DISCUSSION FAPF°. Report No.: ARU 56 On the Determinants of Cross-Country Aggregate Agricultural Supply By Hans Binswanger, Yair Mundlak, Maw-Cheng Yang and Alan Bowers Research Unit Agricult'ire and Rural Development Department Operational Policy Staff World Bank September 1986 The views presented here are those of the author(s), and they should not be interpreted as reflecting those of the World Bank. I6- e I The aiuchors a:e staff members and consultants of ;he Ijrld Bank. However, che World 3ank Ioes not accept responsibility for the views expressed - lherein whidc are :hose of the au:hors and should not be attributed co t.e W;or,li Ba,ir- :)- L. ts affiliated orranizat'.crs. The findings, interpreEa- -.r. a!-o con'Ius';ns a-e the results of research supported in part by rhe 3ank; theti da not i eessarioy reDresent official policy of the Bank. The !esi-na!:io,.-s emplayed and :e?-nttion of material in this documenc 3re s o esig ria:Or s t -e c p ½ ye ~ n d n e ot)ezt o r ~ t r a ± i h s d c m n r solely fcr the cor.venien.e the reader and do not imply the expression o- any ooinion twhat,-,ever on :he part of the World Bank or its affiliates concernftg the .egal statcs of any councrv, territory, area or of ics author iies, or concerning the delImitation of its boundaries, or nacionaj affiliation.. N. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ON THE DETERMINANTS OF CROSS-COUNTRY AGGREGATE AGRICULTURAL SUPPLY Hans Binswanger, Yair ,iundlak, Maw-Cheng Yang and Alan Bower I. Introduction The nature of sup:ly response is a subject often encountered in evaluating the effects of economic policies. The concept of supply response concentrates on the output-orice relationships. In general, the empirical analysis of supply is conducted within the framework of a competitive firm and thereby ignores important features of the analysis, namely: the nature of factor supply and the determination of technology.2 For agriculture as a whole, factor supplies cannot be taken to be perfectly elastic. In fact, the supply of most factors is fairly inelastic in the short run and as such lead to inelastic supply. The rs'e.varce of the elasticity of the factor supply can be Inferred by comparing the fairly high supply elasticities of individual crops with the low ones obtained from a direct elimination of aggregate supply.3 While resources are fairly fixed 1Hans Binswanger (AGRES), Maw-Cheng Yang and Alan Bowers (EPDCS), and Yair Mundlak, Hebrew University of Jerusalem. Our thanks go to Fataneh Semsarzadeh and Hye-Sook Chung for assistance with the painstaking task of filling in mi.sing values, and to Ron Duncan who managed the project. Our gratitude also to those persons and institutions who supplied data. 2The issue of inelastic factor supp y was appt ,priazely treated :y Johnson in 1950, but largely ignored in later literature. 3See Mundlak (1985b), fo- a critical review of the empirical analysis of aggregate supply. For individual estimates see Bapna, Bond, Coleman, Johnson, Griliches and Pandey et al. -2 - for agriculture as a whole, they can more easily be allocated amcng products.4 An exceotion to the findings of inelastic aggregate supply response frcm a direct estimaticn of the supply function is Peterscn wno used cross- country data and obtained an aggregate supply elasticity of 2.0. He argued that the lcw estimates, which have been based on time-series data for individual countries, are invalid because they do not reflect the investment impact of changes in long-term price strategies. The relevance of the technology choice issue for supply analysis is less recognized. As explained in Mundlak (1984, 1985a), the problem of producers is to choose simultaneously the techniques to be implemented and the intensity of their utilization. The traditional supply analysis deals only with the latter aspect. The choice of techniques depends on various state variables, and public inputs, both human and physical are impor.ant among Lhem. To analyze their impact, it is desirable to select a sample with a wide spread in such variabies. Naturally, a wide spread is fourd in cross country comparisons. This seems to be the comparative advantage of a cross country analysis. The purpose of this paper is -o analyze the effects of price and various public inputs, or shifters, on the aggregate agricultural supply. The sample consists of annual observations for 58 countries for the period 1969-1978. This paper summarizes some of the findings which are reported in greater detail in Bins,anger et al (1985) (BMYB). The major findings of the study are: A weak positive supply response is obtained from the variations over time for the individual countries 4See Banpa, 3inswanger and Quizon for pertinent results. For a review of some empirical analyses of individual products, see Askari and Cummings. -3- (within-country variations). An implausible negative supply response is obtained from the between-country variations. The shifters, as a group, account for most of the variations in supply in the within-country and between-country analysis. Does it mean that prices have no role in the determination of supply? The answer is no but the role is somewhat more difficult to detect. We indicate that the within-ccwntry estimates measure the short-run response and as suc,, our estimates are consistent with estimates obtained from similar data using for the agricultural production function. On the other hand, the cross country analysis does not provide estimates for the long-run supply response. It, The Conceptual Framework Exogenous technoloav To a large extent, studies of supply response are carried out within the micro-framework, applying resu-ts which hold for an individual firm dir3ctly to the sector as a whole. Tu p'ace the discussion within such a framework, write the restricted p.-ofit function: (1) i (p, w, k, T) z max (py-wv : y,x £T) y,v where y is a 'ecto- of I oLitputs; x is a vector of J inputs decomposed to (v) variable and (k) fixed components: x = (v,k) with dimensions (a ,, arb = J; T is the feasible tecnnologi set; p is the vector of product prices; w is the vector of factor pr,ces decomposed to conform to the decomposition of x. However, where aroi;uity does not exist, such a decor,osition is not made explicit. -4- The outccme of this optimization are the restricted or short run product supplies and factor demands: (2) yj(p, w, k, T), vj(p, w, k, T) Repeating the analysis for the urrestricted case results in the long run suppl,es and factor demands. (3) y*(p, w, t), v*(p, w, T), k*(P, W, T) Empirical analyses are based on dated data. The decomposition of x to v and k is dcne accord-ng to the ease of changing the inputs within the period of analysis, usually a year. Consequently, tne empirical analysis of (2) produces a restricted or short-run response. The relationships between the restricted supply as given in (2) and the unrestricted supply as given in (3) is given by the identity. (4) y(p, w, k*, T) _ y*(p, w, T) Hence, differentiating logarithmically5 b a In k* (5) cu .' =r + 2 0 *. (5) iUi gri j=1 i In pi where .ui and .ri are the unrestricted and restricted supply elasticities, respectively, and 3j = a In yi/a In k*j are the production elasticities cf the fixed factors . Sometimes E u and cri are referred to as the long-run and short-run elasticities. Note that the relationships in (5) are obtained under the identity in (4). The quantitative importance of the foregoing discussion can be illustrated by a simple example, using a single output production function. 5See Mundlak 1967. For simplicity, let the productitn function be homogeneous of degree p < 1. Then it can be shown that6 d ln y p (6) d in p 1 - p p is related to the prcc.~cticn elasticities (0j), and under equilibrium to the factor shares (Sj): p = 20j = 2Sj Applying it to the aggregate agricultural suDply we consider as a first approximation labor land and capital to be largely fixed in the short run. Using a value of .16 for the factor shares of the variable inputs, we obtain a supply elasticity of .19.7 The division to variable and fixed inputs is to some extent arbitrary. Such a dichotomy suggests a zero supply el-.ticity for the fixed inputs and infinite elasticity for the variable inputs. The iatter assumpticn is made, generally implicitly, in many of the p-oduction analyses using derivatives of the profit function. I1 holds true for the individual firm but not for the industry ;as a whole. The analysis is therefore generalized by introducing the factor supply functions: (7) xj = Sj(wj, Tj) where Tj represents the technology (or tastes when dealing with labor supply) in the sec.or producing input j. Let sj = a ln xj/a ln wj. The 6See EMYB. For non-ho-ngereous function use the s:ale elasticity instead of p. 7Using results from Mindlak and Hellinghausen for the cross-country production function, the ccmbined elasticities for land and labor is .62, leaving .38 for capital ard raw material of which aboit .16 can be attrTbuted to the variable inputs. smaller are the factor supply elasticities, the smaller is the product supply elasticity.8 This can bE :een by incorporating (7) in the analysis to yield under sj /0 for all j.9 (8) E = [( - - 1) + 2 Ejy . j/s where aj is the factor share in total cost and Ejy = 31n xj/ i In y are the expansion elasticities of inputs with respect to output. A value of zero for sj is ruled out in (8) because it is assumed that all inputs are allowed to vary. However, sj can be small for some j, thereby making the supply elasticity small. Taking the production function to be linear homogeneous, so that p = 1, and Ejy = 1 for a'l j reduces (8) to (9) g = 1/E aj/sj If some inputs are fixed so that sj = 0 for some j, the expansion path is not linear anymore. In this case, the result is modified. To simplify suppose there are only two factors, x1 = v, x2 = k and p =1. In this case, (10) £ = [(( -1 s1 This expression is con-parable to (6), and shows clearly hcw the supply elasticity is modified by the factor supply elasticity of the variable factor. Thus, for 01 = .16 and s1 1 the produ-t supply elasticity is .09 instead of .19 obtained above under tre assumption of s1 = . 8For a discussion cf this, see Friedman and for application see Brandow, Floyd. 9See BMYB (1985) for details. -7- Endogenous technology10 The foregoing analysis was conducted conditional on the existence of an aggregate agricultural production function. This assumption is now removed in order to allow for the fact that at any time there are numerous production functions. We now outline the implications of this approach for our analysis. The elementary component of the analysis is a technique, described bv a oroduction function Fj(x). Technolcgy (T) is defined as the collection of all possible techniques, T = {(Fj(x)} The optimization problem is given by (11) max L = 5j pjFj(vj,kj) - 2j wv;j X(k - 2j kj) such that Fj(-) E T. The necessary conditions imply: (12) 0 = 2j (pjFvj - wj)vj t lj(pjFkj - X)kj - 0 - 0 for not implemented technique O + 0 + for implemented technique where the signs under (12) give the joint equilibrium conditions for the implementation of the jth tochnique. For a technique j to be implemented, it is necessary that vj and kj are strictly positive therefore their respective bracketed terms must be zero. The technique is not implemented if a respective bracketed term has negative elements. That leads to the definition of the implemented technology: (13) IT(k, p, w, T) = {Fj(vj, kj) I Fj(v*j,!*j) i O, Fj - T). The optimal output of technique j is: y*j = Fj(v*j, k*j), where V*j = vj(s), k- = kj(s) are the optimal values subject to the state 10This discussion is based on Danin and Mundlak and Mundlak (1984, 1985a, 1985b). variables s = (p, w, k, T).11 Thus: the supply depends ot, T and the constraints k. Of particular importance is capital the accumulation of which leads to the employment of capital intensive techniques. The relevance of this issue for long run supply is discussed in Section VII. This formulation is very general in that the concept of a technique is broadly interpreted. It uovers differences in products, firms, regions, etc. Thus, regions may i;ffer in basic agrotechnical factors such as climate and soil type, as well as by accessibility. Thus, production in farms with poor roads is different from one which is close to the market centers in almno.t every possible respect such as product composition, method of production, use of inputs and so on. This illustrates the importance of identifying as much as possible the determinants of the technology set, as it is explained in the next section. III. Data, Variables and Estimation Techniques We estimate suppl) functions for two outputs, aggregate crop output and aggregate livestock outpit and two input demand functions, fertilizer and tractors. Data constraints do not allow an immediate extension to more inputs. ,he system deals with only a small fraction of the agricultural input. Yet, fertilizer is often cons-iered as representing the advanced technology inputs. Two additional alternatives are added. First, output is aggregated to total agricultural output. In this case the system consists of one output and two inputs. Second, crop output is decomposed into area and yield. IIT i; the technology set. It is written here as an element of the vector s. -9.- Thus, the model consists of several equations which have the same presentation: (14) Yhit = Yhit + xit3h Uhit i= 1, ..., I country index, t 1, ..., T time index and h = 1, ..., H equation index, and y is the intercept which is decomposed with a period and country effect. The explanatory variables consist of prices and of different measures of human capital and infrastructure which affect the choice of the implemented techrclogy. The prices include the two output prices, the price of fertilizer and in some cases a measure of wage rates. A detailed discuss ,r of tht data sources, data transformations and gap filling procedure is given in Appendix 1 of SMYB. A brief description of the variables is provided here to facilitate the piesentation of results. Outouts & inputs. The individual output, inputs and price data come from FAO. Outputs and prices were aggregated using multilateral Fisher indices 12 Fertilizers are measured as tons of total nutrients N+P+K and their price is an index where the three nutrients are weighted by their relative importance in t, ? aggregate. Traotors are number of two-axle tractors with more than eight horsepower. Prices were deflated by purchasing power partiy exchange rates (PPPR) and thereby were converted into dollars.13 12Multilateral translog in.exes as proposed Lv Caves et al could not be used because many countries produce zero quantities of individual commodities. See B?"VB fcr details. 131he source is Kravis et al. See BMYB 'or more details and for a discussion of experimenting {ith fertilizer orices and urban wage rates as defiators. The exogenous inputs are land potential, irrigation and several other measures of capital. Land was used for two pirposes, one for normalizing some variables to a per hectare basis. For that purpose, we used agricultural land which includes arable land, land under permanent crops and meadow and pasture land. This however is an endogenous variable, and therefore, is an inaopropriate measure of the size or potential, of the country. The country acricultural potential is measured by the potential dry matter production in eac' zountrv 'MPDM). It is based on the work of Buringh et al and is constructed from three types of information: photosynthesis capacity, moisture and soil types. *rhe variable is country specific and does not vary over time. Caoital. The capital variables can be classified in two ways: private vs. public and physical vs. human. The information on private physical capital is non-existent for -any countries in the sample. The human capital variables are: Adult Literacy - as a measure of schooling. Life Expectarcy - as a measure of health investment. Research - is introduced in two alternatives.14 First, man-years of research and IL cost of research per man-year are entered as separate variables. Second, the stock of research expenditure per hectare is used, without a dec:omposition to the two components. Extension - is medbured in the n.mber of extension agents per capita of farm population. In a number of cases we have also included the following interaction terms: Rese ch and Extension and of Extension and Literacy. Irrigation measures th; proportion of agricultural land irrigated at least once during the year. It represents to a large extent :.ie supply of public irrigaticn project. The physical infrastructure is represented by i4The data on research and extension were kindly supplied by Ann Judd and Robert Evenson. Ro3d Dens't' - the road length normalizel by the potential agricultural land. Pavement - is the percentage of roads paved. In addit1on to these, we use a measure of ccmprehensive capital as suggested in Mundaak and Helinghausen. Let K be a vector if all the comocnents of caoital. Then, we could write an aggregate prcduc;ion function, using t:e prooe vy of constant returns to scale (15) y = f(k) wrere k is K divided by the labor force. Note that y serves a natural aggregaite of the various capital components and as such it se:'vos as a measure of comprehensive capital. In order to eliminate transitcry variations, a three years moving average centered at t-2 h:s used for year t. The variable was approximated by per capita income. This is the measure of the comprehensive capital for the coun;ry as a whole. It is assumed that the comDonent relevant for agriculture is mcnotonicaily increasing with the total. When the variable is used in the regression toqether with the other components as specified above, its coefficient re-resents the effect of that ,art of comprehensive capital not accounted for by the variables which appear in the regression. As such it .-c-lements those components appearing exnlicitly in the regressions. Finally, we aCd Rural Poculation Density as a separate variable to allow for the Boserup hypothesis that an increase in population density leads to agricultural growth. Higher density is believed to produce gains from specialization as iell as to reduce the unit cost of infrastructure. Estimation Equation (14) is estimated conditional on the classical assumDtions that uhit are distributed normally with 0, s2n)j are independent across -12- countries and over time but not across equations. However, the vector of explanatory variables, x,t, is basically the same in all equations. These variables are taken to be exogenous and therefore the cross-equation correlation of u contributes no additional information and the system is cstimated by single equation techniques. The Intercept is decomposed into country and time components and the regressions were estimated alternatively by allowing for a country and year effect (CT), a country effect (C), a year effect (T) and poolea data (P). The implication of such a decomposition are used in the subsequent discussion.15 Finally, the large number of variables produces multicollinearity. This is overcome by using the principal component technique as appeared in Mundlak (1981). To dete mine the empirical importance of the country and time effects, Table 1 reports the values of the R2 for four equations: area, yield crop output and aggregate output. The cxplanatory variables appear in Table 3. Table 1: Degree of Fit (R2) a/ Selected Equations Country Pooled Country Time & Time Effects Effects Effects Equation P C T CT Area .8792 .9986 .8573 .9986 Yield .7925 .9805 .7896 .9810 Crop Output .8883 .9954 .8527 .995; Aggregate Output .9173 .9977 .8957 .9978 a/ The P specification includes MPDM which is omitted from the T-specification. It is for this reason that the R2 of p is larger than t. The variables included in the equations are listed in Table 3. 15See Mundlak (1961, 1978). -13- The. most general model is CT. Imposing no time effect on this model leads to the C model. An F-test of the null hypothesis of no time effect does not lead to the rejection of this hypothesis in the four equations presented in Table l. This is a rema- U.? result. It is well known that agricultural production has undergone a considerable increase in productivity. It thus appears th.at our specification captures this increase and has left no time systematic residuals. In what follows we therefore concentrate on the two specifications C, the "within", and the P estimator. There is a strong correlation between tne country effects in the various eauations and the various variables. The analysis of the C equation, eliminates this as an econometric problem. By so doing, we do not utilize the between-country variations. At the same time, we also avoid the country-specific errors associated with the construction of the data which are elimi'ated under the within-country estimation. However the correlation between the country effects and the explanatory variables is informative and it is taken up when the C and P regressions ar- compare ,n Section V. IV. Within Country Estimates Often empirical results depend strongly on the specification of the estimated equation. Th.s section summarizes results obtained under various specifications which should help to place the discussion within an apprcpriate empirical framework. The detailed results on which this summary is based appear in BMYB. -14- Table 2 shows the detailed results of one single npecification. But we vill distinguish between robust and less robust results according to whether or not they differed with changes in specification. Prices The various specification explored the sensitivity of the results to alternative methods of deflating prices: Purchasing pcwer, parity prices, deflation by fertilizer prices and deflation by the urban wage. Alternative specifications included (1) simply lagging prices once, (2) introducing the price of t-1, t-2 and t-3 in free fron, and (3) using the Nerlovian distributed lag technique and replacing shifter variables by the lagged dependent variable. The main results can be summarized as follows: 1. The key results are not sensitive to the price normalization 2. Short-run aggregate output supply appears to be remarkably price inelastic. Own elasticities of output supply in no case exceed 0.06 for any of the output variables. 3. Extending the lag structure beyond t-1 produces slightly higher "long-run" elasticities. The largest elasticities are estimated using the distributed lag technique. But even these "long-run" elasticities do not exceed a value of 0.3, a point to which we return in Section VI. 4. All short-run crop supply elasticities have the correct sign and are in general statistically significant. Tihlt- 2: thlupt!Sjjy u. n l......Ii t5i 18iming Withd I-Esi;tIr-ito'rs WIti CA o r Efl~fev i / t-Arva ro fa~atCrop Yield Tracor Nman cro1s PlI'rlc 1.024 0.1118 -0051.11 it) - ( ((.5.') ~~~~~~~~~~~~~~~~~~~~~~(.).4) (2.219) ( 1.94) XJ4rgaoe I'II. I.. 0I.1(4f -(0o(.1)4 1 Fert I I I zt-r 1'r t,.t, I ? iUAI.I 0~~~~~~~~~~~~I.65 (-.1) (0.6I4) I.19)(04) ,(.Mo) (o.50 2 0 ("%.r(6) I .2) (-2.11) (~~~~~~~-4.0 1) (4.2 1) (1.411) (-4. 11) (2. 21) (5.417) (-('.144) (1.6 1) (1.84) (1.34 ) (1.80) (-0.16) (-1.26 ) Afe. KxIxwct.,uoj-y I .01 iI 0. 596 -4). 7 22 (1. Of (1.411 I I. 2 2 (5,dill) (1 .44) (- 2.%) ) (O.h4 ) ( I. I1 I5)) (4 .74) (I I lig)It tmI/ I Wv 011191.164 (1.22 2 (1. 1/42 11.159 0.7238 (-2.9)) (1.05) (~~~~~ ~ ~~~~~~ ~~3.10) (1.49) (1()1.5) (8.60) (od,l8 ((~~~~~ ~ ~~~~.016 11.141) 0).029 -(.(I3 11.113 0.X)17 03.064 SN_ (2.46) (.)(3.87) (3. 11 ) (1.08) (0.1 8) (2.03I)a * - '.lvemiit l)~~~~~ ~ ~~ ~~~~.W~2 )~1J-0.oo)4 0.0')2 0((.0I 0.1)66 )).314 (.IS) (.)(-15 3) (5.6 5) (((.2 1) (.! 9. .981l0I ltIu Ili'll It v (1 2 1 0.265 0.121 1. 52/ ((.214 ((.51 (13.4 31 (4 .8/) (4 .50) (2. 18) (I .21) (6.3 (5 .4 3) (3.64s) L1 eracy 20-I Il(.244 0.41)) -(.2bl6 '(.1 I . .39 -1 .14 1 (-1.15) (~~~~~ ~ ~~3. 19) (2 .90) (-4 .5 2) (1I.51)) (9.4 9) 3-3.9) (9' ((~~~~~ ~ ~~~~~.f1)14 1. 16 1 0.151 01.421) ((.2)1 0.37(0 . .446 1.941) (1.9 5) (4._26) (11.4 1 ( 1.94) (',.o4) (6. 72) loit uv8: I / I'III )cu8 .*re (P1' I I cos. 2/ nt.. COOt I lCl,.ht 8 Art e last I c I I I .-. exee;t fo)r I I te.-.!c, aiudi I rrlIgart Ioiu* where I livy wmatiIre t hie propo, t Iona II - Isureis, ot out [hot w tIth reHJ8wtt t o a oile 1wefleota )(, .1i-it lIncrease in Ls .ilt- 111T lte*racy rate or fIn Llth propo)rtlloo .)f the are.i I rrtgat !,!. (I K2 ;I,re .d.r,eIh. fu,.m varial,ues traiisIorin .1 to within-c(Hitit ry varlatlIowi. flit. overall lit wItht original v,trI.,1vq. (j hi)'h, as ca,i ls- learned tI-ow Tabh,. I. 5. The short-run livestock supply elasticities have a statistically significant negative sign and vary b-tween - -0.3 and -0.16. This result reflects .,e stock adjustment process and requires further analysis. 6. The short-run fertilizer demand elasticity i^ remarkably stable at about -0.16 level. It is the largest, in absolute terms of the short-run elasticities obtained. The tractor elasticity with respect to th. output price varies from 0.09 to 0.13 and is significant. Thus, both inputs are seen to be more price responsive than the resulting output levels. This can be attributed to the fact that the input elasticities represent expansion as well as substitution effects, whereas the aggregate supply elasticity represents only expansion effects. 7. On the whole the cross price effects were of secondary quantitative importance. The Shifters The experimentation with shifter variables were primarily concerned with the treatment of research and extension. The other shifter variables were introduced in the same mar,ner in all equations. The variations with research and extension involve (1) the shift from man-years and research cost per man-year to total research expenditures, and (2) whether interaction terms were present between tne research variable (manyears or total expenditures) and the extension variable and between the extension and the literacy variables. -16- GDP. The largest GDP elasticities appear in the tractor stock, the livestock supply and the fertilizer demand equations, respectively. The elasticities are lower in the crop output and crop yield equations. The lowest elasticity, arises in the area equation. Such a result is consistent with an interpretation of the GDP variable as a capital-availability variable. Population Density. Rural population density clearly has a positive and significant effect on all output variables, except perhaps for crop area. It also increases the fertilizer input. Ignoring the tractor equation, the population density elasticity is the highest for fertilizer demand and livestock output (0.53 for both), followed by aggregate and crop output (0.28 and 0.27). The crop output effect is partitioned into nearly equal area and yield effects (0.12). The high impact on fertilizer demand is consistent with the idea that distribution of fertilizer is cheaper in densely populated areas where commerce is well developed. This, however, does not explain the high elasticity of livestock. Alternatively, it is likely that population density affects the cropping pattern -- the more dense the population the larger the proportion of crops which are nitensive in inputs other than land (in particular, fertilizer). This explanation also accounts for the high elasticity of livestock. The strength and consistency of the population density variable is completely consistent with Ester Boserup's hypothesis on the relationship between population density and agricultural productivity. For all other variables the results are less stable. Research. The only robust conclusion that one can draw from the various formulations is that research increases fertilizer demand. A somewhat less rcbust conclusion is that research increases aggregate crop yields, but perhaps tends to reGuce crop area. Because of these apparently contradictory impacts, the results do not show a positive effect on crop output. As the resuls with respect to livestock output are also contradictory, research appears not to influence aggregate outDut in a Dositive way. These results appear to contradict a lot of earlier research on the impact of research on agricultural output. Further work will be required to discriminate among three hypotheses: (i) Our find4ngs are correct and earlier studies falsely attributed to research the effect on agricultural production cf left-out variables, such as population density, capital availability and infrastructure. (ii) Research is not measured accurately enough by our variables. (iii) There is too much heteroyeneity in our data set to allow the effect of research to show up. Groupings of countries into more homogeneous groups is required. Extension. With our crude extension variable we cannot show an effect of extension on either crop area or aggregate output. Extension may, however, have an effect on crop output, crop yield and livestock output. Note that our extension variable cannot measure quality of extension. Irrigation. Irrigation clearly has a positive effect on aggregate output and tractor demand. The latter effect is easily explained by the extra poeier requirements of more intensive croppin'. It also appears to increase livestock output. The effects on yield and crop output are not consistently positive, but the exceptions to positive response can be explained out. Irrigation may have a negative effect on crop area, however. If aggregate demand is limited, newly-irrigated areas may compete with non-irrigated areas and the latter may decline. Just as in the case of research, the higher yields may substitute for area under crops. The aggregate output effect of irrigation may therefore not be as dramatic as usually assumed. Literacy. This variaale appears to have a positive effect on crop yield and crop output as well as on aggregate output and fertilizer demand. Literacy clearly has a negative effect on crop area. Moreover, it :onsistently seems to reduce livestock output and tractor demand. Life Expectancy. An improvement in this variable appears to lead to increased area and greater tractor demand. Positive effects on crop output, livestock output, aggregate outpuc and fertilizer demand appear in most cases. Roads and Pavement. The most robust effect of both these variables is on tractor demand, with pa ement having a particularly strong effect. Livestock output is also unambiguously associated with improvements in this measure of road quality. The following effects are not always statistically significant, however: road density tends to increase crop area, crop output and fertilizer demand, and these effects accord well with a priori expectations. Less in accordance with 3 :riori expectations are the findings that road quality reduces crop aeea but increases yield, with the neg3tive area effect dominating the yield effect so that crop output and aggrega' output are also reduced. To summarize the results fo. the "within-country" estimators, fairly clearly and positive effects on agricultural output are found for GDP, population density, irrigation and life expectancy. Research, extension, literacy and the road variables either have little effect, once the other shifter variables are introduced, or are so poorly measured that this effect does not show up clearly. V. The Imoortance of Country Effects We now return to a comparison of the C and P specifications. Note, however, that P includes a country-specific variable, MPDM. This variable is not included in the C specification because it remains constant over time. Thus, the country effect, which is statistically significant represents variables over and above our measure of the physical potential, MPDM. The estimated coefficients of the two soecifications are presented in Table 3. These regressiors differ from those in Table 2 by the inclusion of interaction effects for research and ext-nsion variables. Only output supply equations are showr. The own price coefficien'.s in the output equations are negative in the P equations and positive, but weak, in the C equations. Since the difference between the coefficients in the P and the C equations is attributed to ccrrelaticn in the sample between the explanatory variable and the countrv effect it is inferred that a negative correlation existed between the country effect and the price. Tracing this to tne two comDonents of crop output. it is seen that blel. 1: Cgrn'x4I-IstI .f EI;lsti eftle. - Tlle Importance of Cointry Effect _~~~~~~~~~~~~~~~r i . I I( _ ~Area __ _- _ Crop Ohutput _ ____ Yleld __ _Aggre~gate Ouitput_ ComIt ry Coullt ry Coutlt ry Couut ry riab le Ef fect Iloo1eŽl Effect Pooled Eftect Ploole(d Effect Pooled o)p Price I/ (.04()8* -0.46:3*** 0.0703* -0.6005*** -0.o (78 -0.1032** vestock Price 0.0(93 0.5126*** 0.0021 0.4313*** -0.0528 -0. l0l7*** gregate Price (.0452 5991 -tLilzer Prl e 0.0221** ').06() 0.0093 0.0829* ().((24, (.()337 0.01(7 0.0391 lenitf tt Mihan Years )0.0491 -0.6553*** -0.06b9 -0.S067*** -.09()03** (.1181*** 0.03)5 -0.4920*** scare!i Cost -0.0357 0.0390 -0.1275** (.1510*** -0.0606 ().1(028*** -0.0287 0.0586 tenst oin -0.(7841** -0.1304 0.0113 -(0.3711*** (. 1064A** -0.200( 1*** -0.0298 -(.3249*** tetisl on x I.ttera.v ().005*** ~'.()II 0*** -(.(((I 0.0 1 32)** -0.0006A** (1.40214* 0.(((0( 0( )127*** search x Fxteislon '.0(()9() -0.0931*** --0.002, -0.(8(9*** -0.01124*** (.()005*** 0.0087 -(.(8(8*** fe :Expectancy i.;18(*** 2.8399*** 0.4924 0.8722*** -0.4129** -2.1584*** 0.4131) -0.3720 r iga t ton -0.7593** 0.9842*** 1.3566*** 1.8644*** 2.6444k** 0.7431*** 1.2428t*x 1.5014*** A1d1 0.0025 -0.0168 0.0037 0.0387 0.0242 0.03231 (0.0069 (.05'3 ** v,menlt -0.(104 0.0033*** -0.0001 0.00/,9*** (.(()( (i.(022*** 0.0003 0.0046k** ulat ioll RI sLty 0).1428*** 0.33)16*** 0.2713*** 0.4976*** 0.1556**A () IH68*** 0.39161** ().881* ilt Literacy 0.0032*** (.(718*** 0.0048*** 0).11()*** (.(((5 0.(41 5*** (.0028** 0. ln)6 P -0(.0199 -0.0197 0.177G*** 0. 1836*** (.2(99*** 0.1760*** 0.2248A** 0.6103*** D1) 0.24-5*** 0.296'3*** (.030 3*** 0.24468**A aik 73 17 73 17 71 17 72 16 atistical Rank 68 10 h7 15 60 12 66 13 Va uies 2/ (.99 (.87 (.99 (.98 (.98 0.79 0.99 0.91 Ls: * **, statistical ly S1g ..1fic at l.(l, 5% aI)(1 1 level. cespec( tvely. / Prices are unastired tit constant dollars using pur-hashlig .ower pa rity exchl.tige ra tes. 2/ R2 values are tLiosc of OLS regro-,sloos . wle onies for 1rwinc ipal componlenits regrcisionorw are w,arly fIleniticica -20- the difference in the price coefficients is considerably larger in the area equation. On the surface it may suggest that large area and high vield cause lower prices. This suggests a simultaneous equation bias due to the omission of a demand equation from the analysis. This possibility is not pursued here. The present discussion deals with other issues. The coefficient of fertilizer price is positive in all equations, though in most c?sc, ;t is not significantly different from zero. Furthermore, a comparison of ,;e C and P equations indicates that in all cases the bias due to country effect was positive. It thus aopears that low output prices, high fertilizer prices and positive court;y effects go together. This is particularly so for The acreage equation as indicated above. This implies that low price ratio of output to fertilizer was combined with strong acreage response. Since international Prices are common to all countries, and abstracting from differences due to transportation, it appears that countries with larger ccuntry effect in the area equation taxea agriculture more heavily. It is clear that we were unable to reproduce Peterson's results. In fact our cross-country comparison gives implausible negacive coefficients.16 are vari, equa anal - l _ -21- different from zero and their. etfect on output is positive. However, when country effects are taken into account the results chaige qualitatively and quantitatively. Many of the coefficients bec insignificantly different from zero, some change signs, and the othe^s change magnitude. Thus, clearly there is a correlat,on between the country effects and the various explanatory variaoles. Only the role of population density and GDP are largely unaffected by the introduction of country effect. Furthermore, quantitatively, the value of the coefficients of GDP is ch2nged relatively little in all equations except for aggregate oLtput whose coefficient in the P equation is considerably higher This is not the case for crop output. It is therefore inferred that the difference is due to livestock output indicati::g a positive correlation between the overall availability of capital and the production of livestock. On the other hand, population density is positively correlated with the country effect in the area equation and that effect carr.es into the crop output equation. In the within-country regressiors, irrigation has a large positive coefficient on yield but a negative one on area; the net effect on crop cutput and aggregate output is --ill a positive coefficient. Comparison with the pooled regrension indicates that irrication is positively correlated with the c)untry effect in the area equation, but negatively correlatid with the rountry effect in the yield equation. What this means is that countries where irrigated area constitutes a large proportion of their total arable land tend to have a negative shift in the yleld equation, i.e., other things equal, have lower yields. Conversely, other thi,gs equal, 7 L -22 they utilize more area. One interpretation of this is that countries with low yield poter,tal have responded by constructing more irrigation. Finally, paved roads and literacy are positively correlated with the country effects of all equations. To conclude, the cross country variations in output are largely accounted for by the infrastructure and capizal avail-bility a.; represented by the various shifters. We have been unt le to detect a positive supply response from the cross count-y comparison. Moreover, as explained in the following section, we question the interpretation of a price coefficient obtaired from such equation to represent the long run response. Thus, as far as price response is concerned, what we have is the short run effect as explained in Sectioi, IV. VII. On the Supply in the lona Run Because shifter variables have such strong effects on c.tput, explaining the long-run suply requires an analysis of the determinants of change in the shifters thernselves. This is not done he,'e empirically. However, we can use the theoretical framework to comme-c on the process. With some qualifications, equation (8) can serve to link between the short jnd the lcng run. The qualifications are: 'I' the equation only applies to the private input ; (2) it is obtained within a comparative statics framework. Since comparative statics ana.lysis is timeless whereas observations are dated, there is a need to link between the two. -23- Cnanges in aggregate supply involve chaniges in resources. Much of the increase in agricultural output was obtained by expansion of capital inputs. In relating such an expansion to prices it should be noted that the total investment in the economy is bounded by overall savings. Therefore an increase in agricultural investment in the economy is bounded by overall savings. Therefore an increase in agricultural investmert takes place at ,e expense of investment in non-agriculturs and as such it depends on the differential rate of rct!'nn between agriculture and non-agriculture, as well as on other variables. This process is not -evealed by comparative statics analysis. To captLre it, a different analysis is required as illustrated in Mundlak (1979) and Cavallo and Mundlak. A change in agricuitural prices affects the rate of returns in agriculture and thereby its share in total investment. The essence ol this process is that an increase in the pace of agricultural investm'it has an increasing cost due to the increase in the shadow price of capital goods. Referring tc (8) that can be expressed in terms of low values for the factor supply elasticities (sj) related to capital inputs. But it should be noted that within this framework sj is a function that depends on the oace of agricultural investment which, by itself, is endogenous within the economic process. Also, the level of employment in agriculture cannot be explained within the framework of comnArative statics. Since agriculture is a declining industry, it generates surplus labor as reflected by lower returns to labor in agriculture than in non-agririlture. As a result, a Jynamic process of off-farm migration determines the labor supply in agriculture (Mundl-'k 1979). The time rate of migration -24- depends, among other things, on the magnitude of the intersectoral wage differential. This is the major channel through which changes in agricultural prices affect labor supply. Again, putting it in terms of elasticity of labor supply as required by (8), th2 value of such elasticity will depend on the length of period under consideration, and in any case will be endogenous. All this leads to the conclusion that using (8), tne move to long-run elasticities cannot be done by innocently assuming all factor supply elasticities to become very large, and consequently obtain a large product supply elasticity. The above discussion dealt with the long-run response arising from the supply of private inputs. The empirical analysis indicated that the public inputs have had a very substantial effect on supply. The levels of such public inputs are not determined within the same framework which applies to the private inputs. We do n. have, at the present, a framework to explain the leve of public inputs and it is therefore taken '.o be exogehous in this analysis.Note, howVever, that if indeed countries follow policies of applying public inputs in agriculture to compensate for discriminating price policies toward agriculture, it is possible to generate negative supply response from crcss-country analysis, as was the case in our analysis. This can be derived immediately by a proper interpretation of (5). Finally agricultural output has historicallv been strongly affected by technical change. Recall that our empirical findings, once account was taken of shifter variables, detected no time effect on supply. Also GDP, our measure of comprehensive capital, performed empirically as well as a time trend. This is not a coincidence. As -25- explained in Mundlak (1985b), the process of technical change is closely related to that of capital accumulation. his process nas two major ascects -- tFe invention of new techniques and their implementation. Both aspects, but especially implementaticn, deoend to a large degree on the pace of capital accumulatior. To conclude the discussioti, the study of long-run supply response requires a substantially more detailed investigation of private agricultural investment and migration. It also requires an analysis of the process by which governments allocate resources to the shifter variables which appear in the single-equation supply function of the sort analyzed above. The simple aggregate supply functions, when estimated from the within-country variations, produce very low short-run and apparent "long-run" elasticities with respect to price. The short-run input demand elasticities with respect to output and own-pricR are higher than the "shcrt-run" aggregate output supply and these results are consistent with conventional theory as presented in Section iI. References Askari, H. and Cummings, J.T., Agricultural SuDoly Resoonse, New York, Praeger Publishers, 1976. Bapna, S.L., Binswanger, H. P., Quizon. J.B., "Systems Output Supply and Factor Dem&id Equations for Semi-Arid Tropical India," Indian Journal of Agricultural Economics, 39(1984): 179-202. -26- Binswanger, H., Mundlak, Y., Yang M.C. and Bowers, A., Estimates of Aggregate Acricultural Suooly Resoonse frcm Time Series of Cross Country Data, EPO;C Division Working Paper No. 1985-3, The World Bank, Washinaton, D.C. 1985. Boserup, E., Cond"tions of Agricultural Growth, Chicago, Aljine Publicaticns, 1965. Brandow, G.E., "Demand for Factors and Supply of Output in Perfectly Competitive Indus,ry," Journal of Farm Economics, 44(1962): 895-9. Buringh, P., van Heemst, H.D.J. van Heemst and Staring, G.J., "Computation of the Absolute Maximum Food Production of the World," Agricultural University, Wageningen, Netherlands, January 1975. Cavallo, D. and Mundlak., Y., Agriculture and Economic Growth in an Ooen Economy: The Case of Argentina, IFPRI, Washington, D.C., 1982. Caves, Douglas W., Christensen, L.R. and Diewert, W.E., "Multilateral Comparisons of Output., Input and Productivity Using Superlative Index Numbers," Economic Journal, 92(11982): 73-86. Colman, D. and Rayner, A. J. "Feed Demand Flasticities and Agriculture's Aggregate Supply Elasticity," Journal of Agricultural Economics, 22 (1971): 125-9. Danin, Y. and Mundlak, Y. "The Introduction of a New Technique and Capital Accumulation," Rehovot, The Centre for Agricultural Economic Research, Working Paper No. 7909 (1979). -27- Floyd, G.E. "The Effects of Farm Price Supports on tna Return tc Land and Labor in Agriculture," Journal of Political Econony 73 (1965): 148-58. Friedman, M. Price Theory, Chicago: Aldine Publishing Co., 1962. Griliches, Zvi "Estimates of the Aggregate U.S. Farm Supply Funiction," Jourral of Farm Economics, 42 (1960): 232-93. Johnson, D. Gale "The Nature of the Supply Func:ion for Agricultural Products," American Economic Review, 40 (1950): 539-64. Kravis, I. V., Heston, A.W. and Summers, R. "Real GDP Per Caoita for More than One Hundred Countries," Economic Journal, 88 (1978): 215-42. Mundlak, Y. "Empirical Production Function Free of Management Bias," Journal of Farm Economics, 43(1961): 69-85. . "On the Pooling of Time Series and Cross Section Data," Econometrica, 46 (1978): 69-85. . "Long Run Co.fficients and Distributed Lags Analysis: A Reformulation," Econometrica, 35 (1967): 278-293. . "Endogenous Technology and the Measurement of Productivity," Rehovot, The Center for Agricultural Economic Research, Working Paper No. 8410, 1984. _ "Capital Accumulation, The Choice of Techniques and Agricultural Output," P-hovot, The Center for Agricultural Economic Research, Working ' er 8504, 1985a. Peterson, W.L. "International Fa 'rices and the Social Cost of Cheap Food Policies," Amer j Journal of Agricultural Ecoromics, 61 (1979):12-21. DISCUSSION PAPERS AGR/Yesearch Unit Report NC.: ARU I AgricuLtural .xachanizacion: A Cotwparative His:orical ?erspective by Hans P. 3inavanger, October 30, 1982. Retort No.: ARU 2 7he Acquisition of tnformation and the Adoption of New Technology by Gershon Feder and Roger Slade, September 1982. Revort No.: ARU 3 Selecting Contact Farmers for Agricultural Extension: The Training and vNsi: System in Haryana, Tndia 3y Gershon Feder and Roger Slade, August 1982. Reoort No.: ARU 4 "he impacc of Atcitudes Tovard Risk on Agricultural Decisions in Rur:3 Indla by Hans P. 3'nswanger, Dayanatha Jha, T. 3alaramaiah and Donald A. Stllers, May 1982. Ptport No.: ARU 5 3ehavior and Material Determinants of Production Relations in Agr';ul:ure by Han" ?. Binsvanger and Mutrk R. Rosenzweig. June 1982 (Revi-ed 'uLy 22, 1985). Revised Dece=ler 1985. Report No.: ARU 6 The Demand for Food and Foodgrain Qualicy In :nd'a b' ians ?. 3inswanger, Jaine 3. Quizon and lurushri Swa=y, November eoort-: No.: ARL 7 ?:si.cv ltmOLcations of Research on Energy lincae 3nd Activity Leve5 att:" 'eference co the Deoate of :he Energy Adecuacy if Existing L.e:; Deveo:pntg Countries ti Sho=o Reu:1tnger, May !983. Reoort No.: ARU 4 More Effec:ive Aid to the World's Poor and Hungry: A Fresh Look at *ni:ed Staces Public Law 480, Title rr Food Aid by Shlomo Reuctinger, June 1983. Report No.: A.Rt 9 .actor Gains and 'osses in the Indian Semi-Arid Trooics: A Didactic Aporoach to Modeling che Agricultural Sector by Jaime 3. Quizon and Hans P. Binswanger, September 1983 (Revised May 1984). Revort No.: ARU 10 rhe Distribucion of Incoze in india's Northern wheat Region 3y Jaime 3. Quizon, Hans P. Binswanger and Devendra Gupta, August ;'93 (Revised june 1984). Discusstcn ?1:8rs (ConMC!ued) Report No.: ARU 11 Populacion Densicy, Far-ing Intensity, Patterns of Labor-Use and Xtechsalgtison by Prabhu L. Pingali and Hans P. Binswanger, Sepcem6er 1983. Report No.: ARU 12 The Nucricional Impact of Food Aid: Critria for the Selection of Cosc- Effective ?oods by Shlomo Riurlinger and Judith K.%tona-Apte, Serceber 1983. Report No.: ARU 13 Project Food Aid and Equitable Growth: Income-Transfer Effic'ency 't.st' by Shlomo keutlinger, August 1983. Report No.: ARU 14 Nucricional Impacc of Agricul:ural Projects: A Conceptual Framework for Modifying the Design antd -plementation of Projects by Shlomo Reutlinger, Augusc 2, 1983. Revort No.: ARU 15 Pacterns of Agricultural Projection by Hans P. Binswanger and Pasquale L. Scandizzo, November 15, 1953. Report No.: ARU 16 Factor Costs, Incom and Supply Shares in Indian Agriculture by Ranjan Pal and Jaime 4ut,ij , December 1983. Reoor: No.: AR' 17 Behavioral and Macerial Determinants of Productcon Relacions in Land Abundant 7ropical Agriculture by Hans P. 3inswanger and John Mcrntire, January 1984. Renort No.: ARU 1I The Rela:ton 3etween Far= Size and Farm Productivity: The Role o: :i- Yv ov Gershon reder, December 1983. Reoort No.: AR' 19 A Comoarac:ve Analysis of Sore Aspects of :-e Tr3islng 3nd '.stI Sv;e2 of Agrlcul:ural Extension in India by Gershon Feder and Roger Slade, February 198 Reoort No.: ARtS 20 Distributional Consequences of Alcernactive Food Policies in India by Hans P. 3inswanger and Jaime 3. Quizon, August 31, 1984. Report No.: ARU 21 Income Distribution in India: The IDpact of Policies and Growth in :'e Agriculcural Sector by Jaime 3. Ouizon and Hans P. Binswanger, Movember 1984 (Revised October 1985). -30- Dtscuss'or. 3:^erS (continued) Reoorc No.: ARU 22 Population Density and Agricultural Intensificartion: A Study of che Evolution of Technologies in Tropical Agriculture by Prabhu L. Pingali and Hans P. Binswanger, October 17, 1984. Revort No.: ARU 2' The Evolucion of Farming Systems in Agricultural Technology in Sub-Saharan Africa by Rens P. Binsvanger and Prabhu L. Pingali, October 1984. Reporc No.: ARU 24 Populacion Density and Farming Syscems - The Changing 'Locus of Innovations and Technical Change by Prahbu L. ?'ngali and Hans P. 3inswanger, October 1984. Report fo.: ARU 25 The Training and Visitc Extension System: An Analysis of Operations and Effects by G. Fader, R. H. Slade and A. K. Sundaram, November 1984. Reoort No.: ARU 26 Tha Role of Public Policy in che Diffusion of New Agricultural 7echnology by Gershon Feder and Roger Slade, October 1984. Report fo.: AR1U 27 Fertilizer Subsidies: A Review of Policy Issues with Special Enohasis on Western Africa bv Haim Shalit and Hans ?. 3inswanger, November 1984 (Revised Nove.-ber 1985). Reoorc No.: kRU 23 From Land-Abundance co Land Scarcitcy: The Effeccs of ?oulaction ;rowc, on Production Relactions in Agrarian Lconomies bv Mark R. Rosenzweig. Hans P. Binswanger and john Mcrntire, .ove ber 1994 Reoorc "o. -Rt 29 The !mDacc E Rural Electrification and Infrascrucc-re on AgricuL:ura! Changes in India, l9 ;-1980 by Douglas F. Barnes and Hans P. 3inswanger, December 1984. Report No.: ARU 30 Public Tractor Hire and Equipment Hire Schemes in Developing Countries (with special emphasis on Africa). A study prepared by the Overseas Division, National lnsticute of Agricultural Engineering (OD/NIAE) by ?. J. Seager and R. S. Fieldson, November 1984. Report No.: ARU 31 Evaluacting Research System Performance and Targecing Research in Land Abundant Areas of Sub-Saharan Africa by Hans P. 3inswanger, January 1985. .P * * A (conctnued) Reoort -No. ARU 32 On che I ovision of Extension Services in Third World,Agriculture by + cair J. Fiacher (Consultant), January 19.85. Report No,: ARU 33 An Economic Appraisal of Withdrav'ng Fertilizer Subsidies in India by Jalme A. Quizor, April 1985 (Reaised August 1985). Report _o.: A1D 3A The Impact of Agricul:ural Extension: A Case Study of the Training and Visit Method (T&V) in Haryana by Gershon Feder, Lavrence J. Lau and Roger H. Slade, March 1985. Report ye.: ARUI 35 Managing Water Managers: Deterring Expropriation, or Equity as a Concrol Ischanism by Robert Wade, Aprill 1985 Report No.: ARU 36 Common Property Zesource Management in South Indian Villages by Robert Wade, April 1985. Report No.; ARU 37 On che Sociology of irrigacion: How do we Know the Truth about Canal Performance? by Robert W-de Mway 1985. Report No.: ARU 38 Some Organizactions Concerned with Animal Traction Research and Develoomert In Sub-Saharan Africa by Paul Starkey, April 1985. Reoort No.: ARU 39 The Economic Conseouences of an Open Trade PoLicy for Rice in t.nJ a by 'a'-e 3. 'u'zon and James 3arbieri, 'une 935. Repor: No.: ARU 'O Agriculcura' Mechanizacion and :he Evolution of 7arming Syscems in Sub-Saharan Africa by Prabhu L. Pingali, Yves 3igot and Hans P. 3inswanger, May 1985. Report No.: ARU 41 Zastiasian Financial Systems as a Challange co Economics: The Adventures of -Rigidiry,' with particular reference to Taiwan by Rober Wads, June 1985. Reporc No.: ARU 42 Educacion, Experience and Imperfect Processing of Information in the Adoption of Innovactions by Alastair J. Fischer, June 1985. Report No.: ARU 43 A Review of che Litceracure on Land Tenure Systems in Sub-Saharan Africa by Raymond Noronha, July 1985. 1 - . -32- Discussion Pacers (con:inued) Resor: No. A.RU 44 Poiicy *o:ions f. r - 3od S . u-r . ny Shioro Rulilnger, 'uiy 1985. Reoor: No.: AR!! *5 Credi;: arKecs in Rur31 Sou:h 'ndia: Theorecical *ssues and Z:mvl:ca! Analys is. vy i. 3inswangar, 7. 3a!ara3aiah, V. 3ashkar Rao, ''.J. 3hende and K<.'. Kas.n,rgagar, :_v;?i5. Reoor: No.: AR" 46 .he inpac: Df Agricul:rali Ex:ension: The Training and Visi: Svs:an in india. by Gershcn .eder and Rogar Slade, June 1985. Re-or: No.: AR!U 47 etchodological Issues in :he 7valuacion of Ex:ension Impact. by Gershon Feder and Roger Slade, July 1985, Reror: No.: A.?U 48 stinacion of Aggregaze Agricul:ural Supply Response. by ?Hans 3inswanger, Y;ir Mundlak, Maw-C'eng Yang and Alan 3owers Augusr 1985 (Revised Oc:ober 1985). 7enorz Nfc.: URU 499 land Values and Land ':zle 5ecurity in Rural Thailand. by Yongyu:h Chala=wong and Cershon Feder, Tune ;985 (Revised Oc:ober Reoor: No.: ARU 50 Land 3wnersmi:) Securi:y and Capital For=a:ion in Rur3l Thaaiand. bY 5ershon Feder and Tongro' Onchan, December :985 fRevised February 1986). Rae ort 'No.: A.RU 5 1 -and Ownership Security and rar- ?roduc:ivi:v in Rural -hai.land. by Gcershon Feder, April 1986. Resort No.: ARU 52 Social and Culzural As-ects of Land >rnheri:ance and ransactions in ?.ural lr-ailand. by Charles 3. Yell. ,_.e 1956. Discussion Papers (continued) ReDort No.: ARU 53 Land 0-nership Securi 3'nd Accrss -o Credi. in Rural Thailand by Gershon Feder, Tongroj Onchan and Tejaswi Raparla, April 1986. Reoort' No.: .AFU 54 The Management of CGomon Property Resources: Collective Action as an Alternative to Privatizacion ^r Statc Rezilation by Robert W4ade, May 19i6. ReDort No.: ARU 53 land PoLicies and Farm ? r ).ictvi:v in ThailarndJ's Forest Reserve Areas by Gershori Feder, Toi.zrOj ½nchan and Yongyu,h Chalanwong, August 1986. Report No.: ARU 5f On the Determinants of Cross-Country Aggregate Agricultural Supply by Hans P. Binswanger, Yair Mundl.ak, Maw-Cheng Yang and Alan Bowers, September 1986.