Journal of Food Products Marketing(USA), 12(4): 59-77

Seasonal and inter-market differences in prices of small ruminants

in Ethiopia
Gezahegn Ayelea, Mohammad A. Jabbarb, Hailemariom Teklewoldc,
Elias Mulugetab and Getahun Kebedec

a. Ethiopian Development Research Institute, PO Box 2479, Addis Ababa, Ethiopia

b. International Livestock Research Institute, PO Box 5689, Addis Ababa, Ethiopia.
c. Ethiopian Agricultural Research organisation, PO Box 2003, Addis Ababa, Ethiopia

Abstract

Hedonic price models were fitted to a sample of 1397 sheep and 1293 goats respectively for
which data were collected from nine markets in Ethiopia over a 12 month period. The
objective was to determine seasonal and inter-market differences in prices after controlling
for the effects of different attributes of the animals, the buyers and the sellers. Results
indicate that, controlling for attributes of the animals and of the buyers and sellers, there were
significant differences in prices between seasons and markets. Seasons in which farmers
faced severe cash shortages exhibited the lowest adjusted prices for animals they sold,
indicating that although livestock may provide a fall back position for cash in times of crisis,
terms of trade may be worst when farmers need cash the most. In general, there was no clear
progression in price of sheep along the primary to terminal market chain ending in Addis
Ababa as would be normally expected except that the farthest market had the lowest price.
The reason for higher prices in some intermediate terminal markets could be partly explained
by the fact that exporters and processors buy animals in these markets and they pay premium
prices for best quality animals, and left over second or third grade animals may end up in
Addis Ababa market, which then virtually becomes a sink market. In case of goats, price
differences between markets followed to some extent the expected differences between
primary, secondary and terminal markets. One possible reason is that in general highland is
not a major production or consumption area for goats, so supplies come mainly from the
lowlands, so the price movement followed the market chain from primary markets in pastoral
areas to the terminal market in Addis Ababa,.

Key words: sheep goats, hedonic price, Ethiopia, Africa

1
Introduction

In the highlands of Ethiopia, livestock as an important component of the mixed farming

system perform multiple functions providing high quality food, draft power and manure for
crop production, and cash income. Field studies in different parts of the country in the 1980s
showed that livestock account for 37-87% of total farm cash income of farmers, indicating
the importance of livestock in rural livelihood, especially as one moves from mixed farming
in the highlands to agropastoral systems on the highland-lowland margins (Gryseels, 1988).
However, despite the reasonably high share of cash income coming from livestock as a
source of smallholders’ livelihood, the production system is not adequately market-oriented
and it is not yet a primary livelihood activity or a primary source of cash income for any
significant number of communities or households. There is little strategic production of
livestock for marketing except some sales targeted to traditional Ethiopian festivals. For
example, a survey of 82 communities in the Oromiya region showed that livestock ranked 3
on average out of top five cash income sources in the sample communities, and it appeared as
a secondary/tertiary livelihood activity or source of overall income for about 40% of the
households in the region (Jabbar et al., 2002). Therefore increased market orientation of
livestock production is essential for producers to be able to improve their livelihood through
livestock.

The primary reason for selling livestock in the highlands is to generate income to meet
unforeseen expenses. Sales of live animals are taken as a last resort and animals are generally
sold when they are old, culled, or barren (Kebede and Lambourne, 1985; EARO, 2003).
Prices depend mainly on supply and demand, which is heavily influenced by the season of the
year and the occurrence of religious and cultural festivals on the one hand and occurrence of
drought or other weather shocks on the other. For example, Northern Ethiopia’s livestock
supply is heavily influenced by the severity of the dry season; supply peaks after the main
rainy season then drops rapidly. In the South, low sales volume characterize the July-
September main rainy season, and the Lent fasting period (February-April), but trade peaks
immediately following these periods (Tilahun, 1983; Kebede and Brokken, 1993; Davies,
2003; World Bank, 2001). Fachamps and Gavian (1997) found drought and pasture
availability as important determinants of price variation in Niger.

Yearly price variation may also be triggered by general crop sector performance and weather
patterns. It is often argued that in mixed farming systems livestock may serve as a hedge
against risk of crop failure as livestock can be sold to derive cash in the absence of crop
output and income. However, crop sector performance may act as a double-edged sword for
livestock. A bumper crop harvest may reduce crop price in the absence of market
stabilization mechanisms, hence reduce rural people’s real demand for livestock, and a crop
failure may also have the same effect. A fall in grain price is, however, beneficial for
pastoralists because of more favorable terms of trade (Jabbar and Ayele, 2004). Thus, most
producers remain vulnerable to the forces of market and do not always get good prices for
their animals. In a survey of sample markets and traders in Oromiya, Amhara and Tigray
regions in 2002, seasonal variation and unstable price, multiple taxes, non-transparent tax
system, limited access to credit and weak demand for the types of animals offered for sale
were perceived by traders as major problems of marketing (Jabbar and Benin, 2004).

In this paper, the results of a survey on seasonal and inter-market variation and other
determinants of prices of small ruminants in selected highland and mid-altitude zonal markets
are presented. Small ruminants play a major role in the livelihood of smallholder farmers in

2
the highlands. There are about 28 million small ruminants in the country, of which 75% are
owned by smallholder mixed crop-livestock farms in the highlands and 25% by pastoralists in
the lowlands. A little over 50% are sheep and others are goats. Understanding the sources of
price variation may be helpful for producers to understand buyer preferences for specific
characteristics of animals and target breeding, fattening, time and place for sales to gain from
important market opportunities.

In section 2, the sampling method and data collection procedures are described. In section 3,
the theoretical and empirical model for price determination is described. In section 4, the
results are discussed with conclusions at the end.

Selection of Markets and Collection of Data

The capital Addis Ababa is the domestic terminal market for most of the small ruminants
originating in the central highlands. On the other hand, majority of the marketed livestock
from the pastoral lowlands in the east and south of the country are exported to the Middle
Eastern countries and to Kenya. In fact, economic exchanges between the highland and the
lowland are rather minimal as livestock export earnings from outside Ethiopia are used for
importing consumable and other goods into the pastoral areas. Given the different supply
hinterlands and structures of the highland and lowland markets, this study concentrated only
on the markets in the central highlands taking Addis Ababa as the terminal market. Also
markets in one transect covering up to 200 km towards south and southeast of Addis Ababa in
Eastern Oromiya region were considered for sampling. In this transect nine markets were
purposively selected. These are Addis Ababa, Akaki, Dire, Debreziet, Nazareth, Arerti,
Methara, Meki and Shashemene (Figure 1).

There are several market locations for sheep and goat in Addis Ababa. Some of the formal
markets are equipped with better facilities while others are road side temporary exchange
points. For the purpose of this study, Gulele market located in the northern part of the city
was purposely selected to represent Addis Ababa based on volume of transactions and
diversity of sources of animals transacted. .

Arerti and Dire are primary markets and Akaki is a secondary market located close to Addis
Ababa. Debreziet is a secondary market located halfway between Addis Ababa and Nazareth,
the second biggest market in East Shewa zone, and supplies animals to both Addis Ababa and
Nazreth. Methara is a primary market supplying animals to Nazareth. Meki is a primary
market for both Nazareth and Shashemene, which is a large market located on the borderline
between the highlands and the lowlands, and supplies animals to both Nazreth and Addis
Ababa in one direction and also towards northern part of Ethiopia destined for export through
Djibouti and Somaliland ports. There may be also two way movement, for example, between
Nazreth and Debre Zeit, and between Meki and Shashemene. Meki also directly supplies to
Addis Ababa

Each selected market meets at least two days per week, mainly on Tuesday, Wednesday,
Thursday, or Saturday. Data were collected in each week in one of the main market day for a
period of one year during September 2002 to August 2003, except for Nazareth where data
were collected on two main market days. Five traded animals were randomly selected from

Figure. 1 Location of sample small ruminant markets studied in East Shewa zone,

3
 Oromia region, Ethiopia

a. Administrative regions of Ethiopia

b. Locations of sample markets in relation to Addis Ababa

4
 each market every week. A total of 2690 animal transactions were recorded by jointly
interviewing the seller and the buyer using a questionnaire. Of these, 1397 are sheep and
1293 are goats. Animals traded in Ethiopian markets as elsewhere in the developing countries
are not standard products to allow collection of price data on a per unit output, e.g. per kg live
weight basis. Actually weighting animals is not practiced in the sample markets. Traders
targeting animals for purchase usually make good guesses about weight and the ability to
guess the weight helps them in making a good bargain (Jabbar, 1998). Agreement on price is
reached through a long bargaining process between the buyer and the seller either directly or
through a broker, who may charge a fee from both parties. Attempts were made to use
weighing scale but it turned out to be difficult in the crowded market place. Since prices are
negotiated for an animal, it was hypothesized that different attributes of the animal will
contribute to the price ultimately negotiated. Therefore data were collected on the agreed
price and the following attributes of the animal, the buyer and the seller: species, sex, age
(months), body condition, skin condition, breed, origin, color, animal type, tail type, height of
the animal (cm), heart girth (cm), expected price of skin, seller type and seller sex, buyer type
and buyer sex, purpose of selling and buying.

Age was approximately estimated by examining the number and type of teeth. Information on
body and skin conditions were graded into three categories (good, average and poor) based on
the assessment of certain observable physical characteristics. Height and heart girth were used as
proxies for weight. Sometimes, weight in cattle is extrapolated from heart girth by using some
standard tape or by using a conversion formula suggested by Payne (1990) but no such suitable
tape is available for small ruminants, and suitable conversion formula using heart girth as the
base is also not available.

The Theoretical and Empirical Model for Price Determination

It is generally hypothesized that products have attributes that confer utility and that the values
of those attributes contribute to the price of the product. Therefore, a composite of the
implicit prices of the product's attributes is reflected in the observed price of a product. An
implicit price of a product in a competitive market will be a function of the product attributes
alone. This implies that only products are differentiated, while their markets, buyers and
sellers are not (Rosen, 1974; Lucas, 1975; Ockowski, 1994). However, in most empirical
studies, price has been found to be related to the product attributes as well as attributes of the
buyers and sellers, implying some non-competitiveness in the market (e.g. Brorsen et al.,
1984; Francis, 1990; Andargachew and Brokken, 1993; Parker, 1993; Parker and Zilberman,
1993; Williams et al., 1993; Oczkowski, 1994; Rodriguez et al., 1995; Jabbar, 1998; Jabbar
and Diedhoudu, 2003).

In this study, an implicit or hedonic price function was estimated to relate the price per
animal to its various attributes, and the attributes of markets, buyers and sellers. The general
form of the implicit price function is as follows: P = F(Q,C) + e , where P is the observed
price of the product, Q is a set of qualitative (discrete) variables or factors each with more
than one category, C is a set of quantitative variables (covariates), and e is an error term.
Interaction variables may also be incorporated. The partial derivative of the estimated
function with respect to a quantitative variable is the implicit marginal value of the attribute.
Qualitative attributes are represented by dummy variables so the estimated parameters
measure the impact of the presence or absence of the attribute. Therefore, the predicted price
cannot be directly obtained from the partial derivatives, and hence additional manipulation
would be required (Gujarati, 1988).

5
Since data on weight of animals could not be collected, the price model was estimated with
price per animal as the dependent variable and the following explanatory variables: species
(sheep, goats), characteristics of the animals (age, heart girth, height, sex, tail type, breed and
origin, body and skin conditions, and color), market locations, time (week or month) of
transaction in the year (representing season and presence or absence of a major festival), sex
and types of sellers (traders vs producers), reasons for selling the animals, sex and type of
buyer of the animal and purpose of buying. The SPSS Analysis of Covariance procedure
(SPSS, 2004) was used to estimate the model parameters. The model estimated the price
differences between categories within a factor, e.g. between male and female animals, after
adjusting for the effects of all other factors and covariates. Bonferroni confidence intervals
were used in the hypothesis tests in order to reduce the likelihood of false rejection of null
hypotheses.

Of the total sample animals, 53% were sheep and 47% goats. Plot of average weekly nominal
price data for sheep and goats show that prices of sheep were generally higher than those of
goats throughout the year (Figure 2). Initially, the entire sample of sheep and goats was
analyzed together using species as a factor, and significant price differences were found
between sheep and goats. However, certain physical attributes of the two species are quite
different, and the effects of those factors could not be properly captured in the aggregate
analysis. Therefore, separate functions were fitted for sheep and goats, which gave better fit.

The weekly nominal price data plot in figure 2 shows that during the year there were about
five different peaks and troughs, some large and some small, in price movement. Two
approaches were tested to define time in the equation to see which fits these cycles better.
First, transaction data were recorded each week so week was defined as a covariate and a
fifth degree polynomial form for this variable was specified. Second, prices usually build up
over several weeks toward a peak surrounding a major festival or religious occasion, e.g.
new year, Christmas, haj, and then falls sharply after the occasion. These occasions also fall
in some calendar month though the build up period may overlap with another month.
Assuming that average for a month might capture these significant occasions, month was
defined as a factor with individual months as categories. Of these two options, the month
option gave the better fit.

Another problem was with respect to the use of reason for selling as a factor. Although every
sale generates cash, farmers sold animals due to some specific underlying reasons other than
cash generation, e.g. dispose of fattened animals for profit, to meet problems arising from
draught, due to feed and water shortage, while traders sold only for profit. It was
hypothesized that, other things being equal, price received for an animal might differ
significantly according to the reason for sale. However, traders sold only for one reason so
the data matrix had empty cells for other reasons for selling, which made accurate estimation
of parameters for those variables difficult.

Therefore, for each species two equations were estimated: one for the total sample using all
factors and covariates excluding reason for selling as a factor, the other for only animals sold
by farmers using all factors and covariates including reason for selling as a factor. For sheep,
R2 for the overall and farmer seller equations was respectively 0.66 and 0.67, while for goats
this was respectively 0.68 and 0.65. The significance of different parameters also remained
largely similar in the two equations except in a few cases. Therefore, detailed discussion is
focused only on the results of the overall equations, and any significant result from the farmer
seller equations are highlighted at the end.

6
 Figure 2. Average weekly nominal prices of sheep and goats per animal,
September 2002 – August 2003

Animal type
220.00
sheep
Goat

200.00

180.00
Price per animal (in Birr)

160.00

140.00

120.00

100.00

80.00
4

40
43
46
49
22
25
28
31
34
37

52
7
1

10
13
16
19

Transaction
Figure 1. Weekly nominal time
price per animal inselected
in the weekmarkets

7
Results and discussion
Overall equations for sheep and goats
The results of best fit equations for sheep and goats are shown in Table 1. The specified
variables explain 66% and 68% of price variation in case of sheep and goats respectively.
Among the covariates, square terms were introduced for age, heart girth and height in order
to capture the price premium or penalty for over aged, larger heart girth and taller animals.
Other things being equal, age, heart girth and height of the animals had significant influence
on the prices but with some difference between sheep and goats. Price per animal increased
with age but declined for older or over mature animals for both sheep and goats. Height and
heart girth are proxies for overall size of an animal and some breeds have larger height but
narrow heart girth, so likelihood of correlation between these two parameters is low. In case
of sheep, price per animal was very low or there was price penalty for animals with very
small heart girth or dwarf size as indicated by the negative sign of the coefficient but price
increased significantly as heart girth became larger and height increased. In case of goats, the
influence of heart girth had the same pattern as in the case of sheep but coefficients of height
and height square were not significant. This may indicate that marketed goat breeds were of
more uniform height or that differences in height of did not affect price of goats. Expected
average price of sheep and goat skins was Birr 22.8 and 10.8 respectively with standard error
of 0.16 and 0.07 respectively. Expected price of goat skin significantly increased price of a
goat but not so in case of a sheep.
Among the sample animals, the proportion of male was 76% in sheep and 66% in goats.
Other things being equal, price per animal was significantly higher for males compared to
females in case of both sheep and goats. One possible reason is that most marketed females
are old culled animals which passed their productive age. Also some consumers don’t want
to buy female animals for slaughtering due to the possibility of pregnancy, as it is ethically
unacceptable in Ethiopian society to slaughter pregnant animals.

Among the sample sheep, 56.8, 38.7 and 4.5% had good, average and poor body conditions
respectively compared to 54.9, 35.8 and 9.3% in case of goats. In case of both sheep and
goats, there was significant price penalty as body condition became poorer compared to the
good condition.1

1
Data on skin condition were also collected for each animal and this was used as a factor in the equation.
However, because of high correlation between body and skin condition, this variable did not improve model fit,
so was left out.

8
Several sheep breeds in Ethiopia are fat tailed but goats are generally thin tailed though some
may have slight fat tail. Apart from inherent breed characteristics, the degree of fatness of the
tail may also indicate general health condition of the animal: a more fatty tail indicates a
better body condition and better health of the animal. Among sample sheep, 58.1, 38.2 and
3.7% were respectively thin tailed, fat tailed and fat ramped (very fat) tailed. Among goats,
88.2% were thin tailed and 11.8% were slightly fat tailed. Other things being equal, compared
to fat ramped sheep, those with thin or fat tail commanded significantly lower prices. In case
of goats, tail type did not significantly affect price.

Some breeds have inherent exclusive or dominant color while in other cases, this may not be
so, especially when crossing among breeds is very common. Among sheep dalacha (28%),
white (20.5%) red (21%) and black (18%) were major colors while among goats the same
colors represented respectively 21.3, 20.6, 18.7 and 26.1%. Some times, buyers prefer a
particular color while buying an animal either because of cultural reasons, e.g. animals
bought for sacrifice may require to be of certain color, or because of one’s personal liking for
a particular color. In this study, other things being, equal, no significant price difference
between color of goats was observed but black colored sheep commanded significantly lower
price compared to red and other colors.

Among sample sheep, 97.8% were various local breeds but exact breed names could not be
ascertained and 2.2% were crosses between some local and imported exotic breeds. Among
goats 99.5% were local breeds. Crossbred sheep commanded significantly higher prices than
local breeds, but prices did not differ significantly between breeds of goats.

No scientifically based breed classification is available yet so all the breeds are generally
considered local nondescript. However, based on the origin or location of their habitat,
various names are used by producers, traders and consumers to identify a particular type of
animal. Each of these types has specific phenotypic characteristics as well as other quality
attributes for differentiation, e.g. meat of animals from certain locations may be more tasteful
than others due to differences in ecological conditions including feeds. Among the sample
sheep, Arsi (45%) and Adal (15.6%) were major know origins and 19.9% came from
unspecified places, while Arsi (37.7%), Simit Sheleko (14.2%) and Somali (12.3%) were
major origins of goats and 27.8% of the goats came from unspecified places.. Other things
being equal, in the markets surveyed, Somali lowland sheep commanded significantly lower

9
prices than Arsi another highland sheep but sheep of unknown origin commanded
significantly higher prices than Arsi and other highland sheep. This may be because the
highland markets prefer sheep of highland origin compared to those from pastoral lowland in
the Somali region of Ethiopia and Somalia. On the other hand, goats originating in Somali
and Jimma commanded significantly higher prices than Arsi and other highland origin goats.

Among the buyers of sheep 44.3% were consumers, 33.2% were traders, 16.4% were
producers and 6.1% were butchers and restaurants; for goats buyers, the percentages were
respectively 45.3, 27.2, 20 and 7.4. Among different types of buyers of sheep, other things
being equal, farmers paid significantly lower prices compared to traders, consumers, and
butchers and restaurants. One possible reason is that farmers as buyers are more prevalent in
local markets while traders operate in all markets and dominate the secondary markets. In
addition, farmers usually purchase animals for breeding, rearing and fattening and not for
resale and consumption like traders and other consumers do. Thus they may prefer small
animals with lower prices. Farmers supplied only 5.2% animals in Addis Ababa market but
99.6% and 97.3% of animals in Dire and Shashemene markets and 83% and 89% in Arerti
and Meki markets. In case of goats, consumers paid significantly higher prices compared to
other types of buyers. Farmers paid lower prices but the difference was not significant.

Fifty four percent of the sellers of sheep were farmers and 46% were traders while 67% of
sellers of goats were farmers and 33% were traders. Other things being equal, there was no
significant difference between prices received by farmers and traders as sellers of sheep. In
case of goats, farmers received significantly lower prices than traders, which could be partly
because farmers operate at the bottom end of the market chain and traders being profit
motivated, may try to pay the lowest price possible in any bargain or negotiation. Farmers are
often forced to sell their animals for reasons other than profit motives in situations of dire
cash needs, so may accept lower prices than traders.

About 89% of buyers of both sheep and goats were male, the remainder was females. On the
other hand, about 92% of the sellers of both sheep and goats were male. Buyer sex or seller
sex had no significant influence on price in either species.

Among the nine markets, Methara is the farthest from Addis Ababa, the domestic terminal
market. Nazareth and Shashemene are also semi-terminal markets in the sense that exporters

10
of live animals and export slaughterhouses operate in these markets and these markets get
supplies from other primary/secondary markets around them. Therefore, Addis Ababa was
used as the base for comparison of prices in other markets as, other things being equal,
prices would be expected to be lower in other markets at least to the tune of marketing and
transaction costs between Addis Ababa and each of the other markets.

Other things being equal, sheep prices were equally and significantly higher in Akakai and
Nazareth, and significantly lower in Debre Zeit and Methara compared to prices in Addis
Ababa. Methara had the lowest price of all. Prices in other markets were not significantly
different compared to prices in Addis Ababa. In general, there was no clear progression in
price along the primary to terminal market chain as would be normally expected except that
the farthest market, Methara, had the lowest price. The reason for higher prices in Akaki and
Nazareth could be partly explained by the fact that exporters and processors buy animals in
these markets and they pay premium prices for best quality animals, and left over second or
third grade animals may end up in Addis Ababa market, which then virtually becomes a sink
market. Also animals from all over the country are supplied to Addis Ababa, so supply and
prices in markets in a particular transect of the country may not significantly determine
prices in Addis Ababa.

In case of goats, prices in Debre Zeit, Nazareth and Methara were significantly and equally
lower compared to prices in Addis Ababa, but prices in Meki were significantly higher.
Prices in Dire and shashamene were higher but statistically significant. Thus price
differences between markets followed to some extent the expected differences between
primary, secondary and terminal markets. One possible reason is that in general highland is
not a major production or consumption area for goats, so supplies come mainly from the
lowlands, of which the markets studied are one of major sources of supplies for Addis Ababa.

In the forgoing comparison of prices between markets, the comparison is for the same time
point or period. In reality, a major source of price variation across markets would be the lag
in price transmission between interlinked markets. If market A supplies animals to market B
and the price in market B on a particular day depends on the price in market A x number of
days earlier due to the time required to move animals from A to B, then price comparison
between these two markets need to take this lagged relationship into account. In the present
anlaysis, the process of transmission and degree of market integration was not tested.

11
Other things being equal, prices showed variation between months during the survey year.
Assuming inflationary effect during the periods was minimal, the monthly price differences
could be largely explained by occurrence of festivals. In case of sheep, compared to prices in
August 2003, which coincided with the Ethiopian new year, prices in most previous months
were lower, and they were significantly lower in October- December 2002 and in February
and June of 2003. Prices in April (coinciding with Ethiopian Easter) were higher than in
August but the difference was not statistically significant. A possible reason is that the month
captured both the build up to the peak at the time of Easter but also the drastic fall following
the Easter festival (Figure 2), thereby canceling the higher price segment of the month. In
case of goats, compared to August 2003 prices were significantly lower during previous
October-November but were significantly higher in January (coinciding the Ethiopian
Christmas) and April (Ethiopian Easter).

Farmer seller equations for sheep and goats

Although in the overall equation, prices differed significantly for sheep and goats of certain
origin, in the farmer seller equations for sheep similar differences were observed but in the
farmer seller equation for goats, prices did not significantly differ between goats of different
origin, indicating that farmers pay equal prices for all kinds of goats for rearing.

In the farmer seller equation, prices did not significantly differ between markets either for
sheep or for goats though there were significant differences between certain markets in the
overall equations. Price differences between months and between markets had fairly similar
pattern as in the overall equations.

Producers sold more than 66% of the animals due to cash shortage, 10% as a result of drought
and 4% due to feed and water shortage and 20% due to various other reasons including
targeting a festival for profit. Other things being equal, prices received by farmers for sheep
sales were significantly higher when they were sold for ‘other reasons including festival time
market targeting’ compared to sales for cash shortage as well as drought and feed/water
shortage. In case of goats, prices received were significantly lower when sold due to drought
compared to sales for cash shortage and other reasons.

Central highlands (Menz and Arsi areas) are the main sheep producing areas in the market
transect studied. In this zone, about 40% cash income from sales of animals was attributed to

12
sales of sheep (Gryseels, 1988). Before harvest time, producers usually sell animals as a
result of cash shortage to maintain livelihood. For instance, most of the farmers who
mentioned cash shortage as a reason for sale came from Arsi and Menz areas, where the main
crop harvest season ‘meher’ is from November to January and short harvest season ‘belg’ is
from April and May. Usually cash needs are higher before the harvest season (Andargachew,
1993). In the markets surveyed, there were no animals originating from Menz during Meher
in November, December, January and during ‘Belg’ in April and May due to the availability
of cash from crop sales (Table 2). Farmers in these locations faced serious cash shortage in
June, July and October and as a result they increased sales of animals in these months.

Input constraints like feed and water shortage could significantly affect the supply and sales
of animals. The occurrence of feed and water shortage often overlap and occur at the same
time. In the transect studied, feed shortage occurs during February to June in dry season. Feed
shortage was significant in Arsi from March to August during the survey year and sales of
animals increased and reached peak in May and June (Table 2). In Simit Sheleko, feed
shortage peaked in September and October when most animals were sold. Sales of animals in
this zone was also affected by situations in the adjacent pastoral regions such as in Afar and
Somali regions due to feed and water shortage. Sales of animals from Adal and Somali
(Shinile zone which is close to Methara through Bike) pastoral origins are significant in
number because of high potential for small ruminant production. In Afar and Somali, the feed
shortage period stretches longer than in the neighboring highlands. Feed shortage was severe
during February to July so sales peaked in June (Table 2).

Natural factors such as drought also affected the supply of sheep and goat, especially from
Afar or Adal and Somali pastoralist areas. In bad years, pastoralists adopt coping strategies,
increased livestock sales, increased slaughtering of own animals (SC-UK, 2001). Gryseels
(1988) noted that small animals were the first to be considered for sale when food is depleted.
Pastoralist in the highland markets sell goats in order to cope up with the drought. Sales from
Afar origin during August to November was highest (Table 2).

Summary
Small ruminants play a major role in the livelihood of smallholder farmers in the highlands.
Producers and traders of small ruminants generally consider seasonal and intermarket price
differences as major problems constraining smallholder benefits from market participation. In

13
this paper, the results of a survey on seasonal and inter-market variation and other
determinants of prices of small ruminants in selected highland and mid-altitude zonal markets
are presented. Understanding the sources of price variation may be helpful for producers to
understand buyer preferences for specific characteristics of animals and target breeding,
fattening, time and place for sales to gain from important market opportunities.

Hedonic price models were fitted to a sample of 1397 sheep and 1293 goats respectively for
which data were collected from nine markets in Ethiopia over a 12 month period. The
objective was to determine seasonal and inter-market differences in prices after controlling
for the effects different attributes of the animals, the buyers and the sellers. Results indicate
that, controlling for attributes of the animals and of the buyers and sellers, there were
significant differences in prices between seasons and markets. Seasons in which farmers
faced severe cash shortages exhibited the lowest adjusted prices for animals they sold,
indicating that although livestock may provide a fall back position for cash in times of crisis,
terms of trade may be worst when farmers need cash the most. In general, there was no clear
progression in price of sheep along the primary to terminal market chain ending in Addis
Ababa as would be normally expected except that the farthest market had the lowest price.
The reason for higher prices in some intermediate terminal markets could be partly explained
by the fact that exporters and processors buy animals in these markets and they pay premium
prices for best quality animals, and left over second or third grade animals may end up in
Addis Ababa market, which then virtually becomes a sink market. In case of goats, price
differences between markets followed to some extent the expected differences between
primary, secondary and terminal markets. One possible reason is that in general highland is
not a major production or consumption area for goats, so supplies come mainly from the
lowlands, and the markets studied are one of the major sources of supplies for Addis Ababa,
so the price movement followed the market chain.

Acknowledgements: The authors are grateful to the Debre Zeit Station of the Ethiopian
Agricultural Research Organization, International Livestock Research Institute and the
Government of Italy for financial assistance in completing this study. However, the authors
alone are responsible for the content.

14
Table 1: Determinants of sheep and goat prices in eastern highlands of Ethiopia
Covariates and factors Sheep Goats
Price Standard t-values Price Standard t-values
margin error margin error
Covariates
Age (months) 2.90 0.28 10.45*** 1.28 0.24 5.37***
Age2 -0.03 0.004 -8.18*** -0.01 0.004 -3.02***
Heart girth (cm) -3.54 1.35 -2.63*** -12.93 1.325 -9.760***
Heart girth2 0.04 0.01 4.62*** 0.11 0.01 11.82***
Height of animals (cm) -2.35 1.58 -1.49 -1.30 1.38 -0.94
Height2 0.02 0.01 1.88* 0.02 0.01 1.51
Expected skin price (birr) 0.03 0.23 0.12 1.83 0.45 4.06***

Sex
Female 0.0 0.0
Male 20.29 2.38 8.53*** 13.69 2.06 6.64***
Body condition
Good 0.0 0.0
Average -20.33 2.27 -8.95*** -13.25 2.08 -6.36***
Poor -34.29 5.18 -6.62*** -30.58 3.71 -8.25***
Tail type
Fat ramped 0.0
Fat tailed -14.58 5.08 -2.87*** 0.0
Thin tailed -12.25 5.23 -2.34*** -0.29 2.78 -0.11
Color
Red 0.0 0.0
Dalacha -0.98 2.63 -0.37 -3.12 2.72 1.15
White -3.23 2.79 -1.161 -2.54 2.68 0.95
Black -9.62 2.84 -3.39*** -2.70 2.56 1.06
Wessera -4.75 3.48 -1.37 -3.96 3.69 1.07
Brown 4.66 6.25 0.75 -5.50 4.68 1.17
Bora -0.50 8.39 -0.06 -2.16 5.93 0.36
Breed
Local 0.0 0.0
Crossbreeds 14.14 6.78 2.09*** 5.45 12.22 0.45
Origin
Arsi 0.0 0.0
Adal -2.95 3.35 -0.88 -4.28 10.28 -0.42
Simit Sheleko 6.96 5.14 1.36 3.81 4.64 0.82
Menz 4.92 4.46 1.10 -3.19 5.32 -0.60
Somali -17.58 6.88 -2.56** 6.53 3.744 1.75*
Others 0.80 5.24 0.15 -4.04 5.85 -0.69
Unknown 7.30 3.11 2.35** -1.22 2.96 -0.41
Jimma 7.35 7.36 1.00 21.45 9.49 2.26**

15
Table 1 (continued)

Factors Sheep Goat

Price Standard t-value Price Standard t-value
margin error margin error
Buyers Type
Trader 0.0 0.0
Consumer -0.89 2.53 -0.35 6.14 2.63 2.33**
Farmer -14.85 3.20 -4.65*** -3.91 3.04 -1.28
Butcher and restaurants 3.03 4.15 0.73 2.07 3.87 0.54
Buyers sex
Female 0.0 0.0
Male -5.57 2.99 -1.86* 0.83 2.87 0.77
Seller type
Trader 0.0 0.0
Farmer -1.93 2.41 -0.80 -9.40 2.42 -3.89***
Seller sex
Female 0.0 0.0
Male 0.98 3.52 0.28 -2.07 3.21 -0.64
Market
Addis Ababa 0.0 0.0
Akaki 10.86 5.08 2.14** 9.54 5.38 1.77
Debreziet -13.89 5.27 -2.63*** -15.43 5.69 -2.71***
Dire 3.95 5.65 0.70 5.45 5.45 1.00
Nazareth 11.40 4.41 2.59** -13.01 5.04 -2.58**
Arerti -6.38 5.50 -1.16 -3.201 6.19 -0.52
Meki -5.67 6.68 -.85 11.90 5.87 2.03**
Shashemene 1.89 5.32 0.36 6.31 5.52 1.14
Methara -27.54 6.33 -4.35*** -16.20 5.26 -3.08***
Month (occasions)
August 2003 (new year) 0.0 0.0
July -4.39 4.43 -0.99 -2.65 4.43 -0.60
June -9.37 3.82 -2.46** 3.17 3.79 0.84
May -2.81 4.04 -0.70 -3.73 4.00 -0.93
April (Easter) 3.68 4.14 0.89 8.18 4.10 1.99*
March (Christian fasting) -4.68 3.964 -1.18 1.82 3.87 0.47
Feb (Christian fasting) -11.86 3.83 -3.10*** -3.53 3.71 -0.95
January 2003 (Christmas) -2.56 4.360 -0.59 8.86 4.22 2.10**
December 2002 -11.67 3.98 -2.93*** 0.48 4.01 0.12
Nov (Ramadhan fasting) -16.70 4.06 -4.11*** -6.87 3.94 -1.74*
October -11.01 3.77 -2.92*** -6.99 3.73 -1.87*
September 2002 -3.94 4.02 -0.98 0.83 4.04 0.21
R2 0.66 0.68
Adjusted R2 0.65 0.66
N 1397 1293
*, ** and *** indicate significant at 10% , 5% and 1% level.

16
Table 2: Percentage of sample animals sold by reason for sale, origin of animals and
month of transaction in Eastern Shoa zone
Cash Shortage Drought Feed shortage
Months Arsi and Menz Adal Somali Arsi Simit Shelko Adal Somali
September 0 13.3 0 0 33.3 7.1 0
October 42.9 13.3 0 0 33.3 0 0
November 0 21.7 16.7 0 0 0 0
December 0 0 16.7 0 0 7.1 0
January 0 4.3 0 0 0 0 0
February 14.3 4.3 0 0 0 7.1 16.7
March 0 4.3 0 11.1 16.7 7.1 16.7
April 14.3 4.3 0 11.1 0 7.1 8.2
May 0 4.3 8.3 22.3 0 14.3 16.7
June 14.3 8.7 8.3 33.3 0 21.6 25.0
July 14.2 0 33.3 11.1 16.7 14.3 16.7
August 0 21.5 16.7 11.1 0 14.3 0
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Source: Field survey

References

Andargachew, K., and Brokken, R.F., 1993. Intra- annual sheep price patterns and factors
underlying price variation in central highlands of Ethiopia. Agric. Econ. 8, 125-138.
Brorsen, B. W., Grant, W. R., and Rister, M. E., 1984. A hedonic price model for rough rice
bid/acceptance markets. American J. Agric. Econ., 66:156-163.
Buvanendran, V. W., Umoh, J.E., and Abubakar, B. Y., 1980. An evaluation of body size as
related to weight of three west African breeds of cattle in Nigeria. J. Agric. Sc
(Camb.), 95:219-224.
Gujarati, Damodor N., 1988. Basic Econometrics. 2nd edition. McGraw Hill, New York,
USA.
Gryseels, G. 1988. Role of livestock on mixed smallholder farms in the Ethiopian Highlands.
A case study from Baso and Worena werda near Denreberhan. PhD. Dissertation,
Agric University of Wageningen, Netherlands.
Jabbar, M.A., 1998. Buyers preferences for sheep and goats in southern Nigeria: A hedonic
price analysis. Agric. Econ. 18, 21-30.
Jabbar, M A, and Diedhiou, M L. (2003) Does breed matter to cattle farmers and buyers?
Evidence from West Africa. Ecological Economics (USA) 45(3):461-472.
Lucas, Robert E.B., 1975. Hedonic price functions. Econ. Enquiry, 13:157-178.

17
Oczkowski, E., 1994. A hedonic price function for Australian premium table wine. Aust. J
Agri. Econ. 38(1), 93-110.
Parker, D.D., 1993. Retail price response to quality characteristics of fresh peaches by store
type. Agribusiness., 9:205-215.
Parker, D. D. and Zilberman, D., 1993. Hedonic estimation of quality factors affecting the
farm-retail margin. American J. Agric. Econ., 75:458-466.
Payne, W. J. A., 1990. An introduction to animal husbandry in the tropics. 4th edition.
Longman, London, UK.
Rodriguez, A., Ali, I., Afzal, M., Shah, N. A., and Mustafa, U., 1995. Price expectations of
sheep and goats by producers and intermediaries in Quetta market, Pakistan. Agric.
Econ.,12:79-90.
Rosen, S., 1974. Hedonic prices and implicit markets: Product differentiation in pure
competition. J. Pol. Econ., 82:34-55.
Williams, C.H., Rolf, J., and Longworth, J W., 1993. Does muscle matter? An economic
valuation of live cattle characteristics. Rev. Marketing and Agric Econ., 61(2):169-
189.
Save the Children UK, 2001. Report on Shinile Agropastoral Food Economy Zone. Shinile
Administrative Zone, Somali Regional State, Ethiopia.

18