AN ABSTRACT OF THE THESIS OF for the DOCTOR OF PHILOSOPHY PAL YONG MOON (Name) (Degree) in AGRICULTURAL ECONOMICS presented on (Major) /972 a (Date) Title: AN ANALYSIS OF FOODGRAIN MARKET IN KOREA Abstract approved Redacted for privacy / John A Edwards In order to effectively use foodgrain prices as tools of government policy, it is necessary to know the possible effects of price incentives on different economic variables. The present study attempts to measure the effects of foodgrain prices, upon various economic variables, including production, consumption, both at the farm and urban levels, and farm sales of two major foodgrains, and to explore possible measures to improve the current foodgrain price policy. The results of the empirical study of farm producers' response indicate that Korean farmers respond significantly in their grain production to price changes and that they are more responsive in barley than in rice production in terms of both input use and planted acreage. This suggests that policy designed to raise barley price to farm producers would contribute relatively more to increasing overall foodgrain production. In studying the consumption behavior of farm and urban con- sumers, and sales d.ecisions of farmers, a simultaneous equation model was used. The system comprises eight equations: six behav- ioral and twomarket identity equations. In specifying the model, special attention was given to a pecuflar feature emerging from the dual role of farmers in semi-monetized agriculture, that is, as consumers on the one hand and as sellers of products on the other. Two types of analyses were carried out on the basis of the estimated behavioral parameters of the model. First, an analysis was made of the partial response to price changes by treating each behavioral equation as an independent single equation under the usual cetéris paribus assumption. Secondly, the total behavioral responses were analyzed by taking account of simultaneous changes in all endogenous variables in the system. The partial response analysis indicates that both farm and urban consumers have a upotentialu tendency to respond negatively to price changes in their consunption of rice and barley and also a "potential" tendency to substitute one grain for another in the face of changing relative prices. It also shows a positive response in the foodgrain marketings of farm producers. But the total response analysis showsthat the responses measured in the partial analysis are substantially offset by the interdependence of the prices of rice and barley on the open market, resulting in positive changes in the quantity demanded or no substitution at all. The empirical results also provide counter-evidence concerning the validity in the Korean economy of the so-called "target cash requirements hypothesis, " advanced by a number of economists. One important policy implication that can be drawn from the study is that if the government's objective is to reduce foreign exchange spending on rice imports by restructuring foodgrain consumption (in addition to increased domestic production), it can be done through the use of price incentives by inducing the consumers to reveal their "potential" responses on the market. This is equivalent to forcing the ceteris paribus assumption made in the partial response analysis to operate in the real world through an appropriate governmental operation. An Analysis of Foodgrain Market in Korea by Pal Yong Moon A THESIS submitted to Oregon State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy June 1973 APPROVED: Redacted for privacy Professor of Agricultural Economics in charge of major Redacted for privacy Head of Department of Agricultural Economics Redacted for privacy Dean of Graduate School Date thesis is presented Typed by Clover Redfern for /Q'72 Pal Yong Moon ACKNOWLEDGMENT The author owes a special debt of gratitude to Dr. John A. Edwards, major professor, for his ideas and valuable suggestions at all stages of this study and guidance throughout the author's graduate program at Oregon State University. He would like to thank Drs. Richard S. Johnston, Bruce R..Rettig, Joe B. Stevens and David R. Thomas, members of his graduate cornmittee, and Dr. William G. Brown for their helpful comments on the dis sertation. The author is under a deep obligation to the Agricultural Development Council, Inc., New York and its staff for providing financial assistance and encouragement throughout the entire period of his graduate study. Special thanks are also due to Professor Herman M. Southworth, Associate with the Agricultural Development Council, Inc., for his valuable comments at the initial stage of the study and Mr. Hae Sung Yoo and other members of theEconomic Research and Statistics Bureau, Ministry of Agriculture and Forestry, Korea,for collecting and arranging the data which made this research possible. Indebtedness to my wife, Haeduck, for her encouragement throughout this study, should be particularly mentioned. TABLE OF CONTENTS INTRODUCTION 1 Problem Objective of the Study PRODUCER RESPONSE TO FOODGRAIN PRICES Some Conceptual Definitions Actual vs. Intended Response Normative vs. Positive Approach Methodological Basis Estimation Procedures Yield Response (E) Derivation of Yield Elasticity Acreage Response (EAP) Empirical Results Production Function Input Demand Function Deviation of Price Elasticity of Yield Response to Planted Acreage Aggregate Supply Response Conclusion MARKET ANALYSIS Model Construction Farm Demand Farm Sales Urban Demand Market Identities Complete Model Identification of the Model Order Condition Rank Condition Empirical Results Data 1 4 6 7 8 8 10 15 15 19 20 22 22 24 33 34 37 37 40 44 55 59 62 64 67 67 68 69 69 Estimation Procedure Possible Source of Errors Predictive Tests Interpretation of the Estimated Results Analysis of Partial Effects Farm Level Demand Farm Sales 74 74 75 Urban Demand 96 80 89 89 93 Page Analysis of Total Response Summary of Findings 103 115 POLICY IMPLICATIONS 119 BIBLIOGRAPHY 127 APPENDICES 130 130 134 137 139 Appendix A: Estimation of Foodgrain Supply and Demand Appendix B: Estimated Seasonal Response Model Estimated Nonlinearized Model Appendix C: Basic Data LIST OF FIGURES Figure 3-1. Effect of price changes on the consumption of the seller-consumer. 49 3-2. Monthly per capitafarm: demand for rice, 1963-7l. 81 3-3. Monthly per capita farm demand for barley, 1963-71. 82 3-4. Monthly per capita farm sales of rice, 1963-71. 83 3-5. Monthly per capita farm sales of barley, 1963-71. 84 3-6. Monthly per capita urban demand for rice, 1963-7 1. 85 3-7. Monthly per capita urban demand for barley, 1963-71. 86 3-8. Monthly wholesale price of rice per kilo (polished), 1963-71. 87 3-9. Monthly wholesale price of barley per kilo (polished), 1963-71. 88 LIST OF TABLES Page Table 1-i. Supply and demand for major foodgrains for 1967-71 and projection for 1972-76. 2 2-1. Per farm planted acreage for rice, common barley and naked barley, by year, 1963-70. 13 Total availability and actual consumption of government supplied fertilizer by year, 1963-1970. 26 2-3. Total supply and consumption of agricultural chemicals, by year, 1963-1970. 27 2-4. Acreage of paddy land by irrigation status, by year, 1961-1970. 36 3-1. The estimated structural coefficients. 76 2-2. 3-2. The coefficients of responses to price changes; partial versus total. 114 Appendix A- 1. Estimation of rice demand and supply. 13 1 A-2. Estimation of barley demand and supply. 132 A-3. Estimation of wheat demand and supply. 133 C-i. Relative prices of rice and input factors used for 1/10 Ha, by size of farm, 1963-70. C-2. Relative prices of common barley and input factors used per 1/10 Ha, by size of farm, 1963-70. C-3. Relative prices of naked barley and input factors used per 1/10 Ha, by size of farm, 1963 -70. C-4. Data used for estimating simultaneous equation model, 1963-71. 142 144 146 149 AN ANALYSIS OF FOODGRAIN MARKET IN KOREA INTRODUCTION Problem The Korean Government has been making various efforts for a past decade to increase foodgrain production. And yet the gap between supply and demand has been widening in recent years. This gap has been filled by the importation of foreign grain; partly by commercial imports and partly by imports under the U.S. FL 480 program. As shown in Table 1-1, the foodgrain shortage is estimated as more than 2,000,000 metric tons for the 1972 consumption period. The shortage of rice alone is estimated to exceed 600, 000 metric tons. If the current consumption and production trends continue, the overall foodgrain deficit, it is estimated, will reach nearly 3,000,000 metric tons by 1976. As is projected in Table 1-1, there will be a deficit of over 800, 000 metric tons of rice and a deficit of over 2, 000, 000 metric tons of wheat in 1976. This increasing imbalance between domestic production and demand is ascribable to numerous factors which govern grain pro- duction and consumption. Historically, government policy has been directed toward maintaining foodgrain prices at low levels in favor of the growing industrial sector. Such a negative price policy has Table i-i. Supply and demand for major foodgrains for 1967-71 and projection for 1972-76. Unit: 1000 M/T (Polished). consumptionb Wheat Barley Ricea Year Domestic Production Barley Wheat Rice 1967 3919 2002 310 3780 1999 1032 113 0 922 1035 1968 3603 1940 324 4032 1984 1460 216 105 1127 1343 1969 3195 2037 356 3899 2042 1529 760 64 1205 2029 1970 4090 2045 360 4320 2105 1562 540 0 1254 1794 1971 2929 1917 352 4474 2166 1684 535 0 1333 1868 1972 3998 2375 361 4736 2226 1825 629 0 1464 2093 1973 4190 2565 380 4786 2287 1979 596 0 1609 2205 1974 4254 2565 380 4950 2349 2149 666 0 1769 2435 1975 4378 2660 389 5119 2413 2335 741 0 1946 2687 1976 4475 2755 398 5296 2478 2540 824 0 2142 2966 Rice Imports Rice Barley Wheat Total a.Rice-year begins on November 1 in the previous year and ends on October 31 in the current year bconsumption includes staple food, non-staple food, seeds, and feedsuse. Source: Table A-i, 2 and 3 in Appendix. N) 3 naturally created disincentives for increased production of grain in the farm area on the one hand and stimulated the consumption of rice in the urban area on the other. Under the present policy, the foodgrain gap will continue to widen, possibly forcing the country to spend nearly 300 million dollars of foreign exchange every year in order to meet the foodgrain shortage alone in the near future. Since a large portion of the foodgrain short- age has been met in the past by local currency purchases under the U. S. PL 480 the food gap itself did not impose a serious problem on Korea1s foreign exchange position. But in anticipation of a change in U.S. policy to cash, or credit, salesinUS. currency, the foodgrain situation will be directly related to the countryT s balance of payments position. It is an accepted notion that foreign exchange plays an essential role in industrialization efforts of the less developed countries by making it possible to import the capital items not produced domestically as well as technical know-how from the developed countries. Considering the foreign exchange earning capacity and the magnitude of the foreign debt servicing obligation of the country, the spending of 300 million dollars of foreign exchange on foodgrain imports will bring about a continuous deterioration in an already weak balance of payments position. A shortage of foreign exchange will eventually impose a serious restraint on industrial growth. ru Thus, a policy designed to achieve the goal of industrialization may have a negative effect upon industrialization itself. The solution to this paradox is a complex one. A positive price policy that favors a general rise in the prices of all foodgrains relative to nonagricultural commodities may foster some increase in aggregate grain production. it is, however, likely to have a conflicting impact upon other target variables in the overall development plan. In this context food grain price policy should be integrated into overall development policy and it should be used as a tool of govern- ment to induce the relevant economic variables to move in such a direction that Korea reduces to a minimum the foreign exchange spending on foodgrain imports, with minimum discouraging effect on other economic goals. In order to effectively use grain prices as tools of government policy, it is necessary to know the possible effects of price incentives on different economic variables.. The present study is undertaken to provide an empirical foundation that will contribute to improving the current foodgrain price policy. Objective of the Study The primary objectives of this study are to measure the effects of foodgrain prices on farm production decisions, consumption behavior, both at farm and urban level, and farm sales of foodgrains 5 and to explore possible measures to alleviate the overall grain shortage in Korea. Specifically, the study intends: 1. to analyze the effect of price changes on farm producers' decision-making in the production of major foodgrains, rice, common barley and naked barley; 2. to analyze consumer response to price changes with particular emphasis on the substitution effect between different foodgrains for both farm and urban populations; 3. to examine the farm producers' response to price changes in market sales of major foodgrains, and 4. to draw appropriate policy implications on the basis of the empirical results. 6 PRODUCER RESPONSE TO FOODGRAIN PRICES From the standpoint of public policy, it is important to know the degree of producers7 response to price changes; the degree of response measured in terms of price elasticities. The price elasticity of supply measures the willingness as well as the ability of farm producers to adjust to changing price condition, one of the most important aspects of a dynamic economy. A change in the relative level of grain prices induces farm producers to change the level of input use as well as its allocation among different crops, resulting in changes in aggregate production. The effect of grain prices can b& analyzed from two different aspects; the effect on the aggregate output through increased or decreased use of input resources and the effect on the short-run decisions of the producers with respect to market sales from already produced output. The present chapter deals with the former effect, that is, the effect on the aggregate production of major foodgrains and the latter effect upon market sales will be discussed in the next chapter. It is widely questioned whether or not the farmers in the traditional-oriented agriculture, where a large portion of their own products goes to family consumption, respond to changes in market pricesin their production decision making [29,32]. It is generally said that the extent to which farmers are sensitive to external 7 economic condition varies depending on their stage of agricultural development and by different regions. For example, farmers in relatively more commercialized agriculture or those living close to urban areas respond more sensitively to changing market condition than those in subsistence farming or those living in remote areas. How- ever, where the rural structure is such that the commercial and subsistence farmers co-exist side by side in terms of both residence and cultivated land, there is every reason to believe that the behavior of subsistence farmers cannot but be influenced by that of commercial farmers. This simple fact has been seldom recognized by the development economists in the past. The main hypothesis to be tested here is that Korean farmers do respond to price changesin their production plan for major foodgrains. The secondary hypothesis is that Korean farmers are more responsive to price condition in barley production than in rice production. The practical implication in terms of price policy of this hypothesis is that the potential for increasing barley output is greater than that for rice. Some Conceptual Definitions We need to distinguish between two conceptually different supply responses; the Hactualhi or hlrealizedU and the HintendedTl or planned" responses, and also between two different approaches; "normative" and "positive." Actual vs. Intended Response The measurement of "actual" production response involves the estimation of the relationship between the variation in price and realized output as observed in the past, irrespectiveof farm producers' intention in production activities. Since actual production is subject to various factors which are beyond the control of decision makers (farmers and policy makers), the degree of response thus estimated reflects only a part of the behavioral intention in farm investment. The "intended" production, however, emphasizes the farm investors' efforts and plan as decision makers. This distinction is important and the knowledge of the "intended" or "planned" response of farmers appears to be more relevant in a country where the government policy is directed towards increasing total grain production by inducing farmers to increase their efforts. In this context, the present study centers on the measurement of supply elasticities in the latter sense with a view to gaining some information and insight as a basis for foodgrain price policy. Normative vs. Positive Approach Normative analysis refers to what ought to exist under certain conditions, while positive analysismeans prediction of quantitative relationships among variables as they do exist at a point in time or have existed over a period of time [12]. Normative analysis intends to identify the optimum relation between the quantity of a product and its price under the assumption of optimizing behavior of farm investors. The normative supply function i& thus an estimate of the optimum supply reaction to product price changes in terms of a given norm, 1. e., profit maximization or cost minimization. On the other hand, positive analysis attempts to estimate the production response to price changes from observations drawn from the "actual operating world, " Where subsistence farming is still the prevailing type and the decision-making of the average farm producers deviates from the concept of profit maximization in their resource allocation, the equating of marginal values to prices could only be conczeived of as a desirable goal but not as something accomplished in production planning. The analysis in this chapter thus follows one of the positive approaches based on the actual observations in the past, ignoring the derivation of the first order condition. In short, the present study attempts, having in mind the above two sets of conceptuai differences, to measure the "intended" response to price changes of farm decisionmakers by means of positive approach. 10 Methodological Basis In his paper published in the Journal of Farm Economics, May, 1959, Zvi Grilliches [ii], presented the result of his analysis on input demand and the estimates of derived supply elasticities for the aggregate U.S. farm products. As he did not state the exact procedures for deriving aggregate supply elasticity there is no way of making appropriate review of his analysis here. In any case, the basic idea of the method used in this study owes its origin to Professor Zvi Grilliches and it is well represented by the following statement in his article!! [11, pp. 309-322]: if inputs respond to price changes, so must also farm output. This is the basic idea underlying the assertion that a study of input behavior can also provide insight into the supply response of farm products. . . The elasticity of aggregate output with respect to its price is the sum of the elasticities of planted acreage and that of yield per unit acreage, both with respect to the product price. That is, EQ = + EAP Heady and L. G. Tweeten [ 3], and L. G. Tweeten and C. L. Quance [343 used essentially the same method in estimating the aggregate supply elasticity of the U.S. farm products. This can be shown as follows: Q=YA Where: 11 Where: = Price elasticity of aggregate output. = Price elasticity of yield. Price elasticity of planted acreage. EAP If each component of total elasticity is to be estimated by fitting stochastic models to the observed data, the above identity holds under the following assumptions: 1. A decision to expand or contract the planted acreage of a certain crop does not per se affect the efforts to achieve the increased yield of that crop, or conversely the efforts to increase yield does not affect the plan to plant a certain area of land. In short, planted acreage (A) and yield (Y) not interdependent in decision making. Aggregate output Y = Yield per unit acreage A Planted acreage Differentiating both sides with respect to product prices we obtain Q dQaQdY+aQdA dYA+dAY dP aA dP dP dP Multiplying through by P/Q, we get dP a dPQdPYQ dPAQ Which is EQ = + EAP dPY dPA are 12 2. Environmental condition is such that expansion or contraction of land scale does not cause a change in yield, i. e. unitary elasticity of production with respect to land. The first assumption relates to the motivation of farm decision makers and the second to the technical relationship between land scale and yield. If they are interdependent, that is, if Y = F(P,A) A = G(P, Y), then and dY/dA, dA/dY / 0 and therefore the mathe- matical derivation of the above identity cannot hold. In econometric terminology the random error terms in the stochastic models for estimating each elasticity will be correlated, resulting in biased estimates. As to the first assumption, it is very doubtful that Korean farmers adjust intentionally in foodgrain production the amount of input application simply because he plans to alter the planted acreage, or vice versa. This could be true if farmers had a flexible choice of planning planted acreage with a fixed amount of investment capital. Considering, however, the limited availability of land, relatively abundant labor supply, high proportion of non-purchased inputs, there is no reason to believe that average Korean farmers substitute or complement planted acreage for the efforts to achieve desired yield. The existence of such substitution or complementary relation, both technically and motivationally is only conceivable for some special- ized crops which require intensive cultivation in terms of inputs. But these crops are outside the interest of the present study. 13 The second assumption is also a realistic one in Korean agri- culture. As shown in Table 2-1, the past trend of planted acreage per farm-operator for major foodgrains does not indicate a variation in sufficient scale as to cause the changing economy of size. Even the study, made with the cross-sectional data, gives the estimate of land productivity coefficient as large as . 8 to . 9 which is near unitary elasticity. Table 2-1. Per farm planted acreage for rice, common barley and naked barley, by year, 1963-1970. Unit: 1/10 Ha. Common Naked Year Rice Barley Barley 1963 1964 1965 1966 1967 1968 1969 1970 6.21 6.08 6.09 6.32 6.52 6.52 6.51 6.69 3.09 3.35 3.42 3.35 3.26 3.20 3.16 3.39 3.01 3.45 3.44 3.46 3.56 4.05 3.72 3.54 Source: Farm Household Economy and Cost of Production Survey, 1963-19 70, Ministry of Agriculture and Forestry. The next procedure is to estimate each response coefficient separately. The elasticity of yield with respect to the product prices will be estimated by an indirect method in order to reflect the fintendedu response of farm producers, whereas the elasticity of planted acreage will be estimated by a direct method. 14 The response of yield to price changes is determined from the response of yield to input application and the elasticities of demand for each input factor with respect to "expected price" for the product. In other words, yield is a function of productivities of inputs and the level of inputs used. The level of input use in turn depend.s upon the "expected price" of the product relative to input prices. This relation can be formulated as: ep !/ = Where: Elasticity of yield with respect to product price. Elasticity of yield with respect to input X. ep Elasticity of demand for input X. with respect to product price. Mathematical proof is given by Zvi Grilliches [ii, p. 319] as follows: Given production function Y = f(X , .. . n 1 Differentiating both sides with respect to product price, we get dY dXn ay dpaX dP 8Xn dP 1 Multiplying through by PlY, dY P ay X1 dX1 we get X dX dPYaXYdPX '&XYdPX n n + 1 that is, = ey eXP 15 Hence, the procedure for estimating yield elasticity requires the combination of two separate estimates: production function and input demand functions from which to derive the relevant response coeffi- cients. Estimation Procedures Yield Response (E) The Cobb-Douglas type production function is fitted to the cross-sectional input-output data of a selected normal crop year; normal in the sense that weather conditions and other uncontrollable fac- tors were not unusually good or bad so as to cause a bumper crop or crop failure. Production function thus estimated will reflect a relatively normal input-output relationship compared to one estimated from the time-series data of different years. To put it differently, the output of that year corresponds more closely to farmers' intention or plan in regard to input use than other years. The underlying as surnption is that region-to - region and individual- to - individual varia- tion in uncontrollable factors are smaller than year-to-year variation. TheCobb-Douglas production function is used, as it is generally considered appropriate to agricultural production3 plus the fact that its simplicity makes it easy to use. Production function for yield, say rice, is expressed as: 16 = L1 Fc2 FH3 Where: = Rice yield per unit area. L = Labor input in man-hours. = Commercial fertilizer use in real value (deflated by the Index of prices paid by farmers). FH = Family supplied manure in (imputed) real value (deflated by PPFI). = Irrigation expenses (deflated by PPFI). 10 Other operating expenses including seed, pesticides, repairs and depreciation on farm building and implements and stock labor (deflated by PPFI). U 0. . O5 Random error term reflecting unidentified factors. Productivity parameters to be estimated. Each of the estimated coefficients, O, gives the degree of responsiveness of yield to each individual input factor, namely the percentage change in yield corresponding to one percentage change in the amount of each input factor, assuming all other factors remain unchanged. The random error term, U, reflects the effect of uncontrollable factors such as difference in weather condition, soil conditions, managerial ability between different regions and/or individuals. In short, this term represents the incomplete 17 specification of the model to be fitted to the observed data. As the next step, input demand functions are estimated by relating the amount of input employed to the product price which farmers expect to receive relative to the input prices paid by farmers during the current production period. As the expected price for the current product cannot be observed, the moving average price of the harvest season in the past three years is used as a proxy. Input demand functions are also of the form of power function. (1) Demand for labor: L = a0() U1 (2) Demand for commercial fertilizer: Fc b0( Rt1)1 T2 ±"Conventionally, the price of the immediately previous harvest season has been used in estimating supply function under the assump- tion that farmers anticipate to receive the same price for the current output. This assumption is valid only if the previous harvest season price had been a fairly stable one with normal crops. But where a fluctuation in aggregate production causes corresponding fluctuation in price, it is hard to believe that farmers will expect to receive that abnormally high or low price for their products. The choice of moving average price in the preceding three years is based on the judgment that farmers are trend-conscious. The use of the three year average is rather arbitrary, and each year's price is equally weighted since appropriate weighting scheme is not found. 18 (3) Demand for family supplied FH manure:' zc(Rt')l T2U3 (4) Demand for irrigation: d d1 Rt IRdo(pPFI) T (5) Demand for other inputs: I0e0( Rt 1 e1 ) e T 2U5 Where: Rt...1 = Index of the moving ave-rage rice price of the harvest and post-harvest season--November, December and January--in the past three years. WF = Index of therural wages in the current period. = Index of the fertilizer prices in the current period. PPFI Index of the prices paid by farmers. P0 = Index of the price of agricultural supplies in PPFI excluding fertilizer. T = Time trend. 'Although family supplied manure has no market price, it is of analytical interest to see how its .ise is affected by the price of rice relative to the price of commercial fertilizer which is thought to be substitutes. 19 U1. . . U5 a1. . . e2 Random error term. Elasticity coefficients to be estimated. Time trend variable, was introduced in order to reflect T, the effect of technological change owing to the extension service, etc. Derivation of Yield Elasticity Substitution of the estimated input demand functions into the estimated production function will give A(R1 a1e1 Rt-1 Rt-1 ) WF x b192 AA d1B c103 PPFI ) Rt1)15 T2222 p0 Where: A A o)2 A A A A (A)3 ()4 (A)5 Rearranging, we obtain = A(Rtl PPFI a1o1+b1e A A < 1e311e41e5 A A -(b102103) F PPFI WF A A A A b2+c2+d2+e2 P0 -e105 T PPFI which gives the elasticity of "intended' yield with respect to product price, zo A A A A A a a191 + b102 + Ac103 + d104 +Ae15 A eLPe YL + eFP eYF ee = + + eFF eYF IP + e1 = The coefficients, give the estimated . elasticities of yield with respect to input prices. Acreage Response (EAP) The response to price changes of planted acreage of a given crop is hypothesized to depend on the expected price of corresponding out- put and the planted acreage in the previous year. The model adopted is: At = a0( Rt a1 a A2i u Where: At = Planted acreage for the current output. Ai = Planted acreage in the previous year. Rt-1 = Index of the three-year moving average price of the harvest and post-harvest season in the past. a1, a2 Elasticity to be estimated. The estimated elasticity is conceptually comparable with that of yield in that the planted acreage reflects the farmers' intention or plan 21 for the current production. The estimated coefficient, the short-run elasticity of acreage and (l_.) , gives the adjustment coef- ficient of the current acreage to product price changes [5]. Throughout this study, it is assumed that farm producerst response to price of a given crop is not influenced by the price of other crops both in input demand and planted acreage. The factors for these rigidities which condition the farmerst decision-making are many and complicated. Some of the major factors are: 1. Peculiar to subsistence farming, many small cultivators grow grain crops to meet their family needs and therefore the choice of crops is largely independent of the prices of other crops. 2. Technical feasibility is also of a factor which limits the elasticity of substitution between crops. For example. rice is grown in paddy land where irrigation is required. Cornmon barley is grown upland (dry land) while naked barley is planted in paddy land as a second crop after rice. 3. Although there is a certain degree of competition in labor input between rice transplanting and common barley harvest for a short period, it is not as strong as to preclude the planting of either one or two crops, once planned to grow both. Moreover, traditional labor exchange system among farmers alleviates the labor shortage problem during this 22 peak demand season. 4. Possibility of diversified use of a given amount of fertilizer is limited because a specific kind or type is generally used for a specific crop and in case of mixed fertilizer the government sells only those of given formula already prepared. Empirical Results Production Function The Cobb-Douglas production function was fitted to the cros s- sectional data of 1970 for three major foodgrains. Data used are individual farm records from the Farm Household Economy and the Cost of Production Survey, 1963-7 1, conducted by the Ministry of Agriculture and Forestry.6/ The number oI observations is 283 for rice when irrigation expenses 'R is used as one of the independent variable, but was increased to 500 when 'R was summed together with other operating expenses (I) 309 for common barley and 345 for naked barley. For each crop, two production functions were The Farm Household Economy and the Cost of Production Survey is a well-designed sample of 1,200 grain farmers who are assisted by 80 enumerators in keeping daily records of their total economic life. This is the best source of data available for understanding the Korean farmers as economic entities. These data are frequently used for obtaining national estimates of certain economic aspects of Korean agriculture in spite of the fact that there are limitations to these sample data. 23 estimated, one including family supplied manure (FH) and the other without it and also without 'R in rice. The estimated production functions are: (Figures in parenthesis are t-values) (1) Rice yield C(2 11445 08164 0(5 157) 944) R(6 917) 03610 = 45. 96F 08141 713) FH(2 . 288) = . 152) = . 257) (R2 08972 52. l6L (2) 9 ) F C3568 09146 (R2 °(5.531) Common barley yield Y GB = 14 66L 14447 F 19940 F 03793 (2. 884) C(5 H(2 073 097) 11346 07 783) (R2 .16833 CB (3) 18. 60L(3 .18802 .11262 435) F C(4 828)10 .247) (R2 687) Naked barley yield .30888 NB .20696 .09278 = 9.33L(7677)FCIO (R2 = . 296) Overall, the R2's are very low, indicating that a large portion of the variation in yield is associated with the variation in unspecified factors. Nevertheless, inasmuch as all three crops are 24 believed to be commonly affected by the unspecified factors in the same year, resulting estimates of the productivity coefficients are still valid for the purpose of comparison among different crops. Since the sum of the estimated coefficients, A 0., shows asubstantial 1. divergence from unity, it indicates that all crop-yields are subject to decreasing returns to scale in input use. Noticeable is a fact that the response of yield to commerciaL fertilizer is considerably lower in rice and naked barley than in common barley, implying in some sense that the absolute level of commercial fertilizer use is relatively closer to its ceiling- -given the present technological status- -in rice and naked barley. Input Demand Function In formulating demand function, it is hypothesized that the amount of input demanded depends upon the three-year moving average price in the preceding harvest and post-harvest season when the largest portion of grain is sold. In Korea, approximately 40-45 percent of the total marketed rice is sold during the three monthst period November through January, and 45-50 percent of the total marketed barley is sold during the two months' period, July through August. In the statistical demand function, if the quantity of input demanded is used as dependent and price as independent variable, it requires the assumption of infinitely elastic supply of input factors, 25 that is, the price of input is predetermined independently of demand situation. We will examine whether or not the above assumption is realistic by investigating briefly the supply situation of the factors involved. 1. Labor - Although accurate statistics are not available, approximately one-fourth of total labor time available is believed unutilized. Even if there is a seasonal shortage of family labor during the peak demand periods it is well taken care of, as already mentioned, by the traditional labor exchange practices within the village. The situation may be different in the suburban farming region where more employ- ment opportunities are available. But as this type of region cannot represent the Korean agriculture the assumption of infinitely elastic labor supply largely holds at a point in time. 2. Commercial fertilizer - Supply and distribution of chemical fertilizer areunder the strict control of the government in Korea. Sale price is determined by the government set criteria--such as manufacturing costs, importation, transportation and other handling costs plus subsidy. Owing to the government emphasis, a sufficient amount has been always available for the farmers who were willing to purchase at the subsidized price. As shown in Table 2-2, total amount available was in excess of total consumption in all years. Table 2-2. Total availability and actual consumption of government supplied fertilizer by year, 1963-1970. Unit: 1,000 M/T (in plant nutrient). Year Nitrogen (N) Availability Consumption Phosphate (P205) Availability Consumption Potash (K20) Availability Consumption 1963 200 192 118 94 29 21 1964 180 173 151 154 41 37 1965 237 218 163 123 79 52 1966 264 240 176 125 94 59 1967 297 278 188 133 117 76 1968 329 286 169 121 114 71 1969 365 320 188 131 128 84 1970 483 356 293 124 182 83 Source: Agricultural Statistics Yearbook, 1963-1970, Ministry of Agriculture and Forestry. 27 3. Pesticides and insecticides - Although these commodities are not under government control, supply and demand situations are similar to those of commercial fertilizer except that the government subsidize the price for only major items such as calcium seresan and BHC dust that are most widely used by farmers. Table 2-3 indicates that most items are made available for farm users. Table 2-3. Total supply and consumption of agricultural chemicals, by year, 1963-1970. Unit: MIT. Year Supply Consumption 1963 20,013 26,963 12,071 12,688 11,933 13,064 20,501 18,772 23,355 12,729 12,549 1964 1965 1966 1967 1968 1969 1970 31,325 9989 9,983 17,531 25,024 Source: Agricultural Statistical Yearbook, 1963 -1971, Ministry of Agriculture and Forestry. The evidence presented above thus justifies the form of the statistical demand functions to be estimated with price as predeter- mined variables. From these simplified equations with only a single-or two if time variable is introduced--independent variable, we cornpute the demand elasticity of inputs with respect to product price, relative to each input price. Each input demand function is computed from eight years of data, 1963-1970, from the Farmhousehold Economy and the Cost of Production Survey. In order to increase the number of observations each year's data was divided into five classes on the basis of farm size, resulting in 40 observations. Thus we have a combination of time-series and cross-sectional data for each estimation, but only one price for each year. This method of utilizing all individual observations available instead of class average for each year has been shown to improve the efficiency of parameter estimates to a substantial degree by William G. Brown and Farid Nawas [2, p. 1]. In their research paper, "Improving the Estimation and Specification of Outdoor Recreation Demand Functions, they made the following conclu- sion as the resulLof a Monte Carlo study; The proposed method should give gains in efficiency of several hundred percent over traditional procedures, thus allowing sample several times smaller than those . presently used to yield parameter estimates at least as reliable as those presently obtained. As the result of using this method, the efficiency of the esti- mated coefficient of the price elasticities were certainly improved, 2, 7/ though the R s are low. Since one is more interested in obtain. . ing the efficient elasticities in this study, low R2's are not of great 7/ R. J. Freund [7] shows that the use of group means increase 2 but it does not represent an increase in the precision of the B. regression estimates, 29 concern. The estimated results are: (Figures in parenthesis are t-values) (1) Input for rice Labor Rtl L 14925 (R 2 .088) Commercial fertilizer Rtl Fc 65608 (R2 = .564) Family supplied manure 01830 = 305. Z0( 129) Other operating expenses 10 66. 14( )) (R2 .004) (R2 .341) Irrigation 11064 Rt-1 -.24425 T2583) = 731,00( PPFI (.588) (R2 = . 153) 10619 = 244. OOT(2 540) (R2 = . 145) (2) Input for common barley Labor L Btl 17696 (R2 .261) 30 Commercial fertilizer Fc = 98.sl(:t1)9) (R2 = .087) (R2 = .156) (R2 = .802) (R2 = .085) Family supplied manure FH = lz46,o(:t1)5:)5 Other operating expenses 10 Btl 69459 (3) Input for naked barley Labor L = 8O.5O(:):;) Commercial fertilizer Fc Btl (R2 .119) Family supplied manure FH = 1476. 0( :t 1 (R2 = .465) Other operating expenses 10 = 36.4o(:t1):o;) (R2 = .684) 31 A few noticeable facts.in the above results are: 1. The elasticity of demand for labor in naked barley is consid- erably lower than those in rice and common barley, indicat- ing that farmers give higher priority to rice cultivation than to naked barley that is mostly grown after rice in the same paddy land. 2. The elasticity of demand for commercial fertilizer in rice is much higher than that in barleys. This explains that Korean farmers are in general partial to commercial fertilizer. This high elasticity of demand with respect to rice price makes a noticeable contrast with its low productivity coefficient in the estimated production function. 3. The elasticity for family supplied manure is negative in both barleys and that in rice is not significantly different from zero. As one would expect, if the price of commercial fertilizer relative to product price falls, farmers will shift to purchased fertilizer instead of bothering to make compost or to grow green manure. 4. The elasticities for other inputs are high for all three crops. The introduction of time dummy as an independent variable in fertilizer and other input demand functions did not result in an appreciable improvement. One possible explanation may be that, even if the use of fertilizer and other inputs are steadily increasing over 32 time owing to the rural guidance services, increasing subsidies on these inputs by the government have such a strong influence that regression effects are mostly absorbed by the variation in relative prices. But the reverse seems to be the case for irrigation. Once the government supported irrigation facilities are established, it is very likely that all farmers within the reach of capacity of the establishment make use of it regardless of product price changes. When the demand elasticities for hired labor and family labor are computed separately the results are very similar to those for commercial and family manure in that those for family labor assume negative sign in naked barley and rice. And the estimated elasticities are considerably higher than those estimated with both categories summed together. The results of separate estimation are: (1) Rice Hired labor L Familylabor L (2) Common barley Hired labor L= Family labor L 52382 657) 4. 36( -.07708 l2(1 58.07( WF 04991 )437) (R2 = .067) (R2 = .002) 172) (R2 = (R = .005) . 33 (3) Naked barley Hired labor L 3.96(1)l3) (R2 Family labor L 130,20(t1)1 (R2 = .035) .130) Deviation of PriceElasticity of Yield Price elasticities of yield are computed by combining the estimated productivity parameters and demand elasticities of input factors. Family supplied manure for all crops and irrigation for rice are excluded from computation. (1) Rice yield (. 14925)(. 14671) + (.65608)(. 08972) + (.41260)(. 09146) p11850 (2) Common barley yield (. 17696)(. 16833) + (.39477)(. 18802) + (.69459)(. 11262) = . 18223 (3) Naked barley yield EYP = (.04273)(. 30888) + (.34457)(. 09278) + (.60697)(. 20696) = . 17079 The result clearly indicates that the response of yield to changes in relative product price is greater in barley than in rice cultivation. 34 In other words, the average Korean farmer tends to respond more sensitively in his grain production plan for barley when he is faced with price changes. This does not mean that greater emphasis was placed on barley growing in the past. Fact is rather contrary, the highest priority was given all the time to rice and it will continue to be so. The result of the analysis reported above only implies that the additional contribution made by additional use of input factors will be greater for barley and at the same time changes in relative price of product will induce farmers to add more productive factors to barley, thus leading to greater proportional increases in yield. Response of Planted Acreage Adopted functional form is also that of power function. In order to obtain the short-run (more suitably intermediate) and the long-run elasticities, the current planted acreage is regressed on product price and previous year's planted acreage. Data used are the aggregate planted acreage for 196 1-1971, resulting in 11 observations. The results are: (1) Rice acreage A= 2174(P3z5:181 A4936 At 464.5()) (R2 = .724) (R2 = .686) 35 (2) Common barley acreage )°) A691 20( At (R2 = .679) (R2 At .534) (3) Naked barley acreage At At . 86( 168. 9( (i) A7293 (R2 )1 (R2 = .162) = . 824) (4) All barley acreage At 19. 1( Bt1 .13448 A48084 PPFI (2. 102) t1(245Q) (R2 .717) The estimated price elasticity of rice acreage seems somewhat high, considering the limited availability of land. As shown in Table 2-4, the area of fullyirrigated paddy land that is benefited by the government established facilities has been steadily increasing, while that of rain-fed paddy shows a downward trend. The main fluctuation originates in the area of partially irrigated paddy which are irrigated by the village constructed commoi facilities or indivdua1 farmers' own facilities. r.i Table 2-4. Acreage of paddy land by irrigation status, by year, 1961-1970. Unit: 1,000 Ha. Partially Irrigated Fully Year Irrigated 1961 1962 1963 1964 1965 1966 1967 238 256 264 277 284 288 295 238 298 307 1968 1969 1970 Rainfed Total 230 218 215 208 200 1132 1143 1158 1191 1209 1209 1214 1136 1208 1193 664 669 679 706 725 740 740 705 769 773 181 179 146 141 113 Source: Agricultural Statistics Yearbook, 1970, Ministry of Agriculture and Forestry. The estimated price elasticity of the acreage of partially irrigated paddy land is computed as: A 4.44( Rt-1 .18627 A64813 PPFI (1.652) t1(957) (R2 .776) The estimated short-run and long-run elasticities are . 19 and 53, respectively, for partially irrigated paddy land and those for all barley acreage are . 13 and 26, respectively. The adjustment coef. ficient for all barley acreage, however, is higher than that for rice acreage, indicating a faster response to price changes. In the computation of aggregate supply response, the elasticity estimated with only price variable is taken into account. 37 Aggregate Supply Response Resulting aggregate response for rice is .3Z6, those for corn-. mon barley and naked barley are .470 and .385, respectively. Approximately, 10 percent increase in product price will induce farmers to increase rice production by 3 percent, common barley by 5:percent and naked barley by 4 percent. Conclusion The empirical analysis supports the initial hypothesis that average Korean farmers do respond in their grain production to changes in relative product price and they tend to be more responsive in barley cultivation than in rice. As stated above, this implies a possibility that an appropriate price policy in favor of barley would contribute relatively more to achieving increased foodgrain production. Especially, the higher yield response for barley indicates a potential to improve the average productivity of land and other produc- tive factors. In interpreting the empirical results based on the past observations, however, one must not overlook the fact that, even if the fitted function is accurate and the estimated coefficients are statistically significant1 they can only serve as an indicator in the neighborhood of the statistical population to which they refer. More specifically, the I] estimated elasticity may not be able to predict what the output response would be if the relative price of product doubles, for such an increase in price has never been experienced by Korean farmers and is very unlikely to occur in the near future. In a statistical sense, such an increase in price is too far from the population which an empirical study intends to explain. In measuring the response of farmers in production decisions, the productivity of resources which is conditioned by the prevailing level of technology was held constant by using the cross-sectional production function approach. Therefore, the major responses are revealed by the farmers' intention to increase the use of productive factors when the price relationship between product and inputs becomes favorable. From this we can conclude that even under a static technological environment, the average Korean farmer does respond to changes in price condition through changes in the quantity of input factors used. In the growing agriculture however, the growth of technology is not independent of the price condition for farm products. Farmers will become more willing to adopt the improved methods of cultivation if the terms of trade for their products become favorable than otherwise. In conjunction with other development policy, price incentives play an important role in disseminating the improved technology in the farm area. The improved technology accompanies by definition an increase in productivity of farming resources employed. Within the context of the empirical analysis presented in this chapter, this has an effect of increasing the productivity coefficients estimated in the production function. This implies that the true response of intendedT! output can possibly be greater than those estimated here. Thus, price incentives have double effects in increasing foodgrain production fast adoption of improved tech- nology by farm producers and increased use of productive factors, both non-purchased and purchased. 40 MARKET ANALYSIS In the preceding chapter, the analysis was centered on the effects of grain prices upon the decisions of farm producers in their production activities. The present chapter will deal with various behavioral phenomena that take place via the market mechanism for the realized output of food grains in a given consumption period. Ideally, the function of grain prices can best beanalyzed by incorporating into one system all the behavioral decisions that govern resource allocation in production, consumption and sales. As already stated, however, there usually exists a divergence between production decisions (or plan) and actual production due to uncontrollable factors This fact alone gives a sufficient reason, from the analytical point of view, to separate our grain analysis in this chapter from the production aspects and to carry out the study with the realized output (or stock) as a starting point. Given the output (or stock) of foodgrains, a change in the relative level of grain prices- -whether relative to all other prices as a group or to one another- -affects not only the total amount of grain consump- tion but also causes a shift from one kind of grain to another. If grain price policy is to be effectively used to influence the grain sector of an economy, it is essential to know in what direction and to what extent both farm and urban consumers adjust their consumption to 41 changing price condition and how it affects the decision of farm pro- ducers in their market sales. There has been a debate among a number of economists as well as policy makers whether an increase in foodgrain prices in the subsistence-oriented agriculture would induce farm families to increase their consumption, resulting in the decrease in market sales or it would cause an increase in market sales,' The underly- ing argument for the former is that an increase in the prices of food- grains makes it possible for the farm producers to satisfy their cash requirements by selling a smaller quantity of foodg rains, leading to a reduction in the amount of grain coming to the market. This reduction in sales, in turn, causes an upward pressure on prices and again a smaller amount on the market, and so on, resulting in the so-called spiral phenomenon. Implicit in this argument is the assumption that the farmers have minimum cash requirements and sell only that quantity necessary to fulfill their minimum cash requirements. The latter argument, however, emphasizes the intersector demonstration effect and hypothesizes an opposite phenomenon. -'See, for example,P.N. Mathur and H. Ezekiel (28, pp. 399) andV. Dubey(6, pp. 696). 'Mathur and Ezekiel hold that, although the farmer's demand for cash income is not necessarily completely fixed, it is more nearly fixed compared to demand for food consumption (28, pp. 398), 42 There is no a priori case for judging whether or not such a phenomenon takes place. It can be posed only as an empirical question. 10/ In short, the present study attempts to answer the following basic questions: 1) Do farm and urban consumers respond to an appreciable degree to changes in the price of foodgrains in their foodgrain consumption? 2) What is the magnitude of the cross substitution effect between the major foodgrains, rice and barley, when the relative price of either one of the grains changes? 3) Do farm producers, when they are faced with an increasing price for grains, consume more of their output (or stock) because of an improvement in income, or do they sell more in order to acquire a greater amount of cash receipts? These questions are posed not as hypotheses of purely academic interest but as practical policy questions which the Korean government is facing in formulating foodgrain price policy. The questions are especially highlighted by the recent movement of the Korean govern- ment to reform the foodgrain price policy in the direction of a two the writer's knowledge, essentially no empirical study was made as to the response of farm marketings of foodgrains to changes in cash income due to changes in foodgrain prices. 43 price system, aiming at a reduction in the consumption of rice, and an increase in the consumption of barley, but at the same time an increase in production of both grains. In carrying out an empirical analysis concerning the subject, special attention must be given to a peculiar feature emerging from the dual role of farmers; as consumers on the one hand and as sellers of products on the other. In Korean agriculture where the foodgrains constitute major items in farm production and where nearly one-half of the total grain output is consumed by farm families themselves, the behavioral patterns regarding farm consumption and sales cannot be studied in isolation from each other. The decision to consume from the given stock of grain necessarily affects the opportunity to obtain cash income that is required to purchase the non-agricultural corn- modities and services, such as clothing, education, etc. Conversely, an increase in market sales in order to increase cash receipts means a decrease in family consumption of home-produced foodgrains. In short, the farm household economy is characterized by the integration of two decision units into one. Further, the market prices of food- grains are not determined by the decision of farmers to sell alone. They depend on the amount of foodgrains demanded by the urban con- sumers, too. Therefore, in understanding the impact of grain prices, it is necessary to take into consideration the joint-dependence of farm 44 consumption, market sales and urban consumption)" In constructing a simultaneous equation mode1 an appropriate modification of the conventional theory of demand is attempted in order to reflect the dualistic feature of the farm household in Korean agriculture. Then the effects of foodgrain prices are analyzed on the basis of the estimated results. Model Construction The model adopted is a simultaneous equation system based on the market equilibrium concept. The system consists of eight equations: six behavioral equations comprising two farm demand, two farm sales and two urban demand equations. each for rice and barley respectively, and two market identity equations for rice and barley. The system contains the total of 21 variables, of which eight are endogenous and the rest are treated as exogenous variables. The characteristics of each variable will be explained as we proceed with the specification of each equation. necessity of simultaneoas consideration of related variables is well stressed by M.A. Girshick and T. Haavelmo (10, pp.83) it is impossible to derive in the general context as follows: ". statistically the demand functions from market data without specification of the supply functions involved.. More generally, if we wish to estimate any particular economic relationship on the basis of market data we are forced to consider, simultaneously, the whole system of economic relations that together represent the mechanism that produces the data we observe in the market. . . 45 Farm Demand The ordinary theory of demand hypothesizes that the demand for a commodity is a function of the price of the commodity under consid- eration, the prices of related goods and the disposable income of the consumer. It is also shown that the total effect of a change in the price of the commodity upon the quantity demanded, assuming constant money income, can be analytically decomposed into two component effects: the substitution effect involving a movement along the original indifference curve if a price change is compensated by a simultaneous income change and the real income effect producing a movement along anEngel curve {14,36]. The following Slutzky equation represents the decomposition of the total effect of price changes. (3. 1) aq api = (-) 3q1 The first term on the right hand side is the substitution effect with the Happarent real incomeT' unchanged and the second term the income effect with price unchanged. keep apparent real income constant implies that the consumer is always enabled to purchase the original bundle, and it is a good approximation to real income where price changes are not so great [8]. 46 The basic assumption in the above analytical process is that the consumer is the purchaser of the commodity with given money income which is earned independently of the price of that commodity. The level of money income has no direct relation to changes in the price of the commodity demanded, except that the real income (as defined by the purchasing power of money income) is liable to vary as a result of price changes. However, if the consumer is at the same time the seller of the same commodity, as is the average Korean farmer, and a major portion of his money income comes from the sale of that commodity which otherwise could be consumed bythe family, money income is no longer independent of changes in the price. Hence, in analyzing the consumption behavior of the seller- consumer, it is necessary to take into accountthe effect of price changes upon money income. Given the following demand function, p1 1 2 where Y and q1 g(p1) denotes the quantity demanded, moneyincome, I p1 the price index excluding the price, Y the and the p1 I 47 Lespeyres price index, the total effect of price changes can be decomposed as; aq1 aq q1() i (3. 2) ap1 - = a q1 y i , --'Mathematical derivation is given as follows; given demand function f(, p1 q1 = I L where Y g(p1) Taking derivative for both sides with respect to p1 Y a() a() a q1 Ii p1 ap1 a f 1 a f a Y r a(jX) i 12 af + ap1 _L we get p1, 2 2 a,2 ap1 -_. [ap1 '2 2 f. p1 a() Ii + af 12 12 Y a If we assume that apparent real income remains constant so that the consumer can buy the original bundle of goods (following Slutzky), then we can rewrite aq1 af ap1 a + q1 ap1 Y Ii Further, by setting the price indices 2 I and 12 for the initial The first two terms on the right-hand side represent the substitution effect and the income effect, both in the Slutzky sense, and the third term represents the direct money income effect due to a change in sale price. If the money income price p1 i.e. , if Equation (3. 2) aY p1 0, Y is not dependent upon the then the third term drops out and the reduces exactly to the Slutsky Equation (3. 1). Therefore, the Slutsky equation explains a special case of the purchaser with the given money income. It does not explain the behav- ioral responses to price changes of the seller-consumer. The total effect of a price change on consumption now depends not only on the sign but also on the relative magnitude of each term on the right-hand side of Equation (3. 2). When faced with changing prices of grain at a given point in time, the farmer has three alternative choices concerning the disposal of his grain stock, He has to decide how much to consume and how much to sell, and/or how much to retain for future sales and consumption. Suppose the price of rice rises due to a certain exter- nal cause, then two opposite effects are conceivable. The increase in cash receipts due to a rise in the price of grain may induce an period to unity, we obtain aq1 1. 8q1 aq1 = ()1 apparent real income = constant q( ) p1p10 + 1 p1-p1 I- 49 increase in the consumption of that grain. Or the higher cash receipts due to the higher price may create a desire for a still higher cash income which could only be acquired by reducing home consumption and selling more on the market. The case of the seller-consumer is illustrated in Figure 3-1. M M0 M2 p1 M1 0 q4 q3q2q1q0 q* Figure 3-1. Effect of price changes on the consumption of the sellerconsumer. For illustrative purpose, let us assume that the farmer disposes of his grain stock Oq* in a given period by consuming and selling, that is, he does not retain any grain at the end of the period. Let the vertical axis OM represent all other goods lumped together. Given the preference pattern between consumption and other goods, he will consume 0q1 units of grain at the price of units of other goods by selling of grain rises from 0q2 p1 to p1 and obtain 0M1 units of grain. Now the price q1q* p2. At this price he will consume units of grain (the above diagram is drawn in such a way that the net result is a decrease in consumption, but it could be an increase) and sell q2q* units of grain to obtain 0M2 units of other goods. The total movement from the initial consumption sumption q2 q1 to the new con- is attributable to a dualistic characteristic of this consumer. Let us for a moment treat him as if he were a buyer of grain with a given income 0M0, then he would consume of grain at the higher price p2. This movement from can be decomposed into the substitution effect, i.e. , and the real income effect, i. e. , q3 to 0q4 units to q1 q1 to q4 q3, q4. These two effects are represented by the Slutzky Equation (3.1). or by the first two terms on the right-hand side of Equation (3.2),. For the seller-consumer there is an additional effect due to the increase in monetary receipts. This monetary income effect causes the movement from term ip1 q4 to q2 which is explained by the third in Equation (3. 2). Whether or not the counteracting effect of the increase in mone- tary receipts is dominant to the direct effect of price changes is 51 subject to an empirical test in this study. Another point with which one must take special care in formula- ting the statistical demand function for farm households is the vagueness of the concept of farm income. In semi-monetized and selfemployed farming a large portion of farm income consists of non-cash items. The measurement of farm income is difficult especially when an analytical study is made on the basis of monthly or seasonal observations, for the largest part of income, whether in cash or in kind, is realized only after the crop harvest. During the non-harvest season they consume and sell from the stock of grain, partially depending on the cash income they earn off-farm. Seasonal characteristics of farm behavior should also be considered in specifying farm demand and sales equations. There is a tendency for farm consumers to consume more grain when more is available at the harvest and immediate post-harvest seasons and less when less is available during the rest of the year. It is also reason- able to assume that the farmers respond to price changes differently in different seasons both in consumption and sales, that is, the slope of demand and sales curve may vary depending on different seasons' The discussion thus far leads to the formulation of the farm demand functions as follows: Seasonal responses are not discussed in the main text, but the estimated model is presented in Appendix B for reference. 52 (3.3) q FK FD b0 + bUPRt + blZPBt +bi3Pw + b14q1 FS FD (3. 4) + b16D3 + U1 FK = b20 + bZlPBt + bZZPRt + bZ3PWt + b24q1 + FS + b26D2 + U2 Where FD = Monthly farm per capita consumption of rice. = Monthly farm per capita consumption of barley. = Monthly farm per capita sales of rice. Monthly farm per capita sales of barley. FS Rt = Monthly average wholesale price of rice deflated by the Index of non-grain wholesale prices. Monthly average wholesale price of barley deflated by the Index of non-grain wholesale prices. Monthly average wholesale price of wheat flour deflated by the Index of non-grain wholesale prices. FK = Farm per capita stock of rice on hand at the end of the previous month. qc1 Farm per capita stock of barley on hand at the end of the previous month. 53 NRB Monthly farm per capita income originating in the non- rice-barley sources deflated by the Index of prices paid by farmers. = Random error terms. b.. 13 t Behavioral parameters to be estimated. Current period. if February-May period =1 = 0 otherwise. D3 = 1 if June-September period (barley harvest affected season) = 0 otherwise. All quantity variables are deflated by the. ratio of urban to farm population in order to maintain consistency with the market identities (to be explained later). The above two farm demand equations contain five endogenous FD, FD variables, namely, R' F5 FS Bt and five exogenous variables including seasonal dummies, and D3. 15/ FS +Pq FS The composite term (Pq 15/ Though FS and FS ) which appears in both are endogenous and 'NRB exogenous variables in the model, the whole composite. term is treated as a single endogenous variable. q 54 demand equations represents the cash receipts from sales of rice and barley and it is obvious that the cash receipts is determined by the level of prices for both grains and the quantity sold on the market. The quantity sold of each grain in turn depends on the prices of grains and other relevant variables as will be specified in the sales equation later. Thus the measurement of the money income effect requires a simultaneous consideration of all the variables in the system. It is of analytical interest, however, to show how the estimated coefficient on this composite term relates to the money income effect which is represented by the third term in Equation (3..2), under a certain quali- fication. Assuming other variables remain unchanged the total response of the quantity demanded, say for rice, can be measured by FD = b11 +b5q FS A A one can clearly see, the estimated b A the ordinary demand curve, and b15 FS represents the slope of the response to a unit change in money income due to price changes. To break it down, b15 aq1 corresponds to ()p1 p, and term with the price remaining constant, i. e. q money income effect, to p1 in Equation (3. 2); Whether the (-)p1 p1 , offsets or reinforces the direct effect of a price change depends on the sign and the magnitude 55 FS of each coefficient, and Farm Sales The shortrun decision to sell from a given stock is hypothesized as a function of the current price of the grain under consideration, the size of stock on hand, the amount of liabilities as of the end of the previous period, cash expenditures for non-farm source goods and services, and the value of the stock of major grains rice and barley, evaluated at the current prices. FS (3.5) qRt FK L +b34E+b351 P +b3ZBt P +b33t1 -b 30 +b3lRt + FK + FK + U3 1 FS (3.6) FK b40 + b4JPBt + b42PR +b43L 1+ b44Et + b45J + FK + FK + U4 1 where: L = Farm per capita liabilities as of the end of the previous month deflated by the Index of prices paid by farmers. Farm per capita cash expenditures such as for childrens education, clothing, etc. deflated by the Index of prices paid by farmers. tJ3, tJ4 Random error terms. Other variables are the same as specified in the demand equations. Insofar as the main objective of selling grains is to meet the cash requirements such as for education of children and purchase of goods and services from the nonfarm sector, an ideal thing to do is to use the total cash demand as an explanatory variable in the sales function. But since it is almost impossible to identify the actual cash demand from the observed time-series data, the farm debt and other cash expenses are used as a proxy in order to reflect at least the major part of farmers' monetary demand Theoretically the outlay of cash expenses is determined by the level of farm income rather than vice versa. But in the shortrun it is reasonable to assume that the educational expenses and purchase of clothing s, etc. are relatively stable The introduction of the term what analogous to that of the term 16 / FK FS FK FS is somein the demand equations. In the sales equations it was hypothesized that farmers' decisions for market sales are affected by the value of the --"This should not be taken to mean that farmers' demand for cash for all purposes is fixed in the shortrun. It only means that the cash spendings for certain purposes arerelatively stable. The fact is that the total outlay varies even on a monthly basis depending on the price level for farm products and off-farm earnings. 57 grain inventory, whereas the decision to consume was hypothesized to depend on the direct cash receipts plus the off-farm earnings.JV In understanding the sales decisions of the farmer, it is necessary to understand the saving pattern that prevails in the rural area. Even though the monetization of the rural economy has proceeded along with overall economic development, the average Korean farmer still has a tendency to save in kind rather than in the form of cash, whenever immediate cash need is satisfied.- In the face of chang- ing prices of grain plus some uncertainty about the next crop, the !ZiSince PR and are the averageprices during the curthe beginning stock of the curand rent month while FK FK does not represent the rent month, the term actual value of the current stock. A closer approximation could be obtained by replacing the beginning stock by the average stock holding FK by during the current month, that is, 1 FK FTC FD FD FS FS or and FTC 2 FD FS FTC by FTC 2 Considering, however, that a similar pattern of variation in all three variables is repeated every year, the estimated coefficient may not differ significantly. immediate cash need does not mean the minimum cash need. One may think that a continuous inflation is another factor which discourages the monetary saving in general. But it is doubtful that monthly sales of grain are affected by inflation. It may influence the longer term saving patterns at least longer than a month. 1The 58 farmer may feel safer with foodgrain in storage. When the price of grain, say rice, rises, he may expect a further rise in the future and feel that his saving potentialities increase with the increased value of his stock, thus selling a smaller quantity of rice. Or conversely, the increased saving which is realized by the increased value of stock may make the farmer feel richer and accelerate the sale on the market to obtain more cash so that he can consume more of the non-agricultural commodities and services. In which direction the farmer tends to move depends on whether or not the response to an increase in the value of his grain stock offsets the increased desire for more cash. The total effect on rice sales of one unit change in the price of rice can be measured by FS = b31 + The estimated parameter b3 1 curve (in the ordinary sense) and A FK represents the slope o the supply represents the effect of the stock value. The rate of response in sales is affected by the size of stock on hand. On the other hand, the size of stock which the farmer can maintain is dependent upon the storage capacity. The inadequacy of storage facilities is one of the important factors which force the average Korean farmer to sell, particularly in the iminediate post-harvest season, a large portion of his output. Thus the size 59 of the stock itself tends to have a positive effect on the amount of sales, probably offsetting the propensity to save in kind due to an increase in stock value. This counteracting effect of the grain stock is measured by FS A FK aq1 A = b35 + b36PR Notice that it is a function of the price level for rice. This implies that the sales of rice respond to a change in stock differently at different price levels. Urban Demand The urban demand for foodgrairi is hypothesized to depend on its price, the prices of related goods and disposable income. In order for the demand function to be empirically useful, one must be able to make statements about the preference function of the consumer at varying level of prices and money income.-' More specifically the urban consumer's preference pattern may be such that he may respond differently to a given change in price at different levels of income and !2i'Based on the lecture notes for Agricultural Economics 517: Product and Factor Market taught by John A. Edwards, Department of Agricultural Economics at Oregon State University, 1971. also respond differently to a given change in income at different levels of prices. An inference about such preference function is analytically possible by introducing an interaction term between the price of the foodgrain consumed and disposable income. The urban demand equations for rice and barley would be written (3.7) b50 + b51PR + b5ZPBt + b53Pw+ bs4Yut + b55(PRtYUt) + TJ5 (3. 8) q UD b60 + b6lPBt + b6zPRt + b63Pwt + b64Yut + b65(PBtYUt) + b66DG + U6 Where: qTJD UD = Monthly urban per capita consumption of rice. = Monthly urban per capita consumption of barley. Monthly urban per capita income deflated by the Index of urban consumer prices. U5, U6 Random error terms. 1 if 1963-64 period 0 if otherwise. The dummy DG in the barley demand equation reflects a special grain situation that lasted for two years, 1963 through 1964. Due to the two successive, crop failures for 1962 rice and 1963 barley, 61 the Korean government initiated a nationwide campaign to encourage the increased consumption of imported barley and wheat flour in the urban area. The estimated effect of a change in the price of rice, for example, on urban consumption is given by UD A aPR b51 + bssYu Which implies that, assuming other things remain constant, the urban consurners response to price change varies depending on the level of income Y. Similarly the effect of income changes is estimated by UD A aY A = b54 + b55PR As already stated, the response of consumption to a change in income dffers at a different level of the price of rice. From the above results, we can approximate the level of income and the level of the price of rice at which the urban consumer reaches the HsatietyT point in rice consumption; the htsatietyu point defined as the potential consumption at which the consumer does not respond to changes in income and price. These levels of income and price can be derived by setting eac1i of the above partial derivatives equal to zero. 62 b55 Market Identities The total disappearance from the farm stock is the realized supply of domestic grain available for both farm and urban consumers during the period concerned, assuming that the grain held by whole- salers and that in pipeline remain constant over time. This is a somewhat rigid assumption because there are always some whole- salers who buy a lump sum quantity from farmers at the harvest time and sell in varying quantity during the period of several months depending on the price condition. This inventory operation of whole- salers causes a time lag between the release of grain from farmers and the actual consumption by urban consumers. The inclusion of the marketing function that specifies the operational behavior of wholesalers would improve the quality of our equation system. But it could not be done due to a lack of statistical information. Besides, a distortion caused by the exclusion of such a marketing function may not be big enough to make the whole system of equilibrium invalid. 63 Let The total farm stock of rice at the beginning of the MQ current month. MQK = The total farm stock of rice at the end of the current month. The total consumption of rice by farm population MQ during the current month. The total consumption of rice by urban population MQ during the current month. = The total imports of foreign rice during the current month. = The total purchase of rice by government during the MQ current month. GS MQRt = The total sales from the government stock during the current month. MQS = The total salesof rice from the farm stock during the current month. MQ The total exports of rice during the current month. Then, the following identity relation must hold by definition: MQ1 MQ = = MQ + MQ + MQ Since MQ + MQS = MQ + MQ + MQ + MQ appears on both sides, we can write FS I + MQ = MQ MQRt + MQRt + MQ + MQt Reducing to urban per capita basis, we obtain FS (3. 9) I +q + UD GS = +q GD + E Likewise, for barley FS (3. 10) +q I + UD GS = + GD Notice that the farm demand variables E +q q FD and q FD do not appear in the market identities. This does not mean that the farm demand is wholly unrelated to determining equilibrium condition in the market. As already specified in the demand and sales equations, the quantity demanded and that sold are jointly dependent upon the prices of both grains and the size of grain stock on hand. Complete Model To rewrite the whole system of simultaneous equations: (Farm demand) FR FD (3.3) b10 +bllPRt +blzPBt+bl3Pwt +b14q1 Is +YNRB) + b16D3 + U1 + FD FR -b 20 +b21PBt +b22PRt +b23PWt +b 24Bt_1 IS + NRB )+b26D 2+U (Farm sales) FS (3.5) b30+b3lPRt FR 3ZPBt+b33Ltl +b34E+b351 + FS (3.6) i+ FR Bt i + b40 + b4lPBt + b4ZPRt + b43Lt + FR FR 1 FR + b44E + b451 + U4 (Urban demand) (3.7) = b50 + bSlPRt + bSZPBt + b53Pwt + bS4YUt + bsS(PRtYUt) + U5 (3.8) q UJD Bt b 60 +b 61PBt + b 62PRt + b63Pwt + b64Yut + b6S(PBtYU) + U6 I__ 66 (Market identity) FS (3.9) q (3. 10) q FS +q + I + I + GSUD UD GS = + + GD GD + + E E Since the structure is nonlinear in variables, though linear in parameters, the reduced form obtained by solving it will be non- linear in variables, and the solution may be complicated involving multiple roots. It has been suggested by Klein [16, pp. 120-121] that replace- ment of these nonlinear terms by their observed sample means makes it possible to obtain an explicit solution. The nonlinear terms in the model were transformed into linear approximations according to the formula given by Klein. The formula is: xYYx+xY-xY Linearized terms are: FS + FS + NRB = FS FS+ FS + + FS FS---PB+YNRB FK FK + Pq1 = FK 1R + FK + FK 1B FK FK FK- FS The market price of rice, can be determined by substitut- ing the farm sales Equation (3. 5) and the urban demand Equation (3.7) for rice into themarket identity (3. 9), and likewise that of barley, by substituting the farm sales Equation (3.6) and the urban B demand equation for barley into the market identity (3. 10), and then solving (3.9) and (3. 10) simultaneously. Next, we can solve for FS q FS , q UD and UD by substituting the obtained in the first step. Last step is to substitute R' pqFS into Equations (3.3) and (3.4) in order to get and qFD and and Hence, the system is complete as an econometric model. Identification of the Model Order Condition The identification status of each behavioral equation is determined by the following rule [161. Let G = Number of endogenous variables which appear in the equation. = Number of exogenous variables which are excluded from the equation. Then (1) If K* >G - 1, we have overidentification. (2) If K** = - 1, (3) If K** < 1, we have exact identification. we have unidentification. The composite variable which appears in each equation must be treated as a single endogenous variable, since they all contain multiplicative terms in one or more endogenous variables. The result of identification status for each equation is given by: Equation - 3. 3 3. 4 3 3 3.5 3 Status Over identified Overidentified Overidentified Over identified Overidentified Overidentified 1 3. 6 3 3.7 3 6 6 6 6 7 3. 8 3 7 Rank Condition The general rule is that an equation in a linear model- -linear in parameters- -of G equations is identified if and only if at least one nonzero determinant of G - 1 rows and columns is contained in the array of coefficients formed by omitting the rows of coefficients of the equation in question and omitting all columns not having zero in that equation [411. Let FS + Yv FK, FS + FK NRB OS OS GS GD E I GS qB+qB GD E I Then, the array of all the coefficients in the model is as on the following page, each column being headed by its variable (the column of constant terms is excluded). For Equation (3.3), for example, the matrix of coefficients relevant to the rank conditionsi.s shown on page 71. From this matrix, eight 7 x 7 non-zero determinants can be formed. Hence, the rank condition for the farm demand Equation (3.3) is satisfied. By the same procedure one can easily show that the rank condition is satisfied for the rest of the equations. Empirical Results Data The price variables employed are the monthly wholesale price series for rice, barley and wheat flour developed by the Bank of Korea. The data are deflated by the Index of non-grain wholesale prices so as to reflect the real prices relative to the prices of nongrain commodities. The common use of the wholesale price series causes an upward bias for thefarrn-level demand and sales equations, and downward Equ Fb FD FS PS TJD LID F1 Os Os b22 21 (3.5) o o -1 o o (3.6) o o o -1 o (3.7) o 0 0 0 -1 (3.8) 0 0 0 0 0 (3.9) (3.10) 0 0 o o b3 o o b3 b32 o o b o o b42 b41 0 0 b o b5 -10 0 o b b o 0 1 o 0_i 0 1 -1 23 0 6S 24 o b b3 o b b43 b b52 b53 0 b b 61 62 63 o 0 0 0 0 o 0 0 0 0 o 0 b3 o 0 0 o o o o o o 0 o 0 o b5 0 0 0 o 0 b 0 o 0 o b 64 0 0 o 0 0 0 0 0 66 o 0 0 0 0_i 0 0 0 -1 0 FD FS FS UD UD V -1 o 0 o 0 -1 0 0 0 0 RU BU 0 Q 0 0 FK as L E 0 0 0 0 b34 0 Y q as D2 DG 0 1 0 0 0 0 0 0 0 0 b36 0 0 0 -1 0 0 b46 0 0 b45 b43 b44 0 0 0 0 0 0 -1 0 0 b55 0 0 0 0 b54 0 0 0 0 0 0 0 b66 0 0 0 0 -1 0 0 b65 0 0 0 b64 0 -1 0 1 0 0 0 0 0 0 0 0 -1 0 0 0 -1 0 1 0 0 0 0 0 0 0 0 -1 0 0 0 0 1 72 bias for those in the urban demand equations, for the prices actually received by farmers tend to be lower and the prices actually paid by urban consumers (consumer price) tend to be higher than the whole- sale prices. Provided that the marketing margin is constant over time, as one can expect it to be, if there is no particular demand for marketing services for foodgrain, a certain proportional relationship can be derived between the wholesale prices, the prices received by farmers and the urban consumer prices. Based on the hitorical trend from 1963 to 1971, the following relationships have been obtained. = 1. O4P (R2 = .995) P = . (R2 = .991) P = . (122. 79) P = 1. O4P (15805) 96P (R2 (334. 02) 94P 966) .989) (238.3) Where the superscript W denotes the wholesale price, prices received by farmer, and (R2 . U, F the the urban consumer price. 'Since the regression line was estimated forcing through the origin, R2's were corrected in order to make comparable according to the following formula: 2 Y-bX.Y. 11 1 73 By using the above relationships, one can transform the original equations expressed entirely in terms of wholesale prices to equations expressed in terms of the prices received by farmer and the urban consumer prices. Per capita consumption data were drawn from the records of The Grain Consumption Survey, 1963-71, conducted by the Ministry of Agriculture and Forestry. Farm per capita sales of foodgrains, stock on hand, liabilities, cash expenses and non-rice-barley income data are based on the records of Farm Household Economy Survey, 1963-7 1. Farm per capita liabilities, cash expenses and non-rice- barley income are deflated by the Index of Prices Paid by Farmers. Urban per capita income data were obtained from the records of the Urban Household Expenditure Survey, 1963-71, of the Bureau of Statistics. It was deflated by the Index of urban consumer prices. In fitting the model to the observed data, the farm-level quantity variables, i.e., per capita consumption, sales and stock, were deflated by the ratio of urban to farm population so that all the quantity variables contained in the behavioral equations were consistent with those in the market identity equations. T1-ie model is fitted using monthly data for the period from January 1963 through September 1971, totaling to 105 observations. 74 Estimation Procedure Two-stage least squares (TSLS) was used to estimate the behavioral (structural) equations under the following assumptions that 1) the observed data have no measurement errors, 2) the exogenous variables in each equation are not correlated with the error term in the equation, 3) the structural error term in each equation is serially independent and 4) each structural error term follows a normal distribution and has common variance for all observed periods TSLS method is a most widely used single-equation method for estimating an overidentified structural equation in a joint-dependent system of equations. The estimation process consists of two successive applications of the ordinary least squares method. In the first stage the reduced form equations for each endogenous variable are estimated and in the second stage the least squares method is again applied using the predicted value of the endogenous variables obtained in the first stage as predetermined variables [16]. Since each endoge- nous variable which appears on the right-hand side of the structural equations is regressed on all the exogenous variables in the system in the first stage, this is one of the full-information methods Possible Source of Errors The error term in each structural equation may arise not only 75 due to the omission of relevant variables, both economic and noneconomic, but also to a failure in specifying the important economic relationships in the model. There may be a number of such structural equations which might have improved the predictive accuracy of the model but were omitted. Important of them are 1) the demand equations for wheat flour for both farm and urban consumers, 2) the marketing function that explains the behavior of wholesalers, espe- cially their inventory operation and 3) the industrial demand for grains. The omission was inevitable due to a complete lack or insufficiency of data for the period considered. The estimated structural coefficients are presented in Table 3-1. Predictive Tests The ordinary R2 obtained in the second stage regression is not a workable indicator of the usefulness of an estimated structural equation. H. Theil suggests three different methods for testing the validity of the model--' [33, pp. 56-5811. "The same ixiethods are also suggested by Christ [5, pp. 385- 408] defining the first method as the ex ante prediction, the second as the expost reduced-form tests and the third the ex post structural tests. Footnotes for Table 3-1. The coefficients for "B bf _FS = Cl = + in the original equations are adjusted using the following relationships: and = 1,O4P = = L.04P = -PqFS + - + .96P - _FS q8 FS _FS- - B FK + ,-FK '1B "B + -Bt-i Rt-1R- - FK B + NRB FK FK *** denotes significance at 99% probability level ** U U * U U 95% 90% II II U Figures in parentheses are V-values. Simple correlation coefficient between observed series and predicted series. Table 3-1. The estimated structural coefficients Explanatory varab1e Constant Equation term (3> qD (34) (3.5) (3.6,) qFD FS 19.515 (6.087) (3,8) q R B W Rt-1 qFK Bt-1 L t- 1 (NS) .328850 (8.785) .007342 (2.703) -. 117110 .014225 (4.963) .003991 (1.166) .372224 (6,656) (2.555) .028372 (3.485) * (NS) -21.360 (4.970) .160111 (1.412) *** -.048812 (.722) .101249 (5.162) (NS) (NS) 4.707 (4.352) -.034915 (.116) -.003913 (.116) B qUD -.35414 (6.277) qFK B *** -12.895 (3.809) qFS (37) R *** .047675 (8.902) *** 31. 273 (10. 925) R * UB - 5,760 (1,611) .00879S (9.417) -,443071 (11.014) *** .237600 (12,995) .105982 (4.360) (NS) -.073949 (.990) -.008850 (4,589) (NS) .003014 (1.357) * -.000670 (1.786) Table 3-1. (continued) Explanatory variable Equation E Y.E/ Y PY d/ D D D - (3) (3 4) (3,6) FS (3,7) (3.8) .000914 (1.869) UD UD -2.169 .792 .843 765 861 .834 .913 .751 .855 .825 .822 .759 .782 (5.017) *** qFD FS r!' * qD (3.5) R2 -. 002904 (6.383) -1, 253 (3. 141) .011425 (6,020) -.001043 (2.549) -.005386 (7.063) .000091 (1.637) , * -.002350 (3.152 .00006839 (4.984) (NS) * .000295 (.389) -.00003743 (1.722) (NS) .627 (1,055) (i) One can take the predictions as they are and compare them with the corresponding observed actual data. (ii) One can separate the endogenous variables from the exogenous ones, insert the observed values of the latter category in the equation system, and derive conditional predictions of the endogenous values by means of this system. (iii) One can insert, for each equation of the model, the observed values of the right-hand (explanatory) variables, and compare the implied value of the lefthand variable with the observed value. The first method is a straight forward forecasting by substituting the future year values of exogenous variables. The second method is the one based on the reduced form equations and the third is the structural equation tests. In respect to the relative merits of these alternative methods, C. Christ [4] states the reduced form tests can testforecasts, whereas the structural-equation tests cannot. . . . Thus for checking forecasts, the reduced-form tests are appropriate; and for testing one or more structural equations individually for validity in the postsample period, the structural-equation tests are appropriate. For the quantity variables the third method of testing was adopted for the reasons that the main objective of estimating the structural equations from (3.3) through (3. 8) is to investigate the behavioral responses to changes in the prices of foodgrain in consurnp- tion and sales decisions, not to forecast the future movement of the endogenous variables. Moreover, as long as the predictive test is 79 performed on the basis of the sample period errors, thereduced form equation and the structural equation testsareessentially the same. For the price variables, and B' the second method was used; that is, the reduced form equation for each price was obtained by means of two market identities (3.9) and (3. 10) and then the observed values of exogenous variables were inserted to calculate the predicted series. 22/ Reduced form equations for U O579Ol .272369 1 B U are: .212583 1 R and .057901 V whe r e: .079523 + U OO00398ZY V = .515732 X = -46.31 + .o08850P + Z = . 05Z635q OOOO7lZ4Y FK . 9.92 + 003Ol4P + . + .00235Oy 03303Zq FK - + .008795L + .011425E E GS GD I + . 000670L . 0053 86E E GSGD 004Z42q1+ 050557qT1- 627DG + I(qq+q . The predictive ability of each structural equation is summarized in Figure 3-2 through Figure 3-9 where the corresponding observed and the predicted series in the sample period are plotted. In order to see the degree bf deviation between two series, simple correlation coefficient, r, is given for each set.-1 Interpretation of the Estimated Results In interpreting the empirical results given in the preceding section, the main emphasis will be placed upon quantification of farm seller-consumer and the urban consumer responses to changes in the price of foodgrains. The impact of other variables such as farm debt (Lti), urban income (Yu), etc. will be discussed only to the extent that they are closely related to the effects of foodgrain prices. First, the analysis will be made of the partial responses by treating each behavioral equation as an independent single equation under the usual assumption of ceteris paribus. Next, the total behav- ioral responses will be examined by allowing simultaneous changes in all endogenous variables within the system. The analysis of the partial effects of foodgrain prices on farm demand, sales and urban demand has a practical importance as much as it is of theoretical Hooper [15] proposed a statistic called the "trace correlation" to measure the proportion of the total variance of the jointly dependent variable as a group that is explained by the predetermined variable as a group in a structural model. N 1] :30 20 Ui Ui -J 0 t' 1963 2 6 2 e 2 6 2 )2 6 f2 6 2 6 YEI-F C'1) MN'H 1964 1965 1966 1967 1968 1969 Figure 3-2. Monthly per capita farm demand for rice, 1963-71. (FD); r = Actual . 843. 1970 1971 0 CL in j -J Si I 6 12 6 12 6 12 6 12 6 12 6 12 6 12 6 6 YE(F 1129 1963 1964 1965 1966 1967 1968 1969 Figure 3-3. Monthly per capita farm demand for barley, 1963-71. Actual (qFD); p p 0 Predicted (AFD) r = . 861. 1970 1971 03 NJ 0 30 2 20 - Lr' 0 YER OND MSHTHS 1963 1964 1965 1966 1967 1968 1969 Figure 3-4. Monthly per capita farm sales of rice, 1963-71. Actual (qFS);e n ePredicted (FS); r = .913. 1970 1971 -J in r 6 c:i 1963 2 6 .2 .12 6 .12 6 2 2 ñND MNTH1 1964 1965 1966 1967 1968 1969 Figure 3-5. Monthly per capita farm sales of barley, 1963-71. Actual (qFS); p o opredicted (FS); r = . 855. 1970 1971 CL: 2::; a' Di s fl: -J 1) D{. nr 6 J2 6 1? 1? 6 : 6 Li 6 J2 6 Li 6 Li YEcP D;'C MflH 1963 1964 1965 1966 1967 1968 1969 Figure 3-6. Monthly per capita urban demand for rice, 1963-7 1. Actual (q)); c e 0 Predicted (CD); r = . 822. 1970 1971 Ui N JT G . 2 P 2 .P YE 1963 1964 1965 1966 I? 2 42 1967 1968 1969 Figure 3-7. Monthly per capita urban demand for barley, 1963-71. Actual (qUD); o o Predicted (AUD) r = . 782. 1970 1971 a. ..r '0 ./I j/. c r_. a p T r ;r 1 2 --#----2 1963 1964 1965 --'2 1966 1967 1968 1969 1970 Figure 3-8. Monthly wholesale price of rice per kilo (polished), 1963-71. Actual R' o p e-Predicted r = . 800. 1971 2 'Ic 2 1963 6 t 1964 1965 2 1966 1967 6 1968 2 6 1969 12 6 1970 Figure 3-9. Monthly wholesale price of barley per kilo (polished), 1963-71. Actual B)0 o e-Predicted r 851. 1? 6 1971 interest. Nearly all the past studies made of the foodgrain price policy in Korea have been based on single-equation approaches under the other-things-remain-constant as sumption. The policy suggestions drawn from such analyses tended to be unrealistic and nonoperational when brought into actual implementation. One good example is that in the past the government decided to raise the price of rice through manipulation of government stock in the hope that the urban consumers might reduce the consumption of rice and increase that of barley or wheat flour. The result was contrary to expectations because of the simultaneous rise in the price of other grains. This experience gives a lesson that certain conditions, whether institutional or behavioral, must be fulfilled in order for the cross substitution between different kinds of grains to take place. By comparing the results of the partial response analysis against those of the total response, one may be able to draw policy implications regarding the condition which is required for implementing food- grain price policy. Analysis of Partial Effects Farm Level Demand. A unit increase in the price of rice, other variables remaining unchanged, has the effect of reducing the quantity of rice demanded at farm level by 90 aq -(.354141-. 0009135) -(.354141-. 008231) R aPR (ES = 9. 01 kilo) .345910 units = This change in per capita consumption is caused by two different effects; the direct effect of price changes and monetary income effect 24/ The second effect of monetary due to changes in sales proceeds. income caused by price changes depends on the quantity of rice the farmer can afford to sell, which in turn is affected by the price of rice. Using the linearized equation, if the net effect is evaluated at the mean value of the sales ES over the sample per- = 9.01 kilos iod, a unit increase in the price of rice received by farmer tends to reduce its consumption by approximately . 35 units. Note that the positive sign of the coefficient for the cash income term (Pq FS + Pq IS + . NRB indicates that anincrease in cash income has an effect of increasing farm consumption. But this positive effect is too small to offset the direct price effect so that the result (in the partial sense) is still a reduction in rice consumption when its price rises. Instead, the consumer tends to increase barley consumption by 'The substitution and real income effect due to price changes in Slutzky sense are both entailed in the direct price effect. 91 3q FD F aPR = (.372224-. OOZ9O4 .346059 ) = (.372224-. 026165) units i. e., approximately .35 units (kilos) of barley. Notice that the effect of the cash income is negative for barley. The net result is a substitution of barley for rice consumption by exactly the same amount. The effect of a change in the price of barley can be approximated in a similar manner. When the price of barley goes up by one unit, the consumer reduces its consumption by = -(. ll7llo+.0029o4S) = -(.1l71l0+055466) -. 172576 instead a , units, FS = 1.91 kilos) increases rice consumption by FD = (.32885O+.00O9l35qS) = .330595 (.328850+. 001745) units. The difference of about . 16 units could be explained by a net increase in total consumption of grain by that amount of rice or by a reduction in the consumption of other grain. In either case, the above analysis indicates that the average Korean farm family has a relatively strong preference toward rice consumption. 92 The response to a change in the price of wheat flour presents a somewhat interesting case: the quantity of rice demanded increases whereas that for barley decreases when the price of wheat flour decreases (the same phenomenon is seen in the case of the urban demand function to be discussed later), Does this mean that wheat flour is complementary to rice and substitute to barley? The answer cannot be given solely on the basis of the sign of the response coeffi- cient in the estimated equations. A determination of the commodity relationship requires an estimation of the demand for wheat flour in the same system so that the estimated response coefficients can be compared to one another-' As already mentioned, this could not be done due to inaccurate data on wheat flour consumption. In view of the general consumption pattern in Korea, however, it is more likely that rice and wheat flour are substitute goods for each other. A possible explanation for this is that the price of wheat flour (nearly all the marketed flour is imported) has been controlled by the government and continuously lowered relative to the prices of non- grain items and other grains over the past several years, consequently Henry Schultz [31, pp. 569-582) shows that in the estimated and demand functions for two commodities A and B if are both positive, then A and B are in the substitute relationship, and if they are both negative, then A and B are complementary to each other. 93 leading to increased consumption both by farm and urban consumers. In addition, the wheat flour furnished as wage goods in the rural area through development projects has substantially increased its consump- tion by farmers. Owing to this relatively abundant supply and its low price in combination with a strong preference towards rice, the consumer could afford to consume more rice. Farm Sales The estimated results indicate that the average farm producer is much less responsive to changes in the prices of his products in sales decisions than in consumption. The difference in response between the two decisions is attributable to the fact that the farmer is an inventory operator as well as a seller and a consumer. The quantity of grain he decides not to consume is not necessarily sold in the same month, but can be held for future consumption or sales. The partial effect of a change in the price of rice on the market sale is measured by FS FK (. 160111-. 001043q ) ( . 160111-. 099753) .060358 aPR = 95.64 kilos) As already pointed out, the slope of sales function, i.e. , the farmerTs responsiveness to price changes, varies significantly depending on the size of grain stock he can afford to retain at home. The negative sign for the coefficient of the stock value implies that the farm seller tends 94 to reduce his market sales--or to put it differently, he tends to save in form of rice--whenever the price of rice goes up. If the net effect is measured by setting at its mean value, a one unit change in the price of rice induces the farmer to increase his market sales by .06 units. The mean elasticityof sales with respect to the price is given by FSFR F FS aPR 48 47 (.160111-.099753)( 9.01 = .324696 (.861317- .536621) 95. 74 kilos) Which means that a 10% increase in the price of rice will cause a 3. 2% increase in the market supply of rice. More specifically, when the price of rice rises by 10% the farmer is motivated to increase the market sales by approximately 8. 6% in order to acquire more of other. goods and services, but his feeling of uncertainty for future prices, combined with a desire to increase his family consumption due to improved savings position, has a counteracting effect of reducing his sales by about 5.4%, resulting in 3. 2% net increase in sales. This negative effect on the rice sales of changes in. stock value confirms with the positive effect on consumption of an increase in cash income. One of the most significant results of the partial analysis is that the average Korean farmer tends to increase family consumption and 95 to reduce market saleof rice, if the price of rice rises or is raised, but this tendency is not sufficiently strong enough, as against the expectation of some economists and policy formulators, to cancel out the farmer's desire to earn more cash which would enable him to 26/ purchase better clothing and send hLs children to school. Another explanation is that the need for cash income to purchase non- agricultural goods and services is so great that in spite of his strong preference toward rice compared to other grain, he is obliged to restrict his consumption but increase the market sales. As for the sale of barley, the net response to its price changes is given by FSU B B F FS aPB = (-.0039l3+.0046l6)() .012120 (FS = 1.91 kilos) which is smaller than for rice, indicating that the average Korean farmer is much less sensitive to change in the price of barley. One noticeable result is that a change in stock value due to a rise in the price of barley has a positive effect whereas it has a negative effect in the case of rice. This fact again indicates, as already 'The response of farmers to price changes is different depending on different seasons. The seasonal characteristics of behavioral responses are not presented here, but the empirical results based on the dummy response coefficient are given in Appendix B. noted in the demand equations, that the Korean farmer is partial to rice consumption. The increase in farm liabilities and cash expenditures has a positive effect on rice sales, as is expected, but they both have a negative effect on barley sales. These opposIte effects of cash demand (as proxy) are due to the fact that a major portion of cash needs is met by the sale of rice. The elasticities of rice sales with respect to farm debt and cash expenditure are 1. 64 and . 72 respectively, while those of the barley sales are -. 59 and -. i6 respectively. This gives an interesting sidelight on the possibility of increasing the market supply of rice through credit expansion. Another interesting finding is that the further away from the harvest season, the steeper is the slope of sales curve for both grains, indicating that the farm producers become less sensitive to market price conditions as their stock is drained. Contrary to this, the slope of the demand curve becomes greater during the non-harvest seasons, implying that the farmers respond to price changes more sensitively when their grain stock is reduced. Urban Demand. According to the estimated response coeffi- cients, urban consumers appear to be more responsive to a change in the price of rice than are farm consumers. . 27/ In respect to seasonal differences, See Appendix B. 97 The partial effects of an increase in the price of rice, the barley price and other variables remaining unchanged, upon rice and barley consumption are approximated by &q UD XS apu R = -(.443071-. 00006839Yu) = -(.443071-. 191150) = -. 251921 (at Yu = Yu 2795 won) or -(.443071-. 273560) = -. 169511 (at Yu 4000 Won) UD U 237600 3PR As indicated, the degree of response to price changes varies at different levels of income. As per capita income grows, price changes have less influence on the quantity of rice demanded. For example, at the mean level of income during the sample period, Yu 2795 won (unit of Korean currercy), a unit increase in the price of rice causes about . 25 units decrease in rice consumption, but at the higher level as 4000 won which is the average of the most recent two years, the estimated response is approximately . 17 units decrease On the other hand, a unit. rise in the price of rice tends to increase the consumption of barley by approximately . 24 units. The effect of barley price is measured by UD aq, apU B 073 949+. 00003 743 Yu) = -(.073949+. 104617) -. 178566 (at Yu Yu 2795 won) or -(, 073 949+. l49720 = -. 223669 (at Yu = 4000 won) a UD U B 105982 If the price of barley goes up by one unit, the urban consumer would decrease his barley consumption by . 18 units and instead increase rice consumption by . 11 units.'" Thus, it is obvious that the average urban consumer would respond to price changes in his rice and barley consumption and that he would substitute barley for rice or rice for barley whenever the relative prices of two grains change, if all other variables considered in our system stayed unchanged. The estimated elasticities and cross elasticities of demand with respect to the prices of rice and barley are as follows: Yu (all at Yu) -According to the criteria shown by H. Schultz (31, pp. 569582), rice and barley are substitute goQds in Korea for both urban and >0 and farm consumers, since B at >0 UDU U TJD aq = -1. 11173 48.70 won, -2. 32273 (P1 = 33.69 won, = 10.98 kilos) UD PU aPB UD = 2.59 kilos) IJDU U UD = .325153 = 33.69 won, = 4. 46688 = 48.70 won, 10.98 kilos) UDU U UD aPR = 2.59 kilos) The cross substitution effect between rice and barley due to relative price changes deserves special attention in the context of foodgrain price policy in Korea. The cross-elasticity of demand for barley with respect to the price of rice is 4. 5, while that for rice with respect to the price of barley is .33. The former is far greater than the latter, indicating that the average Korean urban consumer shows a higher responsiveness in barley consumption when the price of rice changes. From these facts we can hypothesize that there is a potential for increasing barley consumption and reducing rice consumption in urban areas, if the appropriate policy measures are taken. The effect of growth in per capita income on per capita rice 100 consumption is estimated by UD - aYu -. 002350 + 00006839R 00098 (at = -. 002350 + .00376 = .00141 (at = -. 002350 + .00333 = . or UU = 48.70 won) = 55.00 won) Which explains that at different levels of the price of rice, the consumer's response to income changes is different. The higher the price level, the greater is the response to income changes, given the initial income level. Using the elasticity concept, a 10% increase in income would cause approximately 2. 5% increase in consumption when the price of rice is 49 won per kilo, but at the higher price of 55 won per kilo the same percentage increase in income would cause approximately 3.9% increase. As the price of rice goes up, the consumer is forced to increase his expenditure on rice purchase if he wants to maintain the same level of consumption, or the same thing, he would become relatively poorer due to price rise so that he naturally becomes more susceptible to changes in market condition. As for barley consumption, the effect of income growth is given by: 101 UD B DYu 000295 B UU 33.69 won) = .000295 .001260 = -.001001 (at = .000295 .001497 = -.001202 (at P = 40.00 won) or B 'B The income effect is negative both at the mean and the recent level of barley prices and the responsiveness to income changes becomes greater as the price of barley goes up (same as in rice consumption) but in negative direction. The estimated income elastici- ties are -1.04 at the price level of 34.00 won and -1.39 at 40.00 won. The positive response in rice consumption and negative response in barley consumption indicate that rice is regarded as normal goods while barley as inferior goods by the average urban consumers. One important analysis can be made of the partial effect upon rice consumption of both price and income changes shown previously. By setting the following derivative at zero, we can approximately estimate the potential income level at which the rice consumption of the average urban consumer would not respond at all to changes in the prices of rice, provided that the present consumption pattern continues. In other words, when such a level of income is attained, the consumer would reach the "satiety" point in terms of rice consumption. 102 UD * = -(.44307 1-. 00006839Yu) = 0 ap Y = 6479 won which is about 1. 5 times as high as the current level of monthly income. A rise in average per capita income beyond this level would rather cause a reduction in rice consumption and might accelerate a shift to other high quality foods, such as meats and processed foods, even if the price of rice falls. We can also conceive of a certain price level which is so low that the average urban consumer would be able to consume as much as he wants with the present income, that is, the 'satiety'T point in terms of price. UD R (00235000006839U) R 8Yu 0 34 . 36 won per kilo The derivation of the potential "satiety' point is not only of theoretical interest. The knowledge of such a boundary level of income and price provide a landmark on the basis of which to judge the degree of urban consumers' preferences toward rice. It is evident that there still exists a possibility that the urban demand for rice will continue to increase in the future as per capita income grows. 103 in the event that per capita income reaches such a level and at the same time the supply of rice is such that its price is as low as 34. 00 won per kilo, then the growth of population would be the only factor for increasing the demand for rice. Analysis of Total Response The preceding analysis of the partial responses on the basis of each individual behavioral equation was possible only under the assumption that other variables and relationships hold constant. But once everything is left to be determined in the free market other things cannot remain constant. When the price of rice changes, for example, due to some external cause or by some kind of government manipulation, the price of barley, farm level demand, sales and urban demand will also change. Therefore, it is necessary to devise a scheme to measure the price-quantity relationship when interactions among the endogenous variables that occur in the market structure are taken into account. Such a measurement will enable us to approxi- mate the simultaneous responsiveness of farm demand, sales and urban demand for rice and barley to a change, for example, in the price of rice, the price of barley allowed to vary correspondingly as the market system requires to reach a new equilibrium level. Since the main focus of the study is on the effects of the prices of rice and barley in the context of foodgrain price policy, the effects 104 of other exogenous variables are not analyzed here. The total response to a change in the price of rice for example, on' the farm demand for rice is estimated as follows;" Let = FS ES-- ESPB+YNRB) ES P+Pq ES+qES Then, from the farm demand Equation (3.3), we obtain FD ED ED + dPR dPB 8PB dPR ED aciR + dYF aYE dPR The process of the total response to a change in can be decomposed into three different component behavioral phenomena according to the way in which the farm demand function was specified. The first term aql)/aP represents the direct price effect which induces the consumer to change his consumption, the second term the repercussion due to simultaneous changes in the price of barley, and the third term the effect of changes in monetary income. term contains three endogenous variables The and qFS (other than qFS which are supposed to vary when Calculation is first based on the estimated coefficients for wholesale prices, R and B' in each equation, and at final stage those corrected for farm and urban consumer prices are used. 105 R changes. Hence, dYF FS dPR FS + R aPR + FS FS dPB dPR + dPB B aPB dPR From the sales equation we can obtain FS RS and aPB The procedure requires that the price of rice, exogenous variable, since we change R R' be treated as an arbitrarily to measure its effect. This implies now that the identity Equation (3.9) which repre- sents the equilibrium condition in the rice market no longer holds. In order to obtain dPB/dPR we need to express A reduced form equation for B B as a function of as a function of can be obtained by substituting the barley sales Equation (3.6) and urban demand equation for barley (3. 8) into the barley market identity Equation (3. 10). The dPB dPR dP dP thus derived is R .281072- 000091qK1 .074907 + 00003982Yu+ . OOOO91q' !QiStrictly speaking, dPB/dPR should be written as since other exogenous variables such as P' Li and E are held constant. But as they are regarded as given at the point in time, the total derivative seems relevant. 106 Note that the rate of change in the price of barley due to a unit change in the price of rice, assuming other exogenous variables do not change, is a function of Yu, FK q1 and FK which are exogenous variables in the system. More specifically the higher the level of income and the larger the stock of barley and rice, the smaller is the influence of a change in the price of rice on the equi- librium price of barley. For analytical convenience, we will hold these three exogenous variables at some constant level. For the size of rice and barley stock, the mean values during the sample period were used. For the / per capita urban income the average of the most recent three years 31/ was used. The estimated result is dPB 272369 dPR .238803 = 1. 140559 Since the monthly variation of the grainstock is repeated every year with almost similar pattern, though slightly different in quantity depending on the crop size, the mean value appears to be reasonable. But the urban per capita income has been steadily increasing in the future. Therefore, the mean value of the most recent years seems more adequate. The values used are: FK 95.64 kilos 50.72 kilos Yu 4000 won 107 We are faced with similar problems in obtaining dYF/dPR; this time with endogenous variables FS FS and Again by the same procedures, we set these variables at mean values. Since the dqlJ/dY is as small as .0009135, the computed result is not much affected by the term (aq/aY)(dY/dP) as shown by the following: a FD = -.354141 aq FD F dPB = (.328850)(L 140559) = .375073 F aPB dPR a q FS = .160111 - (.001043)(95.64)= .060358 FS -.003913 + (.000091)(5072) = .000703 aPB aq FD dYF FdP = .000913519. 01+(44. 81)(. 060358)+(1. 91)(1. 140559) + (30.45)(. 000703)(1. 140559)] = .01274 Therefore, FD dP'R = -.354141 + .375073 + .012714 = .033646 The result is opposite to that estimated in the partial analyses above. Our estimates in the partial analyses showed that a unit increase in the price of rice was supposed to cause a reduction of rice consump- tion by .35 units. However, by dropping the ceteris paribus assumption, a unit increase in the price of rice has the effect of increasing rice consumption at the farm level, even though by small quantity. The result apparently seems to confirm the hypothesis of some economists. Although it is true that an increase in cash income has a positive effect upon rice consumption, a dominant reason seems to lie elsewhere. A careful examination of the joint-dependence between the price of rice and barley will clarify this point. Let us decompose the total response into three analytically different effects, each caused by a unit increase in the price of rice. 1) The consumer tends to reduce his rice consumption by . 35 units and shift to other grain. 2) But a simultaneous increase in the price of barley forces the consumer to shift back to rice consumption by .38 units, more than offsetting the intended substitution. 3) A simultaneous increase in money income also causes an increase in rice consumption by . 013 units. Thus, the net result is a slight increase in total rice cons ump- tion. We must make one point clear, that is, the positive net response 109 is not because of an increase in money income. It is because of the dynamic price-to-price repercussion that occurs whenever either one of the prices changes. The following are the total responses of farm demand, farm sales and urban demand for both rice and barley to a change in the price of rice, estimated by the same procedures. As in the case of the farm demand for rice, the first term represents the direct effect of P, the second the repercussion due to simultaneous changes in and the third term represents the effect of change in money income caused by a change in P. dq FD F dPR dq = -.354141 + .375073 + .012714 .033646 FD .372224 - . 133505 - 040950 = . 197769 . dP dq FS = . 160111 .055656 - . 160059 = -.055604 dP FS -.034915- .004461 + .042954 = .003578 dPR UD -.443071 + .120819 + .273560 = -.048692 dP (repeated) 110 dq,UD = . 237600 - .084302 - . 149720 = .003578 dP The estimated total demand and sales response have entirely different meaning from the ordinary concept of demand and supply in the sense that interaction among different variables is taken into account. The ordinary concepts of demand and supply as defined in the usual texts are based on the strict qualification in the form of ceteris paribus assumptions which do not necessarily hold in the market situation. The total response coefficients given in the above, therefore, must be interpreted as a slope of a locus of equilibrium points after allowing all other variables (endogenous) to reach equilibrium at the same time. The slopes of farm and urban demar1 curves in the ordinary sense are each represented by the first term in and UD U IdPR FD F /dPR respectively,which are negative as expected. And . that of sales is represented by the first term in FS F /dPR which is positive as expected in the ordinary market supply curve. For example, the positive total response in farm demand should not be misjudged as the so-called Giffen goods case. The cross substitution effect in barley demand with respect to the price of rice was estimated as .35 in the partial analysis, but it is only . 20 in the total response analysis. This is mainly due to the 111 "return shift" caused by the increase in theprice of barley. Even if the consumer intends to shift to barley consumption in the first round because of a rise in the price of rice, a simultaneous rise in the price of barley forces him to return to rice consumption. The resulting increase in barley consumption is thus smaller than originally intended. As for rice sales, the total response is negative in spite of the positive slope of the sales curve itself, as shown in the partial analy- sis. The effect of an increase in the value of rice stock combined with the repercussion that occurs inthe barley market seems to offset the initial intention to sell more rice on the market. A similar explanation can be given in regard to barley sales and urban demand for rice and barley. Changes in income level and changes in barley prices both tend to offset the direct effect of a change in the price of rice. It is notable that the quantity of barley demanded increases negligibly when the price of rice rises, in spite of the fact that rice and barley are substitutes according to the conventional (or partial) analysis. The total response to changes in the price of barley of each variable is computed by the same procedure. This time we need to treat the price of barley as an exogenous variable which can be varied arbitrarily. The rate of change in the price of barley dPR/dPB obtained by substituting the rice sales Equation (3. 5) and urban 112 demand for rice Equation (3. 7) into the rice market identity (3.9) and expressing as a function of R The computed total responses are: O0l043q" 001043FK 6l5485-. 00006839Yu-. dPR . dPB 159682+. (at Yu = 4000 won) FD .328850 - .310864 + .013392 .004594 dP dq FD -.117110 + .326738 - .042727 = .166901 dPB dq FS = -.048812 + . 140545 - 134549 = -.042816 dP dq FS -.003913 - .030648 + .014604 = -.019957 dPB UD = . 105982 - .388928 + .240130 = -.042816 dP UD = -.073949 + .208565 - .149720 = -.015104 dP 113 The results are quite different and opposite in many cases to what would have been expected on the basis of partial analysis The reasons for these phenomena are about the same as already discussed concerning the effect of the price of rice. The cross substitution of rice for barley due to a rise in the price of barley is very small in farm demand and negative in urban demand whereas it is appreciably positive in the ordinary analysis. The quantity of barley demanded by farm consumers moves in the same direction as the price of barley while that by urban consumers responds negatively though by small amount. One is again likely to think that barley is a Giffen good for farmers. It was shown that barley is an inferior good but not a Giffen good in the ordinary sense, as is clearly indicated by the negative sign of the first term in each of the demand response equation. The reason for the positive slope of total demand response at farm level lies again in the dynamic price-to-price relation which is liable to change whenever either one of the prices of the two grains changes. To use the elasticity concept a. 10% increase in the price of rice causes a 17% increase in the price of barley while a 10% rise in causes a 6% increase in The estimated partial responses and total responses to both prices are summarized in the following table. 114 Table 3-2. The coefficients of responses to price changes; partial versus total.!' Equation Farm demand for rice (3.3) Farm demand for barley (3.4) Farm sales of rice (3.5) Farm sales of barley (3.6) Urban demand for rice (3.7) Urban demand for barley (3.8) Barley price Response to Changes in Total Response Partial Response Mean Direct Mean Direct Coefficient Elasticity Coefficient Elasticity -.346 -1.538 .034 .151 .346 2.497 .198 .143 .060 .325 -.056 -.278 -.026 - .187 -.170 - .719 4.484 .238 b/ .004-.049 .004-hi 1. 141 .094 -.207 .072 1.677 Response to Changes in Farm demand for rice (3.3) Farm demand for barley (3.4) Farm sales of rice (3.5) Farm sales of barley (3.6) Urban demand for rice (3.7) Urban demand for barley (3.8) Rice price 1.000 .331 -. 173 - .848 -.100 - .338 .001 .012 .106 .325 -.224 -2.937 .005 .015 167 .819 -.043-b/ -.145 . 020 .319 -.043 b/ -.132 -.015 -.197 . . 878 . 597 All calculations are based on the mean values for q1, q1, 50.72kilos, and Yu at 4000 won. 95.64 and qFK The same magnitude of the total response coefficients in the farm sales of and urban demand for barley is due to the fact that the barley market is set at equilibrium while the rice market is not. The same is the case for the rice sales and urban demand for rice. 115 Summary of Findings The following is a summary of the major findings in this chapter. 1) Overall, the analysis based on the total response approach which takes into account the simultaneous changes in all endogenous variables demonstrates a result which is entirely different from that based on the ordinary (or partial) approach under the usual ceteris paribus assumption. It was shown that theresults obtained by the formermethod reflect the quantity-price relationship in the real market condition. Z) The demand for rice at the farm level is supposed to be highly responsive to a change in the price of rice when the price of barley is fixed. But if the price of barley is allowed to change at the same time as the price of rice, the average farmer tends to increaserice consumption becauseof the Hreturn shift" caused by a simultaneous rise in the price of barley. An arbitrary change in the price of rice by 10% causes a higher rate of increase in the price of barley; approximately 17%. The cross substitution elasticity of rice for barley consumption due to a change in the price of barley is unitary if 116 the price of rice were fixed, indicating that a 10% increase in the price of barley induces farm consumers to increase rice consumption by 10%, but when the price of rice is allowed to change, there is almost no substitution effect between the two grains. 3) The demand for barley at the farm level is also negatively related to its price if the rice price were fixed, but the total response is positive in the real situation due to a simultaneous change in the price of rice. The rate of shift to barley consumption, when the price of rice rises, is much smaller than that widely believed under the other-thingsremain-constant assumption, because of a rise in the price of barley at the same time. 4) The farm sales of rice shows a positive response in the partial analysis, but negative response in the total response analysis. The farm sales of barley follows the same pattern as in the case of rice. 5) Urban consumers respond negatively to a change in the price of rice in both the partial and total response analyses, but the degree of responsiveness is substantially offset by a change in the price of barley that occurs simultaneously with the price of rice. The substitution effect between rice and barley is either negative or almost negligible against the 117 general expectation based on the partial analysis. One noti.''eable fact is that, given the present consumption pattern, the average urban consumer would show no response at all when his income (real) reaches as high as 6800 won which is approximately one and a half times the present level. In the event that this situation arrives in the future, the growth of per capita income alone would have no effect upon the rice consumption and increases in income would be absorbed in the consumption of higher quality foods or nongrain commodities. 6) Although the results of the total response analysis indicate that farmers increase their consumption of rice and reduce their market sales if the. price of rice rises, the primary cause for this does not Li.e in the increase in cash income or stock value resulting from the price rise as hypothesized by a number of economists. The empirical results show that an increase in cash income due to a rise in the price of rice does have the effect of increasing family consumption, but this effect is not strong enough to offset the farmer's desire to shift to consumption of other grain. It is also shown that an increase in the value of grain stock due to a rise in the market price tendsto induce the farmer to increase his saving in kind 118 and reduce the market sale, but this effect alone does not cause a reduction in sales. The net decrease in sales is attributable mainly to the pr ice-to- price correspondence that takes place via the market mechanism. 119 POLICY IMPLICATIONS From the foregoing empirical analysis, one can draw the following policy implications which are addressed to improving the present foodgrain price policy in Korea. Problems related to institutional or detailed operational aspects are not discussed. The evidence is clear that Korean farmers respond in their grain production to changes in relative product price and they tend to be more responsive in barley cultivation than in rice as already stated. This implies a possibility that the price raising measures for barley would contribute relatively more to achieving increased foodgrain production. A price policy which aims at increased production of major foodgrains may have some conflicting effects on the production of other agricultural products, especially cash crops, due to the competition in land use and family labor input. Such competition may be relatively greater in common barley cultivation as most cash crops are grown upland (dry land). An increase in foodgrain production is not costless; nor can it be achieved automatically. In Korea, where it is imperative to step up the rate of growth in foodgrain production, a certain amount of expense is inevitable. Since the empirical measurement of farmers' response is based on the average iarm household data for three major foodgrains, the 120 relatively high elasticities of production should not lead one to judge that the production responses are high for all situations and for all farmers, or that the elasticity for aggregate agricultural production is also high. As already pointed out, the production response differs by different region and by different individuals. The relatively more commercialized farmers will in general show a higher response to changes in price relationships. Also farmers living in relatively more progressive areas will be more-responsive. These differences in response have important policy implication considering the limited government resources. In many cases, it is more effective to con- centrate on certain areas or certain classes of farmers, for example in input distribution or farm credit program, rather than a little-bitfor-everyone approach. From the above-reported results one might conclude that a decrease in input price such as by increased subsidy, with product price remaining unchanged, would have the same effect as an increase in product price, leaving input price constant, as long as the result- ing price ratio is the same. This may be only partially true but not in general for foodgrain production. In agriculture where non- purchased inputs constitute a large portion of the total cost of produc- tion, a government subsidyon purchased inputs may not affect the relative price sufficiently to influence farm decisions. Since most farmers are more familiar with product price than input price, 121 especially when new inputs appear on the market, farmers are likely to be more sensitive to changes in grain prices. Grain price-raising measures will be more effective in increasing the use of, for example, family labor, assuming the existence of substantial amount of idle labor in rural areas. One of the drawbacks in using product price policy is that the marginal or submarginal farmers may spend extra income on consumption items rather than for improved cultivation. But this kind of phenomena is limited to only small farmers whose products do not add much to market supply. The aggregate effect is still an increase in production and market supply. As agricultural development proceeds, the purchased inputs will of course assume an increasing importance. For this reason, input price policy (subsidy policy) should be viewed in a long-run context. Where an achievement of increased foodgrain production is viewed as an immediate goal, more emphasis should beplaced upon foodgrain price policy, though some combination with input price policy is necessary. The results of the study shows that the higher price of rice induces the average Korean farmers to increase family consumption and reduce market sales when total responses on the open market are taken into account. Superficially, it appears to confirm the Utarget cash requirementT' hypothesis. But the net result of increasing con- sumption and decreasing sales are only partially associated with the increase in money income due to the higher price they receive. 122 Although marginal, or submarginal, farmers who are in the underconsumption status would show such reaction when the price of rice rises, it is not the general case in Korea. A knowledge of the real causes for an increase in family con- sumption and resulting decrease in marketings is crucial from the standpoint of policy formulation. It makes a great deal of difference whether this reaction of farmers is due to an increase in money income brought by a price rise or itjg attributable to other causes, even if the result is exactly the same amount of reduction in marketings, If a reduction in market sales of rice were caused by an increase in money income, thefl price-raising policy would serve to foster the aggregate foodgrain production, but it would not help in alleviating the overall market shortage because the additional output would be absorbed by increased family consumption. It was found in the analysis of consumers' responses to price changes, however, that both farm and urban consumers have a "potential" tendency to respond negatively to price changes in their consumption of rice and barley and that they have a "potential" tend- ency to substitute rice for barley or barley for rice whenever either one of the prices changes, but that this negative response to price changes and the tendency toward cross substitution between two grains are offset by the simultaneous change in the price of the other grain on the open market, resulting in positive changes in the quantity 123 demanded, or no substitution at all. The fact that such potentials exist, though they are not actually displayed, provides an optimistic view in respect to the effectiveness of price policy in restructuring foodgrain consumption. If the government's objective is to reduce foreign exchange spending on rice imports by reducing rice consumption and increasing the consump- tion of barley or other grain, it can be done, through the use of price incentives, by inducing the consumers to reveal their potential responses on the market. This is equivalent to forcing the ceteris paribus assumption, which is made in the partial analysis, to operate in the real world, that is, the price relationship which leads to the results obtained by the partial analysis must be maintained by the government operation. More specifically, a policy designed to raise rice prices and lower barley prices to consumers would decrease rice consumption and increase barley consumption, and at the same time higher prices for both grains at harvest time would stimulate increased production of both grains, especially of barley. The actual operation of this policy requires a two price system for barley as well as some degree of manipulation of the price of rice. In order for such a pricing policy to be effective, a sufficient stock of grain should be held by the government. The operation of a two price system for barley will create a considerable amount of loss in the government budget. But if the 124 government sells rice in the urban area at higher prices, it may be able to generate a substantial amount of net revenue with which a loss in barley operation can be compensated When the government used low price policy for rice in the urban area in the past the government cost was borne by farm producers and the urban tax payers. The higher rice price policy will shift this burden directly to consumers so far as rice consumption is concerned, but if the net revenue generated by such operation is plowed back into barley stock manage- ment, there will be little or no burden in aggregate on the urban consumers. Individual consumers who still prefer to consume rice even at higher price will bear the cost. There is a movement within the Korean government to reform price policy in this direction. Although the adoption of such a policy is a step in the right direction which is consistent with the results of empirical analysis in this study, determination of the price level for rice and barley still remains as the problem. The instrument of price policy can be successfully used in achieving the objectives only when the prices are determined at the appropriate level. Finally, the impact of wheat flour prices deserves special attention in formulating foodgrain price policy although the demand situa- tion for wheat flour was not analyzed in detail due to a lack of data As was already pointed out, the government policy to keep the price of wheat flour at low level (relative to the prices of rice and barley) 125 caused a rapid increase in its consumption. There is no doubt that such a low price policy served to reduce rice as well as barley con32/ sumption. As discussed earlier, however, the relative price of wheat flour has been lowered to such a level that further decrease in its price would cause an increase in rice consumption, whereas it would cause a decline in barley consumption. The policy based on a simple judgment that a lower price for imported wheat will continue to serve to reduce rice consumption does not seem to be workable in all situations any longer. If the govern- ment is determined to reform its foodgrain price policy with a view to increasing barley production as well as increasing its consumption in the urban area, a higher sales price for imported wheat flour would better serve this purpose because it may promote ashift to barley consumption and reduce rice consumption. The consequence of such a policy would also enable the government to create a considerable amount of revenue which in turn can be injected to cover the cost incurred for implementing two-price system for barley. Of course, a change in this direction must be a step-wise one in relation to the The overall assessment of the effect of imported wheat and that of low price policy requires an extensive research. The discussion here is confined only to its impact in conjunction with the prices of rice and barley. 126 rate of growth in rice and barley production In any case, the pre- sent ceiling price system for imported wheat should be reconsidered. 127 BIBLIOGRAPHY 1. Bauer, P.T. and Yamey, B.S. ACaseStudyofResponseto Price in an Under-Developed Country. Economic Journal 69:800-805, Dec. 1959. 2. Brown, W. G. and Nawas, F. Improving the Estimation and Specification of Outdoor Recreation Demand Functions. Corvallis, Oregon, Agricultural Experiment Station Project 850, 1971. 3. Buse, R. C. Total Elasticities A Predictive Device. Journal of Farm Economics 40:881-891. Nov. 1958. 4. Christ, C. F. Econometric Models and Methods. New York, John Wiley, 1966. pp. 292-294, pp. 314-331, pp. 564-575. 5. Aggregate Econometric Models: A Review Article. The American Economic Review 46:385-408. June 1956. 6. Dubey, V. The Marketed Agricultural Surplus and Economic Growth in Underdeveloped Countries. Economic Journal 73:689702. Dec. 1963. 7. Freund, R. 3. Some Observations on Regressions with Grouped Data. The American Statistician 25:29-30. June 1971. 8. Friedman, M. Price Theory; A Provisional Text. Chicago, Aldine, 1968. pp. 48-55. 9. The Marshallion Demand Curve in Essays. In: Positive Economics. Chicago, The University of Chicago Press, 1966. pp. 47-99. 10. Girshick, M. A. and Haavelmo, T. Statistical Analysis of the Demand for Food; Examples of Simultaneous Estimation of Structural Equations. Econometrica 15:79-110. April 1947. 11. Grilliches, Zvi. The Demand for Inputs in Agricultural and a Derived Supply Elasticity. Journal of Farm Economics 4 1:309322. May 1959. 128 12. Heady, E.O. Uses and Concepts inSupplyAnalysis. In: Agricultural Supply Functions - Estimating Technique and Interpretation, ed. by E.O. He.ady and others, Ames, Iowa State Univ. Press, 1961. pp. 3-25. 13. Heady, E.O. and Tweeten, L.G. Resource Demand and Structure of the Agricultural Industry. Ames, Iowa State University Press, 1963. pp. 433-435. 14. Henderson, J.M. and Quandt, R.E. Microeconomic Theory. New York, McGraw-Hill, 1958. pp. 24-30. 15. Hooper, J. W. Simultaneous Equations and Canonical Correlation Theory. Econometrica 30:324-331. April 1959. 16. Johnston, J. Econometric Methods. New York, McGraw-Hill, 1972. pp. 352-356. 17. Klein, L.R. A Textbook of Econometrics. Evanston, Ill., Row, 1953. pp. 120-121. 18. Korea, Government of. Bank of Korea, Monthly Economic Statistics, 1963-71. Bank of Korea, Price Statistics Summary, 19. 1963 -71. Bureau of Statistics, Monthly Statistical 20. Review, 1963-71. Bureau of Statistics, Urban Living Expenditure 21. Survey, 1963-71. 22. Ministry of Agriculture and Forestry, Agricultural Statistics Year Book, 1963-7 1. 23. Ministry of Agricu1tcre and Forestry, Foodgrain Consumption Survey, 1963-71. 24. Ministry of Agriculture and Forestry, The Farm Household Economy and Cost of Production Survey, 196371. 25. Korea, National Federation of Agricultural Coops. Monthly Review, 1963-71. 129 26. Korea, National Federation of Agricultural Coops. Rural Prices and Wage Survey, 1963-71. 27. Krishna, R. Agricultural Price Policy and Economic Development. In: Agricultural Development and Economic Growth, ed. by H. M. Southworth and B. F. Bruce, New York, Cornell University Press, 1967. pp. 540. 28. Mathur, P.N. andEzekiel, H. SurplusofFoodandPrice Fluctuations. Kykios 14:396-408. 1961. 29. Mellor, J. W. The Economics of Agricultural Development. New York, Cornell University Press, 1966. pp. 196-213. 30. Nerlove, M. and Addison, W. Statistical Estimation of Longrun Elasticities of Supply and Demand. Journal of Farm Eco- nomics 40:861-880. Nov. 1958. 31. Schultz, H. Theory and Measurement of Demand. Chicago, The University of Chicago Press, 1938. pp. 565-582. 32. Schultz, T.W. Transforming Traditional Agriculture. New Haven, Yale Univ. Press, 1964. pp. 162-174. 33. Theil, H. Economic Forecasts and Policy. Amsterdam, North-. Holland, 1961. pp. 56-58. 34. Tweeten, L.G. and Quance, C,L. Positivistic Measures of Aggregate Supply Elasticities: Some New Approaches, The American Economic Review 59:175-183, May 1969. 35. IJSAID/Korea. Foodgrain Demand and Supply Projection for 1967-1971. Mimeograph. 36. Vickrey, W.S. Microstatics. New York, Harcourt, Brace and World, 1964. pp. 59-66. APPENDICES 130 APPENDIX A Estimation of Foodgraiii Supply and Demand Footnotes to Table A-i. aRY 1967 beginning stock as of November 1, 1966 estimated as 200,000 MT and those for RY 1967-70 based on net change in stock during the year and those for 1971-76 assumed to be constant. bActual production adjusted to Rice-Year for RY 1967-71 and those for RY 1972-76 projected based on linear trend (1960-71). imports on the arrival basis for RY 1964-70 and those for RY 71-76 are shortages based on total distribution minus production and beginning stocks. dEstimation for RY 196 7-71 based on daily per capita consumption data surveyed by Ministry of Agriculture and Forestry (MAF) and those for Population growth rate: urban 2 8%, farmer 1. 4%; Per capita income growth 1972-76 based on population growth and income growth rate (1960-71): urban 5. 5%, farmer 2.8%; Income-elasticity of consumption (1960-71): urban 0. 35; farmer 0. 78. effects eimation for 1967-71 based on the residual (total distribution minus staple food use and feed/seed/waste) and those for 1972-76 assumed to be constant at 100,000 MT. lncludes all uses on farm, and estimating method same as d. use based on MAF plan and waste estimated as 3% of total production. hEstimated as carryover of "old graixl. J)es not include fall harvest, the total of which is included in production. Data Source: I Agricultural Statistics Yearbook, 1963-1971, Ministry of Agriculture and Foresty 2 Grain Consumption Survey, 1963-1971, Ministry of Agriculture and Forestry 3 Monthly Statistical Review, 1963-1971, Bureau of Statistics. Table A-i. Estimation of rice demand and supply. 1. II. Unit: 1,000 MT RY 67 RY 68 RY 69 RY 70 RY 71 RY 72 RY 73 RY 74 RY 75 RY 76 200 452 239 140 450 450 450 450 450 450 3,919 3,603 3,195 4,090 2,939 3,998 4,190 4,254 4,378 4,472 113 216 760 540 535 629 596 666 791 824 4,232 4,271 4,194 4,770 4,924 5,077 5,236 5,400 5,569 5,796 3,780 4,032 3,899 4,320 4,474 4,627 4,786 4,950 5,119 5,296 1,717 1,904 1,834 2,014 2,079 2,146 2,216 2,288 2,362 1608 1,704 1,806 1,914 1,979 2,046 2,416 2,188 ?262 2440 2340 109 200 28 100 100 100 100 100 100 100 1,910 1,986 2,065 2,148 2,234 2,317 2,403 2,494 2,584 2,680 153 142 155 158 161 164 167 170 173 176 452 239 140 450 450 450 450 450 450 450 4,232 4,271 4,194 2,770 4,924 5,077 5,236 5,400 5,569 5,746 Supply A. Beginning stocka B. Production c. ImportsC D. Total supply Distribution A. Consumption 1. Commercial a. Food use d b. Non-staple 2. On farm 3. B. Exports C. Ending stock D. Total distribution fOOde Data Source: 1 Agricultural Statistics Yearbook, 1963-1971, Ministry of Agriculture and Forestry 2 Grain Consumption Survey, 1963-1971, Ministry of Agriculture and Forestry 3 Monthly Statistical Review, 1963-1971, Bureau of Statistics. use based on MAF plan and waste estimated as 3% of total production. hFor RY 70-76 surplus represents the amount necessary to balance total distribution with total supply. 1This carry-in appears high in view of the short barley crop in 1963. No reasonable basis is available for making adjustments. 1Estimating methods same as d. '1Estimation for RY 1967-71 based on daily per capita consumption data surveyed by Ministry of Agriculture and Forestry (MAF) and those for 2.8%, farm 1.4%; Per capita income growth RY 1968-76 based on popuiation growth and income growth effect: Population growth rate: urban (-)0. 925, farm (-)0. 560. rate (1960-71): urban 5. 5%, farm 2, 8%; Income-elasticity of consumption: urban This item reflects any errors eEstjmation based on residual (total consumption minus staple food use and feed/seed/waste) for RY 67 to RY 70. of estimation in production and other consumption categories during this period. imports on the arrival basis for RY 1964-70. year and those for RY 1971 to 1976 assumed to be constant. bActual production estimates adjusted to Rice-Year for RY 1967-71 and those for RY 1972-76 projected based on linear trend (1960-1971). based on net change in stock during the aRY 1967 beginning stock as of November 1, 1966 estimated as 850,000 MT and those for RY 1967-70 Footnotes for Table A-2. Table A-2. Estimation of barley dmand and supply 1. II. Unit: 1,000 MT RY 67 RY 68 RY 69 RY 70 RY 71 RY 72 RY 73 850 853 914 973 913 913 913 2,002 1,940 2,037 2,045 1,917 2,375 105 64 - - 2,852 2,898 3,015 3,018 1,999 1,984 2,042 Commercial 820 790 a. Food use 202 RY 74 RY 75 RY 76 913 913 913 2,470 2,565 2,660 2,755 - - - - - 2,830 3,288 3,383 3,478 3,573 3,668 2,105 2,166 2,226 2,287 2,349 2,413 2r478 834 878 922 966 1,010 1,054 1,098 1,142 199 196 193 190 187 184 181 178 175 618 591 638 685 732 779 826 873 920 967 987 998 1,009 1,020 1,031 1,040 1,049 1,058 1,068 1,078: 192 196 199 207 213 220 228 237 247 258 Supply A. Beginning stocka 13. Production c. ImportsC D. Total supply - Distribution A. Consumption 1. b. Non-staple 2. Onfarxn 3. fOOde B. Exports (surplus - - - (84) (114) (149) (183) (216) (247) (277) C. Ending stock 853 914 973 913 913 913 913 913 913 913 D. Total distribution 2,852 2,898 3,015 3,018 2,830 3,288 3,383 3,478 3,573 3,668 Footnotes for Table A-3. aRY 1967 beginning stock as of November 1, 1966 estimated as 200,000 MT and those for RY 1967-70 based on net change in stock during the year and those for RY 1971-76 assumed to be constant. bActual production estimates adjusted to Rice-Year br RY 1967-71 and those for RY 1972-76 projected based on linear trend. cActual imports on the arrival basis for RY 1967-70 and those for RY 1971-76 are differences between total distribution and supply. dEstimation for RY 196 7-71 based on daily per capita consumption data surveyed by the Ministry of Agriculture and Forestry (MAF) and those for RY 196 8-76 based on population growth and income effect. Income elasticity of consumption: 0. 24. eEstimation for RY 1967-71 based on the residual (total consumption minus staple food use and feed/seed/waste) and those for RY 1972 -76 based on 12. 3% average annual increasing rate for 1960-67. 1A11 domestic production assumed to be consumed on farm, in addition to imported wheat as food use which is included in "commercial food Increases in consumption on farm for RY 1968 and 1969 are due to relief program in drought-stricken area. use based on MAF estimates, waste estimated as 3% of production and 0.3% for imported wheat. Data Source: 1 Agricultural Statistics Yearbook, 1963-1971, Ministry of Agriculture and Foresty 2 Grain Consumption Survey, 1963-1971, Ministry of Agriculture and Forestry 3 Monthly Statistical Review, 1963-1971, Bureau of Statistics. uses Unit: 1,000 MT Appendix Table A-3. Estimation of wheat demand and supply. I. II. RY 67 RY 68 RY 69 RY 70 RY 71 RY 72 RY 73 RY 74 RY 75 RY 76 pply A. Beginning 200 400 391 423 475 475 475 475 475 475 B. Production 310 324 356 360 352 361 370 380 389 398 c. IniportsC 922 1,127 1,205 1,254 1,464 1,609 1,769 1.946 2,142 1) Total supply 1,432 1,851 1,952 2,037 2,160 2,300 2,454 2,624 2,810 3015 1,032 1,460 1,529 1,562 1,685 1,825 1,979 2,149 2,335 2540 Commercial 660 959 1,041 1,139 1,244 1,367 1,504 1,656 1,825 2,013 a, Food use 359 372 382 399 413 434 456 479 503 528 301 587 659 740 831 933 1,048 1,177 1,322 1,485 310 433 414 342 352 361 370 380 389 398 62 68 74 81 89 97 105 113 121 129 400 391 423 475 475 475 475 475 475 475 1,432 1,851 1,952 2,037 2,160 2,300 2,454 2,624 2,810 3,015 Distribution A. Consumption 1. b. 2. Non-staple foodC On farm 3. B. Exports C. Ending stock D. Total distribution (J.) 134 APPENDIX B-i Estimated Seasonal Response Modela! (Farm Demand) qFD = 19.71818 - . 37891P (4.429) (4.634) R + * *** + .09112D2P . 19285D1P R R (3.317) (1.515) (NS) (r'4S) + .28441P + .00504q1 + .00O4l(P (6.324) B (1.290) (.585) ** FD - 2. 29762D2 - . 00584P (.799) (1.832) B (NS) (NS) . + . 01967P *** (. 283) R + (R +YNRB) 2 = . 815) (NS) = 4.4200 - 02975P (.854) (.369) B (NS) FS * (NS) - 5. 8340ZD (2,095) q FS - . O178OD1PB + (.431) 0416;qFK Bt-1 (6. 117) 2639lD3P (5. 995) *** 00199(P qFSp BB (4 261) R R . (NS) .08842D - 5.42871D + .00194P (.063) 1 (3.786) (.535) NRB (NS) + .46364D (.444) G (R2 = . 866) 135 (Farm Sales) FS * = -16.42861 + . 10343P (3.966) (1.401) + .00662L (NS) (NS) R - .06336D P (.638) + .01020E + 1 R + .02841D2P R (294) .07998q<1 1 (7.306) (4.963) ** - . OOO7Z(Pq FK 1+Pq FK (5.462) (NS) l+YNRB) + 5. 86913D (2.088) 1 (1.116) (NS) 1.15583D2 (R2 . 855) (.244) (NS) (NS) (NS) (NS) FS = .56891 + .01763P - .00508D P - .01446D3P q B B 1 B B (.534) (.508) (.493) (.175) *** FTC 00479E + . 03720q B- (NS) + . 00023L 1 (.669) (5.482) (9.284) ';** + . OOOl8(Pq FK 1+Pq FK l+YNRB) (3.408) (NS) . 03374D (.350) (NS) 1 (NS) + 1. 21469D2 - 07429D G (1.217) (1.531) . 2 (R . 802) 136 (Urban Demand) qUD R = * ** 28. 84979 - . * ** 40033P (5.224) R (7.729) -F .05579P (1.166) B + *** 13043P (3.060) . 85843D (1.098) .00172Y (1.898) U - .00663P (2.859) (R2 = .770) * - . 03494DP B (1.523) 000019PBYU (.784) R 10288D2P + (4.380) * (NS) . - 00l35Y (1.773) U . (NS) - 00232P (.911) . *** (NS) + 67l35D P (1.901) 2 R . (1.484) * *** 5. 86330 - . 20029P (1.365) (2.499) B . + - 2. 69047D2 (2.251) + - * * - 4. O8876D = . *** .000056PRYU (3.233) ** qUD + * ** 10238D P (2.764) 1 R 1 2. 57700D + . 166548D (2.423) G (3.365) (R2 Figures in parenthesis under each coefficient are t-values. * Denotes significance at 90% level of probabiflty. ** Denotes significance at 95% level of probability. *** Denotes significance at 99% level of probability. .751) 137 APPENDIX B-2 Estimated Nonlinearized Mode1 (Farm Demand) qFD *** = 19. 249 (6. 053) + . 345828P + . 322775P - OO714ZP (6.409) R (8 956) R (2. 651) . . + 003264qRt +YNRB) RR .b d. -p -p -p - 2. 107D3 (4. 881) qFD B (R2 -12. 050 - . 118751P + .356392P (3.634) (2.765) B (6.816) R + . 02729 1q (3.429) 1.262D (3. 254)2 1 = . 795) + (4.869) (6.887) (R2 .775) 138 (Farm Sales) q * FS (NS) = -20.456 + 147469P - .058110P + .008562L R (5552) (1.547) R (.916) B (10.029) + . 011960E + i9l7q" Rt(6. 339) (6. 027) 00ill2(Pq 1 1+Pq< i (3. 056) (R2 = .838) q (NS) FS = 4.627 B (4.257) - * (NS) . 002457P - . 033260P - . 000653L B (.077) (1.010) R (1.733) *** - *** 047835qFK . 005449E + (7.067) (9.003) 1 * + . 000097(Pq 1+Pq13 1L (1.713) (R2 = .751) (Urban Demand) qUD *** *** * ** = 31.273 - .461532P + .112747P R B (10.925) (11.014) OO235OY + (3. 152) * -5.769 (1.611) (4.589) 00007124(P Y *** (NS) .078669P B (.990) + (NS) 247500R (12.995) * - °0003982BU (1.722) .825) (R2 R U) (4. 984) (NS) (.389) (4.360) * * * * * - UD * ** + .003014pw (1.357) (NS) + 627D (R 2 = (1.055) Figures in parenthesis under each coefficient are t-values. * Denotes significance at 90% level of probability. ** Denotes significance at 95% level of probability. *** Denotes significance at 99% level of probability. . 759) 139 APPENDIX C Basic Data 140 Notation for Table C-i (1) Rt-11'F Index of the three-year moving average rice price of the November-January period relative to Index of fertilizer prices. (2) Rt1'O Index of the three-year moving average rice price of the November-January period relative to Index of the prices of agricultural supplies excluding fertilizer. (3) = PRt1/WF Index of the three-yearmoving average rice price of the November-January period relative to Index of rural wages. PRt1/PPFI = Index of the three-year moving average rice (4) price of the November-January period relative to Index of prices paid by farmers. (5) LH Hired labor per 1/10 Ha. (6) LF Family labor per 1/10 Ha. (7)=L Expenses on commercial fertilizer per 1/10 Ha deflated by Index of prices paid by farmers excluding fertilizer (PPFI nf ). (8) = FC/PPFIf (9) = FH/PPFIf = (10) = FA/PPFInf = (11) = IR/PPFI = (12) = 10/PPFI = Imputed value of home manure per 1/10 Ha deflated by PPFI nf (8) + (9). Irrigation expenses deflated by Index of prices paid by farmers (PPFI). Other operating expenses including seed, pesticides, repairs, and depreciation on farm building and implements, and stock labor, etc. deflated by PPFI. 141 Source: 1) Farm Household Economy Survey and Cost of Production Survey, Ministry of Agriculture, 1963-71. 2) Rural Prices and Wage Survey, Korean National Federation of Agricultural Cooperatives, 1963-71. Table C-i Relative prices of rice and input factors used for 1/10 Ha, by size of farm, 1963-70. Year Size of Farm (1) (2) (3) (4) (5) (6) (7) (8) (10) (11) (9) Ha rn/h rn/h rn/h won won won won 1963 .5 or less 53y 79.1 55.5 32.8 35 120 498 155 431 379 209 G57 .5 1.0 79.1 36.5 82.8 31 111 142 523 496 1319 229 1.0 1.5 1.5- 2.0 1964 2.0 or more 5 or less .5 1.0 1i4.L. 1.0- 1.5 114.4 114.4 114.4 1.5-2.0 2.0 or 1965 more .5 or less .5 1.5 - 20 1.0 1.5 1.5- 2.0 2.0 .5 .5 1.0 ilS.i 115.1 3.15.1 3.32.5 86.5 36.5 86.5 30.4 90.+ 40 10.2 lOi.2 82.8 82.8 82.3 92.5 32.5 92.5 92.5 92.5 130.2 22 103.2 133.2 130.2 103.2 lQo.2 132.6 132.6 102.6 102.6 102.6 35 48 56 1U.2 103.2 9.4 9.4 93.4 3.30.2 100.2 113.2 j3.3 1Jj.3 94.5 94.5 94.5 34.9 94.5 73.9 93.4 33.4 93.4 38.4 95.4 3j.... 81.0 73.9 81.3 1.0- 1.5 lt,1.7 3.49.? 53.4 53.4 53.4 1.5 19.7 1'.7 53. t5.2 53.4 73.3 73.3 /3.3 73.3 73.3 63.4 1.0 158.? 1.0- 1.5 i5.? 61.3 61.3 1??. 1.5 2.0 133.5 2.0 or more .5 or less 3.497 3.5 1.0 -2.0 2.Oormore .5 or less .5 1.52.0 1970 ti.2 79.1 79.1 79.1 99.3 99.3 99.3 99.3 99.3 73.9 73.9 .5 1969 .o or more 115.1 or less i22. 1.0- 1.5 1968 .i. 2.0 or more 1J.3 .5 or less 115.1 .5 1.0 1967 IOT.2 1.0 1.0- 1.5 1966 93.9 93.9 93.9 114.4 2.0 or more 3.0.0 .5 or less i.''. 1 .5 1.0 - 1.0 1.5 1.5- 2.0 2.0 or more i4.j 1+.1 IP..1 t.1 /3.9 61.3 51.3 ,1.3 64.0 54.0 64.0 64 81.. 6.4 61.4 58.4 9j.t 91.0 91.0 31.3 91.3 85.5 89.8 85.8 55.9 85.8 35.3 89.3 88.3 38.? 96 69 41 33 41 54 75 70 43 37 45 93 77 61 130 115 133 130 156 1.1.2 1.45 93 51 134 126 128 121 149 141 136 134 156 149 131 129 122 114 153 3.11 146 135 136 125 72 49 99 3.14 93 80 64 116 112 39 3.05 1.44 32 43 lb 76 141 133 132 122 129 138 131 124 56 118 959 811 846 957 395 534 53 45 62 27 12 34 45 92 38 63 102 132 98 92.1 92.1 29 16 18 32.1 31. 133 136 13? 131 38 36 124 32.1 53 55 115 Sb.IL 781. 73 43 86 74 66. °2.1 730 729 96 32 55 55 71 3P 35 38.! 55.3 S5. 731 847 768 770 795 772 829 329 836 886 876 961 869 932 932 935 976 578 954 935 3M... 66... 513 575 563 657 612 624 712 741 754 52 33 '35 395 353 413 431 376 372 395 311 288 301 265 227 295 309 252 266 199 357 363 345 327 320 333 397 341 356 349 293 343 313 339 281 340 331 348 312 428 951 930 916 1067 1033 1373 1174 1096 1065 1319 1331 994 1008 1142 1377 1322 1061 971 1.1.86 1197 1181 1213 1196 1263 1266 1.273 1288 1284 1269 1221 1267 1274 124c 1151 1177 1235 1207 1332 241 297 267 188 267 265 304 315 212 211 261 276 302 263 245 312 283 310 283 292 340 383 416 239 237 344 340 366 229 228 286 325 31.2 358 243 266 123 360 (12) won 394 385 360 366 381 486 486 473 496 513 433 448 420 485 560 358 342 363 349 385 389 361 374 379 380 250 293 277 330 306 393 375 391 425 405 483 413 435 447 433 A 143 Notation for Table C(1) Bt-1'F Index of the three-year moving average barley price of the July-August period relative to Index of fertilizer prices. Bt-l"O = Index of the three-year moving average barley price of the July-August period relative to Index of the prices of agricultural supplies excluding fertilizer. (3) = PBtl/WF Index of the three-year moving average barley price of the July-August period relative to Index of rural wages. (4) = PBt1/PPFI Index of the three-year moving average barley price of the July-August period relative to Index of prices paid by farmers (PPFI). (5) = LH = Hired labor per 1/10 Ha. (6) = Family labor per 1/10 Ha. LF (7)=L (8) = FC/PPFIf (9) = FH/PPFIf (10) = FA/PPFI (11) = 10/PPFI = Expenses on commercial fertilizer per 1/10 Ha deflated by Index of prices paid by farmers excluding fertilizer. = Imputed value of home manure per 1/10 Ha deflated by PPFInf = (8) + (9). = Other operating expenses including seed, pesticides, repairs, and depreciation on farm building and implements, and stock labor, etc. deflated by PPFI. Source: 1) Farm Household Economy Survey and Cost of Production Survey, Ministry of Agriculture, 1963-71. 2) Rural Prices and Wage Survey, Korean National Federation of Agricultural Cooperatives, 1963-71. Table C-2. Relative prices of common barley and input factors used per 1/10 Ha, by size of farm, 1963-70. Year Size of Farm (1) (2) (3) (4) Ha 1963 .5 or less 1.5 1.0 2.0 1.5 1964 L5 2.0 143.5 2.Oormore 1.43.5 .Sorless 1.0 1.5 134.6 134.6 134.6 134.6 2.0 or more 134.6 .5 or less 151.2 .5 1.0 lSe.2 15.2 1.5 2.0 15.2 2. 0 or more 15.2 1.5 .5 or less .5' 1.0 1.0 1.5 1.52.0 2.0 or more .5 or less .5"' 1.0 1.0' 1.5 1.5 2.0 144.a 144.2 144.2 144.2 144.2 113.3 113.3 134.6 134.6 134.6 115.9 115.9 134.6 134.6 134.6 6 T346 134.6 128.4 128.4 128.4 128.4 134.6 133.8 133.8 133.8 133.8 2.0 139.2 139.2 11.4 2.0 or more 131.4 .5orless .5 1.0 1SC.1. 1.0 15.1 5 3.50.1 1 1.5" 2.0 126.5 126.5 134.6 134.6 134.6 1 &T 134.6 123.3 123.3 123.3 123.3 1Z3.3I?84 80.4 80.4 80.4 80.4 8.4 13Z53.4 139.2 53.4 2.0 or more 139.2 .Sorless 131.4 1.0 .5 I31..4 1.0"' 1.5 131.4 1.5 1970 113.3 143.5126.5 1.0 1969 11.3.3 126.5 7.3 1:5Z.O 1968 1.26.5 51.5 81.5 81.5 115.9 115.9 97.3 1.0 5 1967 85.1 65.1 85.1 85.1 85.1 97.3 143.5 143.5 1.0 1966 77.3 77.8 77.8 77.8 77.8 973 2.0 or more .5 or less .5 1.0 1.5 1965 97.3 15.1 2.0 or more 1SC.i 53.+ 53.4 53.4 51.0 51.0 51.3 51.0 51.3 52.2 52.2 52.2 52.2 53.2 88.3. 88.1 88.1 85.1 88.1 61.1. 68.1 68.1 68.1 66.1. 56.8 6.5 56.8 56.5 56.8 53.8 53.8 53. 53.8 53.8 81.5 8I. (5) (6) (7) (8) rn/h rn/h rn/h (9) (10) (11) won won won won ii 99 110 595 1293 1502 1804 1726 1539 1474 349 374 400 379 341 523 537 1276 1293 1289 1288 1266 547 506 407 529 488 1.390 465 25 372 tl93 18 79 35 40 1.3 14 25 41 16 19 22 75 47 99 86 61 47 102 76 85 1 97 110 87 1.12 100 86 88 118 95 107 698 52677 604 898 IZ2 658 607 467 551 1146 1119 1072 923 532 559 553 744 734 736 586 705 680 670 746 736 641 564 1205 1325 1306 1284 387 386 53 41J 42 82 12 13 16 100 78 112 104 94 38 50 88 535 579 654 665 720 99.1 99.1 99.1 99.1 99.1 7 79.7 79.7 79.7 79.7 73.3 15 11 13 21 39 103 118 98 90 86 92 733 769 819 767 875 862 678 655 639 698 1595 1447 1474 1406 1573 354 311 320 293 270 8 78 70 86 81 85 79 91 799 726 658 604 667 884 1528 1453 1553 1639 257 240 247 217 309 73.3 73.3 73.3 75.1 75.1 75.1 75.1 75.1 7 1446 1348 1359 1675 1604 274 288 265 293 210 1315 1295 204 203 ii. 20 26 T37 6 91 81 77 65 53 65 51 85 80.3 849 886 755 rB,,,,,,',,85 85 66 61 81 76 74 92 93 85 775 781 814 777 816 18 60 23 51 78 74 760 860 22 34 1.2 9 8 88 95 52832 671 567 545 898 788 614 555 435 1.461 t4462fl5 145 Notation for Table C-3 Bt.1"F = Index of the three-year moving average barley price of the July-August period relative to Index of fertilizer prices. PBt1h1)O = Index of the three-year moving average barley price of the July-August period relative to Index of the prices of agricultural supplies excluding fertilizer. (3) PBt1/WF = Index of the three-year moving average barley price of the July-August period relative to Index of rural wages. (4) PBt1/PPFI (5) L = All labor per 1/10 Ha. (6) Fc/PPFI = Expenses on commercial fertilizer per 1/10 Ha deflated by Index of prices paid by farmers excluding fertilizer. (7) = FH/PPFIf (1) Index of the three-year moving average barley price of the July-August period relative to Index of prices paid by farmers (PPFI). = Imputed value of home manure per 1/10 Ha deflated by PPFI nf (8)= (9) = I/PPFI = Other operating expenses including seed, pesticides, repairs, and depreciation on farm building and implements, and stock labor, etc. deflated by PPFI. (10) = LH = Hired labor per 1/10 Ha. (11) = Family labor per 1/10 Ha. LF Source: 1) Farm Household Economy Survey and Cost of Production Survey, Ministry of Agriculture, 1963-71. 2) Rural Prices and Wage Survey, Korean National Federation of Agricultural Cooperatives, 1963-71. Table C-3. Relative prices of naked barley and input factors used per 1/10 Ha, by size of farm, Year Size of Farm (1) (2) (3) Ha .5 or less 1963 .5 1.0 1.5 1.0 1.5 2.0 2.0 or more .5 or less 1964 .5 1.0 97.3 Q7.3 7.3 97.3 97.3 1435 11+3.5 I 1.5 -2.0 1965 1966 1967 1968 1969 1970 143.5 77.8 85.1 77.8 35.1 77.8 85.1 77.8 85.1 77.8 85.1 126.5 113.3 126.5 113.3 126.5 12b.5 134.6 134.6 134.6 81.5 81.5 81.5 81.5 81.5 115.9 115.9 113.3 115.9 113.3 115.9 134.6 134.6 134.6 134.6 134.6 134.6 2.0 or more 1'+3.5 .5 or less 134.6 .5 1.0 134.6 1.0 1.5 134.6 I 2.0 or more 134.6 134.6 13k.6 .5 or less 1511.2 123.3 128.4 .5 1.0 150,2 123.3 128.4 1.0 1.5 150.2 123.3 128.4 1.5 -2.0 151.2 123.3 128.4 20 6tthbe1.2 .5 or less 144.2 811.4 88.1 .5 1,0 11.4.2 80.4 88.1 1.0- 1.5 144.2 80.4 88.1 1.5-2.0 11+4.2 83.4 88.1 2.Oormore 144.2 80.4 88.1 .5 53.4 .5 1.0 139.2 53.4 68.1 1.0 1.5 139.2 53.'. 68.1 1.5- 2.0 53L 139.2 68.1 2.0 or more 139.2 53.4 68.1. .5 or less 131.4 51.0 56.8 I 1.0- 1.5 51.0 56.8 1.52.0 131.4 131.4 51.0 58.8 2.0 or more 141.1+ 51.11 56.8 .5 or less 151.1 52.2 53.8 .5 1.0 1,0_is 151.1 52.2 53.8 1.52.0 151.1 52.2 53.8 2.0 or more 151.1 52.2 53.3 oles1Z2 (4) 134.6 133.8 133.8 133.8 133.8 99.1 99.1 99.1 99.1 99.1 6.IT79.7 7 79.7 79.7 79.7 73.3 73.3 73.3 73.3 75.1 75.1 75.1 75.1 (5) (6) (7) rn/h (8) won won won 127 116 106 103 838 1028 975 837 11+40 1J4 115 (9) (10) rn/h m /h 1866 620 Br4go5l71g5a5 791 858 1649 15101 17 109 447 522 20 44 49 91+7 19 15 86 58 55 96 481+ 21+15 781 833 823 838 799 877 725 844 819 592 567 139164021 1441+ 621 26 112 110 106 111 1u20 997 867 891 955 558 733 715 632 687 1578 1730 1582 1523 1642 769 540 484 518 539 107. 106 95 89 86 1053 933 997 1099 925 1804 1719 579 1121 1001 751 786 797 695 662 714 798 636 611 1.070 61+7 85 90 99 87 85 91 1tZ 93 81 84 80 95 87 86 83 102 97 102 126 (11) won 1247 936 717 97 1963-70. 729 1.045 958 1088 2084 1665 1498 1558 1667 1717 953 91+7 842 677 39 46 18 46 22 82 46 41 81 66 59 z5 21 90 89 30 41 76 1791+ 20 52421 521 25 87 85 70 1587 515 2160 27 57 655 1757 1612 1717 1743 342 366 391 366 395 671+ 1689 381 402 384 16 1+76 1+36 16 71 62 53 86 11+ 83 392 452 21 35 81 91 1794597 17538B 1756 1015 1033 974 1163 1125 711 696 670 61+9 1670 1830 1774 1168 1180 637 587 1805 1767 171+1+ 70 2960 38 48 25'77 18 75 37 19 21+ 30 76 147 Notation for Table C-4 (l) = (2) Monthly farm per capita income originating in the non-rice-barley sources deflated by the Index of prices paid by farmers (PPFI). NRB (3) (4) Monthly urban per capita income deflated by the Index of urban consumer prices (CPI). Lti Farm per capita liabilities as of the end of the previ- E Monthly farm per capita cash expenditures on nonfarm goods and services deflated by PPFI. (5) ous month deflated by PPFI. = Monthly average wholesale price of wheat flour per bag (22 kilos) deflatedby the Index of non-grain wholesale prices (WPIn). Monthly average wholesale price of rice per kilo deflated by WPL n (7) = qD (8) Monthly urban per capita consumption of rice Monthly urban per capita consumption of barley. = (10) Monthly average wholesale price of barley per kilo deflated by WPI q = Monthly urban per capita supply of rice from non-farm sources, i.e., qOS (11) = Monthly urban per capita supply of barley from non- farm sources, i.e., (12) (13) = (14) = + q I F + GS GD = Monthly farm per capita sales of rice. = Monthly farm per capita sales of barley. Farm per capita stock of rice on hand at theend of the previous month. 148 (15) = q1 = Farm per capita stock of barley on hand at the end of the previous month. (16) = = Monthly farm per capita consumption of rice. (17) = = Monthly farm per capita consumption of barley. Source: 1) Price Statistics Summary, Bank of Korea, 1963-71. 2) Urban Living Expenditure Survey, Bureau of Statistics, Government of Korea, 1963-7 1. 3) Foodgrain Consumption Survey, Ministry of Agriculture and Forestry, Government of Korea, 1963-71. 4) Farm Household Economy and Cost of Production Survey, Ministry of Agriculture and Forestry, Government of Korea, 1963-71. 5) Rural Prices and Wage Survey, Korean National Federation of Agricultural Coops, 1963-7 1. Table C-4. Data used for estimating simultaneous equation model, 1963-71. (2) (3) (4) (5) (6) (7) won won won won won won won 1 181.2 530 1.313 1+11. 1313 1313 1817 1817 1817 1890 1833 1 1612 i812 1637 1637 1637 1705 1705 1705 1743 1743 1743 1565 40.83 42.91 43.01 43.16 50.59 60.39 81.63 77.01 67.81 2.26 32.0'. 2 2 3 (1) Year Month 1963 522 597 1019 1112 523 189t, 2253 2253 2253 1453 156 592 594 388 600 331. 580 301 568 238 554 252 536 300 519 438 519 402 508 601 1+88 653 485 534 465 521 1453 681 483 4 1565 1491 411 5 1.1+91 31+9 6 1491 '+26 7 1496 3 4 5 6 7 8 9 10 11 12 1964 1965 +uo 481 49'. 8 i'+9 9 1496 393 364 360 10 1570 1+68 11 1.570 12 1571 1 1491 638 810 776 1453 1622 1622 1622 1645 16'+5 1645 1976 1976 1976 1246 521 124E 2 3 4 5 6 1491 1491 1524 1524 1521+ 7 1821 8 1821. 9 1621 1773 1773 1773 1.890 10 11 12 1966 393 85 352 1 2 3 4 1393 1890 0O6 621 563 51+3 1+81 1+67 494 561 600 825 1.156 71'. 1+91 598 568 656 459 516 '+96 582 535 612 777 751 738 742 731 725 474737 146 537 730 1908 1908 1908 1978 415 44.85 '.48 3L+5 431 273 325 327 53u 391 648 673 731 44.11+ 46.1+6 56.14 56.09 60.68 57.58 52.61+ 44.06 39.1+8 38.69 'iJ.91 '+1.74 4± .E 39.45 40.70 41.50 34.00 34.08 '+9.06 41.61 1+5.29 53.15 53.74 52.88 1+7.36 42.03 43.52 45.19 52.72 55.22 56.60 53.22 40.39 36.33 32.42 31.28 32.63 33.25 31.76 (8) Kg 13.73 8.88 9.14 5.99 7.91 7.34 5.47 5.05 6.15 10.37 9.85 10.61 10.80 (10) (11) (12) (13) (14) Kg Kg Kg Kg Kg Kg 2.30 3.11 8.83 7.22 8.13 7.96 5.99 4.08 2.70 3.29 3.63 4.30 4.81. 5.28 6.91. 1.9O 1.66 1.01 1.03 1.92 3.26 2.53 1.66 1.50 3.94 7.73 6.61 4.33 2.74 -5.83 1.95 -1.3.85 1.54 .62 1.1.691.20 3.36 10.89 10.50 9.83 8.30 7.33 1.79 2.36 3.24 4.45 5.61 1.52 1.32 2.93 3.60 3.93 8.54 9.89 11.11 11.14 11.73 4.67 3.19 2.91 2.49 1.62 -5.07 1.60 -14.15 1.70 -2.91 3±661 1.4s 31.19 31.62 31.76 28.77 (9) 10.97 10.90 1. 2.09 2.20 2.28 2.75 3.78 5 3.1.7 3.75 4.34 3.13 3.40 5.46 5.19 3.11 2.32 1.67 1.26 5.00 6.90 1.99 1.29 .97 15.39 23.60 13.09 .60 8.78 6.62 7.70 5.07 3.73 3.45 3.23 8.62 1.13 1.10 1.15 3.50 2.08 4.35 4.80 4.61 3.69 .03 1.32 1.65 1.35 1.15 -2.70 -1.26 .74 -1.05 -13.87 1.21 .08 1.18 1632 474 10.71+ 669 37.18 2.33 -0.53 10.76 1.30 2.1+9 2.57 1.17 36.74 36.23 38.39 3.81. 2.91 2.12 1.83 1.94 28.391. ii32.3H 26.07 1.37 37.0 2.49 1.50 37.08 27.76 38.21+ 4.34 2.15 3.51 2.27 1.42 .92 .42 .22 673 42.29 26.17 25.97 26.24 25.78 27.06 26.35 .1+7 6.44 4.39 3.17 498 1+4.83 27.L.1. .55 95.8 74.5 49.6 39.8 1.98 2.91 1.13 1695 43.75 13.02 8.97 8.47 9.53 9.67 13.76 11.38 12.35 .1+6 .35 .30 3.03 4.93 i.07.7? 1.47 1978 2620 2620 2623 1695 1635 1.978 1+3.30 44.1+0 143.5 116.5 89.1 69.2 46.1 .26 297 285 319 562 460 597 773 588 661 1.69 3.39 4.65 6.43 15.68 24.52 9.50 1.87.6 27.6 18.3 13.1 126.0 257.9 201.7 729 716 703 7u3 731 699 661. 673 665 667 676 '+53 2.91+ .1+0 .18 .1+9 .77 6.71+ 8.56 13.56 22.58 10.54 .31. (16) Kg (17) Kg 27.9 16.8 13.7 16.16 12.97 1.1.0 12.39 11.60 9.86 6.18 4.72 3.87 4.13 5.77 6.73 8.1 17.2 68.1 57.0 43.9 32.3 24.8 21.1. 17.8 1.2.94 5.9'. 15.36 17.63 15.7B 13.96 .12151.31.0.815.62 .29 125.2 7.91. 10.79 15.16 15.22 13.27 4.54 2.05 1.60 1.47 8.8 7.0 3.1 11.0 95.7 13.93 13.38 9.17 5.32 2.21 2.55 4.12 10.39 12.39 61.7 .1.3 20.2 14.0 121.2 248.2 .51+ 1.99.7 25.7 6.90 13.61 14.77 13.38 13.13 10.69 4.73 2.59 2.61 3.03 .69 .50 142.3 118.9 92.3 73.1 53.6 42.1 33.3 23.5 79.5 212.4 164.0 19.9 16.9 13.6 10.? 111.4 99.6 80.9 66.8 53.8 47.5 39.3 11.53 3.98 4.59 4.95 .81+ 1.43 5.80 5.00 2.56 2.17 .81 .95 .57 .678.36iI6 9.69 .89 7.04 (15) Kg 1.13 143.3 119.2 96.1 34.9 2.7 14.1+6 11.?'. 11.23 9.19 6.64 6.29 10.13 13.00 12.51 12.31 10.29 10.01 4.14 4.87 5.0'. 35.10.19 31.0 24.8 7.81 10.42 11.23 10.15 6.36 3.61 3.25 2.94 341 . '.0 Table C -4 Continued. (1) (2) Year Month won won 5 2316 536 6 216 4l 9 2135 2135 2135 451 10 21+37 11 .437 12 3 2437 2767 2767 2767 4 5 2506 6 28 609 719 525 702 689 619 639 662 535 7 2939 2939 2939 339k 7 8 1967 1 2 8 9 10 11 12 1968 60) 530 626 57 33q 1393 104 10 3394 3181 3181 3181 3159 3159 3159 3189 3189 3189 3293 11 1290 12 3293 3734 3704 3734 3497 1 2 3 4 5 6 7 8 9 1969 ?86 480 524 1 2 3 4 5 6 7 8 3497 37 3948 3945 571 755 679 678 626 519 519 1+79 555 730 556 1176 775 671 644 685 717 642 567 517 (3) (4) (5) won, won won 1632 1632 2356 2396 2396 2196 21% 415 298 1+09 371 576 415 673 743 639 696 536 512 2136 1353 1353 1383 1682 1682 1+45 1652 31.8 1+37 1903 1903 399 19u3 625 2377 452 377 721 2377 306 1397 589 1397 617 1397 576 1556 503 1556 439 1556 361 1870 328 1570 327 1870 458 1951 568 1981 618 1951 '824 1437 61.5 1437 733 1437 587 1394 580 1.394 1394 18+8 1348 654 638 632 637 642 638 63. 629 632 627 610 607 619 615 612 581 563 563 547 533 548 (6) (7) won won (8) Kg 39.11 38.98 39.43 43.00 45.38 25.03 22.35 20.52 20.99 22.89 25.61 27.90 11.95 11.21 13.03 8.93 9.37 10.47 11.83 42.00 36.45 (12) (13) (14) (15) (16) (17) Kg Kg Kg Kg Kg Kg Kg 2.50 5.36 4.30 4.90 4.42 3.27 .65 -5.77 1.50 .49 -5.35 -1.59 .23 .97 1.07 5.67 5.94 4.41 3.85 5.24 8.44 15.13 .86 2.12 7.82 4.76 3.20 1.97 1.19 1.11 79.6 63.7 47.1 33.0 21.7 14.3 41.1 201.8 177.1 152.3 125.1 107.6 83.6 21.4 16.7 102.7 99.0 84.0 69.0 57.8 10.20 9.57 6.26 4.81 6.02 9.80 13.49 6.30 7.97 9.83 10.91 9.76 6.79 4.03 48.0 36.1 26.4 20.7 132.3 100.4 96.5 83.5 67.5 53.6 .9E 3.74 3.92 3.95 3.26 2.46 12.451.93-18.70 12.44 13.23 12.45 12.61 12.06 11.50 13.74 10.23 11.25 11.57 12.44 37.1.3 42.18 64.00 '4.96 44.04 43.25 42.39 39.71 38.83 27.36 27.51 28.77 30.75 31.45 26.78 24.81 24.85 26.05 26.27 27.25 25'28."9 39.33 559 41.13 41.34 42.50 (+2.53 1.2.33 42.05 42.01 45.95 49.14 46.21 28.05 12.45 13.47 2.69 1.65 1.79 1.71 1.84 26 2.73 2.87 2.68 2.09 1.54 1.44 1.2.41 1.43 27.87 26.93 25.82 25.27 24.51 25.75 24.43 25.05 28.01 28.60 12.53 12.47 12.13 11.91 11.38 13.95 11.15 1.65 1.82 2.00 2,11 2.38 2.57 12.117 1.98 1.73 1.62 12.21 12.26 12.04 12.58 11.71. 11.35 10.98 1i.'Oo 10.55 10.33 -3.15 1.25 -0.88 .67 4.49 .B1' 3.44 5.20 4.48 -0.74 -12.08 i'.3'18.72 28.07 49.53 29.23 53.06 28.98 55) 69.05 26.89 547 48.59 27.08 989 543 1.893 25.84 1+36 54j 49.39 25.36' 398 546 1.9.28 25.76 41.7 51+3 49.18 26.8? 555 553 (11) Kg 35.83 36.69 41.88 563 565 565 565 565 563 562 62 (10) 35.5527.18 55? 56 (9) Kg 2.1+8 -0.52 .6426.78 1.30 .12 .08 .14 .16 1.3.66 .70 13.50 11.68 10.46 6.64 1.34 6.39 4.41 5.94 10.75 21.49 8.09 4.79 2.79 1.4? -0.504.93 -6.50 -2.63 -0.50 .42 .63 1.51. 1.39 1.48 .4 .80 7''ai3V'. 3 .45 13.05 .92 .62 .48 11.00 .31j .74 11.41. .32 5.37 6.52 8.41 7.18 5.84 2.28 -0.26 1.06 1.00 7.60 6.31 5.03 .59 -5.13 -1.33 -0.49 -0.38 .38 2.77 2.96 4.95 9.13 11.63 -11.68 .31 21.40 1.76 3.12 1.42 9.02 1.77 3.64 .93 9.03 2.38 3.1.1 .47 8.69 2.33 5.62 1.31 5.79 2.25 2.87 .24 8.19 2.44 4.30' 135' 6.77 2.66 6.34 -5.52 4.25 2.73 6.34 -0.51 4.03 '+8.611.76 3.61 42.0 37.9 33.0 27.9 23.0 3.70 3.21 4.47 4.98 5.43 63.8±?.? 2±o.5'"'' 65. 169.2 38.0 .89 147.5 34,0 .85 .71 .93 1.1+2 7.01 3.64 2.77 2.20 1.26 1.22 1.16.3 95.3 75.2 56.5 43.3 31.9 26.0 19.7 41.1 ils.6 29.0 24.9 21.0 17.2 111.5 1.07.1 93.1 77.6 67.3 51.1 .35 169.1 52.2 .85 149.1 48.1 1.62 1.30.1 43.9 .96 110.? 38.0 2.02 90.9 32.9 1.10 69.3 24.6 8.27 52.7 105.2 3.33 '#1.6 111.9 11.21 12.96 11.02 10.91 10.46 9.92 6.82 5.25 7.12 10.99 12.34 9.24 10.06 9.12 5.80 4.09 12.59 4.05 10.71 11.24 10.14 4.93 5.50 6.10 1.46 ii.3O'33''"'" 10.82 3.83 1U.36 7.27 7.20 5.58 5.57 9.03 9.39 11.36 10.51 9.71 11.33 9.69 9.25 9.17 9.93 7.02 5.75 9.18 9.62 7.10 4.67 4.16 4.49 4.27 4.94 5.99 6.50 6'+1 8.94 9.?'. i Ui Table C-4. Continued. (1) (2) (3) Year Month won won won 3948 4177 4177 4177 +344 7F1. 757 18.8 9 10 11 12 1970 1 910 0i1t3 785 3 4344 809 4 3955 81.7 5 6 7 8 3955 747 608 601 657 873 2 9 10 3955 4190 4193 4190 4389 7'". (5) 735 1870 512 1870 784 1870 li4S 1168 820 1168 833 1168 687 1276 67 1276 640 1276 417 1651 412 1651 460 1651 782 1839 590 1839 74'. 536 532 5311 528 518 510 504 499 495 495 1.95 '+95 492 (6) (7) won won (8) Kg 48.85 48.50 43.23 49.64 26.13 10.54 26.4611.12 1.7.81 27.31. 11.1.7 2.1.8 47.54 47.10 27.65 30.03 10.93 12.98 49.15 1.8.46 4891+7.06 18J9 999 485 1305 483 479 3 1.259 769 1375 1375 1375 381 2 4259 4259 55.39 56.69 56.45 55.91 55.11. 4 4416 1.76 797 31.11 33.45 36.22 35.40 34.80 5 4416 '+69 55.28 1.0.73 6 '+'+16 7 8 4637 4637 772 710 685 613 4637 825 53.91 55.31 9 37.90 35.02 36.99 1 us 1443 1443 1443 1732 1732 1732 834 730 732 621 532 41555.6137.83 54.03 432 512 511 569 810 563 54.83 1+30 12.'.t 1 .5i 12.12 11.83 55:j 2.13 4.67 2.18 39.18 (11) Kg (12) Kg .1.2 1.27 .86 .69 1.07 .76 .02 .54 .52 .64 2.75 1.77 6.57 6.54 6.59 6.94 -2.49 -1.81 12.79 12.72 13.54 12.41 12.77 1.65 1.70 1.61 1.56 1.84 1.12 -5.79 3.02 5.62 4.23 .36 .31 1.05 .96 12.46 11.74 11.02 11.06 1.97 2.18 2.18 .71 3.t3t.1ia.eo 1371 12 11.98 12.24 12.41 12.58 (10) Kg 2.95 5.79 2.45 i.'.9 1.79 -1.36 1.63 -11.91 1.65 .81 1.40 5.09 1.66 3.65 4s.44 48.35 47.89 1.9.U0 115 1.87 (9) Kg 26.81 26.35 27.31 27.13 26.61 27. 1u 27.37 27.25 4389 4389 11 1971 11189 11115 (4) won won 11.42 2.16 13.87 24.40 13.02 8.40 9.62 6.74 6.30 6.82 5.70 -0.02 1 .39 .49 -2.35 -1.84 .65 (14) Kg (15) Kg (16) Kg (17) 2.56 31.2 10.70 22.4 99.2 85.6 5.30 1.19 .95 227.7 181.6 161.3 140.0 75.3 65.3 56.8 53.1 49.1 8.57 12.49 12.06 11.35 11.72 10.15 4.23 3.84 3.89 3.12 4.36 '.1.7 10.475.0k .94 .66 .72 2.11 73.5 59116.1 Kg 7.8'. 1. 8.36 5.93 5.53 4.84 6.77 1.81 1.73 5.58 5.11 1.80 97.1 8.6 62.9 51.4 38.7 35.8 29.0 79,7 104.9 92.7 9.76 9.51 7.16 5.08 6.43 5.62 6.21 7.59 9.31 8.03 13.09 20.76 12.49 1.45 1.56 .63 .68 2.09 62.7 208.1 175.5 153.7 132.8 70.7 63.1 57.6 52.0 47.1 11.92 10.94 10.83 9.74 3.82 3.12 3.57 3.53 4.53 1.78 1.90 5.08 4.85 '37.4 9.35 9.01 6.48 5.45 5.63 5.27 6.11 7.91 8.94 8.57 5.8 i43i35' 2.1.8 .17 Ls.80 9.73 (13) Kg 8.07 9.43 0.8 .89i.5 8.02 5.94 5.61 5.04 6.79 1.69 114.1 kG.'. 78.3 64.8 54.9 35.1 27.6 87.6 97.6 '.4.3 85.'. 9s;l6 12.5'. 101k6 01