Livestock Research for Rural Development 21 (9) 2009 Guide for preparation of papers LRRD News

Citation of this paper

Pattern and determinants of meat consumption in urban and rural Ethiopia

Shawel Betru and H Kawashima

Laboratory of International Environmental Economics
Department of Global Agriculture, Graduate School of Agriculture and Life Sciences, The University of Tokyo, Bunkyo-ku Yayoi 1-1-1, Tokyo, Japan


Per capita meat consumption in Ethiopia has declined from 20 kg/capita/year in 1961 to 8 kg/person/year in 2004. FAO has been the prime source of Ethiopian livestock data though it has been acquiring the information through estimation of related resources. Central Statistical Agency of Ethiopia (CSA) started undertaking nationwide household surveys in 1996 and its result on livestock consumption data differed from FAO. In this paper, patterns of meat consumption in urban/rural Ethiopia between 1996 and 2004 were analyzed using cohort specific model by categorizing households according to level of income. Feasible general least square (FGLS) estimation technique was conducted to identify the key determinants affecting meat consumption in Ethiopia. Comparison of meat consumption by FAO and CSA was also done.


The result showed that the response of meat consumption to income was higher in urban than in rural areas. Rural meat consumption made significant improvement between 1996 and 2000 but lost the momentum between 2000 and 2004.  In urban areas, on the contrary, there was continual improvement throughout this period. The result of economic analysis revealed that urbanization and income have been found to be positively and significantly influencing meat consumption in Ethiopia at 1% and 5 % significance level respectively. On the other hand, level of cereal production and price of meat did not significantly affect per capita meat consumption.  The comparison between FAO and CSA data showed that the former overestimated the per capita meat in Ethiopia by more than 100% compared to the CSA household survey result. The disparity arose from overestimation of rate of livestock utilization than number of livestock. Therefore, the level of consumption in Ethiopia must have been lower than commonly reported by FAO.

Keywords: Cohort model, CSA, FAO, household income, meat consumption


There has been a positive trend of meat consumption in the developing countries. Delgado (2003) called this trend a Livestock Revolution. This Revolution, Delgado (2003) argued, has been driven by population growth, urbanization and income change. He advises governments and development partners in developing countries who want to help the malnourished people to follow the Revolution closely. Observations like this often understate the unequal distribution of income growth and non-uniformity of rate of urbanization in different parts of the developing countries. For example, the projection by Alexandratos (1995) suggested that the poor developing countries will remain dependant more on cereals and less on livestock consumption for daily energy in the future.


However, developing countries with well performing economies will drive the future world meat demand. Steinfeld (2002) reported that in the next 20 years the demand for livestock products will double in developing countries as a result of population growth and income improvement. He pointed out the need to address the problem of lack of competitiveness and higher risk on small holder producers which have the potential to increase livestock productions. After analyzing the impact of population pressure on livestock production in smallholder farmers in Tanzania, Ogle (1990) recommended intensive production that is integrated with crop production system.


Ethiopia has the highest number of livestock population in Africa (Solomon et al 2003). However, consumption of animal source food (ASF) has always been low and declining as a result of the low production and continuously growing population (FAO 2005; UN 2005). Ethiopia’s per capita consumption in 2004 declined by more than 100 % from an average of 20kg in 1961 (FAO 2005, Solomon et al 2003). In contrary, the average world meat consumption doubled during this period (FAO 2005). 


The UN’s FAOSTAT has been the major source of livestock data for Ethiopia and most countries of the world. This database heavily relies on national statistical agencies for source of information and its reliability depends on the quality of statistics in those countries (Speedy 2003). In the absence of reliable data from national statistical sources, FAO makes estimations using related information. This has been the case for Ethiopian livestock data for many years. Aklilu (2002) argued that estimates for Ethiopian livestock were made based on insufficient information. Further assumption of slaughter rates and livestock yields has been performed to estimate the total livestock production. It is, therefore, highly unlikely that Ethiopian livestock products’ consumption estimates from FAOSTAT had been satisfactorily reliable.


Ethiopian household income, consumption and expenditure (HICE) survey was started in 1996 and have been undertaken every four years (HICE 1996, 2000 and 2004). These surveys provide comprehensive nationwide information on livestock consumption beside other data. It appears that FAO has also adjusted some of its statistics such as the number of livestock population based on CSA publications but continues to publish estimate of meat consumption as previous.


To address the aforementioned problems, there are scarce literatures on system wise analysis of meat production and consumption in Ethiopia. Most of the livestock studies were done at smaller scale and focused on livestock yield, pricing and specific projects (For example, Guru et al 2008; Ayele and Peacock 2003) but did not address the problem directly. However, some related studies have been done at country level and provided indispensible information. For example, Jabbar et al (2007) tried to identify and map geographic distribution and market routes of livestock in Ethiopia using CSA’s livestock surveying data.


Nell (2006) made a quick review of the state of livestock production and consumption in Ethiopia for a project by the government of the Netherlands. The report suggested that livestock disease, lack of livestock nutrition, subsistence animal husbandry and weak market were the main constraints to livestock development in the country. Avery (2004), on the other hand, studied red meat and poultry consumption in Addis Ababa, the Ethiopian capital. Although the study relied on unstructured interviews and limited personal observation, it illuminated the state of meat production and consumption pattern in urban areas.


Aklilu (2002) did a survey of livestock marketing and status audit in Sudan, Ethiopia and Kenya. The study focused on export of live animals and meat produces and found that Sudan was performing better than the other two countries in livestock husbandry and suggested regional collaboration to address the gap in internal and external market information.


The CSA surveys provides meat consumption values according to urban/rural residence and income groups which helps in the analysis of factors that determine meat consumption and formulation of policies.  This study tried to achieve the following objectives: -

-         analyzing  patterns of urban and rural meat consumption in Ethiopia using cohort specific model

-         investigating factors that determine level of meat consumption among regional states in Ethiopia and

-         comparing meat consumption data from CSA and FAOSTAT.


The result of this study is expected to help understand patterns of meat consumption in rural and urban households and factors that determine the level of consumption. The cohort meat consumption analysis will contribute to our understanding of the dynamics of present and future nutrition challenges and thereby formulate appropriate policies. The comparison of CSA and FAO data will clarify the discrepancy in the level of meat consumption in Ethiopia.


Methodology and data analysis  

A cohort specific consumption model depicted on Figure 1 was used to estimate the aggregate meat consumption of urban and rural population in Ethiopia. The population was categorized into five income groups or quintiles under urban and rural residence. Average per capita meat consumption for each income quintiles in each residence was estimated from CSA survey data. These average consumptions were used to estimate age specific consumption within a household by adopting the Japanese household consumption pattern (MHLW 2000) in which consumption rises up to 15-19 years age group and stays relatively unchanged up to 65 years age and finally drops gently.  The total meat consumption was calculated for each cohort group under respective income and residence.


Figure 1.  Scheme of the model for estimation of meat consumption for urban and rural cohort population

Periodic household data from 1996-2004 was acquired from Central Statistical Agency (CSA) of the Government of Ethiopia (GoE). CSA undertook three consecutive household income, consumption and expenditure (HICE) surveys in 1996, 2000 and 2004.  These surveys were done nationwide and represented all major urban and rural areas of the country.  In 2004 survey, for example, 9,564 rural and 12,160 urban households were involved. The number of income groups was different among the surveys.  In 1996 and 2000 surveys, households were categorized into 14 household income groups, whereas in 2004 the categories were only five. In the later survey, all households in the country were arranged according to level of income and then divided into five equal parts and ascribed quintile 1 to 5, i.e., quintile 1 represents low income households while quintile 5 represents high income households. Then the households in each income quintiles were separated under urban/rural population. Because of difference in the number of household members and income inequality in urban and rural areas, the number of people in each quintile varies between residence and survey years. The average incomes of each quintile in 2004 were used to regroup the 1996 and 2000 households.


Feasible General Least Square (FGLS) estimation technique was used to analyze factors that influenced meat consumption among the different regional states in Ethiopia using Stata/SE 10.0 statistical software for window. 


Result and discussion  

Population and income change


Rural households account for the majority of the population in Ethiopia (Figure 2).


Figure 2.  Population cohort pyramids of rural and urban households in Ethiopia; Q1-Q5 represent income quintals, and   U and R stand for urban and rural households respectively

Between 1996 and 2000 urban population showed no growth, whereas between 2000 and 2004 it grew by 5.3 % a year (Table 1).

Table 1.  Summary of population and per capita meat consumption in urban and rural Ethiopia



(million people)

Per capita meat (kg/capita/year)





































Sources: Data obtained from CSA 1996, 2000 and 2004

On the other hand, rural population grew continually in both periods at a rate of 1.9% and 3.5% respectively. In the later period the rate of urban population growth exceeded that of rural population. There was change in distribution of household along income groups among surveys and residences. Between 1996 and 2000, there was a 7% and 6 % increase in the first and third income quintiles of rural households respectively. On the other hand, at the same period, the first urban income quintile declined by 13%, while quintiles 4 and 5 increased by 7% and 4% respectively.  In general, more homogenization occurred in rural areas, whereas shift towards upper income quintiles took place in urban households.  Because of change in household grouping method in 2004, direct comparison was difficult and requires caution interpretation of the result. Nonetheless, the trend appeared to follow the pattern observed between 1996 and 2000, i.e., further homogenization of income in rural areas and income accumulation in higher quintiles in urban households.


Household income and per capita meat consumption


There was distinct pattern in urban and rural meat consumption in Ethiopia in 1996, 2000 and 2004. For a given level of income, urban households consumed more meat than rural households. Figure 3 demonstrates the changes in per capita meat consumption response to change in income among the three surveys. In all the surveys, these responses were higher in urban households and the change was steep between quintile 4 and 5.


Figure 3.  Patterns of per capita meat consumption change among rural and urban household income groups in 1996, 2000 and 2004. U and R stand for urban and rural respectively. Note: There were 14 data points in 1996 and 2000, whereas 5 data points in 2004.

In rural households, there is change in the meat consumption response among the surveys although the consumption responses were lower than urban once. Rural meat consumption showed considerable improvement between 1996 and 2000 but lost most of the gain between 2000 and 2004.  The average per capita rural meat consumption increased by 11% between 1996 and 2000, whereas, urban per capita consumption remained relatively unchanged from the 1996 level (see Figure 4).

Figure 4
.  Changes in average urban and rural per capita meat consumption between 1996 and 2004

Both urban and rural households in the lower quintiles have had very low meat consumption because they had little extra income to spend on meat. However, with small change in household income, urban households showed rapid increase in meat consumption compared to rural households. Moreover, the change in consumption between consecutive income groups in rural households was smaller.


One of the factors that limit meat consumption in rural Ethiopia has been the absence of meat retail market. In rural areas people often consume meat during holidays or special occasions and it is considered rather as luxury food than essential component of daily household nutrition. During these occasions, commonly, a group of 10 to 20 people buy a live animal, slaughter and divide the meat among them. Therefore, the change in household income alone can hardly change rural meat consumption unless accompanied by change in market arrangements and culture of food consumption.


The consumption pyramids in Figure 5 show that the total urban meat accounts for bigger share of meat consumption relative to its population size. 



Figure 5
.   Meat consumption in rural and urban Ethiopia disaggregated by income quintile and cohort groups. Q1-Q5 represent income quintals

The total urban meat consumption accounted for 44%, 33% and 40% of the total national consumption in 1996, 2000 and 2004 respectively. Urban consumption declined between 1996 and 2000 because population did not grow and the proportion of upper income groups declined. In contrast, in 2004 urban per capita meat consumption increased by 26% from 2000 level.  This improvement was attributed to increase in proportion of upper income groups especially in the quintile 5 which consume exceptionally high per capita meat (see Figure 3).


In rural households, gain in meat consumption occurred between 1996 and 2000. This improvement was attributed to two factors. In 2000, firstly, rural household’s meat consumption response to change in income increased considerably. Secondly, the proportion of upper income households increased.  This phenomenon was partly reversed between 2000 and 2004 where response of meat consumption in relation to household income returned to the 1996 level. As a result the total rural meat consumption in the country decreased from 2000 level (Table 1). 


Determinants of meat consumption in regions: basic econometric model


Determinants of meat consumption was identified using an econometric model that allows the explanatory variables explaining meat consumption differences in regional states of  Ethiopia (i) at period t and takes the form:

Yit  = Xit β + εit                                   


Y is regional per capita meat consumption,
β is a vector of regression coefficients to be estimated;  
Xit is a matrix of explanatory variables displayed in Table 2; and  
εit is a vector of disturbances or random error terms.


Table 2 present summaries of statistics of the variable used to analyze the determinants. There is greater variability in per capita meat and percentage of urban population among regions as indicated by high standard deviation. The level of per capita meat is by far lower than the world average of around 40 kg/capita/year (FAO 2005).

Table 2.  Summary of variables used for statistical analysis



Standard Deviation



Per capita meat, kg/yr





Meat Price, ETH Birr/kg





Log Income, Birr/person





Log Cereal Production, t





Urban Population, %





The result of Cross-sectional time-series FGLS regression analysis in Table 3 shows regional per capita household income and percentage of urban population had significant effect on level of per capita meat consumption among regions in Ethiopia.

Table 3.  Coefficients of Generalized Least Squares of determinants of Meat consumption among regions in Ethiopia



Standard Error

Z score


Meat Price










Urban Population, %





Cereal Production










a Number of observations is 20

These two variables are highly significant because most of meat consumption occurs in the upper income households in urban areas. Therefore, as urbanization proceeds in the country, the demand for meat will increase to certain level and income growth will take consumption to a higher level if accompanied by other changes as discussed earlier. The result of regional consumption agrees with consumption pattern observed in Figure 3.


The price of meat and cereal production did not significantly affect consumption of meat because, meat is primarily consumed in upper income households in urban areas (see Figure 3) and generally, level of consumption was very low to be affected by price. The amount of cereal feed has also been insignificant to affect meat production and per capita meat consumption.


FAO versus CSA meat consumption data


FAO’s estimate of Ethiopian per capita meat was 8 kg/year in 2004, whereas CSA’s HICE survey estimate was about 3.5 kg/year in 2004, i.e., a 100% difference (Table 4). This difference is significant when considered on the backdrop of the size of Ethiopian population.

Table 4.  Comparison of per capita meat consumption estimate by FAO and CSA


Data Source












Although, comprehensive understanding of the discrepancy is difficult, partial explanation could be found from FAO’s estimation methodology for Ethiopia.  FAO (2004) reported a relatively constant slaughter rate and livestock yield (carcass weight per live animal). However, there is little information on how these indices were calculated. Since, livestock numbers reported on FAOSTAT were similar with CSA’s livestock survey data, at least since 1996, the year CSA started the first livestock survey, the difference must have resulted from over estimation of the livestock utilization indices. The three consecutive HICE surveys of CSA have resulted in consistently lower per capita meat consumption compared to FAO. Therefore, it is highly unlikely that Ethiopian per capita meat was accurately estimated by FAO and must have been different from what has been reported. This assertion could not imply that the average meat consumption in Ethiopia that was reported by FAO was high in absolute terms. However, it illuminates the extent of the level of Ethiopian meat consumption being lower than what has been  commonly reported and expected.


Conclusions and policy recommendations 








Aklilu Y 2002 An Audit of the Livestock Marketing Status in Kenya, Ethiopia and Sudan. Volume II, Issues and Proposed Measures,%20Ethiopia%20and%20sudan%20volume2.pdf


Alexandratos N 1995 The Outlook for world Food and Agriculture to Year 2010. Islam, N 1995 (editor) Population and Food in the Early Twenty-First Century: Meeting Future Food Demand of an Increasing Population. International Food Policy Research Institute


Avery A 2004 Red Meat and Poultry Production and Consumption in Ethiopia and Distribution in Addis Ababa. Borlaug-Ruan World Food Prize Intern, International Livestock Research Institute. Addis Ababa, Ethiopia


Ayele Z and Peacock C 2003 Improving Access to and Consumption of Animal Source Foods in Rural Households: The Experiences of a Women-Focused Goat Development Program in the Highlands of Ethiopia. The Journal of Nutrition, American Society for Nutritional Sciences November 2003; Volume 133: 3981S - 3986S


Delgado C 2003 Rising Consumption of Meat and Milk in Developing Countries has Created a New Food Revolution. The Journal of Nutrition, American Society for Nutritional Science November 2003; Volume 133: 3907S - 3910S


FAO 2004 Livestock Sector Brief: Ethiopia. Food and Agriculture Organization of the United Nations. Livestock Information, Sector Analysis and Policy Branch. AGAL


FAO 2005 Supply Utilization Accounts and Food Balances.


Guru M, Abebe G, Goetsch A, Hundessa F, Ebro A and Shelima B 2008 On-farm performance of Arsi-Bale goats in Ethiopia receiving different concentrate supplements. Livestock Research for Rural Development, Volume 20, Article 12


HICE 1996 Central Statistical Authority Revised report on the 1995/96 Household income, consumption and expenditure Survey, 204 Statistical Bulletin


HICE 2000 Central Statistical Authority Revised report on the 1995/96 Household income, consumption and expenditure Survey, 258 Statistical Bulletin. Volume I Analytical report


HICE 2004 Central Statistical Authority Revised report on the 1995/96 Household income, consumption and expenditure Survey, 394 Statistical Bulletin. Volume II Statistical report


Jabbar M, Negash A and Gidyelew T 2007 Geographic distribution of cattle and shoats population and their market supply sheds in Ethiopia. Discussion Paper number 2, Improving Market Opportunities. ILRI (International Livestock Research Institute)


MHLW 2000 National Nutrition Survey, Ministry of Health, Labour and Welfare, Government of Japan.


Nell A 2006 Quick scan of the livestock and meat sector in Ethiopia Issues and opportunities. Wageningen International, the Netherlands


Ogle B 1990 Suggestions for intensive livestock-based smallholder systems in semi-arid areas of Tanzania. Livestock Research for Rural Development, Volume 2, Article1


Solomon A, Workalemahu A, Jabbar M A, Ahmed M M and Hurissa B 2003 Livestock marketing in Ethiopia: A review of structure, performance and development initiatives Socio-economic and Policy Research Working Paper 52


Speedy A 2003 Global Production and Consumption of Animal Source Foods. Animal Source Foods to Improve Micronutrient Nutrition in Developing Countries. American Society for Nutritional Sciences November 2003; Volume 133: 4048S - 4053S


Steinfeld H 2002 Economic Constraints on Production and Consumption of Animal Source Foods for Nutrition in Developing Countries. Livestock Information Sector Analysis and Policy Branch, Food and Agriculture Organization of the United Nations, Rome, Italy. The Journal of Nutrition November 2003; 133: 4054S - 4061S,


UN 2005 World Population Prospect.

Received 17 June 2009; Accepted 20 June 2009; Published 1 September 2009

Go to top