86182 Policy Brief Issue 5 | September 2013 Caught in a Productivity Trap: A Distributional Perspective on Gender Differences in Malawian Agriculture Talip Kilic, Amparo Palacios-Lopez and Markus Goldstein Summary The vast majority of households in Malawi are involved in agriculture, and improving agricultural productivity – particularly for women, who tend to attain lower yields than men - could lead to significant poverty reduction and improvements in gender equality. This study asks two main questions: (1) Exactly how great are the differences in agricultural productivity between men and women in Malawi? And (2) How much of the gender gap is explained by differences in levels of agricultural inputs vs. differences in returns to these inputs? We trace the varying constraints faced by farmers at different levels of productivity, as well as at average productivity – a level of analysis that is crucial for designing effective interventions aimed at bridging the gender gap. We find that on average, female- managed plots are 25 percent less productive than plots managed by males. Further, the gender gap widens significantly as agricultural productivity increases. More than 80 percent of the mean gender gap is explained by differences in levels of agricultural inputs, suggesting that addressing market and institutional failures underlying these differences could have direct economic benefits. AGRICULTURAL PRODUCTIVITY AND GENDER IN THE Agricultural growth has been shown to particularly benefit the 1 MALAWIAN CONTEXT poorest in the developing world. Poverty in Malawi is Agriculture is critically important to both Malawi’s economy especially widespread among female-headed households, and to its social fabric. The sector accounts for 31 percent of suggesting that investing in agricultural growth has benefits Gross Domestic Product, and 84 percent of Malawian both for poverty-reduction and for gender equality. Yet households own and/or cultivate land. The majority of farming systematic gender differences persist in agricultural households practice subsistence agriculture, and their living productivity across Sub-Saharan Africa, mostly due to standards are directly affected by the inconsistent agricultural differences in (i) access to and use of agricultural inputs, performance that Malawi has seen over the last two decades. including improved technologies; (ii) tenure security and These impacts are especially acute for the poorest related investments in land; (iii) market and credit access; (iv) households. human and physical capital; and (v) informal institutional Malawi Annual Maize Yield Estimates (1990-2010) constraints affecting farm/plot management and the marketing 2 3000 of agricultural produce. Addressing these gender differences could result in tremendous productivity gains. In “The State of Yield (KG/HA) 2500 2000 Food and Agriculture 2010-2011: Women in Agriculture,” the 1500 United Nations Food and Agriculture Organization (FAO) 1000 reports that if female farmers had the same access to 500 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 1 Ligon and Sadoulet (2008) document that a one percent rise in agricultural GDP results in six percent income growth for the lowest income decile of the population. 2 Cultural roles that are assigned to males and females regarding Year domestic duties and gender segregation in crop production (i.e. staple Source: FAOSTAT vs. cash crop cultivation, high-yielding vs. low-yielding variety cultivation, etc.) could be thought of as informal institutional constraints. The World Bank Group | Africa Region Gender Practice Policy Brief 1 productive resources as men, they could increase yields by i. Levels of observable inputs or attributes, such as the 20-30 percent, which could increase total agricultural output in education level of the plot manager, the amount of developing countries by 2.5 to 4 percent and lift 100 to 150 inorganic fertilizers and other inputs, the use of male million people out of hunger. household labor, etc. We refer to the impact of these combined factors as the endowment effect; Previous studies looking at the gender gap and agricultural growth rely largely on data from small-scale surveys, and are ii. Returns to these inputs or attributes, such as the limited in terms of geographic coverage, topic, or attention to monetary return that a farmer earns from applying intra-household dynamics (or, in some cases, all three). Our one kilogram of fertilizer to her/his plot. We refer to study is the first output of a two-year research program on the impact of these combined factors as the structure Gender and Agriculture in Sub-Saharan Africa. The program effect. is primarily based on the data generated by the Living Standards Measurement Study – Integrated Surveys on Thus, the study seeks to quantify not only the difference in 3 Agriculture (LSMS-ISA) initiative , which is funded by the productivity between women and men, but also the relative International Fund for Agricultural Development (IFAD) and contributions of key inputs and returns to these inputs (based the World Bank Research Support Budget. It contributes to a on correlation, not causation). Identifying the factors driving the small but growing body of nationally representative and gender gap is crucial for informing policy interventions aimed at methodologically sound data collected in varied settings addressing the gap at its roots. across Sub-Saharan Africa. The study moves beyond the comparisons of male vs. female DATA farmer at average levels of productivity, and breaks down the This study uses data from the Third Integrated Household key factors driving the gender gap at different points in the Survey (IHS3), collected from March 2010 to March 2011 by agricultural productivity distribution. Since farmers at different the Malawi National Statistical Office, with support from the levels of productivity may face different constraints – or similar 4 LSMS-ISA initiative. The IHS3 data covers 12,271 constraints but to different degrees – we try to tease out the households. The full sample consists of 16,372 plots, contributions of key factors towards the gender gap at the low, 26 percent of which are managed by women. mid- and high-end of the agricultural productivity spectrum. METHODOLOGY FINDINGS Our econometric approach applies a decomposition On average, we find that female-managed plots in Malawi are methodology that has been widely used in labor economics, 25 percent less productive than plots that are managed by starting with studies by Oaxaca (1973) and Blinder (1973), to males. The endowment effect explains 82 percent of this understand the gender gap in agricultural productivity. We gender gap. In particular, female managed plots are look at the average difference in agricultural productivity constrained by: (defined as total value of output per hectare) on male- and (i) Lower use of inorganic fertilizer; female-managed plots, and determine how much of the (ii) Lower use of household adult male labor; gender gap is driven by differences in: (iii) Lower production of high-value export crops; (iv) Restricted access to agricultural tools. 3 Female plot managers try to compensate for the lower levels The LSMS-ISA initiative is a household survey program, established by a grant from the Bill and Melinda Gates Foundation, to provide of household adult male labor with higher levels of household financial and technical support to governments in Sub-Saharan Africa to design and implement nationally representative, multi-topic panel adult female labor, as well as household child and exchange household surveys with a strong focus on agriculture labor, but this is not enough to overcome the differences in (www.worldbank.org/lsms-isa). 4 The IHS3 data and documentation are publicly available through the productivity. This labor difference especially disadvantages LSMS-ISA website. The World Bank Group | Africa Region Gender Practice Policy Brief 2 Policy Brief Issue 5 | September 2013 female farmers because they also tend to have lower access result in the greatest yields, and thus experience lower returns to labor-saving agricultural devices. than men. The remaining 18 percent of the gender gap – the structure NEXT STEPS effect – is driven by differences in returns to the use of Our findings suggest that a large and significant difference in household adult male labor, and the application of inorganic the levels of inputs is the central factor behind the gender gap, fertilizer. Not only do adult males in the household spend less particularly for farmers at lower levels of agricultural time on female-managed plots, but their labor contribution is productivity. On male-managed plots, higher levels of less productive than on male-managed plots. One reason for household adult male labor and area under export crop this may be that female plot managers provide less cultivation widen the gender gap, while household and supervision compared to male plot managers, possibly due to childcare responsibilities restrict the time that female plot having other household responsibilities. Indeed, our study managers can dedicate to farming. Effective policies to 5 found that a greater child dependency ratio decreases the counteract these effects include providing female farmers with productivity of female-managed plots but has no effect on the support to adopt high-value crops and to access inorganic male-managed plots. This points to childcare responsibilities fertilizer, seeds from improved varieties, and labor-saving falling primarily on women, preventing them from providing as agricultural tools. Ensuring that female plot managers have much time and energy supervising labor as male plot similar years of schooling as men and apply similar levels of managers, who do not experience the same limitations. non-labor agricultural inputs - including inorganic fertilizer, pesticides/herbicides, and improved and/or export crop In terms of inorganic fertilizer use, female farmers not only varieties - could reduce the mean gender gap by 50 percent. apply lower levels of this input, but the fertilizer that they do In addition, policies need to address the underlying constraints apply does not yield as many benefits. A knowledge gap along that lead female farmers to experience lower returns to gender lines may account for a relatively less efficient usage inorganic fertilizer use and to household adult male labor. of fertilizer by female farmers, but more research is needed to However, further research is needed to determine why these determine this conclusively. inequalities in time use, as well as access and returns to agricultural inputs, continue to persist. Finally, our analysis shows that the gender gap widens as agricultural productivity increases. While the gender gap in Malawi is 25 percent at mean productivity, it ranges from th 22 percent at the 10 percentile of the agricultural productivity th distribution, to 37 percent at the 90 percentile. At the same time, the returns to inputs decrease progressively for female For questions and more information about the Africa farmers but not for male farmers, meaning that the structure Region’s gender program, please contact Katherine Manchester at kmanchester@worldbank.org. effect explains more and more of this gender gap at higher The World Bank productivity levels. One possible explanation is that even as 1818 H St. NW female farmers use higher levels of productive inputs, they are Washington, DC 20433 USA less effective at determining the combinations of inputs that 5 The child dependency ratio is defined as the number of household members below the age of 10 divided by the number of household members aged 10 years and above. The World Bank Group | Africa Region Gender Practice Policy Brief 3