AN ANALYSIS OF THE DYNAMICS OF TECHNOLOGICAL CHANGE IN AGRICULTURE. Yoav Kislev

Transcription

1 AN ANALYSS OF THE DYNAMCS OF TECHNOLOGCAL CHANGE N AGRCULTURE Yoav Kislev Technological improvements are generally conceived in economics as identical increases in prodctivity for all prodcers. Reality is more complicated-the diffsion of new technologies is generally associated with differential changes in the scale of prodction, with entry into promising new lines of activity, and in the exit of prodcers who cannot keep p with the technological pace and are forced ot by deteriorating terms of trade. These are the economic processes that revoltionize the technological and social strctre of the rral commnities in the process of development. This paper otlines a few assmptions and findings of a theory of strctral change associated with technological improvement and illstrates it with developments in the dairy and the poltry indstries in srael. Theoretical Otline The theory is presented in terms of the dairy indstry. Eqation (1) defines the density of cow distribtion: (1) n(m,t) = N(m,t)N*(t), where: m (O~m~l) is the index of the level of management of the dairy operation; N(m,t) is the nmber of cows in year t in herds of management level m; N*(t) is the total nmber of cows in the contry; and n(m,t) (O~n~l) is the density of cow distribtion by management level. Let milk yield y be a fnction of the genetic level of the herd G(t) (both measred in kilograms of milk per cow per year) and management level m: (2) y(m,t) = mg(t). n srael, all cows are bred by artificial insemination. t is therefore reasonable to assme that the genetic potential of all herds is the same. Cows differ in their genotypes, bt these differences wash ot at the herd averages. Actal, realized yield differences stem from differences in operator management abilities. The indstry's milk spply will, therefore, be: 1 (3) Q(t) 0 y(m,t)n(m,t)dm = G(t)N*(t)J~n(m,t)mdm G( t)n*( t)e(m, t). E(m,t) is the contry average management level (the average weighted by the distribtion of cows by management levels.) 34

2 ncome of the better farmers-those with higher management abilities-will be higher and they will tend to expand their operations; the worst farmers are will lose and have to exit. These developments will be reflected in the factor determining spply expansion, which can be seen by differentiating the second line in eqation (3): 1 (4) Q = N* 0 G(t)n(m,t)mdm [expansion effect] 1 + G(t)N* f 0 n(m,t)mdm [genetic effect] l + G(t)N* J 0 n(m,t)mdm [concentration effect]. The expansion effect is the increase in spply reslting from a change in the size of the national herd, leaving constant the genetic level and the distribtion of cows by management level. The genetic effect is the increase in prodction de to genetic improvement, holding constant the size and the distribtion of the herd. The concentration effect is the increase in milk prodction de to the shift of the prodction between operators of different management levels. Empirical estimates of eqation (4) are presented in the next section. Empirical llstration Concentration of prodcers is affected by the terms of trade, defined here as the ratio of the price of the prodct to the price of feed concentrate. Another factor which is important in determining the dynamics of the indstry is the proportion of prchased inpts in total cost; this last proportion is termed ':he selection stress. The tighter the selection stress, the stronger the financial effect of a change in the terms of trade and the faster the adjstment process to this change. (Government intervention in the credit market- -prevalent in srael--mitigates these effects and modifies the dynamics of the indstry.) Table 1 reports the terms of trade and the selection stress in eggs, broilers, and milk prodction and their changes between two years-1967 and Terms of trade stayed constant in egg prodction, deteriorated sbstantially in the prodction of broilers, and changed slightly in the milk indstry. The selection stress is the tightest in broilers; eggs are second, and milk is last. Figres 1, 2 and 3 depict the cmlative distribtions of prodction by efficiency levels for the years 1967 and 1972 (1971 for milk prodction). Efficiency, a measre of the level of management, has been defined in the dairy indstry as annal yield per cow, and for the poltry indstry as the nmber of eggs per tonne of feed (with meat from the egg prodction flocks converted into eggs at the rate of 1 kilogram of meat eqal to 20 eggs), or the nmbers of tonnes of meat per tonne of feed in broiler prodction. These definitions are jstified by the assmption of constant marginal prodct to feed ratios and constant fixed costs. Figre 1 indicates exit of the lower tail of the distribtion between 1967 and 1971 (compare the broken line to the solid line) and improvement in yields in the pper tail. Average yield rose over that period from 5,039 to 5,312 kilograms per cow per annm. Thogh potential genetic improvement is the same in all herds, it is only the better prodcers that realize yield increases. The relatively low selection stress and the jst slight deterioration in the term of trade (de in part at least to government intervention) permitted many of the weaker dairy operators to stay in the indstry-the graphs for 1967 and 1971 coincide in the moderate efficiency zone. 35

3 Table 1: The Terms of Trade and the Selection Stress. Prodct Terms of Trade (Ratio of the price of 1 nit of the prodct to 1 kilogram of concentrates) Terms of Trade in 1972a Selection Stress (Ratio of otlays on prchased inpts to vale of prndct)b Eggs egg Broilers kg Milk kg a b at the sample mean Sorce: nstitte of Farm ncome Research, Tel Aviv. Table 2: Components of Spply ncrements. Milk Broilers Period Prodction increment a(%) ~ Partition of increment b(%) Total Expansion Concentration Genetic 9.9 Residal Kislev and Rabiner, 1979, and Michal Reiss, Master Thesis, in prepar Sorce: ation. a End of period mins first year over first year's level. b See eqation (4). 36

4 100 '";;; :> r< s ~ " ~ 60 - ~ 0 " 0 " 40 - f '" ' " kg milk per cow Figre 1: Distribtion of Milk Prodction. 'r- '";;; ,_, :> 80 s "... " ( $.,_, " 60 0 } "" 0 " '" 40 " 0 <J " '" 20 '";;;.,_, :>... $ " ,_, " """ '" " 0 40 '" 20 j (eggston feed) Figr~ 2; Distribtion of Egg prodction Figre 3: (ton meatton feed) Distribtion of Broiler Prodction 37

5 The pictre for the poltry indstry (figres 2 and 3) is more complex and more interesting. Average prodctivity in the contry increased in broiler prodction and deteriorated in egg prodction--the last from 4,608 to 4,488 eggs per tonne between 1967 and Examination of the graphs leads to the impression that both high and low efficiency prodcers left egg prodction and only the mediocre stayed. There was a general increase in prodctivity in this prodct. Here it also seems as if the weaker prodcers either exited or improved the efficency of their operation. Faster technological changes may have attracted the more efficient operators into broiler prodction--those who can hope to reap an appropriate rent for their abilities to be technological pioneers. These prodcers may have left egg prodction. On the other hand, the technological stagnation in egg prodction made this line of activity comparatively simple. Everyone can prodce eggs, inclding prodcers with relatively low alternative costs. This permits the sector to operate even at low income levels, bt the efficient operators--those with better alternatives--leave. Concentration of prodction in the hands of the more efficient operators contribtes to increased prodctivity. Table 2 attempts to partition spply increments in the three lines of prodction to the effects of eqation (4). Efficiency is defined in terms of yield per cow or prodction per tonne of feed as in figres 1-3. The calclation will be illstrated with milk prodction. The national milk prodcing herd is divided (and this division was employed in table 2) into registered and nonregistered herds. The registered dairies are generally larger (mostly in Kibbtzim), have higher yields, and their yields-as a grop- increased throgh time. The nonregistered herds are mostly in relatively small family operated dairies with low and stagnant yields. The share of the registered dairies grew over the last two decades, from 30 to 50 percent of the national herd (Kislev, Meisels, and Amir). This growth is the measre of concentration sed in table 2. Total milk spply in 1974 was 210 perent higher than in 1960, at the beginning of the period considered. Of this spply increment (taken as 100), 73.5 percent was de to the mere increase in the nmber of cows. The concentration effect was calclated as the differential growth of the registered herd times the yield differences between the registered and the nonregistered herds. The estimate is that this effect is responsible for 7.4 percent of the spply increment. The genetic effect was calclated nder the assmption that the genetic process was taken from breeders' estimates prepared in progenity tests. The residal, the complement to 100 percent, can be attribted to general technological changes: improvements in hsbandry, veterinary practices, and the like. The calclation can be smmarized in rogh terms as indicating that 30 percent of milk spply increment over the 15-year period was de to prodctivity improvements. Of these, a third was de to each of the prodctivity components: concentration effect, genetic effect, and more general technical change. n the poltry indstry, the expansion effect was taken as the increase in feed inpt times the base year prodct to feed ratio. Unlike the sitation in milk prodction, estimates of genetic progress in the poltry indstry have not yet been prepared. The residals in the broiler and egg colmns inclde, therefore, the genetic effects. These estimates of the residal technical change indicate, if we accept the method sed, that there cold not have been sbstantial' genetic progress in broiler prodction. This finding is somewhat pzzling and will probably not be accepted by breeders who claim to have introdced several sperior types dring the peri-od covered by the analysis. On the other hand, many changes in strctre and eqipment were introdced-mostly by the better 38

6 prodcers-while expanding the scale of operation (part of the concentration process), and it cold be that the efficiency gains reslting from these changes can be attribted to the genetics of the new lines. A better answer to this pzzle can be expected to come ot of a detailed stdy of poltry breeding in srael, now nderway. The empirical findings illstrate, trst, the significance of the dynamic perspective sggested by the theoretical assmptions of the paper. The work in this area is far from completion and several extensions are worth mentioning: (a) the explict analysis of a mltiprodct farming firm; (b) explicit recognition of government's role in shaping the dynamic processes in agricltre; and (c) incorporation of the effects of omitted variables, sch as alternative income opportnities otside agricltre. Reference Kislev, Y.; Meisels, M.; Amir, S. (forthcoming) The dairy indstry of srael, in Animal Prodction Systems (edited by B. L. Nestel). Amsterdam, The Netherlands; Elsevier Scientific Pblishing Co. 39