Market effects of farmer field schools in Sub‐Saharan Africa: The case for cocoa

• Author: Bruce L. Dixon,Jennie S. Popp,Lawton L. Nalley,Francis Tsiboe,Jeff Luckstead
• Published date: 01 November 2018
• DOI: http://doi.org/10.1111/rode.12520
• Date: 01 November 2018

REGULAR ARTICLE

Market effects of farmer field schools in Sub-

Saharan Africa: The case for cocoa

Francis Tsiboe

|

Jeff Luckstead

|

Lawton L. Nalley

|

Jennie

S. Popp

|

Bruce L. Dixon

University of Arkansas, Fayetteville,

Arkansas

Correspondence

Jeff Luckstead, AGRI 217, University of

Arkansas, Fayetteville, AR 72701.

Email: jluckste@uark.edu

Abstract

Billions of dollars are invested in development programs,

all of which have a monitoring and evaluation component

and consider only the programs’intended effects while

ignoring their market effects on prices. This study analyzes

the market effects of the Cocoa Livelihood Program (CLP)

in Ghana. To evaluate both the benefits to cocoa farmers

and market outcomes of CLP on cocoa and other noncocoa

markets, this study simulates a partial-equilibrium farm

household model. The analysis shows that CLP participat-

ing households benefit by U.S.$154 per household, while

nonparticipating households benefit (through market out-

come) by only U.S.$21 per household, at the 2016 partici-

pation rate of 6 percent. The net welfare changes in staple

food markets (maize, rice, cassava, and yam) attributable to

CLP, is estimated at U.S.$850 per household. A rise in the

CLP participation rate leads to a decline in the Ghanaian

cocoa price. However, net gains from the program become

negative at participation rates of more than 60 percent, as

supply increases and the cocoa price falls. This result

suggests that CLP participation could be expanded up to 60

percent of Ghana’s cocoa producing households. However,

expanding cocoa demand is equally as important as

expanding production if participation rates exceed 60

percent.

DOI: 10.1111/rode.12520

e160

|

©2018 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/rode Rev Dev Econ. 2018;22:e160–e184.

1

|

INTRODUCTION

Impact assessment of developmental interventions and investments in the agricultural sector have

generated an extensive literature (Davis et al., 2012; Dexter, Hughes, & Farmer, 2008; Larsen &

Lilleør, 2014; Van den Berg, 2004; Waddington et al., 2014), most of which evaluate the direct

impact of a program on the target area/group, but fail to consider potential market effects outside

of the target area/group. The few studies that consider spillover effects, focused on industry, micro-

finance, and agro-processing rather than agricultural production (Currie, Davis, Greenstone, &

Walker, 2015; Dabalen, Chuhan-Pole, Kotsadam, Tolonen, & Sanoh, 2015; Henderson, 2003;

Khandker, 2005; Kotsadam & Tolonen, 2015; Moretti, 2004; Waldkirch & Ofosu, 2010). Partial-

equilibrium simulation-based impact studies (Kaitibie, Omore, Rich, & Kristjanson, 2010; Kristjan-

son, Swallow, Rowlands, Kruska, & De Leeuw, 1999) focus on the market directly affected by the

intervention. General-equilibrium simulation-based impact studies (Rich, Roland-Holst, & Otte,

2014) account for the impacts of intervention on the market directly affected as well as the indirect

effects on all other markets in the economy. To the best of our knowledge, the present study pro-

vides the first attempt to quantify both the direct and indirect market outcomes of interventions in

cocoa production in West Africa. A key contribution of the partial-equilibrium model developed in

this paper is including horizontal linkages between cocoa intervention and the corresponding farm-

income effects, which then impact noncocoa staple food markets through changes in staple-food

consumption, prices, and production.

Gockowski, Asamoah, David, Gyamfi, and Kumi (2010) showed that farmer field schools’

(FFS) implementation in Ghana statistically improved cocoa farmers’yield by 14 percent per hec-

tare; and Opoko, Dzene, Caria, Teal, and Zeitlin (2009) concluded that the provision of extension

services and credit on inputs resulted in a 43 percent revenue increase and a subsequent revenue to

cost ratio of 2.5 for participating cocoa farmers in Ghana. However, both studies fail to consider

market effects through price adjustments on nonparticipating households. With this in mind, this

paper aims to obtain quantitative evidence on the nature and magnitude of direct and indirect mar-

ket outcomes of interventions aimed at improving the productivity of cocoa farmers, and use the

results to draw conclusions for policy and future research. We do so by developing a partial-equili-

brium farm household model (FHM) in Ghana and simulating the effect of the first phase of the

Cocoa Livelihood Program (CLP-I) on (1) prices and quantities of cocoa and noncocoa agricultural

products and (2) farmer welfare of CLP and non-CLP participants. The model explicitly accounts

for the family-farm structure of cocoa production where households are semi-commercial and

semi-subsistence, which allows for simultaneous elements of both producer and consumer theor y

in farm household decisions.

The Cocoa Livelihood Program, an ongoing developmental program implemented in Sub-

Saharan Africa to double the income of approximately 200,000 cocoa-growing households, has

been shown to directly increase yields. Norton, Nalley, Dixon, and Popp (2013) used ordinary least

squares to show that, for a sample of 138 Ghanaian cocoa farmers exposed to the full CLP-I pack-

age, cocoa yields significantly increased by 75 percent. Tsiboe, Dixon, Nalley, Popp, and Luck-

stead (2016) subsequently applied a difference-in-differences model to a different dataset collected

from 2,048 pre- and post-CLP-I interviews of cocoa producers in Ghana, C^

ote d’Ivoire, Nigeria,

and Cameroon, and showed that yield enhancements attributable to CLP-I are 32, 34, 50, and 62

percent in Ghana, Cote d’Ivoire, Nigeria, and Cameroon, respectively. While the estimates from

Tsiboe et al. (2016) for Ghana were smaller than those of Norton et al. (2013), it should be noted

that this is partly due to both the sample and estimation method used. Nonethel ess, both studies

show that yield gains attributable to CLP-I are statistically significant and substantial.

TSIBOE ET AL.

|

e161