The diversity of postharvest losses in cassava value chains in selected developing countries

Abstract

The extent of physical and economic postharvest losses at different stages of cassava value chains has been estimated in four countries that differ considerably in the way cassava is cultivated, processed and consumed and in the relationships and linkages among the value chain actors. Ghana incurs by far the highest losses because a high proportion of roots reach the consumers in the fresh form. Most losses occur at the last stage of the value chain. In Nigeria and Vietnam processors incur most of the losses while in Thailand most losses occur during harvesting. Poorer countries incur higher losses despite their capacity to absorb sub-standard products (therefore transforming part of the physical losses into economic losses) and less strict buyer standards. In monetary terms the impact of losses is particularly severe in Ghana and estimated at about half a billion US dollar per annum while in the other countries it is at the most about USD 50 million. This comparison shows that there are no "one-size-fits-all" solutions for addressing postharvest losses but rather these must be tailor-made to the specific characteristics of the different value chains.

Full text

Journal of Agriculture and Rural Development in the Tropics and Subtropics
Vol. 115 No. 2 (2014) 111–123
urn:nbn:de:hebis:34-2014121946902 ISSN: 1612-9830 – journal online: www.jarts.info
The diversity of postharvest losses in cassava value chains
in selected developing countries
Diego Naziria,∗, Wilhelmina Quayeb, Bernard Siwokuc, Sittichoke Wanlapatitd,
Tu Viet Phue, Ben Bennettf
aNatural Resources Institute (NRI), University of Greenwich, UK and International Potato Centre (CIP), Lima, Peru
bFood Research Institute (FRI), Council for Scientific and Industrial Research (CSIR), Ghana
cFederal University of Agriculture, Abeokuta (FUNAAB), Nigeria
dCassava and Starch Technology Research Unit, National Center for Genetic Engineering and Biotechnology (BIOTEC), Thailand
eSchool of Biotechnology and Food Technology (SBFT), Hanoi University of Science and Technology, (HUST), Vietnam
fNatural Resources Institute (NRI), University of Greenwich, UK
Abstract
The extent of physical and economic postharvestlosses at different stages of cassava value chains has been estimated in
four countries that differ considerably in the way cassava is cultivated, processed and consumed andin the relationships
and linkages among the value chain actors. Ghana incurs by far the highest losses because a high proportion of roots
reach the consumers in the fresh form. Most losses occur at the last stage of the value chain. In Nigeria and Vietnam
processors incur most of the losses while in Thailand most losses occur during harvesting. Poorer countries incur
higher losses despite their capacity to absorb sub-standard products (therefore transforming part of the physical losses
into economic losses) and less strict buyer standards. In monetary terms the impact of losses is particularly severe in
Ghana and estimated at about half a billion US dollar per annum while in the other countries it is at the most about
USD 50 million. This comparison shows that there are no “one-size-fits-all" solutions for addressing postharvestlosses
but rather these must be tailor-made to the specific characteristics of the different value chains.
Keywords: physical losses, economic losses, cassava, Ghana, Nigeria, Thailand, Vietnam
1 Introduction
Since the onset of the food crisis initiated in
2006/2007 there has been a change in development pri-
orities, bringing an increased focus on agriculture and a
renewed interest in the reduction of postharvest losses
(PHL) as a means of increasing food availability. The
review “The Future of Food and Farming” (Foresight,
∗Corresponding author
Natural Resources Institute, University of Greenwich,
Central Avenue, Chatham Maritime, ME4 4TB,UK
Email: d.naziri@greenwich.ac.uk
Phone: +44.1634.88.30.69; Fax: +44.1634.88.37.06
2011) and the “Missing Food” report (World Bank,
2011) highlighted the reduction of PHL as key response
to global food availability concerns.
Food losses take place at production, post-harvest
and processing stages in the supply chain. Losses at
the end of the chain (retail and consumption) are often
called “food wastes”, which relates to retailers’ and con-
sumers’ behaviours (Parfitt et al., 2010). In this study,
with the term “losses” we refer to both losses and wastes
but clear distinction is made about the value chain stage
where they occur.
PHL reduction offers the particular advantage of in-
creasing food availability without requiring additional

112 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
land, water, labour, energy and agricultural inputs. In
view of these advantages an in-depth analysis of the
cassava value chains in four countries (Ghana, Nigeria,
Thailand and Vietnam) was undertaken to identify the
extent and causes of physical and economic losses and
estimate their impact in monetary terms.
Cassava (Manihot esculenta) is the third most im-
portant source of calories in the tropics, after rice and
maize; and the second one in Africa (FAO, 2004). In
Sub-Saharan Africa (SSA) it is mainly grown by small-
holder farmers often on marginal land where it is pro-
ductive even on poor soils and under drought conditions.
As such cassava is a vital crop for both food security
and income generation in least developed countries. In
South-East Asia cassava is grown almost exclusively as
cash crop, sometimes in large plantations, primarily for
industrial processing into dry chips and starch.
SSA is the most important cassava production region
in the world and Nigeria the world’s leading producer
(FAOSTAT, 2013). South-East Asia ranks second as
largest production region, where the main producers are
Indonesia, Thailand and Vietnam. Thailand and Viet-
nam are the largest world’s exporters of cassava starch
and chips.
Once harvested cassava root is highly perishable and
the rapid postharvest deterioration that cassava incurs
actually restricts its storage potential to two to threedays
(Iyer et al., 2010). As storage of roots is rare, the most
common and sensible way to minimise losses is to con-
sume or process them as soon as possible after harvest-
ing. Unfortunately, this does not always happen and sig-
nificant amount of roots spoil or incur various degreesof
quality deterioration.
These losses have a broad range of negative impacts
such as loss of income and food intake and represent
an obstacle for transforming cassava from a subsistence
to a cash crop, particularly in SSA. Fresh cassava roots
(FCR) are affected by two types of postharvest deterio-
ration: primary physiological deterioration that involves
internaldiscolorationandsecondarydeteriorationdueto
microbial spoilage (Booth & Coursey, 1974). As well
as direct physical loss of the crop, postharvest deteri-
oration causes a reduction in quality, which has impli-
cations on marketing of cassava leading to price dis-
counts. Furthermore, there can be additional losses due
to change in use. For example, in countries where cas-
sava is mainly eaten in fresh form, if the roots cannot
be marketed within two or three days from harvest they
may be processed into dried products of lower value
(Westby, 2002).
Little reliable information is available that quantify
the extent of losses. Several studies have commented on
the high levels of losses in the cassava value chain but
without attempting to estimate them (Scott et al., 2000;
Vowotor et al., 2010). In some documents these esti-
mates seem to be only gross values based on anecdotal
evidence. Wenham (1995) reports that PHL in a num-
ber of SSA and Asian countries range between 5% and
30 % but does not indicate their cause or the value chain
stage where they occur.
FAO (2011) has attempted a systematic assessment
of losses for a number of commodities worldwide. The
estimates refer to physical losses only while economic
losses are not reported. Losses of root and tuber crops
in developing regions have been estimated at 40% in
SSA, 31% in South and South-East Asia and 35 % in
Latin America.
GIZ has assessed cassava PHL in Nigeria focusing
on gari and starch value chains only (Oguntade, 2013).
The study found that most losses occur at the processing
stage and affect almost exclusively FCR rather than pro-
cessed products. Other studies have focused on PHL in-
curred during traditional processing, for instance during
gari (Boahen, 2004) and agbelima processing (Dzied-
zoave et al., 1999) in Ghana.
Our study allows comparison between countries that
differ considerably in the way cassava is cultivated, pro-
cessed and consumed and in the relationships and link-
ages among the value chain actors.
2 Materials and methods
In order to estimate the extent of PHL along a com-
modity chain and to identify appropriate strategies and
technologies to reduce them, the full range of activities
required to bring a product through different stages of
production, processing, and marketing until it reaches
the end-user has to be evaluated. A Value Chain Anal-
ysis (VCA) provides one approach for such an under-
standing in that it is a process of tracing a product’s flow
from the point of production to the point of consump-
tion along with identifying the roles and relationships
of different stakeholders at different points of the chain
(Kaplinsky & Morris, 2001).
Following Yin (2003) a mix of both quantitative and
qualitative approaches were adopted. The VCAs have
been designed in such a way to gather specific additional
layers of information on how much, where and when
PHL occur and what are the main causes and remedies
adopted by the different actors.

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
113
A semi-structured questionnaire was specially de-
signed for data collection. The questionnaire covered
core processes, traded volumes, prices and price mecha-
nisms, seasonality, standards, types and major causes of
losses, and information on loss quantities and mitigation
measures for each cassava product.1The data collection
has been carried out by individual country teams be-
tween July and October 2012. In addition to individual
interviews with value chain actors, focus group discus-
sions were organised for triangulation purposes. 2Key-
expert interviews were heldwith relevant publicofficers,
managers of other cassava projects and initiatives, agri-
cultural inputs and credit suppliers. Furthermore, obser-
vations of production, processing and distribution activ-
ities were conducted to cross check information given
by respondents.
In order to ensure consistency two preliminary work-
shops have been held in Vietnam and Ghana to design
the methodology and develop a common definition of
PHL. In fact the definition of PHL is a contended sub-
ject, often defined on a situational basis, and attempting
to find a universally accepted definition is beyond the
scope of this study. Based on the inputs provided during
the workshops and discussions heldwith postharvest ex-
perts we have defined two distinct categories of losses,
namely physical losses and economic losses.
For the purposes of this study, physical losses are
losses incurred in the value chain which do not have
alternative uses or residual value. Physical losses in-
clude the following: (a) product left behind in the field
after harvesting; (b) spoiled or damaged product that
is thrown away; and (c) product that disappears at any
value chain stage (e.g. eaten by pests).
Physical losses can affect either FCR or processed
products. Value chain maps were developed based on
literature and validated during the survey. In line with
Gustavsson et al. (2011) physical losses have been cat-
egorised according to the value chain stage where they
occur, namely: (i) on-farm losses, including harvesting
and postharvest losses; (ii) losses during trading, trans-
1Cassava can reach the final consumer or user in either fresh form
or processed in a number of ways. The term value chain encompasses
the whole array of different products (and hence processes) a single
commodity can be traded as. Since very different losses can be in-
curred when cassava is consumed and traded in one form rather than
another, for each country, we have broken down the value chain in
his different ramifications (to which we refer to as “sub-chains”) and
estimated the losses for each of them. However, in this study, sub-
chains absorbing less than 1% of the country production have been
neglected.
2Due to the different characteristics of the cassava value chain and
the heterogeneity in the answers of the value chain actors the number
of respondents varied from country to country and ranged between 99
(Thailand) and 145 (Ghana).
port and handling of cassava and cassava products; (iii)
losses at the processing sites;3and (iv) losses at distri-
bution, retail and consumption level.4
Based on the information gathered from the inter-
views, we have estimatedthe most likely level of losses
incurred at each stage of each sub-chain in percentage.
Using as a starting point the national annual production
statistics or relevant literature, we calculated the extent
of losses in absolute terms at each sub-chain stage, net
of losses incurred at previous stages in order to avoid
double counting (therefore we have been more rigor-
ous than simply summing up the relative losses of all
stages and then applying this figure to the overall an-
nual production). Once we have determined how much
is physically lost at each stage, we havethen calculated
the monetary value of physical losses by taking into ac-
count the average market price at the relevant point of
the chain.
As far as economic losses are concerned these, un-
like physical losses, have alternative uses. In this study
economic losses refer to (a) spoiled or damaged prod-
uct whose market price is discounted and (b) spoiled or
damaged product that cannot be used for what it was ini-
tially meant (e.g. damaged roots processed into lower
value products). Since the major problem is the dete-
rioration of the root we have assumed that only FCR
incur economic losses. For FCR we have identified two
typical levels of quality deteriorationsold at discounted
price to either final consumers or processors. For each
level we have determined the level of price discount rel-
ative to the price of good quality roots at that point of
the chain. The market value lost due to quality deterio-
ration has been used to estimate the extent of monetary
losses.5
3 Results and discussion
3.1 Cassava production, processing and products in
target countries
In Ghana, cassava production in 2011 was estimated
at 14.2 million tonnes. Cassava is cultivated by over
90% of the farming population, mostly intercropped in
small plots (0.4 to 12 hectares). About 95% of har-
vested cassava is commercialised and mainly used for
the preparation of traditional food such as fufu,gari,
3Excluding on-farm processing.
4Excluding farmer household’s own consumption.
5In some cases the assessment of the quality losses might be very
difficult, such as when a partially spoiled FCR is used to increase the
organic matter in the soil. In such a case quality losses have been
assimilated to physical losses.

114 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
agbelima and kokonte.Fufu is prepared at household
level and, at less extent, by caterers such as restaurants
and hotels by boiling and pounding peeled FCR. Gari is
a creamy-white, granular flour usually made by small-
scale women processing groups nearbythe cassava pro-
ductionareas. Agbelima is a fermented wet paste (called
fufu in Nigeria) that, similarly to gari, is largely pro-
duced by women groups around cassava farms. Kokonte
consists of dry cassava chips and chunks and is usually
produced by the farmers themselves, particularlyin the
regions where the consumption of fufu is not popular.
Small volumes of roots are used for industrial purposes
such as for the production of starch and high quality cas-
sava flour (HQCF). About half of the harvested roots are
processed into fufu.Gari and agbelima are also very
popular. Table 1 shows the estimated allocation of FCR
to the different sub-chains.
Nigeria is by far the world leading producer of cas-
sava (in 2010 the production was estimated at around 37
million tonnes). However, our study covered only the
South-West zone whose annual production is reported
at about 7.5 million tonnes.6Cassava is largely grown
intercropped by small-scale farmers on small plots (0.2
to 5 hectares) for both food and income purposes but
there are a few farmers who practice larger scale mono-
cropping and target processing factories. About 80 % of
production is marketed. The majority of roots is pro-
cessed into gari (Table 1) by village processing groups
usually located nearby the plantation areas. Substantial
amounts are processed into fufu that, differently from
Ghana, is a fermented wet paste producedby processors
around the cassava farms. Small volumes of roots are
used for industrial processing into HQCF and starch.
In Thailand the most recent figures estimate annual
production at around 22 million tonnes (2011). Cas-
sava is cultivated with extensive adoption of improved
agronomic practices (including chemical inputs, irriga-
tion and mechanisation) in plantations of 4 to 8 hectares
but farms larger than 80 hectares exist too. Cassava is
grown exclusively as cash crop for industrial processing
into starch and dry chips. It is estimated that 55% of
harvested roots are supplied to large-scale starch facto-
ries (Table 1). About 65 % of starch produced is for the
export market. The rest of the roots is processed into
dry chips by medium-scale processors. About 70 % of
chips are exported, almost exclusively to China where
they are used for livestock feed and as feedstock by the
emerging biofuel industry.
6Information about cassava processing and consumption patterns
in other parts of Nigeria is extremely scarce, and often unreliable. As
such we should be extremely careful in extrapolating these results to
Nigeria at large.
In Vietnam annual production is estimated at about
10 million tonnes (2011). The production scale signif-
icantly differs between North and South of the coun-
try. In the North farmers are mainly small-scale with
an average land size of less than one hectare. Im-
proved agronomic practices are not widely adopted
leading to considerably lower productivity than in the
South where cassava value chain is characterised by
higher level of developmentin both production and pro-
cessing. In the latter cassava is mostly produced by
large commercially-oriented farms with a size of 20–
30 hectares and up to 300 hectares. In Vietnam cas-
sava is almost exclusively processed into starch and dry
chips. About 55% of roots are processed by medium
and large-scale modern dry starch processing firms (Ta-
ble 1) primarily located in the South. In this part of the
country there is also some limited rudimental dry chip
production carried out by micro and small-scale pro-
cessors using almost exclusively low-quality roots and
cassava harvested from very small plots (such as in the
backyard) or left in the field after the main harvest. In
the North, starch factories cannot absorb entirely the
local production and a considerable amount of FCR
are locally processed into dry chips or transported to
surrounding provinces, mainly to specialised so-called
“craft villages” producing wet starch in the outskirts of
Hanoi. In the North chip processing is much more de-
veloped than in the South and is carried out by relatively
larger-scale chip processors usually equippedwith slic-
ing machines and coal kilns and exporting mainly to
China for its livestock feed and biofuel industries. Wet
starch (a semi-processed product that is further refined
into dry starch) is produced using simple equipment by
micro and small-scale processors.
3.2 On-farm physical losses
Two main types of on-farm losses have been identi-
fied. The first one refers to losses occurring during har-
vesting due to either the whole roots accidentally left
behind in the field unnoticed or broken roots that are
voluntary left behind in the field because they would not
find a market. The second type refers to the posthar-
vest deterioration when the roots are not immediately
consumed or processed. On-farm losses of processed
products are negligible.
In Ghana, with the exception of the FCR sub-value
chain the roots are always harvested by either the farmer
for immediate use (for own-consumption or processing
into kokonte) or the processor that arrange their prompt
transportation to the nearby processing site (for gari and
agbelima). In these cases the losses are negligible since

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
115
the roots can be left in the ground for a long time with-
out getting overripe or spoiled. On the other hand, in
some cases, fresh roots that have to reach the final con-
sumer without further processing (FCR sub-valuechain)
are harvested by the farmer and sold to an intermediary
at the farm gate. In such a case the collection may be
delayed for several reasons and, therefore, some losses
occur (estimated to affect 0.5 % of harvested roots).
In South-West Nigeria roots for own-consumption are
harvested a few at a time by the farmer and usually im-
mediately processed. Again, losses can be assumed as
negligible in this case. However, for gari and fufu, un-
like in Ghana, most roots are harvested by the farm-
ers themselves and sold at the farm gate to assembling
agents in charge of transporting and selling the roots to
the nearby processors. When delays occur, these may
determine some spoilage of roots (estimated to affect
1 % of roots).
In Thailand physical losses occur during harvest due
in particular to breakages from mechanised uprooting.
Unlike in poorer countries, broken roots are usually left
in the field because they would incur rapid deterioration,
thus not meeting the standard of the processor and be-
ing rejected. Moreover, dueto the relatively large size of
the farms some roots remain unnoticed and unharvested
in the soil. Besides, the roots are often sold to the pro-
cessors through an intermediary that collect them from
the production areas. Whenever the intermediary is not
in charge of harvesting and has to collect the roots at
the farm gate, some physical losses may occur due to
delays. Overall on-farm losses have been estimated at
1.5% for both the starch and chip sub-chains.
In Vietnam, in the dry starch sub-chain typical of the
South of the country, cassava is usually purchased by
the trader well before the harvest. He will be then in
charge to organise the harvest and immediate transport
of the roots to a nearby factory. Even when the farmer
decides to sell directly to the factory he often relies on
local traders for harvesting and transport. Whatever is
left unharvested in the field is usually gleaned by local
villagers and therefore consumed, even though not by
the intended value chain actors. Therefore losses can be
assumed as negligible at this stage. Conversely in the
North, although the harvesters are less likely to leave
unharvested roots in the typical small plots, some losses
occur in the wet starch and chip value-chains because
often there is an intermediary that purchases the har-
vested roots and transport them to the processing site.
As such some physical losses may occur in case of de-
lays. These have been estimated at 0.5% for both sub-
chains.
In conclusion, losses due to unnoticed roots seem to
be of some importance only in Thailand due to the com-
bination of three factors: large farm size; mechanised
harvesting; and lack of poor locals looking for unhar-
vested roots. In the other three countries on-farm losses
are less substantial mainly because of the small plot
size (Ghana, Nigeria and North Vietnam) or the pres-
ence of nearby villagers willing to glean (South Viet-
nam). Breakages of roots during harvesting occur in
all countries but are more substantial in Thailand due
to the widespread use of mechanical harvesters. This
is the only country were broken roots are left in the
field and represent a physical loss. In other countries
they are collected and sold at discounted price or pro-
cessed into lower value products, such as kokonte in
Ghana. Therefore, in these countries, breakages repre-
sent an economic loss rather than a physical one. On-
farm postharvest deterioration of roots affects all coun-
tries but seems more important in Nigeria where roots
are typically harvested by the farmer, rather than by the
buyer,and sold at the farm gate.
3.3 Physical losses in trading, transport and handling
Physical losses at this stage can affect both FCR and
processed products. However, we found that losses of
processed products, for instance due to damaged pack-
aging, are always negligible. Obviously no losses at this
stage are incurred when the roots are for domestic con-
sumption or processed on farm because, in these cases,
the roots never leave the farm. Two types of losses have
been identified at this stage. The first type refers to roots
that either have to be thrown away because of spoilage
during transport and trading (e.g. due to breakdown of
the vehicle or delays of the sale) or drop offduring the
journey. These represent physical losses. The second
type refers to breakages in particular during loading and
off-loading of the vehicle. These broken roots are then
sold at discounted price and, therefore, represent an eco-
nomic loss rather than a physical one.
In Ghana, in the FCR sub-value chain, roots have to
be transported over considerable longer distances than
when processed into gari and agbelima whose process-
ing occur nearby the cassava production area. As such,
it has been found that physical losses at this stage are
higher in the former case than in the latter one and they
were estimated at 1 % and 0.5%, respectively.
In South-West Nigeria the vast majority of roots to be
processed into gari and fufu are transported to nearby
processing sites. Again, since very short distances are
covered, the extent of physical losses is low and esti-
mated at 0.5%.

116 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
Table 1: Estimation of physical losses by stage of the sub-value chains.
Sub-value chain Allocation
FCR by
sub-chain (%) On farm (t) Trading,
transport and
handling (t) Processing (t) Retail and
consumption (t) Total physical
losses (t)
Total
physical
losses (%)
Share by
sub-chain
(%)
Ghana
Production:
14,240,867 t/yr
Own-consumption 5 % 0 (NEGL) N/AN/AN/A 0 0% 0%
Fresh root 48 % 33,822 (0.5 %) 67,306 (1 %) N/A 1,332,657 (20%) 1,433,785 21.2% 82%
Gari 24 % 0 (NEGL) 16,911 (0.5 %) 168,265 (5 %) 0 (NEGL) 185,176 5.5 % 11 %
Agbelima 17 % 0 (NEGL) 12,176 (0.5 %) 121,151 (5 %) 0 (NEGL) 133,327 5.5 % 8 %
Kokonte 6% 0 (NEGL) N/AN/A 0 (NEGL) 0 0 % 0 %
Total 100 % 33,822 96,393 289,415 1,332,657 1,752,287 12.4 % 100 %
Losses by stage (%) 2 % 6 % 17 % 76 % 100 %
SW Nigeria
Production:
7,500,820 t/yr
Own-consumption 20 % 0 (NEGL) N/AN/AN/A 0 0% 0%
Gari 52 % 39,004 (1 %) 19,307 (0.5%) 307,369 (8%) 0 (NEGL) 365,681 9.4% 76%
Fufu 27 % 18,002 (1%) 8,911 (0.5 %) 88,664 (5%) 0 (NEGL) 115,577 6.4 % 24 %
Total 100 % 57,006 28,218 396,033 0 481,258 6.7 % 100 %
Losses by stage (%) 12 % 6 % 82 % 0 % 100 %
Thailand
Production:
21,912,416 t/yr
Starch 55 % 192,829 (1.5 %) 1,186 (0.01 %) 1,186 (0.01%) 0(NEGL) 195,201 1.6% 39%
Chips 45 % 157,769 (1.5 %) 970 (0.01 %) 970 (0.01%) 145,513 (1.5%) 305,223 3.1 % 61 %
Total 100% 350,599 2,156 2,156 145,513 500,424 2.3 % 100%
Losses by stage (%) 70 % 0.4 % 0.4 % 29 % 100 %
Vietnam
Production:
9,870,000 t/yr
Dry starch 55 % 0 (NEGL) 27,154 (0.5 %) 27,018 (0.5 %) 0 (NEGL) 54,172 1.0 % 18 %
Wetstarch 5% 2,455 (0.5%) 9,772 (2 %) 4,788 (1%) 4,740 (1%) 21,755 4.4% 7 %
Chips 39 % 19,248 (0.5%) 19,152 (0.5 %) 190,565 (5 %) 0 (NEGL) 228,965 5.9% 75%
Total 100 % 21,704 56,078 222,371 4,740 304,893 3.1 % 100 %
Losses by stage (%) 7% 18 % 73 % 2% 100 %
Note: In brackets the share of cassava roots and products affected by physical losses; N/A=not applicable; NEGL =negligible.

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
117
In Thailand minor losses are reported during the
transportation of roots to the processing sites. The fact
that the processing sites are within the main cassava pro-
duction areas implies that the roots have to be trans-
ported over short distances. Therefore the spoilage of
roots is extremely rare and physical losses at this stage
have been estimated at just 0.01%.
In Vietnam, in the wet starch sub-value chain, FCR
have to be transported over considerable longer dis-
tances than the roots processed into dry starch and chips.
Moreover, in the wet starch chain, it might take several
hours to sell all FCR, which usually are delivered the
day after the harvest. As such, physical losses in the
wet starch chain are considerably higher than in other
sub-chains and estimated at 2 % and 0.5%, respectively.
In conclusion, unsurprisingly, the extent of physical
losses during transport, trading and handling is strictly
related to the distance that has to be covered to reach
the sites where the root are processed or retailed. Ac-
cordingly, the processing sites are usually located near
the farming areas. However, the wet starch sub-chain
in North Vietnam is an exception and roots have to be
transported over long distances from the mountainous
areas to the outskirts of Hanoi. Other significant losses
occur in the FCR chain in Ghanabecause cassava is of-
ten retailed in fresh form and has to reach usually distant
urban markets.
3.4 Physical losses at processing sites
At processing sites the quantities of processed prod-
ucts that are exceptionally lost (e.g. due to flooding
of the store) are negligible and, therefore, it can be
assumed that only fresh roots are affected by physical
losses.
In Ghana these losses are relevant for gari and agbe-
lima sub-chains only. Significant amount of roots spoils
at the processing storage yard when processing is de-
layed despite attempts to minimise losses by covering
the roots, especially in case of rains (e.g. with polyethy-
lene sacks). Average losses occurring at these sites have
been estimated at 5 %.
In South-West Nigeria the gari and fufu sub-value
chains are the only ones affected by losses at the pro-
cessing sites when something impedes the immediate
processing of the roots (e.g. shortage in peeling capac-
ity or mismanagement leading to excess of FCR on the
yard). The average losses have been estimated at 8%
and 5%, for gari and fufu respectively.
In Thailand, processors apply good management
practices and have the capacity to quickly process a
large amount of roots. It is rare that roots are not pro-
cessed within two days from the harvest and, as such,
spoilage of roots at this stage is extremely uncommon
and estimated to affect just 0.01% of roots.
In Vietnam, good coordination of the actors exists in
the dry starch chain in the South. Traders schedule the
deliveries of FCR together with the factories that have
the capacity to rapidly process them. However,some de-
lays may occur and therefore some physical losses that
have been estimated at 0.5 %. In the wet starch and chip
sub-value chains in the North the considerably weaker
coordination, the lower processing capacities and the
higher humidity determine higher losses. These have
been estimated at 1% and 5%, respectively.
In summary, losses at the processing sites can be
substantial and it appears that they affect particularly
the traditional processing sub-chains in the two African
countries and the dry chip sub-chain in North Viet-
nam. In this country significant losses also incur by
wet starch processors. Losses are minimal in Thailand
thanks primarily to well-executed just-in-time procure-
ments of FCR, excellent coordination among the actors,
large processing capacity and favourable weather condi-
tions.
3.5 Physical losses at distribution, retail and con-
sumption level
These losses are not relevant when roots are used for
own-consumption and are often negligible for processed
products.
In Ghana these losses affect the FCR sub-chain only
and can be extremely high. It has been estimated that
physical losses at retail stage are about 5%. At house-
hold and caterer level, despite the fact that some miti-
gation measures are undertaken (e.g. just-in-time pur-
chases and, sometimes, packaging and refrigeration of
peeled roots), it is estimated that 15% of the roots spoil
and are thrown away.
In Nigeria some losses at this stage can occur in the
gari and fufu sub-chains. However, in those sub-chains
losses are extremely rare (e.g. when packages are dam-
aged) and, therefore, have been assumed as negligible.
In Thailand a substantial amount of dust is produced
during the transport and mechanical handling of chips,
mainly due to the abrasion of the chip surfaces against
the cement floor. It is estimated that an average of 1.5 %
of chips’ weight is lost in the form of dust. Conversely,
in the starch sub-chain physical losses occur only in ex-
ceptional cases, e.g., breakages of the bags or flooding
of the store. These losses have been considered as neg-
ligible.

118 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
In Vietnam losses of dry starch and chips have been
assumed as negligible. Conversely some physical losses
occur in wet starch sub-chain when the starch block,
typically stored in underground pit, is exposed to aer-
obic conditions. In this case the outer part of the block
has to be thrown away. These losses were estimated at
1%.
In conclusion, losses at distribution, retail and con-
sumption level are exceptionally high in Ghana since in
this country large volumes of roots reach the final users
in the very perishable fresh form. Conversely, in Nige-
ria where cassava is usually distributed and retailed in
processed form losses at this stage are negligible. In the
two South-East Asian countries, the processed products
are the ones that incur most losses at this stage: namely,
dry chips in Thailand and wet starch in Vietnam.
3.6 Economic losses
According to our definition, no economic losses are
incurred if the roots are used for own consumption or
on-farm processing (e.g., kokonte) since, in these case,
the roots are not traded and, therefore, not priced.
In Ghana, breakages occur mainly during harvest
(15% of roots are estimated to be affected) and trans-
port (estimated at 3% and 1 % for the FCR and the
gari-agbelima sub-value chains, respectively, due to the
longer distance that has to be covered in the former).
Partial spoilage leading to discounted price mainly oc-
curs at retail level and in the FCR sub-chain only (af-
fecting 10% of roots on average). As such, altogether,
in the FCR and gari-agbelima chains, 28 % and 16% of
roots incur some economic losses, respectively.
In Nigeria it is estimated that between 10% and 30 %
of roots suffer economic losses on farm due either to
breakages during harvest or deterioration when delays
in marketing occur. Moreover, about 2% of roots are
affected by economic losses, mainly due to breakages,
during transport to the processing sites. As such, alto-
gether, in the gari and fufu sub-chains around 20% of
roots are sold at discounted price.
Based on price information gathered during the sur-
vey it has been estimated that, in Ghana and Nigeria,
approximately half of FCR incurring economic losses
receive around 20% price reduction and the remaining
half a 40% reduction.
In Thailand and Vietnam, more often than not, the
roots are sampled at the time of the delivery to the pro-
cessor and their starch content is determined by using
Reimann balances (standard quality is 25%). The price
is then set accordingly but, beyond a certain level of
quality deterioration, FCR are simply rejected. While
starch content is dependent on the variety, agronomic
practices and other production factors, in these countries
a starch content below 25 % is found only when cassava
incur considerable postharvest deterioration.
In Thailand when starch content in FCR is lower than
25% a price reduction is applied for each point below
the standard. In the starch and chip sub-value chains
about 5% of roots delivered have 23% starch content.
Only a minimal amount of roots have lower starch con-
tent (0.35% of roots with 22 % starch content in the
starch sub-chain and 0.5% with 21 % starch content in
the chip sub-chain).
In Vietnam a similar but less rigorous system is ap-
plied (sometimes relying on visual examination). As
rule of thumb, a so-called “point system” mechanism
reduces the price of roots by about 10% and 20 % in
the first and second day after harvest, respectively. In
the dry starch and chip value chain it has been estimated
that 75 % of roots are processed the day of harvest, 20%
the day after and 5% two days after. In the wet starch
value chain, due to the long distances and frequent de-
lays, only 10% of roots are processed the same day of
the harvest. Around 80% are sold the next day and the
remaining 10 % the following one.
3.7 Comparative analysis of postharvest losses
The extent of physical losses at each stage of the dif-
ferent sub-chains was estimated by taking into account
the estimated cassava production in 2011, the share of
roots allocated to the different sub-value chains and the
relevant percentages of physical losses (Table 1).
Ghana, with about 12.5 % of roots annually produced
lost (equivalent to over 1.7 million tonnes of FCR), is
by far the country that incurs the highest physicallosses.
This is primarily due to the fact that a high proportion
of roots (48%) have to be transported, distributed and
stored in the fresh form until consumption. Accordingly,
over 80% of these losses are ascribable to the FCR sub-
value chain. PHL occur at each value chain stage but
are particularly relevant (76 %) at the final one (retail
and consumption), which is similar to what is reported
to occur in developed countries (Table 1 and Figure 1).
This challenges the generally accepted view that in de-
veloping countries losses occur mainly in the first stages
of the chain. While other important losses are incurred
at the gari and agbelima processing sites it must be no-
ticed how the good organisationof these sub-chains al-
lows keeping on-farm losses at an extremely low level
and this should be taken as an example of good practice.

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
119
Fig. 1: Level of physical losses by value chain stage.
South-West Nigeria ranks second in terms of extent of
physical losses with an estimated 7% of harvested cas-
sava physically lost (correspondingto almost half a mil-
lion tonnes of FCR per year). Unlike Ghana and Thai-
land, most root spoil at the processing stage (over 80%
of physical losses occur at the gari and fufu processing
yards).
Vietnam ranks third in terms of extent of physical
losses. It is estimated that about 3% of annual cassava
production is lost (equivalent to 300,000tonnes of FCR
equivalent).7Losses affect almost exclusivelycassava in
the fresh form even though some relatively minorlosses
occur due to the spoilage of wet starch. As shown in
Figure 1, most of these physical losses are incurred ei-
ther at the processing stage (73%) or during transport
and trading of FCR to the processing sites (18 %). The
dry starch value chain in Vietnam is efficient in minimis-
ing the losses and this is proved by the fact that, while
absorbing about 55% of cassava roots, it is responsi-
ble for just 18% of physical losses. Conversely,the dry
chip and wet starch sub-chains in northern Vietnam are
the most affected by losses: in the former because of
spoilages at the processing yards; in the latter because
of transportation over long distances.
Thailand is the country that incurs the lowest physi-
cal losses. Due to the highly efficient chain coordination
only 2.5% of total annual production is lost (equal to
about half a million tonnes of root equivalents). Unlike
in Ghana, physical losses in Thailand occur mainly at
farm level (70%) due to roots remaining accidently un-
7For Vietnam and Thailand it was necessary to convert some losses
into root equivalents since some of them affect the processed products
(wet starch and chips, respectively).
harvested or, when broken, voluntarily left in the field.
In Thailand other significant losses (29%) occur at the
final value chain stage and are due to the dust produced
during the handling of chips. The dry chip sub-chain
contributes to over 60% of physical losses.
In terms of monetary value lost due to physicalPHL,
again, Ghana outpaces the other countries. As shown
in Figure 2, the worth of physical losses in Ghana has
been estimated at almost USD 390 million per annum,
much higher than in Thailand, South-West Nigeria and
Vietnam (USD 45, 30 and 20 million, respectively). The
fact that the value of physical losses in Ghana is about
tenfold that of the other countriesis explained by the fact
that in this country most losses occur at the last stage of
the value chain when considerable value has been added
and, as such, where the price of the roots is at the highest
(actually the vast majority of roots spoil at household
level where each unit lost is worth its retail price).8It
has been estimated that in Ghana physical losses at retail
and consumption level are responsible for about 93% of
the lost monetary value.
In all countries more roots are affected by economic
losses than by physical losses (Figure 2). We have also
found that poorer countries and households have the
ability to reduce the impact of PHL by transforming part
of the physical losses (no residual value) into economic
losses (residual value), e.g. by processing broken roots
into lower value products.
8At the time of the survey the farm gate price of cassava roots was
about 60 to 80 Ghanaian New Cedi (GNC) per tonne (33–43 USD/t),
the rural wholesale market price was about 80 and 120 GNC/t (44–
67 USD/t) while the retail price ranged between 210 and 490 GNC/t
(114–269 USD/t).

120 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
Fig. 2: Estimated volume (left) and monetary value (right) of physical and economic losses.
The volume of roots incurring economic losses is par-
ticularly substantial in Vietnam where about 28% of
harvested cassava (equal to 2.7 million tonnes per an-
num, Figure 2) is sold at discounted pricedue to quality
deterioration. The wet starch sub-chain seems particu-
larly affected (only 10% of cassava reaches the proces-
sor within 24 hours from harvest). In the dry starch and
chips sub-chains the share of roots incurring some eco-
nomic loss is lower and estimated at about 25 %.
In Ghana and South-West Nigeria about 19% and
16% of harvested roots are sold at discounted price
(equal to 2.4 and 1.0 million tonnes per annum, respec-
tively). Again, the FCR sub-chain in Ghana stands out
and, in this chain, about 28% of FCR are sold at a price
lower than what would be paid for a quality root.
In Thailand it is estimated that just over5.5 % of roots
(equal to 1.2 million tonnes per annum) are sold at dis-
counted price.
Apart from the volume of roots affected by quality
deterioration, the monetary value of economic losses de-
pends on the pricing mechanism in place. This appears
to be extremely high in Ghana (over USD 130 million
per annum). This is again considerably higher than for
any other country studied (Figure 2) and is explained
by the fact that about half of FCR reach the final con-
sumers in the fresh form. Furthermore, in Ghana the
buyer is quite demanding in terms of quality and tends
to considerably discount these roots (up to 50% of the
price paid for quality FCR).
In South-West Nigeria the value of economic losses is
estimated at just over USD 20 million. While the share
of roots affected by economic losses is not significantly
different from Ghana, in Nigeria, gari and fufu proces-
sors, that absorb most of the roots, pay an average price
(67 USD/t at the time of the survey) that is much lower
than the Ghanaian retail price for FCR. That means that
an equivalent percentage price discount has, in mone-
tary terms, in Nigeria a much lower impact than in the
case of Ghana.
In Vietnam the value of economic losses is estimated
at about USD 15 million. While this is the country
where the largest share of roots incurs some price re-
duction, the magnitude of the quality deterioration and,
hence, price reduction is considerably lower than in the
two African countries. The price of FCR can be dis-
counted by no more than 10 %. In case of poorer quality
the processors simply reject the root.
In Thailand economic losses are minimal. They are
estimated at less than USD 2 million per year. Firstly,
a small amount of roots (about 5 %) incurs quality dete-
rioration to the point that the processor has to discount
their price. Secondly, alike Vietnam, whenever it oc-
curs the price discount mechanism works until a certain
threshold. Beyond that point the root is rejected. It is
rare that the price is discounted by more than 5%.
Figure 2 shows the total volume and monetary value
of physical and economic losses in the four countries.
In Ghana a total of about 4 million tonnes of cassava
are affected by either physical or economic losses (29 %
of annual production). The combined monetary value
of these losses is extremely high and estimated at USD
520 million per annum.9This is explained by several
factors. Firstly, Ghana is the second largest cassava pro-
ducer in our sample. Secondly, the vast majority of roots
are marketed and about half of the marketed roots reach
9As pointed out by Oguntade (2013), in reality, it is not feasible
to achieve zero losses, as protection measures to secure 100 % of the
harvest will be inevitably disproportionally costly.

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
121
the final consumers in the fresh form. This has two im-
plications: on the one hand, more roots incur total or
partial spoilage; on the other hand, losses further down
the chain have a more profound impact on the total mon-
etary value of PHL because a higher value has become
embedded in the product (i.e., the economic cost of the
loss accumulates along the chain). Finally, the buyer
in Ghana tends to considerably discount lower quality
roots. In South-West Nigeria and Thailand PHL affect
a similar amount of cassava (about 1.5 million tonnes
per annum) and have comparable monetary value (about
USD 50 million). However, in relative terms, the share
of the country annual cassava production incurring PHL
is very different (20 % and 8%, respectively). Finally, in
Vietnam almost 3 million tonnes of cassava incur phys-
ical or economic losses (with strong predominance of
the latter), representing 30% of the annual production.
However, the monetary value of total PHL is estimated
at just about USD 35 million due to the relatively minor
monetary impact of economic losses.
The worth of PHL in Ghana represents about 22 % of
the total potential retail value of cassava products (net
of physical losses). In South-West Nigeria, Vietnam and
Thailand this is estimated at 7 %, 4 % and 2 %, respec-
tively.
It is worth noticing how, despite the fact that in all
countries more cassava is affected by economic rather
than physical losses, the impact of the former in mon-
etary terms is considerably more modest due to the
residual value embedded in the affected fresh roots.
This clearly emerges by comparing the two graphs in
Figure 2.
4 Recommendations
Our study found that verydifferent levels of PHL oc-
cur across different countries, value chains and stages
of the single value chain. Although our findings reveal
considerable lower physical losses than usually reported
in the literature, it is crucial to understand where, when,
why and how losses occur in order to identify possi-
ble solutions for their reduction and maximise the eco-
nomic, social and environmental benefit of any interven-
tion aiming at reducing them.
While providing solutions for PHL reduction is be-
yond the scope of our work, we would like to suggest
some possible technological, commercial and institu-
tional innovations that can contribute to minimise losses
at different points of the chain.
Among the four target countries, on-farm losses ap-
pear to be significant in Thailand only due to widespread
use of mechanical harvesters. These machines have
been considerably ameliorated over the time but fur-
ther improvements would contribute to reduce left-overs
and breakages. In the longer term, this would benefit
also less developed countries where the cassava sector
is likely to face significant consolidation towards larger
plantation, higher inputs and mechanisation in the near
future.
The main causes of PHL at the stage of trading, distri-
bution and consumption are, at a large extent, the same,
i.e., the short shelf-life of the root combined with de-
lays in marketing and consuming it. These losses affect
mainly the FCR sub-chain in Ghana. At this regard a
combination of innovations can be envisaged. Several
technologies for extending the shelf-life of the fresh root
have be validated and adopted in other countries, partic-
ularly for targeting distant and export markets. Some of
these technologies, including waxing, paraffin coating,
and high humidity storage, can be tested and adapted
to the specificities of other countries. Even though an
in depth cost-benefit analysis should be conducted, this
seems particularly promising for Ghana where the retail
price of FCR is extremely high. In this country the pos-
sibility to address PHL by commercial innovation, and
in particular by developing substitutes of unfermented
fufu with longer shelf-life (e.g., instant fufu to be re-
constituted at household level) should be explored and,
if consumers’ acceptability is proved, this could open
up a large market for a less perishable value-added pro-
cessed product. In terms of institutional innovation we
have found out that poorer countries, in spite of their
better capacity of absorbing sub-standard products and
less strict standards applied by the buyers, incur higher
losses, mainly because of the weak coordination of the
different value chain actors. In this regard the best exam-
ple is represented by the Thai starch value chain where
processors coordinate staggered planting, supply farm-
ing input, schedule the harvest and arrange for the trans-
port of roots to the plant in such a way minimising de-
lays and therefore losses. While we have noticed that
also in Ghana a good coordination between farmers and
processors of gari and agbelima exists and contributes
to minimise losses, other sub-value chains in SSA, par-
ticularly for gari and fufu in Nigeria and FCR in Ghana,
still offer considerable room for improvements in order
to synchronise harvest, transport, processing and retail-
ing (including by adopting forward contracts with in-
stitutional buyers and caterers in Ghana). In Vietnam
losses incurred during the transport of the roots to the
wet starch processing centres could be reduced by relo-
cating the processing sites nearby the cassava produc-
tion area but this requires connecting them to the elec-

122 D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
tricity grid and ensuring regular supplies of clean water.
In terms of losses of processed products, Thailandis the
only country where substantial amount of cassava is lost
in the form of dry chips’ dust. While palletisation is the
most obvious solution for reducing the amount of dust
produced this is a technology from which the country
has moved away years ago since the additional cost is
higher than the price premium paid by the livestock in-
dustry and not required for using the chips as feedstock
for biofuel production.
Finally, we have found that losses of fresh roots in-
curred at the processing sites are substantial in the tra-
ditional processing sites in Ghana and Nigeria and in
the dry chip sub-chain in Vietnam. In the former the
main cause is the delay in processing due to shortage
in peeling capacity and mechanical peeling would enor-
mously help in tackling this problem althoughit must be
recognised that considerable effort has been put over the
last years by several research institutes to develop effi-
cient peelers (e.g., the International Institute for Trop-
ical Agriculture – IITA) but with limited success. In
Vietnam delays in processing are mainly caused by lim-
ited processing capacity of small-scale chip processors
in the North of the country. It is likely that, due to the
fast consolidation of the sector, these processors will be
progressively replaced by more efficient medium-scale
processors that are likely to adopt Thai-styleequipment.
The most important recommendation from this study
is that there are no “one-size-fits-all" solutions for ad-
dressing postharvest losses but rather these must be
tailor-made to the specific characteristics of the differ-
ent value chains taking into account their technical as
well as the economic feasibility.
Acknowledgements
This study has been funded by the EC in the framework
of the 7th Framework Programme.
References
Boahen, C. (2004). Estimation of Material Losses and
Power Requirement in Gari Processing. B.Sc Thesis,
University of Cape Coast.
Booth, R. H. & Coursey, D. G. (1974). Storage of cas-
sava roots and related postharvest problems. In E. V.
Araullo, B. Nestel, & M. Campbell (Eds.), Cassava
Processing and Storage. An Interdisciplinary Work-
shop (pp. 43–49). IDRC, Ottawa.
Dziedzoave, N. T., Ellis, W. O., Oldham, J. H. & Osei-
Yaw, A. (1999). Subjective and objective assessment
of agbelima (cassava dough) quality. Food Control,
10(2), 63–67.
FAO (2004). The Global Cassava Development Strat-
egy and Implementation Plan. Proceedings of the Val-
idation Forum on the Global Cassava Development
Strategy. FAO, Rome.
FAO (2011). Global Food Losses and Food Wastes: Ex-
tent, Causes and Prevention. FAO, Rome.
FAOSTAT (2013). FAOSTAT Database of Crop Statis-
tics.
Foresight (2011). The Future of Food and Farming:
Challenges and Choices for Global Sustainability. Fi-
nal Project Report. The Government Office for Sci-
ence, London.
Gustavsson, J., Cederberg, C., Sonesson, U., Otterdijk,
R. & Meybeck, A. (2011). Global Food Losses and
Wastes. Swedish Institute for Food and Biotechnol-
ogy (SIK), Gothenburg, Sweden and FAO, Rome,
Italy.
Iyer, S., Mattinson, D. S. & Fellman, J. K. (2010). Study
of the early events leading to cassava root postharvest
deterioration. Tropical Plant Biology, 3 (3),151–165.
Kaplinsky, R. & Morris, M. (2001). A Hand-
book for Value Chain Research. Prepared for
IDRC. URL http://asiandrivers.open.ac.
uk/Resources.html.
Oguntade, A. E. (2013). Food Losses in Cassava and
Maize Value Chains in Nigeria: Analysis and Rec-
ommendations for Reduction Strategies. GIZ GmbH,
Bonn, Germany.
Parfitt, J., Barthel, M. & Macnaughton, S. (2010). Food
waste within food supply chains: quantification and
potential for change to 2050. Philosophical Transac-
tions of the Royal Society B, 365,3065–3081.
Scott, G. J., Rosegrant, M. W. & Ringler, C. (2000).
Roots and Tubers for 21st Century: Trends, Projec-
tions and Policy Options. Food, Agriculture and the
Environment Discussion Paper No 31. International
Food Policy Research Institute (IFPRI), Washington
D.C.
Vowotor,K. A., Quaye, W. & Yawson,S. (2010). Base-
line on Cassava Processing Technologies in Togo,
Benin, Ghana and Nigeria. A Consultancy Report
prepared Under Songhai Project. Food Research In-
stitute, Accra.
Wenham, J. E. (1995). Postharvest Deterioration of Cas-
sava: a Biotechnology Perspective. Plant Production
and Protection Paper FAO No 130. FAO, Rome.

D. Naziri et al. /
J. Agr. Rural Develop. Trop. Subtrop. 115 -2 (2014) 111–123
123
Westby, A. (2002). Cassava Utilization, Storage and
Small-Scale Processing. Natural Resources Institute,
University of Greenwich, UK.
World Bank (2011). Missing Food: The Case of
Postharvest Grain Losses in Sub-Saharan Africa. Re-
port No 60371-AFR, The World Bank, Washington
D. C.
Yin, K. (2003). Case Study Research: Design and Meth-
ods. Sage, London.