




































In ternationa l
Scholars
Journa ls

 

African Journal of Agricultural Marketing ISSN 2375-1061 Vol. 8 (7), pp. 001-006, July, 2020. Available online at 
www.internationalscholarsjournals.org © International Scholars Journals 

 

Author(s) retain the copyright of this article. 
 

 

Full Length Research Paper 

  

The dynamics of horticultural export value chains on 

the livelihood of small farm households in Southern 

Ghana 

 
Afari-Sefa Victor 

 
Senior Research Fellow, Department of Economic and Technological Change, Center for Development 

Research (ZEF), Walter Flex-Street. 3, D-53113 Bonn, Germany. E-mail: vafaris@uni-bonn.de 
 

Accepted 26 March, 2020 
 
Policy makers in developing countries have been concerned with the economic and political risks associated 
with heavy dependence on few specialized raw materials as main sources of government revenue and foreign 
exchange. Development partners and donor agencies have equally extolled the need for these countries to 
diversify their export base as a poverty reduction strategy. As a result, several African countries have tended to 
focus on non-traditional agricultural exports (NTEs) which reflect their comparative advantage and for many 
countries the export of horticultural crops has been favoured. This study focuses on a household survey 
undertaken in the forest and coastal savannah transition zones of Ghana, where the farming system has 
undergone a remarkable transition from an established system of food crop farming for sale to urban 
consumers to an intensive production of fruits and vegetable crops for export to European consumers. 
Econometric analysis shows that though export horticulture has a positive impact on the wellbeing of the 
majority of households, the chronically poor households are structurally impeded from seizing the available 
opportunities due to poor resource endowment and liquidity constraints. 
 
Key words. Export horticulture, food security, household livelihood, non-traditional exports, export diversification, pro-

poor growth. 

 
INTRODUCTION 

 
Various studies have shown that countries can improve 
their welfare by opening up their borders to freer trade. 
Moreover, there is a worldwide move toward economic 
integration, the EU being the most prominent example. 
Not only is it foreseen that this movement will improve 
welfare of a country but it’s competitiveness could also 
improve by generating foreign exchange earnings and 
fiscal revenues to increase the income of smallholders 
and to provide employment for the rural poor. In response 
to a liberalised trade regime, horticultural exports have 
grown dramatically in many Sub-Saharan African coun-
tries while many other traditional agricultural commodities 
have faced stagnation and declining world prices. Con-
sequently, several reasons have been advanced for the 
recent boom in horticultural exports from Africa (Jaffee, 
1995; Barrett et al., 1997; Dixie, 1999; Malter et al., 
1999). In Ghana, crops such as pineapples, papaya and 

 
 
mangoes appear promising as options to diversify the 
traditional export base comprising the traditional cocoa, 
timber and gold because of their high labour intensity and 
the expanding demand for fruits in industrialized nations, 
most especially during the winter months. Undoubtedly, 
the horticultural industry provides an important source of 
foreign exchange generates substantial employment and 
has contributed to the upgrading of agricultural produc-
tion skills. But has the growth in export horticulture contri-
buted significantly to smallholder food security and gene-
ral livelihood? A number of studies have raised concerns 
about the microeconomic performance of non- traditional 
exports in developing economies. Most of such concerns 
are related to the trade-offs between food and export 
cash cropping systems due to the possibility of compe-
tition for resources between export crops and food crops 
resulting from a potential re-allocation of resources from 

file:///C:\Users\user\Documents\REPUBLICATION\AGRICULTURAL%20SCIENCES\AppData\Local\Temp\www.internationalscholarsjournals.org


 
 
 

 

one to the other and their effect on household food avail-
ability. Indeed potential synergy effects have been identi-
fied between cash-crop investment and food productivity, 
whereby positive spill over benefits of increased input are 
made possible for food crops through cash crop delivery 
channels (Dione, 1989; Goetz, 1993; Goverah and Jay-
ne, 2003; Von Braun, 1995). Consequently, there are cri-
tics of such policies that advocate cash crop production 
(Von Braun and Kennedy, 1986; Weber et al., 1988). 
They argue that the benefits have never materialized with 
the premise that, in areas where cash crop production 
has increased, food consumption and the nutritional sta-
tus of the poorest households have deteriorated.  

Within the Ghanaian context, the main effects of the 
introduction of export cropping has been the significant 
deterioration in access to land as smallholder food crop 
farms are being consolidated into larger scale export crop 
farms. The increased pressure on arable lands for human 
settlement resulting from the expansion of urban activities 
from Accra towards surrounding villages are the pressing 
underlying concerns that need the immediate attention of 
researchers. So far the major concern of the Ghanaian 
government and donor institutions has been macroeco-
nomic growth in terms of physical output and export earn-
ings of these new crops. At the microeconomic level, the 
short and long-term impacts of the booming horticultural 
exports vis-à-vis their distributional effects still remain 
under-investigated. In an attempt to fill this void, this 
paper emphasizes on the linkages between export diver-
sification policies and the microeconomic performance of 
some horticultural export crop producing communities in 
southern Ghana. 
 

 
MATERIALS AND METHODS 
 
The analysis presented in this paper is based on a primary data set 
collected by a survey of 200 farm households in southern Ghana. 
The survey was carried out during the 2003/2004 cropping season. 
In accordance with the importance of the various horticultural crops 
to total export earnings, 7 villages with 20 households each within 
the pineapple cultivated based communities of the Akwapim south 
district and 3 villages with 20 households each from the mango 
cultivated based communities of the Dangme west district were 
selected for the survey using the stratified random sampling 
approach.  

Descriptive statistic tools were initially used to categorize the 
sampled farm households into three main groups based on the type 
of crops grown and other socio-economic characteristics. On the 
basis of this typology, the respondents comprised of 44 non-
horticultural households, 118 horticultural and staple households, 
and 38 horticultural households. The major differences and similari-
ties among the three household categories have been outlined 
based on the extent of participation in export horticulture. Probit 
regression was used to estimate the determinants of household 
participation in export horticulture. Results from the field survey 
gave an initial indication that participation in export horticulture, 
primarily appears to increase the average per capital income of the 
household. However, as explained by McCulloch and Ota (2002), 

 
 
 
 

 
such descriptive statistics do not take account of other possible 
differences in the characteristics of adopters and non- adopters and 
it may be these differences that are giving rise to the disparities in 
income rather than their participation in export horticulture. To 
account for this, the traditional Log-linear mincerian wage equation 
was used to construct a model of income determination to analyze 
the incomes of households producing only horticultural export crops 
with those producing mainly food crops vis-à-vis those indulged in 
the combined scenario, while making use of the same explanatory 
variables employed in the participation determination model. Based 
on the assumption that households typically distribute their labour 
and capital resources over a set of productive activities, a reduced 
form expression for income Y, as a function of the explanatory 
variables is specified and accordingly estimated as: 

 
In( Y ) = 0 + 1 HHCrop + 2 Demog + 3 Educ + 4 Employ + 

5Asset + 6Misc + 7IntNTE + 

 
Where, HHCrop is a vector of dummy variables indicating catego-
ries of non-horticultural export crops grown by the household; 
Demog is a vector of demographic characteristics; Educ is a varia-
ble indicating the educational level attained by the household head; 
Employ is a vector of dummy variables indicating whether at least 
one household member participates in a specific off-farm occupa-
tion; Asset is a vector of asset and access variables; Misc is dummy 
variable indicating miscellaneous income receipt; IntNTE is a vector 
of variables indicating the intensity of cultivation of non-traditional 
export crops and e is an error term. 

 

RESULTS AND DISCUSSION 
 
Inter-household landholdings and capital resource 

distribution 
 
Households decide whether or not to participate in the 
cultivation of horticultural export crops based on percei-
ved benefits, opportunities and constraints considering 
the risks and probabilities of occurrence. Generally, horti-
cultural households had the highest land resource in 
terms of endowment, total cultivated farm size, farm size 
per capita and fallow land area. This was followed by 
horticultural and staple households and finally non-horti-
cultural households in chronological order. The Kruskal-
Wallis test in Table 1 showed statistically significant 
differences for the landholding characteristics of the 
household categories except for the fallow land area. This 
might be expected because land ownership status of 
some households in the various categories are based on 
traditional inheritance patterns or asset symbolization 
status of land as a result of which some households 
possess parcels of land that are not necessarily under 
current cultivation. Per capita income was also highest for 
horticultural households, followed by horticultural and 
staple households with non-horticultural households hav-
ing the lowest income (Afari-Sefa, 2006). It was observed 
that 78% of the sampled households were cultivators of 
horticultural export crops. This is an obvious reflection of 
the magnificent role these crops play in the daily lives of 



 
 

 
Table 1. Land holding characteristics per sampled household category,Southern Ghana, 2003/2004. 

 

Household type  Total land Total farm Farm size per Fallow land 

  Endowment (ha) size (ha) capita (ha) area (ha) 

Horticultural households Mean 5.61 3.56 0.94 2.05 

(N=38) SD 
a
 5.83 4.24 1.12 2.29 

Horticultural and Staple Mean 4.87 2.98 0.54 1.89 

households (N=118) SD 5.52 4.07 0.55 2.61 

Non-horticultural households Mean 2.61 1.52 0.35 1.09 

(N=44) SD 2.40 1.47 0.44 1.42 

Total (N=200) Mean 4.52 2.77 0.57 1.75 

 SD 5.16 3.75 0.70 2.35 

Kruskal-Wallis test 2 16.28 
***

 20.24 
***

 18.03
***

 4.32 
 

Note: SD denotes standard deviation of corresponding variable.*** denotes significance at 1% probability level. 
 
 

 

of most households in the study area. The pooled results 
of the analysis indicate that pineapple cultivation is the 
most widely adopted crop by indigenous small scale far-
mers, whereas the cultivation of crops such as mangoes 
required a longer investment period and hence mostly 
adopted by immigrant large-scale commercial farmers. 

As would be noted from the agronomic point of view, it 
is not all fruits which are considered exportable. Depend-
ing on a farmers’ level of experience, between 50-90% of 
matured fruits may satisfy requirements for export. The 
remaining fruits may be rejected because they do not 
meet the requisite size, weight, and shape specifications. 
The majority of respondents that do not cultivate horticul-
tural export crops within this risk- prone category were 
found to be older household heads that have been used 
to the practice of cultivating staple crops for a minimum of 
twenty years. 
 

 

Determinants of participation in horticultural export 

crop production 
 
In accordance with the major objective of this study, the 
explanatory variables have been chosen from factors and 
characteristics related to profitability of the crop or resour-
ce requirements for production and exogenous factors 
(locational and institutional) that are closely associated 
with input and market access for horticultural export crop 
produce. As expected the estimated coefficients of the 
hypothesized explanatory variables have the priori ex-
pected signs and most of the coefficients are statistically 
significant at the 10% (or better) probability level. The 
likelihood ratio statistic was used to test the significance 
of the entire Probit model. The model chi-square value of 
79.44 at 11 degrees of freedom was found significant at 
the 1% level implying that the independent variables, 
taken together, significantly influences a household’s de- 

 
 
 

 

cision to cultivate a horticultural export crop (Table 2). 

The pseudo-R
2
 value implies that 60.3% of the original 

variation of the dependent variable is explained by the 
fitted model and still there are other determinants of 
participation in export horticulture that are not included in 
the model. Nevertheless, the model correctly predicted 
82% of participation status in horticultural exports for the 
sampled households.  

The results show that the age of head has a strong 
negative effect on cultivation of horticultural export crops, 
implying that households with older heads are unlikely to 
cultivate non-traditional export crops (NTEs). This is anti-
cipated because the cultivation of export crops requires 
much care and risk in terms of new technology and preci-
sion in the application of various cultural practices that 
may not be of paramount interest to older farmers. More-
over, the existence of a large cohort of younger farmers 
in the sample could have attributed to this observation. 
These were mostly resident natives, city returnee natives 
and other immigrants, who have basically opted into far-
ming mainly because of the perceived lucrative nature of 
export horticulture coupled with the lack of equal alterna-
tive income earning opportunities compared with the 
older farmers, who have been engaged in the cultivation 
of traditional crops as their means of livelihood long 
before the introduction of export horticulture in the study 
area. Participation in off-farm occupation has a significant 
positive effect on the likelihood of cultivating horticultural 
export crops at the 10% probability level. Thus variables 
indicating participation in other economic activities (wage 
employment and the operation of a non-farm enterprise) 
appear to be positively associated with being a horticul-
tural small-holder. This relationship could be explained by 
the fact that households with off-farm occupation would 
easily overcome liquidity constraints and make provisions 
for holder. This relationship could be explained by the fact 
that households with off-farm occupation would 



 
 
 

 
Table 2. Probit function for the likelihood of participation in horticultural export crop production, Southern 

Ghana, 2003/2004. 
 

Variable Marginal effect Standard error T-statistics 

Intercept 0.479*** 0.149 3.205 

Age of head -0.009*** 0.002 -3.734 

Educational level of head 0.000 0.006 0.063 

Residential status (dummy) -0.034 0.052 -0.658 

Household size 0.002 0.009 0.211 

Non-farm occupation (dummy) 0.104* 0.062 1.674 

Land endowment 0.050*** 0.009 5.472 

Nature of road network -0.049** 0.024 -2.069 

Capital input access (dummy) 0.092* 0.055 1.689 

Tropical Livestock Unit 0.001 0.003 0.515 

Cultivate local cash crops (dummy) -0.286*** 0.060 -4.803 

Frequency of extension visits -0.005 0.014 0.387 
     
Dependent variable: Adoption of Horticultural Export Crop; Model Chi-Square = 79.44*** ; Log Likelihood function 

= -70.49; Pseudo R
2
 = 0.603 ; Households correctly predicted: 82% ; N = 200; *, ** and *** denotes significance at 

10%, 5% and 1% levels respectively 

 

 

easily overcome liquidity constraints and make provisions 
for the initial investment in capital inputs required for 
producing the labour intensive horticultural export crops. 
On the other hand it is also possible that the lucrative nat-
ure of the export horticulture industry provides house-
holds originally not engaged off-farm with sufficient inco-
me to diversify their income sources as a surety for their 
livelihood needs.  

The total land endowment of households is highly signi-
ficant and positively related to the probability of cultivating 
horticultural export crops. This seems to indicate that a 
decision-maker with a higher land access usually cultiva-
tes horticultural export crops. Thus households who have 
more hectares of land could easily afford allocating sepa-
rate land parcels for cultivating both staples and horticul-
tural export crops. The positive relationship with land 
endowment could be justified because households with 
more access to land can also afford fallowing already 
harvested fields for some years and shift to cultivating 
other land parcels whiles allowing sufficient time for nut-
rient regeneration on the temporarily fallowed fields. The 
same could not be likened to households with a compara-
tively smaller endowed land moreso if permanent tree 
crops are to be cultivated.  

The nature of the road network from farms to major 
marketing centres has a negative significant relationship. 
This seems to indicate that bad roads increase the 
likelihood of adopting NTEs. However, taking cognizance 
of the fact that, responses to this variable constituted an 
ordinal subjective evaluation of households, the interpret-
tation of this rather surprising observation must be done 
with extreme caution. On a comparative basis, fields 
where NTEs are cultivated generally have a bad road 

 
 

 

network than for alternative crops in the study area. 
Moreover, by the nature of the marketing arrangement for 
NTEs, transportation of produce is solely the respon-
sibility of exporters and other subsidiary buyers and 
hence it is rational for farmers to locate staple and other 
crops on plots closer to their abode where road motora-
bility is comparatively better than for the NTEs. Further-
more, the cultivation of some NTEs such as pineapples 
usually requires bringing new land under cultivation as 
compared to the cultivation of other alternative annual 
crops, most of which could be easily cultivated on contin-
uously cropped fields. In addition, the observed distant 
location of export crops such as pineapple might reflect 
household perception of cropping choices by matching 
crops with appropriate soils as confirmed by Goldstein 
and Udry (1999).  

Access to capital inputs has a positive effect on the 
probability of adopting NTEs at the 10% level. This 
means that households with better access to capital in-
puts or credit access to purchase the requisite inputs for 
cultivating NTEs are more likely to participate in the 
sector than those who have relatively poor access to 
capital inputs and credit. Typically, the cultivation of NTEs 
requires a higher initial capital outlay in terms of planting 
material and agrochemicals which cannot be afforded by 
marginal farmers. Therefore, good access to credit either 
in the form of inputs or cash for paying hired labour costs 
or both would increase a household’s probability of culti-
vating horticultural export crops. Finally, cultivating of ot-
her local cash crops has a highly significant negative eff-
ect on the adoption of NTEs. This means that households 
cultivating local cash crops of insignificant export value 
are unlikely to indulge in NTE cultivation be on account 



 
 

 

Table 3. Parameter estimates of the determinants of total household income southern Ghana 2003/04.  
 

Variable  

 
 
Coefficient  

 
 
Standard error  

 
 
T-statistics   

Intercept  
Land endowment  
Food crop cultivation (dummy)  
Other local cash crop (dummy)  
Age of head  
Education of head  
Residential status (dummy)  
Years of cultivating export crop  
Dependency ratio  
On-farm family labour capacity  
Labour hired out (dummy)  
Credit/Input access (dummy)  
Tropical Livestock Units  
Public paid employment (dummy)  
Trade and Services (dummy)  
Miscellaneous income (dummy) 

 
Facilities welfare index  
Ratio of NTEs area to total farm size 

 
14.7535***  
0.0411**  
0.0313  

0.2625**  
0.0014  

0.0276**  
-0.1555  

0.0420***  
-0.1810***  
0.0007***  

0.1697  
0.0367  

0.0141***  
0.4405***  
0.2671***  
-0.2363* 

 
0.0397*  

1.5941***  

 
0.3661  
0.0186  
0.1612  
0.1040  
0.0042  
0.0133  
0.1122  
0.0135 
 
0.0644  
0.0002  
0.1256  
0.0989  
0.0047  
0.1082  
0.0911 
 
0.1402 
 
0.0229  
0.2299 

 
40.2996  
2.2134  
0.1931  
2.5628  
0.3413  
2.0731  
-1.3855  
3.1096  
-2.8116  
3.1588  
1.3511  
0.3714  
2.9732  
4.0727  
2.9311  
-1.6855 
 
1.7348  
6.9353 
  

Dependent variable: Natural log of Total household income; F = 26.31; R
2
 = 0.711; Adjusted R

2
 = 0.684; N = 

200 ; Probability >F = 0.00; Breusch-Pagan chi-squared for heteroscedasticity correction = 98.70; *, ** and *** 

denotes significance at 10%, 5% and 1% levels respectively. 
 
 

 

of fact that some households rather opt for cultivating 
these local cash crops which have a lower profit returns 
compared to the NTEs but at the same time have a lower 
marketing risks in terms of produce rejection and pay-
ment duration. Since these local cash crops are also 
equally labour intensive, there is the possibility for them 
to compete for similar production resources with percei-
ved changes in the comparative advantages for different 
households. 
 
 
Determinants of household income 

 

Three different empirical proxies of the influence of 
human capital are used as explanatory variables. The 
dependency ratio has a statistically significant negative 
influence (Table 3) on total household income. This 
means that the higher the number of children and elderly 
household members in relation to the number of active 
adult work force, the lower the household income. Con-
trary to the findings of the model of participation in export 
horticulture, the age of household head has a positive, 
non-significant relationship with household income, emp-
hasizing the role played by other productive activities 
outside export horticulture in the income earning power of 
the sampled households. 

 
 
 

 

The residential status of households, though negatively 
related to household income is not statistically significant. 
This somehow confirms that, with regards to residential 
status, there is the potential for immigrants to be more 
market-oriented than resident natives. Households invol-
ved in the production of local cash crop have the pro-
pensity to increase their income by 30.1% whereas pro-
duction of staple crops increases household income by 
3.18%, although not statistically significant. This confirms 
the lower incomes of non-horticultural households, many 
of whom earn their major incomes from staple crop pro-
duction. This observation may also explain the observed 
trend of the food security status of the different household 
categories.  
Generally, variables related to household physical capi-tal 
endowment turned out to be important determinants 
influencing agricultural activity choice. As might be 
expected, the index of welfare facilities has a positive and 
statistically significant but very small influence on income. 
Unsurprisingly, the more assets a household has the hig-
her its income, but the dummy variable related to credit 
access is neither strongly nor statistically significantly 
associated with higher income. This is a reflection of the 
inadequate levels of credit required to realize higher inco-
mes from the various productive activities in the study 
area, a fact that was evident with our respondents. Gene- 



 
 
 

 

rally, hiring out of labour is expected to increase hou-
sehold income. Indeed, incomes from agricultural wage 
labour are very low, in comparism with other “hard to find” 
alternatives, thereby leaving most households with the 
only rational option of utilizing family labour for their own 
managed productive activities. This is justified by the high 
positive and significant relationship between family labour 
capacity for agricultural production and total household 
income. Consequently, households with the ability to hire-
in labour also tend to have higher incomes and returns 
from crop production. Different occupations have statis-
tically significant implications for income as households 
engaged in paid public employment and other self emp-
loyed activities tend to have substantially higher incomes. 
Not surprisingly, households who were in receipt of mis-
cellaneous income by way of transfers and remittances 
were found to be those who obtained lower income from 
own farm and non-farm productive activities. Finally, the 
explanatory variables that are related to the intensity 
cultivation of horticultural exports play a very significant 
role in determining household incomes with a unit increa-
se in the ratio of the area cultivated by horticultural export 
crops to total farm size, in percentage terms increasing 
household income by 1.6%. Similarly, an additional year 
of experience in the cultivation of horticultural export 
crops increases household income by 4.2%. 

 

Conclusions 
 
Empirical evidence has shown that the horticultural exp-
ort has increased the opportunity for higher earnings for 
smallholders and that the much higher land sizes owned 
by horticultural smallholders are indeed a cause or con-
sequence for their participation in the sector. Households 
cultivating horticultural export crops are on the average 
better off than those that do not. Notwithstanding the 
enormous contribution of horticultural exports to macro-
economic growth as an insulator from sharp and unex-
pected changes in the terms of trade of Ghana’s tradi-
tional exports, the micro level distributional effects has 
not favoured the chronically poor households who are 
structurally impeded from seizing the existing opportuni-
ties of the export boom by virtue of their poor resource 
endowment and liquidity constraints. These marked dif-
ferences in resource base between the various house-
hold categories further accentuate the imperfections with-
in rural markets.  

The majority of households are particularly exposed to 
the risk of inadequate technological know-how in meeting 
the ever increasing quality standards and health control 
traceability requirements by European consumers, price 
collapse on the export market and a break down of local 
marketing institutions. The findings from this paper there-
fore calls for an integrated policy framework approach 
aimed at improving rural market imperfections. Efforts to 

 
 
 
 

 

achieve the desired impacts requires the strong need for 

investment in infrastructure and a shift towards value-
added export oriented production, whereby small farm 

households are progressively integrated into the chang-
ing preferences of a dynamic global food chain. 

 
REFERENCES 
 
Afari-Sefa V (2006). Agricultural export diversification, food Security and 

living conditions of farmers in southern Ghana: A microeconomic and 
household modeling approach; Farming & Rural Systems Economics, 
Margraf Publishers, Vol. 81, (ISBN: 3-8236-1490-8), Weikersheim, 
Germany.  

Barrett HR., Browne AW, Ilbery BW, Jackson G and Binns T (1997). 
Prospects for horticultural exports under trade liberalization in 
adjusting African economies. Report for DFID, Coventry University, 
UK.  

Dione J (1989). Informing food security policy in Mali: Interactions 
between technology, institutions and market reforms, Unpublished 
PhD Dissertation, Michigan State University, East Lansing, MI, USA. 

Dixie G (1999). Summer citrus: The role and prospects for Southern 
Africa. In: Jaffee S (ed.), Southern African agribusiness: Gaining 
through regional collaboration. World Bank, Washington, DC. 

Goetz S (1993). Interlinked markets and the cash crop-food crop debate 
in land-abundant tropical agriculture. Economic Development and 
Cultural Change 41: 343–361. 

Goldstein M and Udry C (1999). Agricultural innovation and resource 
management in Ghana, Final report to IFPRI under MP17, August 
1999. 

Goverah J, Jayne TS (2003). Cash cropping and food crop productivity:  
Synergies or trade-offs? Agric. Econ. 28: 39–50, Elsevier.  

Jaffee SM (1995). The Many faces of success: The development of 
Kenya’s horticultural exports. In: Jaffee S and Morton J (eds.), 
Marketing Africa’s high-value foods; Comparative experiences of an 
emergent private sector, Dubuque, IA: Kendall / Hunt Publishing 
Company, pp. 319-374.  

Malter A J, Reijtenbagh A and Jaffee SM (1999). Profits from petals: 
The development of cut flower exports in Southern Africa. In: Jaffee S 
(ed.), Southern African agribusiness: Gaining through regional 
collaboration, World Bank Technical Paper, Washington, DC, USA.  

McCulloch N and Ota M (2002). Export horticulture and poverty in 
Kenya. Globalization, production and poverty discussion Paper No.4, 
Overseas Development Group, University of East Anglia, Norwich, 
UK.  

Von Braun (1995). Agricultural commercialization: Impacts on income 
and nutrition and implications for food policy, Food Policy 20: 187-
202. 

Von Braun and Kennedy E (1986). Commercialization of subsistence 
agriculture: income and nutritional effects in developing countries, 
Working Papers on commercialization of agriculture and nutrition, 
IFPRI, Washington DC. Weber M, Staatz J, Holtzman J, Crawford E 
and Bersten R (1988). Informing food security decisions in Africa. 
Empirical analysis and policy dialogue. In: Am. J. of Agric. Econ. 
70(5):1044-1052. 


