Soy harvesting in Brazil
Soy harvesting in Brazil
New technologies in Brazil increased agricultural productivity, but also reduced economic opportunities for women and increased fertility rates.
Editor’s note: For a broader synthesis of themes covered in this article, check out our VoxDevLit onFemale Labour Force Participation.
Economists and policymakers have long promoted agricultural technologies for their potential to drive development (Gollin et al. 2021); however, not much is known about their gendered impacts. This is particularly relevant as men and women often perform different roles in agricultural production (Boserup 1970, Schultz 2001), making them susceptible to the effects of these technologies in different ways.
The introduction of genetically engineered soy led to structural transformation in Brazil
Weed management is a significant component of crop production, as weeds compete for the resources needed for plant growth. Traditional cultivation techniques involved manual weed control, wherein labourers—predominantly women—used hand-held equipment to identify and remove weeds.
This all changed for soy farmers in Brazil in 2003, after the country legalised Monsanto's Roundup Ready genetically engineered (GE) soybean. GE soy is innovative for its natural resistance to glyphosate, a powerful herbicide that kills nearly all crops. This enabled Brazilian farmers to use glyphosate to manage weeds without affecting their soy crop, mitigating the need for tillage and weeding operations in soy production.
This technology was revolutionary for the Brazilian economy, as it promoted structural transformation by freeing up labour in agriculture (Bustos et al. 2016). This enabled farmers to save more, improving urban access to credit for manufacturing and service firms (Bustos et al. 2020). However, given the high degree of occupational sorting by gender in agriculture—specifically the fact that weeding is mostly performed by women—I hypothesised that this technological change disproportionately displaced female workers. My research (Moorthy 2025) thus analyses the gendered effects of soy technological change in Brazil, an early adopter of modern agricultural innovations.
Identifying the effect of technological change on labour market outcomes
There are several empirical challenges to overcome in order to identify the causal effect of soy technological change on labour market outcomes. Most importantly, I had to establish a counterfactual for regions that adopted GE soy. Simply comparing regions that adopted GE soy to those that did not would be problematic, as these regions likely systematically differ along other dimensions that impact male and female labour market opportunities.
To overcome these challenges, I relied on two key sources of variation. First, the 2003 legislation provides a clear before-and-after point for comparison. For variation across space, I use a measure of the predicted increase in soy production from these new technologies across different municipalities in Brazil based on their pre-existing soil and climatic conditions.
My main assumption was that municipalities that happen to have more or less favourable geo-climatic conditions for GE soy technologies are not systematically different along other dimensions that may affect my outcomes of interest. I further adjusted for any state-level policies or socio-economic factors that vary at the state level by only comparing municipalities to others within the same state. Assuming this assumption holds, differences observed after 2003 between high- and low-suitability municipalities would be attributed to GE soy adoption.
Agricultural technological change negatively affected women’s economic opportunities
My main findings are as follows: (1) new soy technologies reduced employment and earnings for women in agriculture; (2) affected women are unable to reallocate into other sectors of the economy and instead shift to unpaid work within agriculture; and (3) consistent with economic theory, these gender-specific labour market effects led to increased fertility rates. While this technological change resulted in large-scale aggregate productivity gains, my findings show that it also worked against two development goals: improving economic opportunities for women and promoting fertility decline.
First, I find that while this new technology increased overall household agricultural earnings, it also significantly reduced female earnings opportunities. The binned scatterplots in Figure 1 show the differential effect on agricultural earnings by gender; while female agricultural workers experience earnings losses, their male counterparts see earnings gains.
Figure 1: Effect of GE soy adoption on agricultural earnings by gender
Effect of GE soy adoption on agricultural earnings by gender
Effect of GE soy adoption on agricultural earnings by gender
Effect of GE soy adoption on agricultural earnings by gender
Effect of GE soy adoption on agricultural earnings by gender
I find that this displacement is driven by agricultural employees, suggesting that larger farms that hire wage labour are among the primary adopters of these new technologies and subsequently lay off their labour force. Further, I find that while men that may have been displaced by GE soy move into other sectors of the economy, women displaced by these technologies do not leave agriculture. Instead, they move from paid to unpaid work within agriculture, working either for another household member or directly for household sustenance.
This suggests that significant barriers exist preventing women from moving to other sectors, such as services or manufacturing. Interestingly, education levels do not explain this difference—both women with high and low education remain in agriculture, while men across education levels successfully move into non-agricultural employment. One possible explanation for this is that men's agricultural work equips them with more transferable skills, either due to the nature of their tasks or from employer/extension-based training programs targeted specifically at men (Waltz 2016). This may leave women less equipped to take on expanding opportunities in urban employment.
How did genetically engineered soy increase fertility in Brazil?
After establishing the labour market effects of GE soy adoption, I examine the demographic consequences, particularly on fertility decisions—one of the most consequential choices individuals make and one intrinsically linked to both development and female economic empowerment.
Economic theory points to two main factors that determine how changes in the labour market impact fertility decisions. Increased earnings relax budget constraints, making children more affordable and incentivising higher fertility. However, a key insight from Becker (1960) reveals that the ‘price’ of children largely depends on the opportunity cost of time required to raise them. Higher earnings opportunities hence increase the price of children, creating competing incentives to lower fertility.
When women assume most childcare responsibilities, changes in female earnings primarily affect the price of children, whereas changes in male earnings primarily act through relaxing the household budget constraint. Recent empirical work confirms this: higher female earnings increase the opportunity cost of childbearing, incentivising lower fertility (Jensen 2012, Heath and Mobarak 2015, Schaller 2016, Kitchens and Rodgers 2023), while higher male earnings increase fertility (Black et al. 2013, Kearney and Wilson 2018).
In the Brazilian context, GE soy adoption generated two reinforcing incentives towards fertility. Reduced female earnings lowered the opportunity cost of women's time, decreasing the effective price of children. Increased male earnings relaxed household budget constraints, making children more affordable. This led to higher fertility in response to GE soy adoption (Figure 2).
Following the 2003 legislation, regions with higher suitability for GE soy experienced increased fertility compared to regions with lower suitability. These effects persist for up to 17 years, indicating a sustained shift in fertility behaviour rather than a mere change in the timing of births.
Figure 2: Above versus below median predicted GE soy—Birth rate per 1,000 women, 16-49
Above versus below media predicted GE soy—Birth rate per 1,000 women, 16-49
Above versus below media predicted GE soy—Birth rate per 1,000 women, 16-49
Policy implications for agricultural technology and female labour market participation
The case of GE soy in Brazil serves as a cautionary tale for the unintended consequences of technological change. While these technologies deliver substantial productivity gains, they also exacerbate gender disparities and work against key development objectives. Recent work further finds that GE soy adoption reduced female land ownership (Araujo et al. 2024).
Similar patterns have been observed elsewhere: in India, Afridi et al. (2022) found that mechanisation in agriculture also disproportionately displaced women. This demonstrates that the impact of new technology on labour markets has specific implications for how family structure evolves, which may not necessarily support female empowerment or fertility decline. Agricultural technology must thus be accompanied by complementary policies that ensure its benefits are shared equitably and align with broader development goals.
References
Afridi, F, M Bishnu, and K Mahajan (2023), “Gender and mechanization: Evidence from Indian agriculture,” American Journal of Agricultural Economics, 105(1): 52–75.
Araujo, R, B Borges, F Costa, and K Santos (2024), “Seeds of disparity: The gender land divide from Brazil’s agricultural transition,” Unpublished manuscript.
Becker, G S (1960), “An economic analysis of fertility,” in Demographic and economic change in developed countries, Columbia University Press, 209–240.
Black, D A, N Kolesnikova, S G Sanders, and L J Taylor (2013), “Are children ‘normal’?” The Review of Economics and Statistics, 95(1): 21–33.
Boserup, E (1970), Women’s role in economic development, Routledge.
Bustos, P, B Caprettini, and J Ponticelli (2016), “Agricultural productivity and structural transformation: Evidence from Brazil,” American Economic Review, 106(6): 1320–1365.
Bustos, P, G Garber, and J Ponticelli (2020), “Capital accumulation and structural transformation,” The Quarterly Journal of Economics, 135(2): 1037–1094.
Gollin, D, C W Hansen, and A M Wingender (2021), “Two blades of grass: The impact of the green revolution,” Journal of Political Economy, 129(8): 2344–2384.
Heath, R, and A M Mobarak (2015), “Manufacturing growth and the lives of Bangladeshi women,” Journal of Development Economics, 115: 1–15.
Jensen, R (2012), “Do labor market opportunities affect young women's work and family decisions? Experimental evidence from India,” The Quarterly Journal of Economics, 127(2): 753–792.
Kearney, M S, and R Wilson (2018), “Male earnings, marriageable men, and nonmarital fertility: Evidence from the fracking boom,” Review of Economics and Statistics, 100(4): 678–690.
Kitchens, C T, and L P Rodgers (2023), “The impact of the WWI agricultural boom and bust on female opportunity cost and fertility,” The Economic Journal, 133(656): 2978–3006.
Moorthy, V S (2025), “Agricultural technological change, female earnings and fertility: Evidence from Brazil,” The Economic Journal, 135(665): 285–320.
Schaller, J (2016), “Booms, busts, and fertility: Testing the Becker model using gender-specific labor demand,” Journal of Human Resources, 51(1): 1–29.
Schultz, T P (2001), “Women’s roles in the agricultural household: Bargaining and human capital investments,” Handbook of Agricultural Economics, 1: 383–456.
Waltz, A (2016), “The women who feed us: Gender empowerment (or lack thereof) in rural Southern Brazil,” Journal of Rural Studies, 47: 31–40.
Brazil Agricultural development Agricultural technology Agriculture structural transformation female labour force participation