Determinants of cattle Market Participation Decision by Pastoralist in Southern Ethiopia: The case of Moyale district in Borana zone of Oromiya Region

The logistic regression analysis was employed to identify determinants of cattle market participation decision by pastoral household in Borana zone. Three kebeles were selected based on secondary data and reviewed literatures. Primary data were collected from 223 sampled pastoralist, 25 traders and 14 brokers. Econometric Approach was used to investigate determinants of pastoralist household cattle market participation. For logistic regression model six explanatory variables used and out of which three significantly affect probability of producers’ market participation decision. The number of cattle owned, number of camel owned and gender are predictors that significantly affect probability of cattle market participation. Both number of cattle and camel owned per household affects positively the probability of market participation. This implies that rearing large size of cattle and camels provoke cattle market participation decision of pastorals and it is also advisable to promote Borana cattle breed conservation. Holding the other predictor variables constant, being male by its-self increase the probability cattle market participation by 2.92 factors as compared to female household heads. This calls for systematic intervention of government to mainstreaming gender in cattle market that supposed to optimize the sector productivity. For one unit increase in number of cattle owned the probability of cattle market participation increase by 1.12 factors, other factors held fixed. For one unit increase in number of camel owned, the probability of cattle market participation decision increase by factor of 1.23 times, given the other variables held constant. Top 1st to 5th ranked constraints of cattle market include lengthy market, brokers’ interference, disease and parasite, clan conflict and unfriendly relation between market actors. Other 6th to 10th ranked market problems include informal trade, lack of reliable market information, recurrent tax, infrastructure constraint and initial capital problem. As remedial measure shortening the supply chain, animal health center development, capacity development, adopting value addition practices, update market information provision, cooperative formation and linking producers to market is believed to optimize the benefits of pastorals and boost cattle productivity.


Introduction
The greater cattle holdings results in greater cattle market participation and the greater capacity of willingness to sell animals. The key practical implication is that active livestock markets depend on pastoralists maintaining sufficiently large herd sizes that they become willing to liquidate animals through the market. It is generally believed that pastoralists sell their animals at least partly in response to demand for cash to meet expenditure needs. Relatively wealthy pastoralists, with greater herd size, have considerably higher expenditure rates and thus use livestock markets more frequently to cash out animals (Barrett et al. [1]). The cattle markets in most parts of the country are characterized by seasonality in flow and prices of animals. In the pastoral lowland areas major factors contribute to seasonal flow include festivals, weather, pasture, seasonality of consumption demand (fasting and ceremonial period) in domestic, drought, disease outbreak, lack of information, for traders to take advantage of all the markets the whole week and fulfill their requirements (Getachew et al. [2]). Moyale and Mandera are the two largest cross-border terminal markets for livestock between Ethiopia and Kenya. Between 70-80 percent of live animals sold in these two markets originate in Ethiopia. The Ethiopia/Kenya livestock trade is very important because it links prime cattle production areas of southern Ethiopia to the region's largest market in Nairobi, Kenya (Awur [3]).

General Objective
a.
To investigate relative factors of pastoralist household choice of cattle market participation

Methodology
The study area The Borena administrative zone is situated in Ethiopia's Oromia regional state at about 570 km (Yabalo town) from Addis Ababa (Niguse [5]). The Borana zone is made up of 13 districts, divided between two agro-ecological zones, the semiarid lowlands to the south and the more humid lands at higher altitudes to the north (Beyene and Yibeltal [6]).

Methods of Data Collection
In this study both secondary and primary data was obtained from different sources. Journals, books, proceedings, CSA, and internet browsing were secondary data sources that used in the study. Primary data sources include Pastoral household interview, traders interview, brokers interview and key informant interview. A preliminary assessment was conducted to collect basic information about the woreda in order to select representative kebeles and the major cattle marketing centers. Structured and semi structured questionnaires was prepared and pre-tested for households interviewed. Key informants interview and household interview were employed for the study. Using the questionnaire, interviews was conducted to gather data on household characteristics, socioeconomic and demographic characteristics, and other sources of income, marketing period, months and seasons, marketing and challenges and threats of cattle trading.

Sampling Techniques and sample size
For this particular study a two stagesampling techniques was used. The main sampling techniques used for this study were probability and non probability sampling. Factors like percentage of pastoral population of the districts, number of livestock kept/income difference, age of pastoralists, major function of actors in cattle marketing: supplier, buyer, traders, middlemen and broker, proximity to major marketing centers, number of pastoral Peasant associations and cultural issues were important issued while selecting the representative households in the districts. The producers interviewed in the study were pastoralists. The random probability sampling techniques was used for selecting the representative households from the area. Two stages sampling technique was used for selection of pastoralists Kebele identification that made through secondary data of pastoralists development office, proximity to marketing centers, gender, ownership of cattle and other important economic variables. The random probability sampling techniques was used for selecting the representative producers. Two stages sampling technique was used for selection of pastoralists Kebele identification that made through secondary data of pastoralists development office. Three Kebeles from the pastoralist area were considered in the survey. Respondent sample size per each Kebele was determined proportionally to the number of total household in the area. The sample size determination techniques employed was Rule of Thumb Techniques that estimate by using the following formula:100% for 0-100 populations,10% for101-1,000 populations, 5% for 1,001-5,000 poulations,3% for 5,001-10,000 population and 1% for more than 10,000 respectively (Yount [9]). Based on this technique the sample size of pastoralists household interviewed from respective kebels, Maddo, Mado Miggo and Laga Sure were 100,63 and 60 respectively. The survey study at woreda level considered 167 male (74.9%) and 56(25.1) female household heads. The sample size of male and female household heads interviewed in Maddo kebele was 78(78%) and 22(22%). The number male and female respondents interviewed in Laga Sure kebele was 45(75%) and 15(25%). From whole interviewed respondents in Mado Migo Kebele, 44 (69.8%) was male and 19 female (30.2%).

Method of Data Analysis: Descriptive and Econometric Analysis
Descriptive statistical analysis was used to define mean, mode, percentage and standard deviation. Econometric analysis was employed to identify factors affection household choice of pastoralists to take part in cattle market. The data collected was analyzed using Statistical Package for Social Science (SPSS) version 20, STATA version 10 and Excel 2007 after editing, coding and the raw data collected from survey.

Results and Discussions
This chapter presents the results of both the descriptive and econometric analysis. The first section deals with the socioeconomic characteristics of producers and traders. The second section looks into the econometric analysis. In the third part of the chapter is about market system in the district.

Socio-Economic characteristics of pastoralists
This section presents the socioeconomic characteristics of sample households in the area. The socioeconomic characteristics looked in to in the survey duration comprised of age, family size and years of schooling (Table 1).
Age Distribution of pastoralist Households: The average age of the Pastoralists household head was 42. However, it ranges in between 20 and 81. The proportion of sampled producers whose age lies in the range between 20 and 65 was 96.9%. The remaining 3.1% sampled pastoralists were aged more than 65 years. This findings show that majority the cattle producers were in the age range of active labor force and only few known to be dependant age category. Thus, this is very important with respect to the quantity supply of cattle and household market participation decision. The study is not in line with Desta et al. [10] result that states roughly three-quarters of pastoral households are of working age. Family Size of Pastoral Households: The family size distribution (Table 1) shows that the average family size of Moyale pastoralists is 7. However, the range of family size for sampled pastoralist is between 2 and 25. From whole sample households in the area, the proportion of households with family size of 2 to 3 family sizes are 9.4%, 4 to 6 are 41.7 %, and 7 to 10 are 41.3% and 11 to 25 are 7.6%. The survey result reveals that majority of households belong to medium and large family size. The proportion of households with low and very high family size is few. The survey result is not agree with Tolossa et al. [8] result that stated the mean family size for pastoralist in Borana area was 8.

Education level of pastoral households heads (N=223):
The result in the Table 2 revealed that 81.2% of sampled households did not attend formal education, whereas; the proportion of pastoralist household who attended formal education (attended from primary grade to high school) was 18.8%. The distribution of pastoralist households with respect to formal education attendance shows that less than one fifth they attended formal education. The study is not in line with Tolossa et al. [8] where the proportion of producers attended formal education in which was 41.7 %.  Age structure of cattle traders: Table 3 revels that the mean age for cattle traders was 39.44 years. However, the age of sampled cattle trader in pastoralist's area ranges in between 23 and 60. The study result confirmed that all the cattle traders are within the age bracket of active age category. This was an implication that almost all the traders are energetic enough to carry out cattle marketing activities. The survey results in Table 4 revealed that 36% cattle marketers were illiterate,20% attended primary school (grade 1 to 4) and 36% attended junior school (grade 5 to 8) and 8% attended secondary high school (grade 9 up to 10) respectively.

Years of Schooling for traders
Gender of Cattle Traders: The survey results revealed that all (100%) cattle marketers were males. Since gender attributes the main criteria for determining entrance into cattle trade and thereby, it is advisable to consider gender mainstreaming.

Family size of cattle traders:
The survey result shows that the average family sizes of cattle traders was 8 (Table 5). Regarding the distribution of persons per household, the result in the (Table 6) shows that, the majority of households have 4-6 persons (48 %) closely followed by those households with International Journal of Environmental Sciences & Natural Resources 7-10 persons (40%). But the proportion of households with very large family size (12-14 persons) is 12% . High school (9-10) 2 8% 100% Factors affecting pastoralists Household's Choice of Cattle Market Participation: In order to estimate the market participation decision of pastoralists in the area, Logistic regression Model was employed by (STATA version 10) Here, the factors influencing the household's discrete choice behavior was modeled using a binary logistic model. Then binary logistic regression model allowed the estimation of a set of probabilities of market participation regimes for households with a given characteristic logit model where only one coefficient is estimated for all the outcomes. To investigate factors embedded in deciding cattle market participation, the following variables were used as explanatory variables: gender, education level, and other sources of income, household size and proximity to animal health center, number of camel owned, source of market information and accessibility of business support services. Similar to OLS regression, the prediction equation of Logistic regression model can be written as: Log (p/1-p) = b0 + b1*gender + b2*number of cattle owned + b3*number of camel owned, where p is the probability of cattle market participation. Expressed in terms of the variables used in this study, the logistic regression equation is Log (p/1-p) = -2.87+ 2.94*gender +.19* number of cattle owned + .20*number of camel owned +error term. These estimates tell about the relationship between the independent variables and the dependent variable, where the dependent variable is on the logit scale. These estimates tell the amount of increase in the predicted log odds of cattle market participation = 1 that would be predicted by a 1 unit increase in the predictor, holding all other predictors constant. Because these coefficients are in log-odds units, they are often difficult to interpret, so they are converted into odds ratios.
The model specified for market participation: where P is probability of explanatory variable to affect market participation decision and G is a function taking on values strictly between zero and one: 0 < G(z) < 1, for all real numbers z. Yi=the dependant variable this defines probability of market participation and xi is explanatory variables. The independent variables defined as, X1=gender of respondent, X2=number of camel owned, X3=market information source, X4=distance in km between pastoralist home and animal health center, Y=1 when the household participated in market and Y =0 Otherwise. In short, the logistic regression model for market participation decision can be written as: log (p/1-p) = b0 + b1*gender + b2*cattle owned + b3*camel owned----------------------Equation 2 Gender: The coefficient for the variable gender is 1.071. Because these coefficients are in log-odds units, they are not as such simple to interpret, so it is converted into odds ratios. The log odds units are converted in odds ration by taking exponentiation of the coefficient manually and excel sheet. As the result, odds ration coefficient for gender=Exp (1.071454) = 2.919622. This is the proportional odds ratio of comparing male to females on cattle market participation given the other variables in the model are held constant. For males, the odds of market participation decision 2.919622 times higher than for female, given the other variables are held constant. This is to mean that being male increase probability decision of cattle market participation by the probability about 2.92 units as compared to females.
Number of cattle owned: the parameter estimate for the variable cattle owned is 0.1935. This implies that for one unit increase in cattle owned, we expect 0.1935 increase in the logodds of dependent variable termed market participation, holding all other explanatory variables constant. The size of cattle reared by pastoralists is not only the variable that positively affects cattle market participation; it also affects number of cattle supplied. As the herd size increases, the probability of pastoral household to take part in cattle market increases while the probability of non-participation in cattle market decreases. This finding is also agrees with Asfaw and Jabbar [11] and Barrett et al. [3] that states the households with larger herd size has higher ability to generate surplus animals and are therefore more likely to sell. The implication is that active cattle markets depend on pastoralists attaining and maintaining sufficiently large herd sizes that they become willing to liquidate animals through the market.

Number of camel owned:
The coefficient for the variable cattle owned is 0.2031. This indicates that for one unit increase

International Journal of Environmental Sciences & Natural Resources
in camel owned, we expect 0.2031 increase in the log-odds of dependent variable termed market participation, holding all other explanatory variables constant. Total size of camel the respondent owned, is continuous variable and the sign is positive. The positive coefficients of parameter estimate implies that the pastoralists' household having one more additional camel have the more higher probability to participate in the cattle market. Due to wealth difference over pastoralists households, the household that had more camel have also more cattle, the probability of market precipitation increased with camel size. As the result households that owned one more camel have more likelihood of cattle participation. In addition to this, as climate adaptation strategy, pastoralists in the study districts has been changing production from cattle to camel there by increasing the camel herd and in opposite decrease the cattle by being participated in cattle market. This study is also supported by other finding Kelemework [12] that states pastoralists have long developed adaptive strategies against environmental shocks through effective management of their resources. Adaptive strategies include the establishment of strong economic and social support networks, herd splitting, and herd diversification. Thus, for a one unit increase in number of camel owned, the odds of market participation 1.225168 times greater, given the other variables are held constant in the model.  Table 7 illustrates that cattle selling seasons identified by pastoralists in the area are dry season for about 40.4% households, wet season for 8.5 % pastoralists and holidays for about 8.5% respondent households. The other 15.7% sampled households usually sell at combination of wet season, dry season and holidays. The remaining 26.9 sampled households sell their cattle in holidays and dry seasons. This implies that institution and climate variability has such a great role in cattle selling and market participation decision in pastoralist's households. As of interviewed pastoral households respondents confirmation in the area, usual buying seasons and occasion of cattle are dry season for 44.8 % respondents, wet seasons for 13% households and holidays for 4.9% respondents accordingly. The proportion of pastoralists buy cattle in wet season, dry season and holidays is 14.3% and that buy in both holidays and dry season equals 22.9% respectively. Even if the proportion of cattle bought at wet season overrides sold, the survey result justified that major cattle selling season are also cattle buying months for pastoral households in the area. The major seasons of cattle marketing identified in the area are based on weather variable wet season, dry season and holidays based on marketing occasions. As result, the study result suggests that the major cattle marketing season skewed toward climate variability especially toward dry season. Change in weather condition understood as one of the important variable that forces the pastoralist household to take part in cattle market and cattle supply to market. Getachew et al. [2] also observed that cattle markets are characterized by seasonality in flow and prices of animals. Shifts in supply could occur as a result of factors including seasonality of consumption demand in domestic and export markets, drought, disease outbreak, lack of information, availability of food aid, clan conflicts and others 2008. Awuor [3] concluded that the volume of livestock traded and pastoralists' willingness to market dependents on the availability of pasture and water and rainfall variability. Pastoralists are reluctant to sell their animals during the wet period when pastures rejuvenate and the animals return to milk production. But expectations of prolonged dry spells and outbreaks of disease increase the supply of livestock to the market. The off take rates for livestock during drought were estimated to be about twofold the off take rates during a normal year.

Major Cattle marketing seasons and occasions in Pastoralist Area:
a.
Months of Cattle selling b. The Figure 1 above points out that the major cattle selling months in study area is highly linked with festivals or holidays (informal institution), seasons (dry and wet seasons) and related household cash need. From Figure  1 above, cattle selling in September, January and June are the top season of marketing that are known dry season in the area. As the study result indicates the double festival in September, dearth period around January and June inflated cattle selling frequency. This study result agrees with other finding of (MOA,2013) that stated cattle supply and demand are heavily influenced by the season of the year and the

International Journal of Environmental Sciences & Natural Resources
occurrence of religious and cultural festivals, as well as availability of feed which drives sales during the dry season.
c. Figure 2 point outs that the producers purchase cattle throughout the year. However, there was variation in level and frequency of cattle bought across months. The proportion of cattle bought around April (first rainy season) was the major purchasing months that followed by second rainy season around October. The proportion of cattle bought around February (first dry season) was the lowest followed by the August (second dry season).  The broker in cattle market confirmed that they collect different level of income from different cattle type. The mean income earned by broker by facilitating ox, cow, bull, heifer and calves was 75.00, 72.86, 65.71, 58.58 and 43.57 respectively. Table 8 presents amount of income earned by broker. f. Table 9 explains the large proportion of brokers obtained income from ox marketing process ranged in between 50 and 70 (42.8%), followed by 20 and 40 (28.6%) and 80 and 100 (21.4%). The remaining 7.1 percent of brokers collected income more than 100 Birr per ox. The range of income earned by brokers from cow market process was 20 to 40 (28.6%), 50 to 60 (50%), 80 to 100 (14.3%) and more than 100 Birr (7.1%) accordingly. The large proportion of income earned by brokers range from bull was 50 to 70 (50%) that followed by 20 to 40 (28.6%), 80 to 100 (14.3%) and more than 100 Birr (7.1%). Majority of brokers collected income from heifer in the range between 50 to 70 (50%), that loosely followed by 20 to 40 (28.6%) and 80 to 100 (21.4%). The income earned by brokers from calve marketing activity lied in the range between 20 to 40 (50%) and 50 to 70 (50%). This study implied that almost all income earned by brokers from various cattle category skewed toward 50Birr. The brokers in pastoralist area obtain different level of income from different market actors. The survey study indicated that broker collect payment for role they play in cattle marketing process from both seller and buyer. An average income collected by brokers in the area from producers was 47 and 96 Birr trader. Minimum amount of payment from producer was 20Birr and maximum income was 80Birr. The minimum and maximum payment from trader was about 50 and 250 Birr respectively. This study result confirmed that brokers collect income from both seller and buyer, but in different amount. h.
Among the interviewed brokers 64% perform their activity in collaboration with other brokers and the remaining 36% undertake the activity individually. Similar to other business activity in the area, the brokering activity faced some sorts of conflicts with other stakeholders in the marketing process. The proportion of brokers faced conflicts with other stakeholders in the area amounts to 86% and that did not 14%. Some of major cattle marketing actors that had conflicts with broker mostly were other brokers for 42%, pastoralists for 33%, and trader for 25% accordingly. The brokers used different sorts of conflict resolution techniques that rose between market actors. It was reported that 83% of respondents' brokers resolve conflicts by having their own customer buyer and seller and 17% used elders from the area that could be traders, brokers and pastoralists as well. i.
Responsible Household member of Cattle marketing. j.
The function of households in cattle market chain of pastoralist's area is different for various households' members. Major reasons for participation of different household members in different cattle market activity listed are institutional related reasons and works division. The institutional related factors such as gender, personality, holidays and similar issues had such an influence in cattle market chain dynamics. As the result institution was responsible for 56% pastoralist cattle marketing and 46% marketing activity accomplished in relation to work division. Form interviewed pastoralists 56% households performed by husbands only, 3% performed by Wife, 35% performed by anybody who owns the cattle and 6% performed by both husband and wife. The cattle marketing activity as any business activity was performed by community in the area by different household members. Important factors that influence carrying out cattle marketing activity listed were the institutional matter and work division. From the influential factors institutional matter highly important over marketing of cattle in the area than work division and it was assured by 56% of interviewed households and the remaining 44% households under take cattle market based on work division of the household member. The proportion of respondents confirmed was 56% for husband, 3% for wife, 35% anybody who owns the cattle and 6% for Husband in line with wife accordingly [12,13].

Conclusion
Regression analysis confirms that greater cattle and camel holdings results in greater cattle market participation and it also has a positive impact upon household choice of market participation. The key practical implication is that active cattle markets depend on pastoralists attaining and maintaining sufficiently large herd sizes of cattle and cattle that make them willing to liquidate animals through the market. It is generally believed that pastoralists sell their animals at least partly in response to demand for cash to meet expenditure needs. Relatively wealthy pastoralists, with greater herd size, have considerably higher expenditure rates and thus use cattle markets more frequently to cash out animals. This study result also pointed out that destocking is not solution for production and productivity optimization, rather Cattle breed conservation promotion. It is also justified that quantity is believed to bear quality. The major market constraints sorted out by market actors include lengthy market channel, brokers interference, disease and parasite, clan conflicts and unfriendly relation between market actors, illegal trade, lack of reliable market information, recurrent tax, infrastructure problem and initial capital shortage. This calls for systematic government intervention that supposed to optimize the cattle production productivity and ease characterized constraints through cooperative formation, market and market route searching, updating market information, capacity building, linking producers to market, adopting value addition tools and integrated approaches of different stakeholders.