Three Essays on Storage Technology and Seed Choice in Maize-Legume Systems


To study hybrid adoption in Mexico:

Hypotheses: 1) Farmers' WTP for hybrid seeds sold by MasAgro Maíz partner seed companies is higher for hybrid relative to creole seeds 2) Farmers' WTP for grain stemming from hybrid seeds is lower relative to grain from creole seeds annex 2 is a draft of the conceptual framework linking MasAgro Maíz’s work with seed companies and how that work connects with farmer’s decision to adopt hybrid seeds.

Aim: To measure farmers’ demand for hybrid vs creole seeds in the Mexican state of Puebla.  To understand how agriculture extension interventions through the private sector like seed companies, as opposed to government initiatives, can achieve an increased presence and use of hybrid maize seeds, and benefits to economic and health wellbeing can be achieved. There is strong evidence of impact of improved maize variety adoption on farm household welfare (Becerril & Abdulai, 2010); Through the MasAgro Maize program there is increased availability of hybrid seed production and distribution that may lead to increased adoption of improved maize varieties. MasAgro Maíz works through seed companies to reach agro dealers and farmers to increase knowledge about agronomic traits with larger advantages over existing alternatives. Through adoption of hybrids, increases in yields are expected, which increases effect of production practices (including production diversity) that affects home consumption and or income; thus, affecting food purchases, food consumption, and diet. Farmers who have been exposed to performance gains of hybrid seeds are expected to have a higher WTP than farmers not exposed. If hybrid seeds are not providing gain in performance farmers WTP will be at or below the WTP for non-hybrid seeds. WTP is expected to be below the amount necessary to cover the cost of seed development.

A random sample from farmers in the state of Puebla who obtained seeds from agro dealers connected to MasAgro Maize seed companies will be taken and used in 2019 to conduct a survey and BDM auction. After several years of hybrid seed availability adoption rates in certain Mexican states like Puebla remain low. There is literature exploring adoption barriers in several states like Chiapas and Oaxaca but not Puebla, which has four seed companies supported by MasAgro, 2,7211 ha of maize planted, with an average of 1.5 tons per ha, and low hybrid seed adoption rates.  Two methods, BDM and Paired Choice Experiment, are used in hybrid adoption studies to evaluate WTP, but with BDM infrequently used in the Mexican studies. This study uses a structured survey along with BDM auctions to measure WTP against consumption traits and agronomic traits of hybrid seeds at the point of sale for a maize growing population with access to hybrid seeds to understand what drives farmers’ choices among varieties.

Data will come from MasAgro Maíz’s partnership with four seed companies and their agro dealer distributors from the state of Puebla. Data on seed varietal difference is already available. Complementary data about household characteristics like wealth and individual level data about diet diversity along with BDM auctions will be collected. The sample will include 400 farmers who purchased seeds from agro dealers connected to the partner seed companies in the state.  I would use quantile regression to test for heterogeneity in response, hypothesizing that demand of hybrid seed is linked to agronomic performance, experienced yield gain and knowledge of hybrid benefits.  The analysis plan for this aim will compare individuals’ WTP for a hybrid seed relative to creole seed. WTP will be regressed on performance difference. Control variables (e.g. wealth, education) will be used in a variety of model specifications. In the case of demand for grain from hybrid seeds, individuals’ WTP for this grain will be compared relative to grain from creole grain. Both seed and grain comparisons would measure prices on a Kg to Kg basis.  WTP will reveal premium relative to agronomic performance in the seed BDMs; WTP will reveal premium relative to consumption traits in the grain BDMs.

Work expectations

Winter 2018

Connect with Jason Donovan about the MasAgro program and learn about the MasAgro Maize sub component to understand what data is available, and what data needs to be collected like WTP for hybrid seeds and dietary diversity of growers. Confirm in which Mexican states data collection will occur. Puebla is a likely candidate due to its low improved seed adopting rate and its topography as one of the high valley and tropic regions. Confirm sample size (e.g. clustering at the seed company level, seed type stratification, gender balance) needed for complimentary data and funds and resources (e.g. seeds, data collection staff, data collection equipment, cars and drivers) available between me and CIMMYT to execute data collection.

Finalize letter of intent including background, objectives, research questions, hypothesis and methodology.

Spring 2019

Preparation of thesis proposal with feedback from thesis committee and extended CIMMYT team. Confirm dates I will work from CIMMYT during the months of April and May. Apply for Tufts IRB and CIMMYT IRB.

Summer 2019

Prepare for data collection to take place in the Fall.

Fall 2019

I will travel to Mexico and participate in data collection for 4-6 months.

Winter 2020

Data cleaning and analysis will take place.

Spring 2020 – Spring 2021

Data will be incorporated into dissertation.



The dissertation work will consist of three essays on farmers' decisions regarding production, storage and consumption in rain fed traditional production systems in Malawi and Mexico. In the former, farmers in the poorest districts combine maize with groundnuts and beans and in the later farmers from a range of socioeconomic levels from the high valleys and tropics combine maize and beans. The first is a collaboration with Catholic Relief Services (CRS) in Malawi on the United in Building and Advancing Life Expectations (UBALE) program, funded by the United States Agency for International Development (USAID) from 2015 to 2019; and the second is a collaboration with the International Center for Maize and Wheat Research (CIMMYT) in Mexico on the Programa de Modernización Sustentable de la Agricultura Tradicional (MasAgro) project, funded through a grant from CIMMYT and the Secretaría de Ganadería, Desarrollo Rural, Pesca y Alimentación (SAGARPA) from 2016 to 2021. The activities outlined here are related to the collaboration with CIMMYT.

The collaboration with CIMMYT responds to the research interests outlined in the Mexican National Council on Science and Technology (CONACYT) fellowship awarded to me in 2016 and builds on work within one component of the MasAgro program, MasAgro Maize, which is focused on high-yielding improved maize seed, referred to as hybrid seeds for the rest of this document.  The CONACTY proposal envisioned a study in Mexico, which looked at knowledge, attitudes and practices of smallholders around agricultural inputs. The overarching themes of interest underpinning the study were defined as agriculture, food systems and nutrition, with an emphasis on vulnerable populations (e.g. subsistence farmers). The work builds on economic models of farm household behavior and other conceptual frameworks linking agriculture to nutrition, such as diagrams used by the International Food Policy Research Institute (annex 1) and others to show how agriculture can directly and/or indirectly impact health. It was proposed that through formative research, collaboration with CIMMYT and Mexican institutions a research question, hypotheses and aims would be further defined at the dissertation stage of the doctoral program.  Specifically, work with CIMMYT was founded on interest around adoption of hybrid varieties and the work done under the MasAgro program. Thus, aim 3 of this dissertation focuses on adoption practices from this collaboration.


Required skills

 Mastery of specific content areas

The work proposed requires an advanced understanding of household survey design and implementation, impact evaluation, and issues facing agriculture and nutrition in the emerging economies with understanding of the local context of Malawi and Mexico. Additionally, understanding of regression analysis and factor analysis is needed. I have substantial field experience implementing agriculture projects in Malawi where I designed and completed several survey-based evaluations of agriculture and nutrition projects. Through this work I have gained practical experience in survey design and implementation and have a strong background in agriculture and nutrition issues facing Malawi and Mexico.

My academic preparation through the course worked described below has provided knowledge in both theory and empirical literature. This coursework includes directed studies focused on the aims outlined in this letter of intent: 1) Food markets and diet quality focused on assessing food markets with an emphasis on economic analysis of commodity production and supply, demand for food, and the organization and performance of the food system; 2)  Agriculture technology adoption in Malawi focused on Storage choices and willingness to pay analysis with cross sectional survey data; and 3) Economics of maize in Mexico which looked at how government policies specific to maize, influence the structure and performance of market systems, with regard to poverty, food security and nutrition among rural and urban households in Mexico.

Additionally, I have a solid grounding in methods using principal component analysis and explanatory factor analysis in previous work and can apply that knowledge to analyzing the relationships proposed above regarding diet quality and other variables like social economic status. I will also continue to advance my knowledge in these areas in order to be able to incorporate the most appropriate methods in this thesis.

Data collection methods

Most of the MasAgro data to be used was collected by MasAgro and CIMMYT. The MasAgro Maiz seed companies will serve as the sampling universe from whcih to draw a set of farmers from whom complimentary data will be obtained. Data will be collected electronically and following CIMMYT’s internal data collection standards over the course of 2019.

Statistical analysis techniques

This study will use OLS regression, factor analysis, and statistical tests to assess correlation, multi-collinearity of variables, and robustness of construct validity for indices. I am highly proficient in Stata and familiar with the application of each of these methods using Stata. Should additional techniques to address questions that arise I will learn these methods and seek out support from Friedman Faculty. I will seek guidance from the thesis committee members and Friedman faculty during the course of the analysis portion of dissertation preparation.