Department of Agricultural, Food, and Resource Economics Food Security Group

IDWP 105. David Tschirley, Luis Flores, and David Mather. 2010. Agricultural and Food Security Policy Analysis in Central America: Assessing Local Institutional Capacity, Data Availability, and Outcomes

Spanish Version: IDWP 105s. David Tschirley, Luis Flores, y David Mather. 2010. Análisis de Políticas Agrícolas y de Seguridad Alimentaria en Centroamérica: Evaluación de la Capacidad Institucional Local, la Disponibilidad de Datos y la Demanda Efectiva para Datos e Información

EXECUTIVE SUMMARY:
Performance of the agricultural sector in developing countries is fundamental to ensuring
robust and equitable economic growth and broad-based food security. Yet donor support to
agricultural development in developing countries has declined continuously for 30 years. This
same period saw dramatic deterioration in developing countries’ institutional capacity to
provide services to their agricultural sectors. These trends may now be changing, due in part
to the global food price crisis of 2007 and 2008 and concerns that it unleashed about the
world’s ability to feed its poorest inhabitants. This paper reports on the results of a two week
trip to Guatemala and Nicaragua made by Michigan State University’s Food Security Group.
The purpose of the trip was to assess two aspects that form the foundation for applied
agricultural and food security policy analysis and outreach: (a) the organizations involved in
research and outreach on these topics, and (b) existing data sets and processes for continued
generation of data sets useful in such analysis and outreach. The team also explored the
extent to which policy makers and designers of public programs solicit empirical data and
analysis for the design and implementation of local food security programs and policies.

While in the region, the team held in-briefings and out-briefings with USAID missions in
each country, made two field visits in Guatemala and one in Nicaragua, and met with over 20
individuals across 16 institutions in Guatemala and with over 20 individuals across 14
institutions in Nicaragua.

The paper reaches five basic conclusions. First, Guatemala and Nicaragua both have better
than average data for applied analysis of agricultural and food security policy. Though
suffering from problems typical of statistical institutes throughout the developing world, INE
in Guatemala and INIDE in Nicaragua have generated what we expect to be reasonably good
LSMS data for two recent years. Common strengths across these surveys include well
designed and quite comprehensive questionnaires, more than adequate sample sizes, high
ratios of supervisors to enumerators in data collection, and good understanding of the
challenges involved in these activities by the individuals in charge. In addition, both LSMStype
surveys have larger and better-designed agricultural sections than is often found in an
expenditure survey. Both countries collect agricultural commodity price data at least weekly
from markets around the country. We know that these price data series have few gaps, and we
expect that data quality will be acceptable for the types of uses to which it would be put.
Household survey data and market price data are the basic data needed to generate better
empirical understanding of a range of food security issues in each country, and could be
complemented with additional focused data collection built around specific analytical
questions.

Our second finding is that each country has research institutions with some degree of
sustainability, some demonstrated track record attracting outside funding and collaborators, a
commitment to good research and active outreach, and great interest in improving their
capacity. IDIES and IARNA with Universidad Rafael Landívar in Guatemala, and Nitlapan
with Universidad de Centroamérica and the agricultural economics program at Universidad
Autónoma de Nicaragua-Managua in Nicaragua stand out in this regard. Yet all of these
organizations are dominated by analysts trained at the M.S. level (very few Ph.D.s), some
struggle to fund the research they wish to do, and none of them has so far been successful
integrating students into their research and outreach activities. In other words, both countries
have institutions that are already doing good and relevant work but whom could benefit
greatly from increased training and from resources for collaborative research and outreach.

Third, a key challenge in both countries is building demand in public agencies for these
analyses. Lack of civil service reform has combined with inadequate tax revenue to result in
low funding, low salaries, inadequate operating budgets, and high turnover of personnel.
These conditions make it difficult for the public sector to generate sustained demand for
objective information and analysis. Yet there are areas that have been less affected by these
problems, and in nearly all cases one can find committed and motivated people doing the best
work they can under the circumstances and anxious to be able to do more. We believe that,
across a range of public sector organizations, meaningful progress can be made by building
research and outreach capacity in other institutions (preferably universities) and pursuing a
process of flexible engagement with government in circumstances and with agencies (and
specific personnel within those agencies) judged to be able to benefit from the engagement.
This kind of university-public sector collaboration is a hallmark of effective policy analysis
and outreach processes in many countries, and we believe it would pay high dividends in
these countries as well.

Fourth, great value remains to be generated from existing household level data sets.
Specifically, we believe that these data sets could be used to generate substantially greater
insight regarding the distributional effects of a range of policies and programs. This is
especially true in Nicaragua, where the Encuesta Nacional de Hogares sobre Medición de
Niveles de Vida (ENMV) to our knowledge has been greatly under-exploited. The capacity to
extract as much information as possible from data sets such as ENMV and ENCOVI
(Guatemala’s living conditions survey) is important, as government and research institutions
in both countries are quite interested in customized territorial development policies and the
strengthening of municipality-level development planning.

We identify two such potential analyses. First, what are the distributional effects of the
existing structure of support to agriculture in both countries? Various studies show the
overwhelming importance of (a) price distortions driven by trade controls and (b) price
subsidies through direct fiscal support (for example on fertilizer in Guatemala) in overall
support to the agricultural sector. Studies also document the limited investment in public
goods in Guatemala, a problem likely also seen in Nicaragua. No studies that we have seen
quantify the distributional implications of these policies – who wins, who loses, and how
much? ENCOVI in Guatemala and ENMV in Nicaragua could be used to generate a detailed,
disaggregated answer to this question, with major implications for policy. Annex B provides
a brief note on analytical approaches to assessing the distributional consequences of
agricultural policies.

A second analysis is specific to Guatemala: what has been the effect of the boom in export
horticulture on poverty in the altiplano? USAID has for many years supported this sector, and
it has been a major success, including for smallholder farmers. Yet the boom has reached
only the southern portion of the altiplano, which enjoys the best infrastructure and market
access, leaving the historically more isolated and poorest areas in the north and northwest
relatively untouched. In the southern areas where the boom has occurred, which farmers have
been able to participate directly in the export boom and what has been the impact on their
standard of living?; which farmers have benefited from increased demand for labor driven by
export production, and how has this affected their livelihoods?; how have both of these
affected farmers’ crop mix and their productivity in staple grains? High quality research on
this topic, quite possible with existing data and in collaboration with local researchers, would
help USAID to explicitly bridge its focus over the past decade on non-traditional agricultural
exports into the new era of emphasis on broader agricultural productivity, food security, and
poverty reduction.

Our fifth and final finding is that both countries urgently need to put into place well designed
impact assessments for the plethora of special projects that have been launched in recent
years. Mi Familia Progresa and Mi Comunidad Produce in Guatemala, the various
components of Hambre Cero in Nicaragua, all have obvious potential positive and negative
effects that need to be understood in order to improve them over time and to design second
generation projects that are better targeted and have larger positive impacts. Most such
analyses would require dedicated data collection focused on beneficiaries and nonbeneficiaries
– ENCOVI and ENMV may be able to play complementary but not central
roles.

The paper closes by outlining an approach to collaborative research and outreach that we
believe would contribute to three objectives: quickly enhancing the empirical content of the
policy debate in each country, building sustainable local capacity for enhanced applied policy
research and outreach, and strengthening the linkages between government and applied
research organizations that are central to good policy making. We also identify the most
promising public sector and research organizations in which to invest as part of this approach.