This paper examines gender differences in agricultural productivity in developing countries. Whereas gender alone is often an insignificant determinant of agricultural efficiency, there are significant differences in the distribution of underlying variables, i.e., education, access to land or other resources. This paper attempts to determine the nature and causes of these differences through analysis of (a) estimates of technical and labor productivity differences, (b) labor supply and earnings functions, and (c) determinants of the adoption of new technologies by male and female farmers.

The Study
The paper reviews studies of agricultural productivity in numerous countries in Africa, Asia, and Central and South America. These include studies by the World Bank, IFPRI, and Yale University Economic Growth Center. In order to address the differences in productivity arising from gender-based biases (such as lower levels of education and physical assets for women), this study relies only on those using regression analysis.

The studies reviewed used several models in their estimations. The technical productivity studies were analyzed through Cobb-Douglas production functions. Most of the wage-earnings studies used OLS-instrumental variables and Tobit analysis. The technology adoption studies used logit equations and probit estimates.

Whereas gender alone is often an insignificant determinant of agricultural efficiency, there are significant differences in the distribution of underlying variables.

Findings
The study found that men and women farmers are equally productive, once individual characteristics and input levels are controlled for. Only one study, in Burkina Faso, estimated that women were less efficient than men farmers. In this case, however, lower inputs into women's crops (which brought lower yields) resulted from highly asymmetric roles and household obligations.

The study found that women's productivity increases dramatically when they are given the equivalent of male inputs. Giving women at least a year of primary education would raise yields by 24 percent, according to studies in Thailand and Korea. Data from Sub-Saharan Africa suggest a 22 percent increase in women's yields on maize, beans, and cowpea plots if women farmers were given the human capital and input levels of male farmers.

Regarding wages, women were found to earn about half what men earned. The study suggested that physical size (as measured by height) may account for some productivity differences, since it may affect efficiency in some tasks such as land preparation.

The study also suggested that the disparity in earnings may reflect women's household obligations, a factor that has not yet been—but should be—studied in depth. Traditional agricultural wages, after all, may measure only a part of women's productivity when subsistence crops and other income-generating activities are included.

In examining the returns to schooling, the study found that education has a positive effect on wages for both men and women in cases where modern technologies have been introduced. Where women are restricted from participating in the market due to cultural factors, however, returns to female schooling are low. Returns to education are also low for rural Indian women, unlike rural women in the other countries studied, because wages in India depend mainly on local market conditions rather than on personal charac-teristics such as education.

Education was clearly the most important determinant of whether farmers adopted new technology, among both male and female farmers. Farmers with larger plots and higher values of tools were found to be more likely to adopt new technology than those with fewer resources. Other factors such as household size and age of the household head were found to be insignificant in affecting technological adoption. Evidence showed a significant gender-specific copying effect, meaning that women adopted technological innovations much more easily once other women living nearby had done so. The prevalence of female extension agents among female farmers was found to help with this process.

Policy Implications
The study suggests that educational attainment, land ownership, and the long-term nutritional status of men and women result from their parents' and families' decisions regarding allocation of resources over the long term. These decisions have life-long implications for the productivity of their children. Underinvestment in girls' education, for example, could lead to lower probabilities of female farmers adopting new technologies. Inheritance laws favoring boys lead to men having larger farms, and cultural biases favoring boys regarding nutrition could have significant effects on height and productivity. Although this paper does not analyze the processes that determine such life-cycle intrahousehold allocations, it strongly suggests that many of these factors can be addressed through policy interventions, and that further research is needed on these issues.