ASSESSING PRODUCTIVITY IN PRECOLONIAL AFRICAN AGRICULTURE AND INDUSTRY 1500-1800

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a widerange of content in a trusted digital archive. We use information technology and tools to increase productivity andfacilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available athttps://about.jstor.org/terms

University of Wisconsin Press is collaborating with JSTOR to digitize, preserve and extend

access to African Economic History

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTC

All use subject to https://about.jstor.org/terms

African Economic lIstory, No. 19 (1990-1991) 3 1 ASSESSING PRODUCTIVITY IN PRECOLONIAL AFRICAN AGRICULTURE AND INDUSTRY 1500-1800 By J.S. Hogendom and H.A. Gemery John Thornton's provocative paper raises questions concerning the productivity precolonial African agriculture and industry that have in general not addressed by the economic historians of that continent. His effort is therefore welcomed, as it is certain to stimulate a Useful debate on an important There are two major thrusts to Thornton's paper. First, he cites data on agricultural output to support a suggestion that the productivity precolonial African agriculture was high. Second, he points to the relatively advanced technology in certain African industries as evidence form of high productivity in that sector as well.1 The definitions of productivity employed by Thornton, and the utilizes, are the subject of this comment. The Contention That Productivity in African Farming Was First let us take up the contention that with the contemporary hoe agricultural yields in Africa were quite high. Thornton cites data that suggest that there was some basis to the earlier assertions that African even without the plow, was more efficient that that of early modern Europe" In another place, Thornton states "I do not believe that African productivity low in the period before 1800" (p. 5). Our concern with these statements is that the observations agricultural yields given by Thornton are possibly misleading, in that they in terms of yield from seed, or yield per unit of land. Neither of these able to capture the productivity of the constraining factor in African which was not seed or land, but labor. Thornton's evidence of high productivity based on two different measures, first that the seed to output ratio was hundredfold" in the seventeenth century, and that yield per hectare was 1 Thornton at several points qualifies his arguments with useful caveats. For consideration of productivity includes the appropriate warning that the productiveness economy involves substantial difficulties in measurement, especially in a context of amount of non-marketed goods and services. In general Thornton appears to believe appropriate concept for measuring the efficiency of Africa's output is labor productivity. with this approach, as is made clear below.This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTCAll use subject to https://about.jstor.org/terms

32 J. S. HOGENDORN AND H. A. GEMERY nineteenth century (p. 7). The problem with use of such measures not give information on the output per unit of labor, when constraining factor. The main limit on African agricultural output historically was relative scarcity of labor. That there was an abundance of un- or underutilized has been generally agreed by virtually every historian and economic has ever examined African economic history before the twentieth constraint has historically been the shortage of labor at times of peak chiefly the times of planting, weeding, and harvesting. At harvest in typical even in twentieth-century Africa to find almost every able-woman, and child who is capable of performing useful work to be The output per unit of labor is thus the key statistic in African agriculture, the agriculture of every society where labor is the scarce input. The productivity per unit of labor is, however, difficult to infer data describing yield from seed or yield per unit of land. Let us cases in sequence. (A.) Yield from seed does not serve to identify labor productivity. make a great difference whether two people could collect a given seed, or a hundred people, a number that perhaps could not even be given village. It must also be pointed out that on its own ground, data on yield from seed can lead to misunderstanding. The number cited, "a hundredfold" increase from seed for maize and millet, may or may not be better than elsewhere. Is this superior to yields from seed in other parts of the tropics? For example, in parts of the Americas, maize yields of 150 kernels from one seed would have been considered low, and much greater yields than that were reported.2 We also note that figures such as "a hundredfold" were most likely not averages, but single observations. Yields probably would have been much lower in years of drought, and would have varied greatly by region. In order to assess the magnitude of the variance, we consulted modem yield statistics for Gambia in the 1970s and 1980s at two-year intervals.3 In terms of yield measured in kilograms per hectare, the early millet yield was 1133 kgs. in 1974, 664 in 1976, 954 in 1978, 910 in 1980, and 1241 in 1982, a variance from lowest to highest of nearly 100 percent. The respective figures for maize were 1973 kgs. in 1974, 1082, 1396, 1057, and 1800, another variance of nearly 100 percent Quite similar variations appear for different regions even within a territory as small as that of 2 Femrnand Braudel, The Structures of Everyday Life (New York, 1979), 161. 3 See Government of Gambia, PPPU Statistical Estimates, 1974-1985, Banjul, various years. These data were compiled by James L. A. Webb.This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTCAll use subject to https://about.jstor.org/terms

RESPONSE TO THORNTON 33 Gambia. For example, the early millet yield in kilograms per 768 kgs. in Western Gambia, 700 in Lower River, 1216 in North in Upper River, the corresponding 1974 figures for maize were 946, and 2033. The variances here are 74 percent from lowest to and 281 percent for maize. Similar huge seasonal and regional drought, insects, plant diseases, soil fertility, and other causes common occurrence in many parts of Africa. It is not possible whether labor productivity, or indeed productivity in any form, the basis of single observations from a few scattered locations. Finally, yields from seed for maize, and, we believe, millet, higher than for grains such as wheat and barley, or yams, or Perhaps observers were getting a glimpse of the impressive revolution, with which they were yet unfamiliar. An agronomic revolution was that it undoubtedly raised the seed-yield ratio above crops such as wheat and barley. That does not, however, mean were higher for the same crop in Africa than they were in Europe the labor productivity was high. (B.) Statistics on output per unit of land in an African setting difficulties of interpretation. For example, a nineteenth-century bushels per acre, as cited by Thornton on page 7, gives productivity per person employed. We cannot tell if these produced by two people or by twenty. With labor the scarce economic problem to maximize the output of the scarce factor and factor, land, the basic question of African productivity is left unresolved. There are certainly reasons to believe that labor productivity rise significantly in precolonial Africa. Two main reasons contribute First, the large imports and domestic production of iron must introduction of better tools, which would have increased productivity. crop revolution that involved the introduction of new plants certainly allowed an increase in the output per worker for some maize as noted above. Yet on the whole the reasons for believing that labor productivity was lower than elsewhere on balance outweigh the arguments for high productivity. Tropical heat and humidity sap human strength.4 The disease environment, including especially the debilitating effects of malaria, has been much commented on by modem development economists, and must have been detrimental to productivity in earlier periods as well. Downpours wash out seed and erode farmland. Insect and 4 The case for intensity of labor input as a critical determinant of productivity in agriculture is made by Gregory Clark, "Productivity Growth Without Technical Change in European Agriculture Before 1850," Journal of Economic History, 47, 2 (June, 1987), 419-32.This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTCAll use subject to https://about.jstor.org/terms

34 J. S. HOGENDORN AND H. A. GEMERY plant diseases cut output. Furthermore, the existence of slavery presumably reduced productivity on two counts. Those farmers who were slaves worked with reduced intensity. They were less motivated than free supervision they received was usually incomplete. Those not in servitude, prey for the slavers, would have had to farm less accessible plots, in hilly country. They also would have had to spend some of the time have devoted to farming on the perpetual watching and waiting, and defenses. Each of these considerations would have detracted from labor productivity as measured by commodity output. The Contention That High-Quality Steel and Textile Production Indicates High Productivity A second difficulty in assessing productivity arises from the use of quality proxy, as when Thornton calls attention to the existence of indigenous high-steel and textile production as evidence of high labor productivity. The problem with making this link is that the existence of an industry capable of producing superior quality output is not evidence that physical output per unit of labor Much production of hand-made "luxury" goods is profitable not because physical output per unit of input is large, but because the price of the product high. Here the difference between the marginal physical productivity of labor and its related concept, labor's marginal revenue productivity, is important. The marginal physical productivity (MPP) of labor can be low, with some given input resulting in a low physical quantity of output, whereas the marginal revenue product (MRP) is high because the resulting output can be sold at a high price. MRP is MPP times the price of the product; if that price is high, then marginal revenue product may also be high, even though few units of output are produced. This is what we suggest may have been the case with precolonial Africa's steel and textile production. Production was profitable not because of high physical output per unit of input That was very unlikely given the shortage of physical capital that could have raised labor productivity. Instead production was profitable because the high quality given the supply and demand conditions led to a relatively elevated price. This would appear to be a good explanation why Africa's steel and textiles were not exported: because their prices were relatively high, these industries did not have a comparative advantage; imports from abroad made large inroads. To the extent that these industries survived international trade, it was not the high physical productivity of the factors used in them, but instead the quality of the output which gave them a continuing place in African consumption patterns.This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTCAll use subject to https://about.jstor.org/terms

ASSESSING PRODUCTIVITY IN PRECOLONIAL AGRICULTURE 35 Even the point that items can be made under conditions productivity of labor but high revenue product needs amplification. any high quality item made in Africa must have been limited, in many cases, by the relative absence of inputs complementary particular, transport and financial markets were underdeveloped. wheeled vehicles; the unavailability of horses and donkeys in sleeping sickness; and the rudimentary nature of credit mechanisms more obvious manifestations of constraints to transport and would have been less effective demand in African markets than would otherwise have been the case, and thereby a lower product price at the source of production and hence lower marginal revenue product of labor. Conclusion We would suggest that not enough is known about labor's physical productivity precolonial Africa to support the conclusion that it was high relative to other such as Europe. Neither occasional observations of yield from seed, nor of per acre, allow an assessment of labor productivity in agriculture, the productivity measure that counts where land is abundant and labor is Furthermore, technical ability or even brilliance that permits high-quality output be produced is not sufficient evidence to conclude, as Thornton does, that "steel making was productive relative to its pre-industrial European counterpart" 9). Nor does high-quality textile production mean that the labor used in output was high in productivity. Great variability in agricultural output, debilitating disease environment, the widespread existence of slavery and raiding, the lack of complementary transport and credit facilities, and the shortage of physical capital, are all prima facie reasons for believing that, on the contrary, African labor productivity was relatively low.This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:34:19 UTCAll use subject to https://about.jstor.org/terms

Permitted Uses of the Content 4.1 Institutional Licensees and/or Authorized Users may search, view, reproduce, display, download, print, perform, and distribute Licensed Content for the following Permitted Uses, provided they abide by the restrictions in this Section 4, Section 5 and elsewhere in these Terms and Conditions of Use: (a) research activities; (b) classroom or organizational instruction and related classroom or organizational activities; (c) student assignments; (d) as part of a scholarly, cultural, educational or organizational presentation or workshop, if such use conforms to the customary and usual practice in the field; (e) authors or other Content creators may at their discretion incorporate their Content (other than Books) into unrestricted databases or website with prior permission from the publisher and other applicable rights holders; (f) linking (g) Data for Research as defined further in Section 14.3 below. and for Institutional Licensees and/or Authorized Users other than Register & Read Users: (h) on an ad hoc basis and without commercial gain or in a manner that would substitute for direct access to the Content via services offered by JSTOR, sharing discrete portions of Content for purposes of collaboration, comment, or the scholarly exchange of ideas; (i) downloading Content (including certain Book chapters where downloading is made available for the Book chapters on the JSTOR Platform) to view, use and display including on a personal digital device (e.g. mobile devices, e-readers, and personal computers); (j) (except for Books) in research papers or dissertations, including reproductions of the dissertations, provided such reproductions are only for personal use, library deposit, and/or use solely within the institution(s) with which the Authorized User and/or his or her faculty readers are affiliated; and (k) fair use under Section 107 of the U.S. Copyright Act, educational exceptions, or other similar provisions of the copyright laws or other intellectual property right laws in the United States. 4.2 The portions of Licensed Content that have been downloaded or printed out by an Authorized User may continue to be used in compliance with these Terms and Conditions even if such license should later terminate or expire. 4.3 Additional Permitted Uses of the Early Journal Content. JSTOR encourages broad use of the Early Journal Content. In addition to all permitted uses described above, users are free to copy, use, and redistribute the Early Journal Content in part or in whole for non-commercial purposes. JSTOR asks that you acknowledge JSTOR as the source of the Early Journal Content; if you use material from JSTOR online, we request that you link directly to the stable URL provided. If you use Early Journal Content offline, we ask that you credit the source as follows: “Courtesy of JSTOR.” Please be considerate of other users and do not use robots or other devices or coordinate activity to systematically download these works as this may be disruptive to our systems. 4.4 Use of Open Access Content. Open Access Content may be subject to different terms of use, such as a Creative Commons license, as indicated in the copyright statement for the content. If an Open Access Content item has different terms posted where the content appears, those terms shall have precedence over these Terms and Conditions of Use.

Nelisiwe Ndlovu

04/26/2019 - 09:37