Agriculture, Poverty, and Postwar Reconstruction: Micro-Level Evidence from Northern Mozambique
| Date | 01 May 2009 |
| Published date | 01 May 2009 |
| Author | Tilman Brück,Carlos Bozzoli |
| DOI | 10.1177/0022343309102658 |
| Subject Matter | Articles |
377
© The Author(s), 2009. Reprints and permissions:
http://www.sagepub.co.uk/journalsPermissions.nav,
vol. 46, no. 3, 2009, pp. 377–397
Sage Publications (Los Angeles, London, New Delhi,
Singapore and Washington DC) http://jpr.sagepub.com
DOI 10.1177/0022343309102658
Agriculture, Poverty, and Postwar Reconstruction:
Micro-Level Evidence from Northern
Mozambique*
CARLOS BOZZOLI
German Institute for Economic Research (DIW Berlin) and Households in
Conflict Network (HiCN)
TILMAN BRÜCK
German Institute for Economic Research (DIW Berlin), Humboldt University
Berlin, and Households in Conflict Network (HiCN)
This article analyzes the effects of household-level activity choices on farm household welfare in a develop-
ing country affected by mass violent armed conflict. The study uses household survey data from postwar
Nampula and Cabo Delgado provinces in Northern Mozambique capturing many activity choices, including
market participation, risk and activity diversification, cotton adoption, and social exchange, as well as income-
and consumption-based measures of welfare. The study advances the literature on postwar coping and rural
poverty at the micro level by estimating potentially endogenous activity choices and welfare outcomes using
instrumental variables. The study finds that increasing the cultivated area and on-farm activities enhances
postwar welfare of smallholders exploiting wartime survival techniques. Subsistence farming reduces income
but does not affect consumption, while market participation has positive welfare effects. This suggests that
postwar reconstruction policies should encourage the wartime crop mix but offer enhanced marketing oppor-
tunities for such crops. Cotton adoption, which was promoted by aid agencies in the postwar period, reduces
household welfare per capita by between 16% and 31%, controlling for market access. This contradicts
previous studies of postwar rural development that did not control for the war-related endogeneity. Hence,
addressing the potential endogeneity of activity choices is important because the standard regression approach
may lead to biased estimates of the impact of activity choice on welfare, which in turn may lead to biased
policy advice. The article discusses and contextualizes these findings, concluding with a discussion of suitable
pro-poor reconstruction policies for national governments and donors.
Introduction
Farmers in rural African war zones are among
the most destitute people in the world. In
2001, one-third of the world’s population lived
in conflict-affected low-income countries, with
two-thirds of these people residing in rural areas
(own calculations). Yet, economic analysis has
only recently started to consider the impact of
war on rural poverty and underdevelopment
(Addison, 2003; Justino, 2006).
* We are very grateful for helpful comments from Tony
Addison, Valpy FitzGerald, Pramila Krishnan, Caterina
Ruggeri Laderchi, Frances Stewart, Henrik Urdal, semi-
nar participants in Antwerp, Berlin, Bogotá, Bristol,
Dresden, Hamburg, Helsinki, Leuven, London, Oxford,
Philadelphia, and The Hague and two anonymous ref-
erees. This study was funded by the UK Department for
International Development and the European Union’s 6th
Framework Programme. The data were generously made
available by the Food Security Project of the Mozambican
Ministry of Agriculture and Michigan State University.
The usual disclaimer applies. The dataset for the empirical
analysis, along with an appendix, can be found at http://
www.prio.no/jpr/datasets. Email: tbrueck@diw.de.
journal of PEACE RESEARCH volume 46 / number 3 / may 2009
378
This study aims to fill this gap in the lit-
erature by estimating the determinants of
income and consumption for war-affected
farm households, using a detailed agricul-
tural and household survey from Northern
Mozambique (henceforth referred to as the
FSP survey). In addition, the article assesses
the endogenous welfare implications of farm
household production choices, as these are
likely to be significant in a postwar setting
characterized by market imperfections, few
public goods, and economic isolation.
The article therefore draws on literature
analyzing the determinants of household
activity choices (Ellis, 2000) and on literature
estimating the determinants of household
welfare. In the context of extreme shocks,
the former typically assesses the effects of
ongoing shocks – such as drought, earth-
quakes, or floods – on the activity choices
of rural households (Corbett, 1988). There
is still little analysis, though, of post-disaster
or of postwar activity choices. Past research
on activity choices addressed the role of risk
(Dercon, 1996), the linkages of the farm
household with the rural non-farm sector
(Abdulai & Delgado, 1999), the farm house-
hold participation in markets (de Janvry,
Fafchamps & Sadoulet, 1991), the potential
of farm diversification for raising export rev-
enues (Delgado, 1995), activity choices and
poverty traps (Zimmerman & Carter, 2003),
informal risk-sharing mechanisms (Dercon,
2002), farm fragmentation (Blarel et al.,
1992) and the role of land abundance for
income strategies (Binswanger & McIntire,
1987). These issues will be addressed for a
postwar economy.
A surprisingly small field considers the
welfare implications of activity choices
(Adams, 2002; Ellis & Mdoe, 2003; Groo-
taert, 1997; Reardon, Delgado & Matlon,
1992) or of cash crop adoption (Bouis &
Haddad, 1990; Kennedy & Cogill, 1987;
von Braun & Pandya-Lorch, 1991). These
studies describe the welfare implications of
activity choices but they do not estimate
the welfare implications directly. Neither do
these studies, unlike this article, account for
the endogeneity of activity choices and farm
income. Addressing the potential endogene-
ity of activity choices is important because
the standard regression approach may lead
to biased estimates of the impact of activity
choice on welfare, which in turn may lead to
biased policy advice.
The literature on the determinants of
household welfare usually assesses the welfare
effects of who the households are and what the
households own but not of what these house-
holds do, especially if they live in extremely
poor, rural, war-affected environments. The
answer to this question is important for two
reasons. First, economic analysis has little to
say, to date, on how people behave under
extremely adverse conditions such as war. Sec-
ond, current policy advice for governments
and donors operating in conflict zones does
not yet know how best to support extremely
poor but numerous victims of war.
The next section of this article defines key
terms and discusses the expected determi-
nants of household welfare. Subsequent sec-
tions describe the case of Mozambique, the
household survey, some descriptive statistics
from that survey, and the estimation strategy.
Next, the results of the multivariate analysis
are discussed. The concluding section sum-
marizes the main findings and presents some
policy implications.
The Determinants of Welfare
Household welfare is usually defined as a
household’s command over market and non-
market goods and services at the household
level. Since welfare is not observable, the lit-
erature has resorted to using proxies such as
income, consumption, and food consump-
tion (Appleton, 2001; Ravallion, 1996). This
definition of welfare disregards, for empirical
reasons, the consumption of services derived
To continue reading
Request your trial