn(x . - u))/nAa r r 093?: 3 ~)1-2 FLEDIIJJJN) o ot.(rL9¢I.J.N) - FLBD(IvJ' fifiltfi L I I 197 CONTINUE FDYLD 195 GONYINUE rDVLg zooo CONTINUE s :gItD c VERAGES 08 YOTAL E NATIONAL A FDYLD g CONPUT 1: FDYLg O 201 J ' 10. FDVL 2AVLD(J) - 6.6‘ rnan 4A?fLB(J) a 0.0 FDVLD STPPIJ) - Bio FDVLD Do 200 L - .3 rDVLD zoo AArx(J.La - 0.6 FDYLg Do 201 N ; 1.: l Egltn N I 201 SArggéJi ' f: 3: . FDYLD ngngJ) . AAVLDtJ) ‘ 4‘3 JI‘VLDII'J AVLDtJ: - AAYLDOJ’IASORGM AAT 'LS‘J) 0 AA H'LO‘ ) t A‘ICJ)07'LIII. J’ IY'LBCJI - AaviLacJ)/u :::::J)Lu singca) . 799i IL J.L) - lav m4 204 AFX(J0L) - AA'I 4.13/189Rl DO 0202 I N 1 . aarua (J.u: - AavLa(J.N) . AC I$J1ctLa(i.J.u1 202 A?LB(J.N) I ALECOCJ9N11‘3°R( (J, ITP VL2 0 87'0“ I 0.6 STOL ' 000 TPVL: - 0.6 SSVC ' 00° STVC ' 00° SUCIS U 0.0 CO ' 0.0 srcosr u 0.6 TIEV I 0.0 aTNrIN - 0. IAT'LB ! 0. .3 ST'OK I ST'OK o OTFOK(II STGL I STOL 0 'TPVLI ITPVLZ I 370087 I “(1) l STPVL O Y'VLl‘II STPVL: ‘ Y'VL2(II 370081 0 "7'???: (I) STREV HSTRIV 203 STUFIN' I STN'IN 0 Y“? ”I (I! SSTPP - Sfpn11a . svprgz) o SIpP13) 0 SYPPlaa . :7PP(6) e BTPPIO) c . c acacecare INPu? USED IN ouguvva c 0 205 L o 1 205 N O O N O V 208 SFX‘I: L) gFXOIIu L) s 35% OI ' 1 00 7L' SSVX(L1 ' IIL ”I FXO‘II JIL) : g! IoJlL’I'X.‘h DO 7I'103 ' 1:1. l 1? I L) t 'x‘Ioh ~13 ssrxit) . srx IoL 531 323 415: J).'X(I:J.L’ SFXO(I¢L) I sSFl°(IpL’ t 'lozIoJaL) SSFXO L) CONTINUE sAAin I Do 259 J SAAFXi I RETURN END 3730(L) o SfXI‘IoL) 101 SAVVLB t ‘177L31JI 0.6 101‘ 3‘Irx1 6 AAVX(4o21 NH 324 ruucrlon raaLtg (VALoSHALLo ntrr.x.nu~nv) nxneusxou nun - AHIN1:AH3§1(DUHNY'SNALLa5.0)oFLoATCKI'DlFFI / ... I ... . o O ”U :1 I- TABLIE I (VAL(I¢1’- KVAL(I_))'(DUH'FLOAT(1'1)'DIFF)IDIVFOVAL(I) RETURN END SUBROUTINE DELDDtRXNR. ROUTRo CROUTR.DEL:IDT.DT.K:ARI DIMENSION CROUTR DEL1 I (DEL-FLOAT(lDT))/(PLOAT(K)'DT) FIDT I FLOAT(ID ROUTR Q 0.0 AR I 0. 0 DO 2 J ! 1.!DT BIN I RINR/VL0A7(ID7) AR I AR 0 DTOCROUTR(K) 001 I'loK ABC 8 CROUTR(II CROUTR(I) I A80 o (FIN ° A8C)/DEL1 SIN I A ROUTR a ROUTR c CROUTP(K) RETURN END 00‘0“...” 3125 gUgEOUTINER 9ELLVFIRIN0ROUT. .RoSTRGoPLR. DEL05ELP10T0K) NSIO ;K I FLOATIK I 10 0 (DEL ' DEL')/(FK'DT) o PLR'DELP/FK 10r- 1. m0 MInELé-AHAX110. A 0 50.07/102L005104f110m HELP IKDEL 00 20 J I 11K107 00101W1 011) a :11) . 0-1011 . 1) - 0.001)) p O 71 u :10) a 010) . 0.1010 . 0.0107) couvxnu 0100 - "0.0’ 00 50 I - 1.0 0700 a 0700 . 0110.051161 nus N o (A O I} O I: 4 a. 060: «500 10/10/74 000 uusrLsn 2 1200 100 so. 07 10100174 1‘031.39.Y‘2°‘2 14.31.40 400 0000- 10/10/74 :14.i:.92 14.51. 40. 125404.00110000. 799. 40300.1100. 14051 ‘0 L‘ST ”C ”‘3' 1 5 .1 7 14.31. 40. 0000- 0572c U050 DOLLAR 0.11005 000111.45 14. 31. 40. 000010 :0“ VALUE 10000000. 01 14.51.40. cr-w 0:0.c .072- .011 0 14.01..40.017100.0109L.v1éL071911 14.31.41.0LDPL M01 lELD11NALPL CH 14031.41.C"PP I50. .0775 .571 9 .00 $000 - _ _ 14.01. 1.00 0000 0005 000300000000 14.31. 55. 000410 canPLeu 14.31. 55. 0h 20 050.1.000- 5.227 0 .53 14.11. 55. Avracuucorvucr.cuconf. 14.51. 55.00 Pvucr . cch0 01. aucopv 14.01. 55. rlLE ATTc wen 14oJ1-55.C”'IECC1.944: 5.505 . 055 141J1. 55. COPYNC'oCOHPILEInaP 14.-1.59.NL 11 000 ______ 14.21. 55 00 00000151 000000000770 14.12.01.NL 4400 14.12.11.00 000000000155 0000fi 001041 14 ‘ . .NL 110000 14.12.11.00 000000000207 000000 14 .1. 71L500005c HA! 0000 0014000. 14. 2. I. 5» 000.0 14. 2. I. 00 0000000002075 .006000002031 14 *2 .c- useu2.551v01021 000.11 14..2. I. 0: use 014. 049 000 VALUE! 000.04 14 *2. Iuse 000. 005 u.uv110£0 000.24 14.52. 1.70711 oonrura VALUEA 002 0 14.34.'|..000037 50050 00107. 001727 LINIS PRINT. r00 5 001102 AT 002. BIBLIOGRAPHY A.l A.2 A.3 AA A.5 A.6 A.7 A.8 A.9 BIBLIOGRAPHY Abkin, Michael H. 1972. "Policy Making for Economic Developnent: A System Simulation of the Agriculttn‘al Econany of Southern Nigeira." Staff Paper Series 72—4. Department of Agricul- tural Economics, Michigan State University, East Lansing. , and Rossmiller, George E. 1972. "Sector Analysis 553 the General Systan Simulation Approach to Agricultural Development Flaming." Paper presented at CENPO Workshop for Agricultural Planners, Islamabad, Pakistan, Novenber 27—Decenber 4. Adams, D. W. and Singh, I. J. 1972. "Capital Formulation and the Firm-Household Decision Making Process." Econanics and Sociology Occasional Paper No. 111. Department of Agricultural Economics and Rural Sociology, Ohio State University, Colmbus. , and Coward E. Walter, Jr. 1972. "Small-Farmer Development Strategies: A Seminar Report." Research and Training Network, the Agricultural Development Council, Inc. , New York. Adelnan, Irma and Thorebecke, Erik, ed. 1966. The Theog and Desi of Ecorgmic Development. Johns Hopkins Press, Elmore. Ahan, Coong Y. and Singh, I. J. 1973. "Comparative Policy Sinnlations: Economic Development in Brazil to 1985." Economics and Sociology Occasional Paper No. 169, Ohio State University, Colmbus. Alcantara, Reinaldo and Prato, Anthony A. 1973. "Returns to Scale and Input Elasticities for Sugarcane: The Case of Sao Paulo, Brazil," American Journal of Agricultural Ecormdcs, 55:577—83. Allen, R. G. D. 1964. Ethanatical Analysis for Economists. St. Martin's Press, New York. Arrow, K. J. , et al. 1961. “Capital-labor Substitution and Economic Efficiency," Review of Econcmics and Statistics, 34:225-148. 326 A. 10 B.1 B.2 B.3 B.4 B.5 B.6 B.7 B.8 B.9 B. 10 327 Auer, Lordwig and Heady, Earl O. 1964. "The Contribution of Weather and Yield Technology to Change in U. S. Corn Pro- duction,1939 to 1961, " Weather and Our Food Supfly. CAED Report 20, Iowa State University, Ames, Iowa. Barker, Randolph. 1961. "Supply Function for Milk Under Varying Price Situation," Journal of Farm Economics, 43:651-58. , 1971. '“I'he Evolutionary Nature of the New Rice Tech- nolo,gy " Food Research Institute Studies, 10: 117-30. Stanford University, ;Stanford, California. Bateman, Merrill J. 1969. ”Supply Relations for Perennial Crops in the less Developed Areas, " in Wharton, Cliftornt R. , Jr. , Subsistence 'culture and Economic Develo pp. 243-5 . 13 Co., 'cago. Beckford, George L. "Strategies for Agricultural Development: Comment," Food Research Institute Studies, 11:149-54. Stanford University, Sword, California. Behrman, Jere R. 1966. "Price Elasticity of the Market Surplus of a Subsistence Crop," Journal of Farm Economics, 48:875-93. Bieri, Jurg, et a1. 1972. "Agricultural Technology and the Distribution of Welfare Gains," American Journal of Agri— cultural Economics, 54. 801-8. Binswanger, Hans P. 1974. "A Cost Function Approach to the Measurement of Elasticities of Factor Demand and Elasticities of Substitution," American Journal of Agricultural Economics, 56: 377—86. Black, John D. , and Bonnen, James T. 1965. "A Balanced United States Agriculture in 1985," National Planning Association. Special Report 42. Washington, D.C. Blakeslee, Leroy L. , et a1. 1973. World Food Production, Demand and Trade. Iowa State University Press, Ames, Iowa. Bonnen, James T. 1963. "Analysis of Long-Run Economic Projec- tions for American Agriculture. Proceedings, TWelfth Annual Meeting, Agricultural Research Institute, October 14—15, 1963. Committee on Agricultural Policy, Agricultural Board on National Academy of Science, National Research Council, Washington, D.C. , and Cromarty, William A. 1958. "The Structure of Agriculture." In Heady, Earl 0., et a1, Agricultural Adjustment Problems in a Growing Econgy; pp. - . Iowa State University Press, Ames, Iowa. B. 12 B. 13 B. 14 B. 15 B. 16 B. 17 C.1 C.2 C.3 C.4 C.5 C.6 C.7 328 Bowden, E. 1970. "A Comparison of Predictions from Static and Dynamic Model of Farm Inrnvation," Rural Sociology, 35:253-60. Boyne, David H. and Johnson, Glenn L. 1958. "A Partial Evalua- tion of Static Theory From Results of the Interstate Managerial Study," Journal of Farm Economics, 40:458-69. Bradford, Lawrence A. and Johnson, Glenn L. 1964. Farm Manage— ment Analysis. John Wiley and Sons, Inc. , New York. Brake, John R. 1972. "Capital and Credit." In Ball, A. Gordon and Heady, Earl O. , ed. Size, Structure and Future of Farm, Iowa State University Press, Ames, Iowa. Brown, Murray. 1968. On the Theory and Measurement of Tech— nologi cal gme. Cambridge University Press, London. Byerlee, Derek and Eicher, Carl K. 1972. "Rural Employment, Migration and Economic Development: Theoretical Issues and Empirical Evidence fran Africa." Rural Employment Paper No. 1, Department of Agricultural Economics, Michigan State University, East Iansing. . 1973. "Indirect Employment and Income Distribution in Effects of Agricultural Development Strategies: A Simulation Approach Applied to Nigeria." African Rural Ehployment Paper No. 9. Department of Agricultural Economics, Michigan State University, East Iansing. Campbell, R. R. 1966. "A Suggested Paradign of the Individual Adoption Process," Rural Sociolggy, 31:485-466. Chattopadhyay, S. N. ‘and Pareek, U. 1967. "Prediction of Multi-Practice Adoption Behavior from Some Psychological Varialbes," Rural Sociology, 32:324—33. Chenery, Hollis B. , ed. 1971. Studies in Developmept Flaming. Harvard University Press, Cambridge. Cochrane, Willard W. 1958. "Some Additional Views on Demand and Supply." In Heady, Earl O. , et al, ed. A§icultural Adjustment Problg_ms in a Growing Economy, pp. - . Iowa State University Press, Ames, Iowa. . 1958. Farm Pricesz Mg? and Reality. University of Minnesota ess, Mimeapo is, Minn. . 1965. The City Man's Guid_e to the Farm Problem. University ofbfimesota Press, Minneapolis, Minn. Computer Library on Agricultural Systems Simulation. 1974. Michigan State University, East Iansing. C.8 C.9 329 Cooper, George R. and McGillen, Clare D.. 1967. Methods of Si ' g1_1a__l and §%stam Analysis. Holt, RJnehart am Winston, Inc., New Yo . Cownie, John, et a1. 1970. "The Quantitative Impact of the Seed-Fertilizer Revolution in West Pakistin," Food Research Institute Studies, 9:57-95. Stanford University, Stanford, orn1a . C. 10 Cromarty, William A. 1959. "The Farm Demand for Tractors, Machinery and Trucks," Journal of Farm Economics, 41:323—31. D.1 D.2 D3 D4 D.5 D.6 D.7 D.8 D.9 Daines, Samuel R. 1972. "Partial Implications of the Analysis for Decision-Making in the Agricultural Sector." Analytical Working Document #6, Colombia Agricultural Sector Analysis. AID-IA/DR/SASS, Washington, D.C. Dandekar, V. M. 1972. "Effectiveness in Agricultural Plaming Development Processes and Plannigg (No. 19) A/D/C Teaching Forum. Agricultural Deve ogment Council, Inc. , New York. Day, Richard H. 1963. Recursive Programming and Production Response. North-HollEmH Publishing Company, Amsterdam. and Singh, I. J. 1972. "A Dynamic Microeconomic Madel of Agricultural DevelOpment.” Social Systems Research Institute, 7135. University of Wisconsin, Madison. Dean, Gerald W. andHeady, Earl O. 1958. " e in Stlpply Response and Elasticity for Hog," Journal of Farm Economics, 40: 845-60 . de Haen, Hartwig. 1973. "Preliminary User's Guide to the Recursive Linear Programming Resource Allocation Component of the Korean Agricultural Sector Model." KASS Working Paper 73-2. Michigan State University, East Lansing. and Lee, Jeung Han. 1972. Dynamic Model of Resource Allocation for Agricultural Planning in Farm Ko .. Application of Recursive Programming within a General System's Simulation Approach. Agricultural Sector Analysis and Simulation Projects. Project Working Paper 72-1. Michigan State University, East Lansing. Denison, E. F. 1962. "The Sources of Economic Growth in the United States." Supplementary Paper No. 13. Committee for Economic Development, New York. Desaeyere, W., et a1. 1967. _l€_ng—Term Develo§t of S7982? and Demand for Agricultura ts 1n gum, 5. SEE'ecentrum Voor Economisch En Sociaal Onderzo sstraat 13, Antwerpen, BelgiLm ek, Prin- 330 D. 10 Dorner, Peter. 1971. "Needed Redirections in Econcxmic Analysis for Agricultural Development Policy," American Journal of Agricultural Wcs, 53:8-16. D. 11 Duloy, J. H. , et al. 1974. "Agriculture and the Energy Crisis: A Case Study in Mexico," Paper presented at the AAEA Annual Meeting, Texas A & M University, College Station, Texas, August 18—21. E.1 Edwards, Clark. 1959. "Resource Fixity and Farm Organization," Jommal of Farm Economics, 41:747-59. E.2 Eicher, Carl K. , et a1. 1970. "Employment Generation of African Agriculture." Institute of International Agriculture. College of Agriculture and Natural Resources, Research Report No. 9. Michigan State University, East Lansing. E. 3 Evenson, Robert. "The Green Revolution in Recent Development Experience, " American Journal of Agricultural Economics, 56: 387- 93. F.1 Falcon, Walter P. 1966. "Programming Nbdels on the Planning of the Agricultural Sector, Comment," In Adelmen, Irme and 'I‘norebecke, Erik, ed. The Theory and Design of Economic Develomt. Johns Hopkins Press, Baltimore. E.2 Feaster, J. G. 1968. "Measurement and Determinants of Innova- tiveness among Primitive Agriculturalists," Rural Sociology, 33:339—48. F.3 Fernando, C. 1963. "Projections of Potential Supply--Tree Crops." FAO, Agricultural PlaIming Comrse, 1963, Agricultural Flaming Studies, Rome. F.4 Ferris, John N. and Suh, Han Hyeck. 1971. "An Analysis of Supply Response on Major Agricultural Commdities in Korea," KASS Special Report 4, Michigan State University, East Lansing. F.5 , et a1. 1972. "Investment Priorities in the Korean Agricultural Sector." Korean Agricultural Sector Study Team. Michigan State University, East Iansing. F.6 Fishel, Walter L. 1971. Resource Allocation in Agricultmtal Research. University of Mimeosta Press, Minneapolis. F.7 Fletcher, L. B., et a1. 1970. Guatemala's Economic Development: The Role of Agriculture. Iowa State University Press, Erie‘s, Iowa. F.8 Fliegel, F. C. , et a1. 1968. "A Cross-National Comparison of Farmers' Perception of Innovation as Relative to Adaption Behavior," Rural Sociology, 33: 437- 49. F.9 F.10 F. 11 F. 12 G.1 C.2 C.3 C.4 G.5 C.6 G.7 C.8 C.9 331 Fox, Karl. 1953. "The Analysis of Demand for Farm Products," U.S. Department of Agriculture. Technical Bulletin 1081, Washington, D Forrester, J. W. 1961. Industrial Dynamics. Massachusetts Institute of Technology Press, Cambridge, Mass. 1968. Principles of System. Wright-Allen Press, C , Cambridge, Mass. Frisch, Ragnar. 1959. "A Complete Scheme for Computing A11 Direct and Cross Demand Elasticities in a Model with Many Secotrs," Econometrica, 27:177-96. Gittinger, J. Price. 1972. Economic Analysis of Agricultural Projects. Johns Hopkins Press, Baltimore. Government of the Republic of Korea. 1971. The Third Five-Year Economic Develomt Plan. Seoul, Korea. Goreux, Louis M. and Manne, Alan S. ed. 1973. Multi-Level Pl ' : Case Studies in Mexico. I‘lorth—Hmishing Campany, Amsterfi. Griffin, Keith B. and Enos, John L. 1970. P1ann__1ng' Develmgt. Addison-Wesley Publishing Co., london. Griliches, Zvi. 1957. "Specification Bias in Estimates of oduction Functions," Journal of Farm Economics, 39:8-20. . 1957. "Hybrid Corn: An Exploration in the Ecoamics of Technological Change," Econometrica, October. . 1958. "The Demand for Fertilizer. An Economic Inter- Pretation of a Technical Change," Journal of Fa__rm Econmtics, 40: 591-606. . 1958. "Research Costs and Social Returns: Hybrid Corn and Related Innovation," Journal of Political Economy, 66:419-31. 1959. "The Demand for Inputs in Agriculture and a Derived Supply Elasticity," Journal of Farm Economics, 41: 309-322. 1959. "Distributed Lags, Disaggregation, and Regional Demand Functions for Fertilizer,‘ 'Journal of Farm Economic____s_, 4].: 90-102. 1960. "Estimation of the Aggregate U. S. Farm Supply Function, " Journal of Farm Economics, 42: 282 -93. G. 12 H.1 11.2 H.3 H.4 11.5 11.6 H.7 H.8 H.9 H. 10 H. 11 H. 12 332 . 1963. "The Sources of Measured Productivity Growth: UiSéBAgZECMUIre, 1900-1960," Journal of Politigal Econcmy, Hahn, F. H. and Mattevs, R. G. 0. 1964. "The Theory of Economic Growth: A Survey," Economic Journal, 74:779-902. Halter, A. N. , et a1. 1970. "Simulating a Developing Agricul- tural Economy: Methodology and Planning Capacity," yuan Journal of Agricultural Economics, 52:272-84. Halvorson, Harlow W. 1958. "The Response of Milk Production to Price," Journal of Farm Economics, 41:1101-13. Harris, John R.. and Todaro, Michael P. 1970. "Migration, Unemployment and Development: A Two-Sector Analysis," American Ecoromic Review, 60:126-42. Hathaway, Dale E. 1963. Goverment and Agriculture. MacMillan Company, New York. Havens, A. E. 1965. "Increasing the Effectiveness of Predicting Innovativeness," Rural Sociology, 30:150-65. Haver, Cecil B. 1958. "Institutional Rigidities and Other Imperfections in the Factor Markets." In Heady, Earl O. , et al., ed. Agicultural Adjustment Problems in a Mg W, Iowa State University, Ames, Iowa. Hayami, Y. 1964. "Demand for Fertilizer in the Course of Japanese Agricultural Development," Journal of Farm Economics, 46:766-79 and Ruttan, Vernon W. 1971. ’cultural Develo t: An International Perspgctive. Johns HopEiE Pfess, Baftimore. Heady, Earl O. 1949. "Basic Economic and Welfare Aspects of Farm Technological Change," Journal of Farm Economics, 31: 293-316. . 1960. Ecormfics of icultural Production and Resource Use. Ffentince—fif, Inc., Englewood Cliffs, N.J. . 1961. "Public Purpose in Agricultural Research and Efication," Journal of Farm Economics, 43:566—81. . 1961. "Uses and Concepts in Supply Analysis." In — Heady, Earl O. , et a1. , ed. Agricultural Supply Funct10ns- Estimating Techniques and Interpretation, Iowa State UdiVersity Press, Ames, Iowa. H. 14 H. 17 H. 18 H. 19 H. 20 H. 21 H. 22 H. 23 H. 24 H. 25 H. 26 H. 27 333 . 1965. Agricultural Policy Under Economic Devel pment Iowa State University, Ames, Iowa. 0 . 1966. "Priorities in the Adoption of Improved Farm Technology." In Iowa State University Center for Agricul- tural and Economic Development. Economic Develogggt of AgEculture, Iowa State University Press, Ames, Iowa. and Candler, Wilfred. 1958. Linear Proggammmng‘ lbthod. Iowa State University Press, Ames, Iowa. 1968. "Processes and Priorities in Agricultural Evelopment." In McPherson, W. W. , ed. Economic Develo - ment of Trqaical Agriculture, University of Florida Press, Gainesville, Fla. and Dillon, John L. 1964. Agricultural Production Functions. Iowa State University Press, Ames, Iowa. and Tweeten, Luther, G. 1963. Resource Demand and Structure of the Agricultural Industry. Iowa State University Press, Ames, Iowa. and Yeh, Martin H. 1959. "National and Regional Demand Functions for Fertilizer," Journal of Farm Economics, 41: 332-48. , ed. 1971. Economic Models and Quantitative Methods for Decisions and PIann1ng 1n Afiiflfifie. Iowa SE—i‘t'e ver31 ess, s, owa. . 1955. "The Supply of Farm Products Under Conditions of Full Employment," American Economic Review, 452228-38. Herdt, R. W. and Mellor, John W. 1964. I'I‘he Contrasting Response of Rice to Nitrogen: India and the United States,” Journal of Farm Economics, 46:150-60. Hertford, Raed. 1971. "Sources of Change in MexicanAgricultural Production 1940—65," Foreign Agricultural Economic Report No. 73, Economic Research Service, U.S.D.A., Washington, D.C. Hirshman, Algert O. 1958. The Strategy of Economy W. Yale University Press, New Haven, Conn. Holland, E. P. 1970. "Discussion: Macro-Sixmllation Madels," American Journal offlic—ultural Economics, 52:284-286. Hsieh, S. C. and Ruttan, Vernon W. 1967. Environmental, Techno— logical, and Institutional Factors in the Growth of R1ce Production: Philippines, Thailand and Taiwan, Food Research Institute Studies, 7:307—42. 334 H. 28 Huh, Sin H. and Lee, Jeung Han. 1971. "Feed Marketing, Live- stock-Feed Price Cycles and Livestock Supply Functions," AERI, Research Report Series No. 35. Agricultural Economics Research Institute, Ministry of Agriculture and Forestry, Seoul, Korea. J .1 Jensen, Harold R. 1958.'Tec1rmologica1 Research in Relation to Adjustment. In Heady, Earl O. , et al., ed. ggg'cultural Adjustment Problem in a GrowingE commy. Iowa tate University Press, Ames, Iowa. J .2 Johnson, Glenn L. 1956. "Classification and Accounting Problem in Fitting Production Emotions to Farm Record and Smrvey Data," In Heady, Earl 0., et al., ed. Resource Productivityy Returns to Scale, and Farm Size, Iowa State University Press, Ames, Iowa. J .3 . 1975. "Sources of Expanded Agricultural Production," Policy for Commerical Agriculture—Its Relation to Econrmic Growth and Stability, Joint Econamic Camittee. 85th Congress of the U.S., November 22. ~14 . 1958. "Supply Function-—Some Facts and Notions,’ In Heady, Earl O. , et al., ed. Agricultural Adjustment P_r9____b1e1e in a Growing Economy, Iowa State University Press, Ames, Iowa. J .5 . 1960. "The State of Agricultural Supply Analysis," Journal of Firm Economics, 42:435-52. J .6 1960. "Review of Nerlove," Agricultural Economics Research, 12: 25— 8. J.7 . 1960. "Value Problems in Farm Management, " Journal oéggicultural Economics, 14: 13-31 J. 8 . 1961. "Budgeting and Engineering Analyses of Normative Supply Fimctions,‘ In,Heady Earl 0., et al., e.d Agricul— tural Supply Emotions—Estimating Techniques and Integre- _t_a_____tions, Iowa State Univers1ty Press, Ames, Iowa. J .9 . 1970. "Discussion: Macro-Simulation deels," American Journal of Aggiucultral Economics, 52:286-8. J. 10 . 1970. "The Role of the University and Its Economists in Economic Development.” Presented as the J. S. McLean Visiting Professor Lecture. Department of Agricultural Economics, University of Guelph, Canada. J .11 . 1972. "Alternatives to the Neoclassical Theory of the Firm," American Journal of Agricultural Eoonemics, 54:295—303. 335 J.12 . Undated paper. "A Review of Work Involving Traditional Projections and Pbdern Computer Simulation." Presented at Dijon Meeting of TACAC. Michigan State University, East Lans' ing. J .13 , et al. 1961. A Stud of Managerial Processes of Midwestern Farmers. State University Press, Ames, Iowa. > J .14 , et al. 1969. "Strategies and Recommendation for Nigerian Rural Development 1966/1985." Consortium for the Study of Nigerian Rural Development. CSNRD 33. Michigan State University, East Lansing. ' J.15 , and Zerby, Lewis K. 1973. What Ecommists Do About Values. Department of Agricultural Economics, Centerfor Rural Manpower and Public Affairs, Michigan State University, East Lansing, J .16 ed. 1972. The Overproduction Trap in U. S. Agriculture. Johns Hopkins University Press, Baltimore. J .17 and Hardin, Lowell S. 1955. "Economics of Forage Evaluation." Purdue Agricultural Experiment Station Bulletin 623. Lafayette, Indiana. J. 18 Johnson, Sherman E. 1965. "Combining Knowledge, Incentives and Means to Accelerate Agricultural Development." In Iowa State University, Center for Agricultural and Economic Development. Econamic Development of Agriculture, Iowa State University Press, Ames, Iowa. J .19 Johnston, Bruce F. and Cownie, John. 1969. "The Seed-Fertilizer Revolution and Labor Force Absorption," American Economic Review, 59:569-82. J .20 and Southworth, Herman M. 1967. "Agricultural Develop— ment: Problems and Issues," In Southworth, Hermand M. and Johnston, Bruce F. , Agricultural Development and Economic Growth, Cornell University Press, Ithaca, N.Y. J .21 and Mellor, John W. 1961. "The Role of Agriculture in Economic Development," American Econamic Review, 51:566-93. J .22 Jorgenson, Dale W. 1961. "The levelopment of a Dual Economy," Economic Journal, 71__309-34. K.1 Kae, BongMyung, Lee, Jeung 11, and Lee, Jeung Han. 1968. "A Study on Optimum Use of Fertilizer Resource in Rape Production in Korea," Journal of Institute for Agricultural Resource Utilization, Ch1nju Agricultural College, Chinju, Korea. K.2 K.3 K.4 K.5 K6 K7 K.8 K.9 K. 10 L.l L.2 L3 336 Kehrberg, Farl W. 1961. "Determination of Supply Functions from Cost and Production thctions." In Heady, Earl O. , et a1. , ed. Agricultural Supply FngtionSnEstimating Techniques and Interpretation, Iowa State University Press, Ames, Iowa. Kelley, Allen C. , et a1. 1972. Dualistic Economic Develgpmit: Theogy and Histgy. University of Chicago Press, Chicago. Kim, Sang Gee. 1971. "The Impact of PL 480 Shipments on Price and Domestic Production of Foodgrains in Korea, " Journal o_f__Agricu1tura1 Economics (Korea), 8: 72- 85. Kislev, Yaav, et al. 1973. "The Process of an Irmovation Cycle,” American Journal of Aggcultural Economics, 55:28-37. Knight, Dale A. 1961. "Evaluation of Time Series as Data for Estimating Supply Parameters. " In Heady, Earl O. et al., ed. Agricultural Supply thctions-—Estimating Techniques and Interpretation, Iowa State University Press, Ames, Kravis, I. B. 1970. "Trade as a Handmaiden of Growth: Similar— ities Between the Nineteenth and Twentieth Centuries," Economic Journal, December. Kresge, David T. 1970. "Discussion: Macro-Simulation IVbdels," American Journal of Aggcultural Economics, 52:288-90. Krishna, Raj. 1962. ”A Note on the Elasticity of the Marketable Surplus of a Subsistence Crop," Indian Journal of Agricul- tural Economics, 17:79-84. . 1967. "Agricultural Price Policy and Economic Develop- ment.‘ In Southwarth, Herman M. and Johnson, Bruce E, ed. Agricultural Development and Economic Growth, Cornell University Press, Itlfica, New York. lau, Lawrence J. and Yotopoulas, Pan A. 1971. "A Test for Relative Efficiency and Application to Indian Agriculture," American Economic Review, 61:94—109. . 1972. "Profit, Supply and Factor Demand Functions," American Journal of Agricultural Economics, 54:11-18. Learn, Elmer W. and Cochrane, Willard W. 1961. "Regression Analysis of Supply Functions Undergoing Structural Change." In Heady, Earl 0., et a1. , ed. Agricultural Supply Func— tions-—Estimatiqg;Techniques and Interpretation, Iowa State University, ‘Ames, Iowa. 337 L4 Lee, Jeung H. 1969. "Factor Demand and Product Supply Functions .1 and Optimal Allocation Patterns Derived from Experimental ‘ Fertilizer Production Functions," Journal of the Institute _fgr_Agricultm‘a1§_e_source Utilization. Chinju National Agricultural College, Chinju, Korea, 3:1—68. L.5 . 1972. "Potemtials of Hog Supply: Adaptability of Some New Production Factors Related to Hog Production in Korea," Journal of the Institute for Agricultural Rgsource Utilization, Chinju. National Agricultural Colelge, Chinju, , Korea, 6:159-204. 1 L6 . 1963. "On Problems of Foodgrain Production," (in Korean). Research Bulletin of Chinju Agricultural College, 2:119—29. Chinju, Korea. L. 7 and Chlmg, Chang Hoon. 1966. "A Study of Substitution Relationship Between Mixed Concentrate and Sweet Potato Ensilage in Pork Production," Research Bulletin of Chinju Agricultural College, 5:29—38, Chinju, Korea. L.8 and Kim, Hoo Keun. 1970. "Optimum Combination of Wmter Crop Enterprises in Rice Paddy in Honan District-- Application of Linear Programming," Journal of Institute for Agricultural Resource Utilization, Chinju Agricultural College, 4:21—26. Chinju, Korea. L.9 . 1973. "A Suggested Dynamic Input-Output deel for Policy Simulation," Unpublished Mimeograph. Department of Agricultural Economics, Michigan State University, East Lansing. L. 10 . 1973. "Farm Resource Allocation Submodel of a Generalized Systems Simulation Model for Korean Agricul- tural Sector Analysis." Discussion paper for World Bank Colloquim on Advanced Methodologies for Agricultural Investment and Policy Analysis, January 29-39. L. 11 , et al. 1973. "El Salvador Credit monDevelopment Project," Unpublished mimeograph. Department of Agricul- tural Econamics, Michigan State University, East Lansing. L. 12 Lee, Teng—hui. 1971. Intersectoral %ita1 Flows in the Econanic Development of Taiwan, 1 95- 6 . Cornell Univer— sity Press , Ithaca. L. 13 Lehker, John N. and Manetsch, Thomas J. 1971. "Systems Analysis of Development in Northeast Brazil: The Feasibility of Using Simulation to Evaluate Alternative Systems of Beef Production in Northeast Brazil." Technical Report G-13. Midwest Universities Consortium on International Affairs, Division of Engineering Research, Michigan State University, East Lansing. L. 14 L15 L. 16 L. 17 L.18 L. 19 L. 20 L. 21 Ml M.2 M3 M4 338 Leibeistein, Harvey. 1963. Economic Backwardness and Economic Growth. John Wiley and Sons, Inc., New York. Lele, Ihe J. and Mellor, John W. "Estimates of Change and Causes of Change in Food Grains Production, India, 1949-50 to 1960—61." Cornell International Agricultural Developtent Bulletin 2. Cornell University, Ithaca, N.Y. Lewis, W. A. 1954. "Economic Development with Unlimited Supplies of Labor." Manchester School of Economics and Social Studies 22:139-91. 1972. "Reflection on Unlimited Labor." In Luis Eugenio Di Marco, ed. International Economics and Development, Academic Press, New York. 1955. The Theory of Economic Development. Urwin, London. Llevellyn, Robert W. 1965. Fordyn: An Industrial Dynamics Simulator. Raleigh, North Carojina. Privately printa. Lu, Yao-Chi and Fletcher, Lehman B. 1968. "A Generalization of the CES Production Function," Review of Economics and Statistics, 502449—52. . 1974. "A Production thction Approach to Measuring Pro- du—ctivity Growth in U.S. Agriculture." Paper presented at the AAEA Annual Nbeting. Texas ABM University, College Station, Texas, August 18—21. Malone, Carl C. 1965. "Some Responses of Rice Farmers to the Package Program in Tanjore District, India," Journal of Farm Economics, 47:256-69. . 1971. "Improving Opportunities for Low—Income Farm Occupied People: Some Indian Experiences." Paper presented in A/D/C Seminar on Smell—Farmer Development Strategies, Ohio State University, Columbus, Septerber 13-15. Manetsch, Thomas J ., et a1. 1968. "Computer Simulation Analysis of a Program for Modernizing Cotton Production in Northeast Brazil." Systems Analysis of Developlent in Northeast Brazil. Working Paper No. 3, Systems Science Group. College of Engineering, Michigan State University, East Lansing. , et al. 1971. "A Generalized Simulation Approach to Agricultural Sector Analysis with Special Reference to Nigeria." Michigan State University, East Lansing. 339 M5 and Leiclrmer, Sanford L. 1971. "Systems Analysis of Development in Northeast Brazil: The Feasibility of Using Simulation to Evaluate Alternative Policies for Development of the Brazilian Textile Industry." Technical Report G 13. Midwest University Consortium on International Affairs, Division of Engineering Research, Michigan State University, East Lansing. M6 and Park, Gerald L. 1970. System Analysis and Simula— tion with Application to Economic and Social System. Preliminary Edition. Department of Electrical Engineering and System Science, Michigan State University, East Lansing. M. 7 . Undated. "Use of the VDEL Subroutine to Simulate Project Implementation in Economic Development." Unpublished Paper. Michigan State University, Fast Lansing. M8 Phadows, Dennis L. 1970. 'cs of Commodi Production Cycles. Wright-Allen Press, Inc., Edge, Mass. M9 Meier, Gerald M. 1968. The International Economics of Develop- ment. Harpers Row, Publishers, New York. M. 10 Mellor, John W. 1966. The Economics of Agricultural DevelopIent. Cornell University Press, Ithaca. M. 11 . 1967. "Toward a Theory of Agricultural Developrent." In Southworth, Herman M. and Johnston, Bruce F. , ed. Aggicultm'al Development and Economic Growth, Cornell University Press, Wea, New York. M12 . 1969. "The Subsistence Farmer in Traditional Economies." In Wharton, Clifton R. , Jr. , ed. Subsistence Agriculture and Econggic Development, Aldine PGBIishing Co. , Chicago. M. 13 . 1969. "Production Economics and the Pbdernization of Traditional Agriculture," Australian Journal of Aggicultural Economics, 13:25-35. M. 14 . 1972. "Madels of Economic Growth and Land Augmenting TeEhnological Change in Food Grain Products." Paper presented at International Economic Association Conference on the Place of Agriculture in the Development of Underdeveloped C'omtries. Bad Gadesberg, Germany, August 26-Septe1ber 4. M. 15 Michigan State University Agricultural Sector Simulation Team. 1973. "System Simulation of Agricultural Development: Some Nigerian Policy Comparisons," American Jomrnal of Agricultural Economics, 55:404—19. M16 M18 M19 M21 M. 22 M. 23 M. 24 M. 25 M26 M27 340 Mighell, Ronald L. and Black, John D. 1951. Interre ional iculture withS ecial erenc enceto ' F_arming in e States and New Eng Univer- sity Press, Cambridge, lMass. Mikesell, Raymond F. 1968. The Economics of Foreign Aid. Aldine Publishing Co. , Chicago. Miller, S. I. and Halter, A. N. 1973. "Systems-Similation in a Practical Policy-Making Setting: The Venezuelan Cattle Industry," American Journal of Agricultural Economics, 55: 420-32. Mitroff, Ian I. and Turoff,Mm'ry.1973. 'Technological Fore— casting and Assessment: Science and/or Mythology?" Technological Foreca_st1nLand Social m, 5: 113- 34. Nbseman, Albert H. 1965. "Research Needed for Technological Knowledge in Agricultural Development." In Iowa State University Center for Agricultural and Economic Development. Economic Development of Agriculture, Iowa State University Press, Ames, Iowa. . 1970. BuildingAgficultural Research Sgtems in the Developipg Nations. Agricultural Development Counch Inc. , New Yor Nbsher, A. T. 1965. "Research Needed on the Development Process for Agriculture." In Iowa State University Center for Agricultural and Economic Development. Economic Developrent of AgE‘cultur , Iowa State University Press, Ames, Iowa. 1966. Getting Aggiculture Maving. Praeger, New York. . 1969. Creating a Progressive Rural Structure. Agricultural Development Council, Inc. , New York. . 1972. "Projects of Integrated Rural Development. " ND D/C oReprint. The Agricultural Development Council, Inc. , I’bulik, T. K. and Lokhande, M. R. 1969. "Value Orientation of Northern Indian Farmers and Its Relation to Adoption of Farm Practice," Rural Sociology, 34:374—81. Midahar, M. S. 1973. "Dynamic Mndels of Agricultural Develop- ment with Demand Linkages." Occasional Paper No.59, ktEfMt of Agricultural Economics, Cornell University, I ca, N.Y. Mynt,ma.1969.'rne Economics of the Developing Countries. Praeger Publishers, New York. M.29 N.l N.2 N.3 N.4 N.5 N.6 N.7 N.8 N.9 N. 10 0.1 0.2 341 Myrei, Delbert T. 1971. "The Puebla Project: A Developmental Strategy for Low Income Farmers." Paper presented in A/D/C Seminar on Small-Farmer Development Strategies. Ohio State University, Columbus, Septenber 13-15. Nakajima, Chihiro.1969. "Subsistence and Commercial Family Farms: Some Theoretical l’bdels of Subjective Emmilibriim." In Wharton, Clifton R. , Jr., ed. Subsistence Agriculture and Economic Development, Aldine Publishing Co. , Chicago. National Council of Applied Economic Research. 1962. @g Term Projections of Demand for and Supply of Selected Agricultural Comodities,1960-61 to 1975- 76. New Delhi, India. Naylor, Thomas H. 1970. "Policy Simulation Experiments with Macroeconomic Mndels: The State of the Art," American Journal of Agicultural Economics, 52:263-70. . 1971. Cmter Simulation Experiments with Models of Economic System. Wiley & Sons, Inc., New York. Nelson, Richard R. 1964. "Aggregate Production Functions and Me dium Range Growth Proj ections,‘ 'A_m_n_erican Economic Review, 54: 575— 606. Nerlove, Marc. 1958. The Dynamics of Supply: Estimate of Farmers' Rgponse to Price. Jfis Hopkins Press Balt timore. . 1961. "Time-Series Analysis of the Supply of Agri- afltural Products} In Heady, Earl O., et al., ed. Agricultural Supply Functions--Estimatipg Techngues and Intgpgetations, Iowa State Univer31ty Press, Ames, Iowa. and Backman, Kenneth L. 1960. "The Analysis of Changes in Agricultural Supply. Problems and Approaches," J_c_>____1mnal of Farm Economics, 42: 331—54. Nicholls, William H. 1964. "The Place of Agriculture in Economic Development, " In Eicher, Carl and Witt, Lawrence, ed. Agriculture in Economic Development, McGraw—Hill Book Co., New York. Nurke, Ragner. 1957. Problems of Capital Formulation of Under develogd Countries. Oxford University Press, Nev YOrk. OECD. 1972. Development Cooperation: 1972 Review. OECD, Paris. Ojala, E. M. 1967. "The Programming of Agricultural Development." In Southworth, Herman M. and Johnston, Bruce R, e . Agricultural Development and Economic Growth, Cornell University Press, Ithaca, N.Y. 0.3 0.4 0.5 R1 P.2 P.3 F.4 P.5 P.6 Q.1 R.1 R.2 342 Olayide, S. O. 1972. "Improving the Data and Information Base." Paper presented A/D/ C Conference on the General Systems Analysis Approach in Agricultural Sector Analysis and Planning Airlie House, Virginia, May 1- 3. Oshima, Harry T. 1971. "Labor Absorption in Fast and Southeast Asia. A Summary with Interpretation of Post War Experience," Malagap Economic Review, 16: 55-77. . 1971. 'Tabor—Force'Enqnlosion‘ and the Labor—Intensive Sector in Asian Growth," Economic Develomt and Cultural ME, 19:161-83. Parker, F. W. 1965. "Priorities in Agricultural Development and Investment. " In Iowa State University Center for Agricultural and Economic Development, Economic Develomt of Ag' culture, Iowa State University, Ames, Iowa. Parvin, D. W., Jr. 1973. "Estimation of Irrigation Response from Time-Series Data on Nonirrigated Crops," American Journal 9§_Ag§cu1tural Economics, 55:73—6. Pen, J. B., et al. 1974. "Structural Input—Output Madeling of the Food and Fiber System Under Conditions of Fuel Scarcity." Paper presented at the AAEA Annual Meeting, Texas A & M University, College Station, Texas, August 18—21. Perry. A. , et a1. 1967. "The Adoption Process: S Curve or J Curve?" Rural Sociology, 32:220—22. Peterson, Willis L. 1971. "The Returns to Investment in '— cultural Research in the United States." In Fishel, Walter L. , ed. Resource Allocation in Aggicultural Research, University of Minneosta Press, Mimeapo is, M1rm. Posado, Alvaro. 1974. "A Simulation Analysis of Policies for the Colombia Beef Cattle Industry." Unpublished Ph.D. disseration, Michigan State University, East Lansing. Quance, Leroy. 1967. "Farm Capital: Use, MVP's and Capital Gains or Losses, United States 1917-64." Ph.D. disseration, Michigan State University, East Lansing. Ranis, Tustav and F11, John C. H. 1964. "A Theory of Economic Development." In Eicher, Carl and Witt, Lawrence, ed., Agriculture in Economic Develomt for Latin America, St. Martin 3 Press, New Yor . Ray, Daryll E. and Heady, Earl O. 1972. "Government Farm Programs and Commodity Interaction: A Simulation Analysis," American Journal of Agricultural Economics, 54:578—90. J__— i 343 R. 3 . 1974. "Simulated Effects of Alternative Policy and Economic Enviroments on U.S. Agriculture." CARD Report 46T, Center for Agricultural and Rural Development, Iowa State University, Ames, Iowa. RA Reynolds, Lloyd G. 1969. "Economic Development with Surplus Labor: Sane Camplications," Oxford Economnic Papers, 21:89-103. R.5 Rogers, Everett M. and Shoemaker, F. Floyd. 1971. Communication of Innovation A Cross—Cultural Approach. The Free Press, New York. R.6 Rosenstein-Radan, Paul N. 1961. "Notes on the Theory of the 'Big Push'," in Ellis, Howard S. and Wallick, Hencry C. , ed. Economic Development for Latin America, St. Martin's Press, New York. R.7 Rossmiller, George E., et al. 1972. "Korean Agricultural Sector Arnalysis and Reccmmended Development STrategies, 1971-1985" Korean Agricultural Sector Study Team. Michigan State University, East Lansing. R.8 Ruttan, Vernon W. 1960. "Research on the Economics of Tech- nological Change in American Agriculture," Journal of Farm Econanics, 42:735-754. R.9 . 1965. The Eoornomic Demand for Irrigated Areas. Johns Hopkins Press, Bafiimore. R. 10 . 1968. "Strategy for Increasing Rice Production in Southeast Asia." In McPherson, W. W., ed. Economic Develo - ment of Trgnical Agriculture, University of Florida Press, Gainesville, Fla. R. ll and Hayami, Y. 1972. "Strategies for Agricultural Development," Food Research Institute Studies, 11:111-28. Stanford University, Stafford, Calif. K12 and Stout, T. T. 1960. "Regional Differences in Factor e in American Agriculture, 1925-57," Journal of Farm Economics, 42:52-68. S.l Schaller, W. Neil. 1969. "Discussion: The Supply Function in Agriculture Revisited," American Journal offigricultural Economics, 51:367—9. S.2 Schultz, Theodore W. 1964. Transforming Traditional Agriculture. Yale University Press, m. 8.3 3.4 3.5 3.6 3.7 5.8 3.9 3.10 8.11 S. 12 8.13 8.14 8.15 344 1971. "Knowledge, Agriculture and Welfare." Papere presented at Pugwash let Conference Meeting at Sincia, Romania, August 25-31. 1971. "The Allocation of Research to Research." In Fishel, Walter L. , ed. Resource Allocation in AE'cul— tural Research, University 0 Minnesota Press, Minneapo is, Minn. Seers, Dodley. 1970. "The Meaning of Development." A/D/C Repr1nt. Agricultural Development Council, Inc. , N.Y. Seol, In Joon. 1972. "Supply Response to Expected Price of 1911?; 13 Korea," Journal of Agricultural Economics (Korean) , : -7 . Shapiro, Harold T. 1973. "Is Verification Possible? The Evaluation of large Econometric Models," American Journal 9f_Agricultural Economics, 55:250-8. Shechter, Nbrdechai and Heady, E. O. 1970. "Response Surface Analysis and Simulation Pbdel in Policy Choices," American Journal of Agricultural Ecornomnics, 52:41-50. Sidhu, Surj it S. 1974. "Economics of Technical Change in Wheat Production in the Indian Punjab," American Journal of Agr_i cultural Economics, 56: 217—26. Solo, Robert M. 1956. "A Contribution to the Theory of Economic Growth," Quarterly Journal of Economics, 20:65—94. 1957. ”Technical Change and the Aggregate Production anction," Reviev of Economics and Statistics, 30:312-20. . 1962. "Technical Progress, Capital Formulation and Economic Growth," American Economnic Review, 52:76-86. Smith, Victor E. 1955. "Perfect vs. Discontirnuous Input Market: A Linear Programmirng Analysis," Journal of Farm Econom1cs, 37:538-45. Srivastava, Ilna K. and Heady, Earl C. 1973. Technological Change and Relative Factor Share in Indian Agriculture: An Empirical Analysis," American Journal of Agricultural Economics, 55:509—14. Staniforth, Sydney D. and Diesslin, Howard G. 1961. "Summary and Conclusions," In Heady, Earl O. , et al. , ed. Aggicul- tural Supply Functions—-Estimating Techniques and Interpre- tation, Iowa State University Press, Ames, Iowa. 8.16 8.17 8.18 T.1 T.2 T.3 T.4 T.5 T.6 T.7 U.1 V.1 V.2 345 Sternfiolliobert End 1:25. la's'Malayan Rubber Production, Inventory dings t E ticity of Export Sup 1 ," Southern Ecornamic Journal, 312314-23. p y Stevens, Robert D. 1971. "Camilla Rural Development Program—- Results from East Pakistan for International Testing." Paper presented in A/D/C Seminar on Small-Farmer Development Strategies, Ohio State University, Columbus, SepbeIiber 13-15. Sung, Bai'Y. , et al. 1973. "Projection of the Demand for Fert11izer—-Ti.me Series Data Analysis," Journal of Aggicul- tural Economics (Korean), 15:21-32. Talpaz, Hovov. 1974. ‘Milti—Frequency Cobweb Nbdel: Decompos— ition of the Hog Cycle," American Journal of Agricultural Ecommics, 56:38-49. Thorbecke, Erik. 1973. "Sector Analysis and Models of Agriculture in Developing Countries," Food Research Institute Studies, 12:73-89, Stanford Universifi, Stanford, Calif. Todaro, Michael P. 1971. Develoment Planning: IVbdels and Methods. Oxford Unviers1ty Press, London. Tom, A. El. 1964. "Projections of Agricultural Production- Seasonal Crops," FAO Agricultural Planning Course 1963, Agricultural Planning Sutdies No. 4, Rome. Tweeten, Luther G. and Quance, C. Leroy. 1969. "Positivistic Measures of Aggregate Supply Elasticities: Some New Approaches," American Journal of Agricultural Economics, 51:342-512. Tyner, Fred M. and Tweeten, Luther G. 1965. "A Methodology for Estimating Production Parameters,” Journal of Farm Economics, 47:1462—7. . 1968. "Simulation as a Method of Appraising Farm Programs," American Journal of Agricultural Economnics, UNCI‘AD, Commodity Division. 1970. "Econometric Analysis of the World Rubber Market." Draft, Geneva. Vaurs, Rene, Condos, Apostolos, and Goreux, Louis. 1971. "A Programming Model of the Ivory Coast." Development Research Center, IBRD, Washington, D.C. Vernon, Raymond. 1966. "Comprehensive lbdel-Building in the. ... Planning Process: The Case of the Less-Developed Economies, Economic Journal, 76: 57-69. V.3 W.l W.2 W.3 WA W. 5 Y.l Y.2 2.1 346 Vincent, Warren H. , ed. 1962. Economics and Mana ement in Agg’culture. Prentice—Hall, Inc., Englenood Cliffs, N.J. Wah, F. Chang Kwang. 1962. "Economic Aspects of Supply of Malayan Rubber." Rural Social Science MmographuMalayan Series No. 3. Council on Economnic and Cultural Affairs, Inc. , New York. Wharton, Clifton R. , Jr. 1964. "Malayan Rubber Supply Condi— tions, " A/D/C Reprint No. 3. Agricultural Development Council, Inc. , New York. . 1969. 'The Green Revolution: Cornucopia or Pandora's Box?" Foreigr_n Affairs, 47:465-76. Wipe, Larry J. and Bawden, D. Lee. 1969. "Reliability of Supply Equations Derived From Production Function,” American Journal ciAgriicultural Economics, 51:170—8. Wood, Garland P. 1972. "Public Institutions: Their Response Characteristics to Agricultural Development Programs in the LDC' s. " Paper presented at the Purdue Workshop on Small-Farm Agriculture, November, Michigan State University, East Lansing. Yamado, Subura. "Changes in Output and in Conventional and Non— conventional Inputs in Japanese Agriculture Since 1880," Food Research Institute, Studies, 72371—413. Yotopoulos, Pau A. 1968. "On Efficiency of Resource Utilizaton in Subsistence Agriculture," Food Research Institute Studies, 8:125-35. Stanford University, Stanford, Calif. Zarembka, Paul. 1972. Toward a Theory of Economic Development. Holden-Day, Ind., San Francisco, Calif.