Pequena ou Grande Propriedade Rural, Quem é Mais Produtivo? - Artigo de Tomas Rotta

23/05/2010 16:51


Agrarian Structure, Institutions and Agricultural Productivity

Tomas Rotta


Agricultural productivity is the result of both technology and institutions. Agricultural productivity is a function of techniques, soil quality, level of mechanization, form of labor-contracting (family labor, sharecropping, hired wage labor, fixed rent tenancy) and of social interactions between workers, tenants and landowners. But how do these factors interact to engender higher productivity is a more complicated question. It is usual to think that large farms are more productive than smaller ones, especially due to increasing returns to scale stemming from increased mechanization. In many countries this argument is used to justify the status quo of huge inequality in land distribution. Ironically, in this respect, both old-fashioned Marxists1 and new-fashioned Neoliberals have one point in common: both sustain the claim that larger mechanized holdings are more productive and, hence, constitute the units to be prioritized by economic plans. Empirical investigations, however, suggest that it might not be necessarily the case. Sen (1996, p.462) states that in India the smaller family farms have higher productivity than the larger farms employing hired wage labor and Ray (1998, chap.12) provides extensive empirical evidence of an inverse relationship between farm size and land productivity.

To begin with, it is important to stress that “productivity” can be measured at least using two ratios: (a) production per land acre/hectare or (b) production per labor input. Therefore, one needs to make clear if agricultural productivity is land productivity or labor productivity, as each measure will yield different and even antagonistic results: “land productivity and labor productivity … move in opposite directions as the size of farm increases” (Griffin, Khan and Icowitz 2002, p.286). In the next paragraphs we will make explicit which one we refer to.

Following the arguments raised by Ray (1998, chap. 11-12), Griffin, Khan and Icowitz (2002), Griffin (1999), and Sen (1996) in what concerns land productivity, we should expect to observe two countervailing sets of forces: technological forces and labor incentive forces. The first set, which includes the application of advanced technology in agricultural production, accounts for increasing returns to scale in production (at least in a relevant production range). That is, in technological terms, mechanization operates to increase the output per labor input ratio and operates also to make the agricultural production process less labor intensive and more capital intensive.

However, the incentive set of forces poses a complex question: what form of land rental contract would yield higher labor productivity? Output is known to vary according to incentives deriving from who hires whom and from what specific form the labor-contracting assumes. Arguments were raised, for example, to show the hypothetical supply distortion of sharecropping via the so-called Marshallian inefficiency (Ray 1998, p.424-428), or for the superiority of fixed-rent tenancy for larger farms. This question is challenging because larger farms are usually non-owner occupied while family farms are owner-occupied. Non-owner occupied land implies that a form of rental will take place, such as fixed rent, sharecropping or hired wage labor. In this case it is known that at least three potential problems can arise: (i) informational problems (moral hazard, adverse selection etc.); (ii) principal-agent problems; (iii) non-verifiability, non-enforceability, or non-observability of the agent’s effort (Ray 1998, chap.11). Besides that, the institutional framework in agriculture matters because labor is the only productive input that has preferences of its own. Thus for example, risk-averse workers might prefer outcomes with lower variances, such as a wage contract or sharecropping in detriment of the uncertainties of the fixed-rent contracting. This happens because landlord and tenant are not just sharing output and input but also the risk of production. On the other hand, the landlord herself can be risk-averse and, therefore, would prefer to receive a fixed rent over the payment of a fixed wage for an uncertain output. In other words, when other dimensions of production other than just output is taken into account, such as the risk of the project, sharing of costs of inputs, tenant’s limited liability, supervision costs, or the need to insure the worker, the form of contracting will directly influence the productivity of both labor and land. This is why sharecropping is more pervasive than what economic theory predicts (Ray 1998, p.439).

Ray (1998, p.443-449) points to the important saying that landlords ultimately do not aim at increasing productivity per se but actually aim at increasing their own returns. These two goals are sometimes compatible, but not always. Often the increase in productivity implies a rise in the returns for the workers or tenants, but not for the landlord. This happens because there is a potential trade-off between the provision of incentives and the provision of insurance to workers and tenants, so maximal productivity will in general not be attained in non-owner occupied farms. In this way, it would be suitable to ask why small farms do not pool their small land together in order to benefit from the increasing economies of scale stemming from mechanization in large arable areas. The answer hinges on the fact that to some extent it would recreate the incentive problem of larger farms, and that it would also be potentially undermined by coordination problems and free riding. Accordingly, why large farms do not divide their lands into smaller holdings in order to benefit from labor incentives? Because tenancy contracts, necessary for non-owner-occupied farms, would recreate the labor incentive problems mentioned above and they would erode the productivity gains.

To sum up, even though larger non-owner-occupied farms can enjoy the benefits of increased mechanization and economies of scale, they would face the labor incentive problem. On the other hand, while smaller and family farms have better labor incentives, they lack economies of scale and mechanization, and also suffer from imperfect credit markets. Therefore, at least in theoretical terms we have distinct sets of forces that would both enhance and undermine productivity in both the larger non-owner-occupied and in the smaller owner-occupied farms. But theory itself cannot answer completely the question of which land size would yield higher productivity. In this sense, the final word would be given by empirical investigations.

As shown by Ray, empirical studies tend to favor the finding that the smaller family holdings, despite their disadvantage in profiting from technological economies of scale, deal better with the labor incentive problems and can, therefore, do better than the larger non-owner-occupied holdings in terms of land productivity. In other words, labor incentive outweighs economies of scale and, therefore, family farming in fact has productivity advantages that cannot be mimicked by contractual hiring of labor (Ray 1998, p.449, 453-455).

Moreover, according to Griffin, Khan and Icowitz (2002, p.286), empirical investigation demonstrate that: (a) labor productivity and farm size exhibit a positive correlation; (b) land productivity and farm size have a negative correlation. What factors account for this divergence of results? Here they point to the following main elements: (i) smaller farms are more labor intensive, and larger farms are more capital intensive; (ii) smaller farms are usually owner-occupied while larger farms usually employ hired wage labor; (iii) if smaller farms are in fact family farms, then the cost of labor for them is lower than the market wage for agricultural workers; (iv) labor-saving mechanization in larger farms contribute to higher labor productivity; (v) supervision costs are higher when labor is hired; (vi) small farmers cultivate land more intensively (i.e. higher cropping ratio) and generate more employment per unit of land; (vii) smaller farms favor a crop mix with higher value added; (viii) smaller farms have higher physical yields for individual crops. The price-of-labor argument listed above is an attempt to explain why the price of family labor is lower than the market wage, even when non-monetary income (such as share of the crop and in-kind payments) in taken into account (Sen 1996, p.456-457, 462). Several elements have been identified: existence of unemployment in labor markets; family labor protects entitlements, as family workers are claimants to the family’s property; discrimination against women and children in labor markets; preference of workers in favor of self-employment and flexible hours, and to the detriment of hired labor working for a boss; and avoidance of migration to other areas in order to receive a market wage (Griffin, Khan and Icowitz 2002, p.286). Moreover, the supervision-cost type of argument tries to explain why the laborer exerts more effort in smaller farms. In this case we can point to some important aspects: supervision costs are smaller in smaller farms; shirking is more common when labor is hired, and larger farms tend to use more hired labor; the ‘social distance’ (i.e. inequality in social status) between workers and bosses is lower in smaller farms.

Furthermore, even if we accept all the theoretical and empirical findings raised previously we could still expect market forces to reallocate land ownership from large to small farmers. That is, if indeed smaller units are more productive in relation to the bigger ones land markets could then reallocate ownership rights and, thus, increase aggregate levels of productivity, therefore equalizing marginal costs and marginal benefits across rural properties. But we have, again, both theoretical and empirical arguments to expect the opposite. Let us first remember that land owners are concerned about the overall rate of return, nor productivity per se. And that in reality we observe the opposite trend, namely that market mechanisms historically tended to concentrate land ownership even further in the hands of large landowners, such as in the extensive cases of ‘reverse leasing’ (Ray 1998, p.423). Alternatively, if land productivity is higher for smaller farms then why in so many countries land ownership and land holding are highly concentrated? The main point is that there are additional institutional factors at play other than the ones mentioned above that shape the way in which the means of agricultural productions are distributed and this, in turn, has direct effects on agricultural productivity.

A few arguments can be raised to explain this vicious circle. First, smaller farmers have fewer assets and fewer entitlements than larger farmers, implying that they are also credit-constrained in credit markets (Sen 1981, p.5). Therefore, in cases of crop failures, rise of input costs or other adverse external shocks, the smaller farmers can be compelled to sell their lands to the larger landowners. Second, agricultural policy has discriminated against small farmers. This ‘landlord bias’ occurs under many forms: research policies in favor of export crops and superior grains, agricultural price support policies, regional policies that favor more fertile and more accessible regions, public irrigation systems that favor large landowners, credit policies, institutional settings that discourage the organization of the poorer etc (Griffin, Khan and Icowitz 2002, p.284). Third, the market for land is fragmented, highly localized and its volume of sales is low (Griffin, Khan and Icowitz 2002, p.285). In fragmented markets the possession of large holdings of land gives landlords a monopsony power in local labor markets. Thus, land concentration implies both asset dispossession and lower wages for rural workers, higher rents for tenants and smaller share for sharecroppers. In other words, land concentration is one element in a broader system of labor controls. Monopolization of land in land markets allows large landowners to become price makers in labor markets (Griffin, Khan and Icowitz 2002, p.285-289 ; Ray 1998, p.410, 412). Fourth, in many countries the political power of large landowners, in the absence of a social revolution, is large enough to impede confiscatory redistribution of assets. Additionally, assets confiscation in one single country would entail international hostility (Griffin, Khan and Icowitz 2002, p.320).

Summing up, the previous paragraphs tried to show how theoretical and empirical arguments interact to explain the relationship between agricultural productivity and institutions. This relationship is clearly bidirectional and consists of a vector of forces that can both support and undermine agricultural productivity. Even though, at least one general conclusion was drawn, namely that labor incentives outweigh economies of scale as land size increases, thus favoring smaller farms in terms of land productivity, but not in terms of labor productivity. However, value added (net income) per hectare tends to fall as the size of farm increases, and these tendencies remain even when adjustments for soil quality are made (Griffin 1999, p.141).



1Marx, Lenin, Kautsky and Stalin were the main Marxist thinkers that tried to demonstrate that land and labor productivities and farm size exhibit a positive relationship. In the Soviet case this was used as one of the arguments in favor of the Stalinist forced collectivizations and extermination of the petty-burgeois kulaks during the 1930s (Nove 1993, chap. 7).



Amsden, A. H. 2001. The Rise of “The Rest”: Challenges to the West from Late-Industrializing Economies. Oxford: OUP.

Amsden, A. H. 1992. Asia’s Next Giant: South Korea and Late Industrialization. Oxford University Press.
Boyce, J.K. 1993. The Philippines: The Political Economy of Growth and Impoverishment in the Marcos Era. London: Macmillan

Chandler, A. D. 1994. Scale and Scope: The Dynamics of Industrial Capitalism. Belknap Press of Harvard University Press.

Chandler, A. D. 1993. The Visible Hand: The Managerial Revolution in American Business. Belknap Press of Harvard University Press.

Griffin, K.B. 1999. Alternative Strategies for Economic Development. London: Macmillan.

Griffin, K. and Icowitz, A. 1997. The Distribution of Wealth and the Pace of Development. UNDP Working Paper Series.

Griffin, K.B., Khan, A.R. and Ickowitz, A. 2002. Poverty and the Distribution of Land. Journal of Agrarian Change 2(3): 279-330.

Nove, A. 1993. The Economic History of USSR. Third Edition. Penguin Book.

Ray, D. 1998. Development Economics. Princeton University Press.

Rodrik, D. 1994. ‘King Kong Meets Godzilla: The World Bank and the East Asian Miracle,’ in A. Fishlow et al., Miracle or Design? Lessons from the East Asian Experience. Washington: ODC.

Schumpeter, A. 1997. Capitalism, Socialism and Democracy. Routledge.

Sen, A. K. 1996. Employment, Institutions and Technology: Some Policy Issues. International Labour Review 135(3-4).

Sen, A.K. .1983. ‘Development: Which Way Now?’ EJ 93: 745-62.

Sen, A.K. 1999. Development as Freedom. New York: Knopf.

Sen, A.K. 1981. Poverty and Famines. Oxford: OUP.