Working Paper

No. 3/2004

 

 

Food insecurity and Poverty in Ethiopia:

Evidence and Lessons from Wollo

 

 

 

EEA/Ethiopian Economic Policy Research Institute (EEPRI

 

 

 

 

P.O. Box 34282

Tel. 251-1-234363

Fax 251-1-234362

Addis Ababa, Ethiopia

E-mail: eea@telecom.net.et
 

 

 

 

Food insecurity and Poverty in Ethiopia:

Evidence and Lessons from Wollo[1]

 

 

 

 

 

 

 

By

Samuel Gebre-Selassie (Ph.D)
EEA/EEPRI

 

 

 

 

 

Working Paper No. 3/2004

EEA/Ethiopian Economic Policy Research Institute

December 2004

Addis Ababa
Acknowledgment

The author would like to thank all research staff of the EEA/Ethiopian Economic Policy Research Institute for their comments. Special thanks go to the director of the EEA/EEPRI, Dr. Assefa Admassie, for his insightful review of the draft document.


Content

1. Background. 5

2. Objective of the study. 6

3. Data and methods of data analysis. 6

4. Brief review of poverty and food security in Ethiopia. 7

5. Major causes of poverty and poverty indicator 8

5.1 Causes of poverty and food insecurity. 8

5.1.1 Shortage of productive resources. 9

5.1.2 Low access to non-farm and off-farm activities. 10

5.1.3 Low level of productivity. 10

5.1.4 Low income and food consumption. 11

5.2 Livestock as indictor of poverty trend. 12

6. Agricultural production and food security. 14

6.1 Analysis at household level 14

6.2 Analysis at community level 16

6.3 Resettlement program – a speedy pathway to food security. 17

7. Determinants of grain production and food insecurity. 18

7.1 Determinants of grain production. 18

7.2 Determinants of food security 22

8. Summary and recommendations. 23

 

List of Tables

Table 1: Average ownership of productive resources in South and North Wello. 9

Table 2. Income from non-farm and off-farm activities in the study areas. 10

Table 3. Land productivity in cereals production during normal year 11

Table 4: Estimated level of poverty in the study areas. 12

Table 5: Change in the size of livestock in the past seven years. 13

Table 6: Production and consumption of grain (Food balance sheet) in the study areas. 15

Table 7: Food security situation during normal and drought years 16

Table 8: The degree of food security (aid requirement) in the study areas. 17

Table 9: Summary Statistics of variables estimated in the regression model 20

Table 10: Tobit estimates. 20

Table 11: Logit estimates. 23

 

Annexes

Annex 1: Distribution of farmland among sampled farmers 25

Annex 2: Drought affected population. 25

Annex 3: Crop production in the study areas 26

Annex 4: Partial correlation of level of education of HH head and ownership of productive                  
       resources (Land and oxen ownership) 27

Annex 5:Partial correlation of land, labor and oxen owned by sampled households 27

 


1. Background

Total grain production in Ethiopia has improved in recent years. However, the increase is so small that it could not affect the level of per capita production and consumption. The country produced, for instance, only 162 kg and 148 kg of grains in 2000/01 and 2001/02 mehr season, respectively, on per capita basis.  Even the 2003/04 record level production of 117.5 million quintals of grains is only a modest improvement when looked in terms of per capita production (i.e. 165.3 kg./person) which, together with the level of per capita income, indicates the progress or the lack of progress towards realizing the objective of food security.    In general, average per capita grain production fluctuates between 106 and 165 kilograms in recent years which on average indicate a deficit of 60 to 100 kilograms of grain per person[2].

 

Likewise, the number of people vulnerable to drought has been increasing. The drought of 2002/03 has shown that people in some parts of the country where drought or drought induced problems were manageable at local or household levels increasingly need food aid to prevent widespread famine. At the same time, people in areas where transitory or weather induced food insecurity has been the predominant problem increasingly suffer from chronic food insecurity[3], which is related more to poverty rather than to temporary shocks.  This worsening trend is manifested by government statistics on the number of people affected by drought.  For example, only about 1.5 million or not more than 5% of the total population of the country suffered from drought induced food insecurity problem during the imperial regime in the 1960’s or early 1970’s; by mid 1984, the figure increased to 7 million or 17.4% of the total population. In 2003, it increased to 14.5 million or 22% of the total population that was estimated at 69 million (see Annex 2). About five million of these people have been suffering from chronic food insecurity.

 

This trend has led to a higher level of food aid to save life and narrow the ever growing gap between food supply and demand. However, the non-stopping food aid program has also become increasingly controversial. In some parts of the country where relief has been provided for more than three decades, the problem of dependency syndrome has reached its highest level and emerged as one of the problems that hinder personal or collective motivations to challenge poverty and to find ways to get out of the prevailing despair situation. Despite saving life and preventing widespread malnutrition, some say the non-stopping relief program has also contributed for peasants to give up hope. The current debate on the role of aid and how to relate it to development programs through safety-net programs and the on-going effort to find the best alternative way to use relief resources for development has partly emanated from the growing dependency syndrome and other problems related to relief programs which have been carried out for so long in some areas of Ethiopia like Wollo.

                     

The justification for the continuous food aid in Ethiopia emanates partly from the public works and employments that have been created using relief resources. However, many decades of aid related development activities such as food-for-work programs in chronic food insecure areas have also contributed too little to reduce the process of environmental degradation and the rehabilitation of natural resources including agricultural lands, soil and forests which are the basis for a sustainable agricultural system.

 

Food aid has also become controversial even at international level. Its supporters claim that it plays a major role in feeding the poor, prevents severe food insecurity and saves lives when emergencies arise. Its delivery is justified by the view that it is a valuable macro-economic resource to fill the gap between the demand and the local supply of food and to improve the balance of payments and budgetary deficit. However, an increasing number of critics argue that food aid has contributed to dependency at the institutional and household levels. They point to the disincentive effects of food aid on agricultural innovation, intensification and diversification (Masefield, 1996).

 

2. Objective of the study

The study tries to assess the food security situation of farmers in South and North Wollo of the Amhara National Regional State (ANRS). It describes and analyzes the level and dynamism of food insecurity and poverty, and tries to understand what explains these problems (mainly at household level) and recommends what should be done to reduce food insecurity and poverty in the study area.

 

 

3.      Data and methods of data analysis

Data collected for an impact assessment study of a cash for relief (CfR) project were used in the study[4]. The CfR program emerged from the experience of using food both as relief resource and incentive to undertake environmental rehabilitation programs and other public works such as feeder road and water development. The CfR project was carried out in 2002 by SC-UK in six Weredas of North and South Wollo Zones, including Legeambo and Mekete Weredas, where the CfR was started in 2000 as a pilot program. 

 

The sample size considered in this study relates to 646 households residing in 12 peasant associations of Legeambo and Meket Weredas of South and North Wollo. About 75% or 495 sampled households were project participants (i.e. beneficiaries of the CfR program) whereas 151 households were non-beneficiaries of the program.

 

Descriptive methods were employed to explain the level and extent of food insecurity and poverty in the area, while regression was used to identify the contributing factors to food insecurity.

 


4. Brief review of poverty and food security in Ethiopia

 

A significant proportion of the Ethiopian population has been suffering from poverty and poverty related problems like malnutrition and disease for a very long period of time. The proportion of people who are absolutely poor (unable to meet their basic needs) during the year 1999/00 was 44.2 percent (about 30 million people). In urban areas, the proportion of poor people as defined by the national poverty line was 47%, while it was 33% in rural areas (MOFED, 2002). In addition to these, the progress so far achieved in reducing poverty is only marginal.  Consumption poverty measured by the head count ratio, for example, has witnessed only a 1.3 percentage point decline (2.9 percent) between 1995/96 and 1999/00 (MOFED, 2002).  Even this trend seems short-lived. The recent report by the Ethiopian Economic Association on the performance of the Ethiopian economy, for instance, indicates that real per capita income was on average falling both in rural and urban areas in the last three years. Real GDP growth averaged 1.7% during the 2000/01-2002/03 period. This translates into a 1.2% decline in per capita income (EEA, 2004). Then in the following fiscal year, 2003/04, the performance of the economy has improved. The government and the IMF reported that the Ethiopian economy grew by 11.6% during the 2003/04 year, which implies a growth rate of  8.7% in per capita terms when compared to the preceding year, when per capita income declined by 6.6% (NBE, 2004).  

 

In addition to GDP and population growth, poverty level is affected by the gap in income inequality or the distribution of newly created wealth among citizens. According to the government the increased gap in income inequality continues to contribute for a high and steady level of poverty in Ethiopia. MOFED, for instance, claims that the degree to which economic growth affected poverty in Ethiopia was counteracted by a rise in income inequality (MOFED, 2002). Bourguignon (2004) also reported that growing income inequality offset the favorable effects of growth in Ethiopia. Growth could have reduced the poverty headcount by some 31 percent between 1981–95. Yet, because of changes in the distribution that contributed to a 37 percent increase in poverty, the final effect has been a net increase in poverty of 6 percent (Bourguignon, 2004). In general, the level of poverty in Ethiopia has been stagnating or deteriorating slightly because of low and unstable economic growth, high population growth[5] and growing income inequality. There are different explanations for the question why the economy has failed to show a steady and dependable trend in terms of economic growth and poverty reduction.       

 

The major factor that hinders a sustained economic growth and poverty reduction in Ethiopia is the poor performance of the agricultural sector which provides employment, income and food for the majority of the population. Rain-fed agriculture has continuously suffered from high population pressure, uncertain weather conditions and low and declining labor productivity. Poverty has also induced the cultivation of marginal and degraded lands that contributes to low land and labor productivity. On the other hand, the non-farm sector has been unable to expand at a faster rate to provide employment and income for excess labor existing in the agricultural sector. Thus, the level of food security and poverty in Ethiopia is closely related to, and determined principally by the performance of the agricultural sector. However, this should not imply that development interventions should be restricted to the agricultural sector. Rather, interventions in the non-farm sectors that facilitate the growth of the farm sector should get the weight they deserve.

Despite the relative high share of agriculture in aggregate output (GDP), the per-capita value added in the agricultural sector has been on the decline since the 1960’s. In other words, the additional labor force that comes into the sector does not have a positive value added to output (Daniel, 2003). Getahun’s (2003) estimates have indicated that unless the declining trend of the sector achievement is arrested and reversed, the poverty situation in the country will rapidly aggravate as a result of which, by 2015, close to about two-third of the population will be in absolute poverty.  At the same time, it was shown that with little improvement in agricultural productivity, poverty could be reduced substantially by 2015 (Getahun, 2003). The government needs to find ways of raising the productivity of rural labor in general, and farm labor in particular (MOFED, 2002).

 

Food insecurity which currently affects over 40% of the population is the major challenge for Ethiopia. A combination of different factors has contributed to this growing problem.

Drought, high population growth and environmental degradation on the one hand, and technological, policy and institutional factors that lead to declining labor productivity, on the other hand, have been the principal causes of poverty and food insecurity especially in rural areas. The government has continued with its effort of tackling these chronic problems. However, to-date achievement has been modest and insignificant. Even though various initiatives and programs have emerged at different times since the change of government in 1991, the strategic direction to achieve food security and economic development has been the Agriculture-Development-Led-Industrialization strategy which was developed in the mid-1990s. The most recent initiatives by the government are the new Coalition for Food Security and the Ethiopia Strategy Support Program (ESSP).  Both the Coalition and ESSP aimed to bring the commercialization of Ethiopian smallholder agriculture, improve farmers’ productivity and achieve other objectives stated in the ADLI strategy. The ESSP which was developed by IFPRI and EDRI in response to the direct request by the government seems to be a broader strategy that aims to generate policy research results and knowledge that will inform policy makers on shortcomings and opportunities on existing rural development strategies (IFPRI and EDRI, 2004).

 

5.  Major causes of poverty and poverty indicator

 

5.1 Causes of poverty and food insecurity

 

The food security situation in the study areas has not improved in recent years. According to the DPPC office of the Amhara region, for instance, about 1,306,976 people or 35% of the total population of North and South Wollo Zones received food aid every year between 1997 and 2001. A recent study carried out in Northeastern Highlands of the Amhara region by IDS and SC-UK also pointed out the growing number of rural households that appeared to be unable to make ends meet, even in good rainfall years (IDS and SC-UK, 2002).

 

Despite a considerable improvement in the amount of food aid and development assistance in recent years, so far its impact on the level of food security is very limited. The massive food aid that has been coming to the country has been successful in preventing famine and widespread malnutrition. The second objective of food aid is to prevent asset depletion of households participating in relief programs and contribute to the rehabilitation of natural resources through relief-resources-sponsored environmental rehabilitation programs (e.g. food-for-work program). However, the program has not been successful at least in the study areas where aid and aid related environmental programs have been carried out for the last 3 or more decades.    

Hunger and poverty are closely related. While the lack of sufficient income to purchase food is clearly a major factor causing household food insecurity, hunger itself contributes to poverty by lowering labor productivity, reducing resistance to disease and depressing educational achievements (FAO, 2001, cited by Getahun, 2003). Poverty in the study area exhibits itself in many forms but mainly in terms of lack of access to sufficient food and high vulnerability even to minor weather related shocks. Some of the main causes of poverty include lack of productive resources, low productivity and low income.

 

5.1.1 Shortage of productive resources

 

As non-farm activities have an insignificant role in the local economy, farming and farm resources have important implications on the level of food security and poverty. Therefore, size of farmland, labor and livestock and fertility of soil have important implications on households’ food security status and poverty level, especially during normal agricultural years. During drought years, livestock, a major asset that can be easily liquidated, is more important in terms of implying better access to food.

 

Table 1: Average ownership of productive resources in South and North Wello, Ethiopia
_____________________________________________________________________

Variable            | Obs     Mean      Std. Dev.       Min       Max

--------------------+-------------------------------------------------
Land (ha./HH)       | 644     0.68       0.30          0.13      2.00 

Per capita Land     | 577     0.23       0.19          0.02      1.75 

LABOR (ME)          | 644     1.84       1.30          0.40      6.10 

HH size (AE)        | 644     3.01       1.52          0.60      8.98 

Livestock (TLU)     | 646     1.33       1.41          0.00     10.53 

OX (No.)            | 443     0.47       0.65          0.00      4.00 

Percent of HHs having no ox =  61%

 

Of the sampled households, 94% owned one hectare or less, while the average farm size was 0.68 hectare. On per capita basis, average farm size was as low as 0.23 ha. About 61% of the farmers reported that they have no ox, while average ox ownership was only 0.47. The problem is not only the shortage of these resources but also the technologies used to convert these inputs into farm outputs as evidenced by the low input-output ratio.

 

In general, available farm resources are too small to provide adequate food and income for the average household. A study conducted in the farming system of the study area, for example, reported that households that owned less than 0.5 ha are unable to meet basic needs and are labeled as destitute households. Households with farm sizes in the range of 0.5 to 1 ha and above 1 hectare are classified as vulnerable and viable, respectively (IDS and SC-UK, 2002).  Based on this single indicator, data collected for this study indicate that about 14.9% of sampled households are destitute, 79% are vulnerable and 6.1% are poor farmers. Even though measuring the level of food insecurity of different households based on a single factor (land) has its own limitation[6], it can roughly indicate the extent of poverty given the current level of productivity and household consumption requirement.

 

 

 

 

5.1.2 Low access to non-farm and off-farm activities

 

Population pressure and reduced farm size and soil fertility has forced farmers in the study area to work harder without being able to maintain their income or standard of living, measured in terms of food consumption.   However, sustainable livelihood has not only been threatened by reduced farm size and productivity, but also by undeveloped non-farm economy in the area and the environs. Low access to non-farm and off-farm activities is, therefore, another reason why people in the study area are very poor and food insecure. Almost every household looks for non-farm employment to supplement family food requirements. Data collected from farmers, however, indicate that not more than 50% of sampled households could get employment opportunities in any year, including a year when food-aid-induced employment opportunities are available. During the survey year, about 44% of households had access to off-farm or non-farm employment. The average annual gross income from these activities was Birr 550 and 239 for valid cases (i.e. households having access to non farm activities) and an average household, respectively. This income from non-farm and off-farm activities constitutes only 11.3% of the total household income (see Table 2).

Table 2. Income from non-farm and off-farm activities in Legeambo and Mekete  
             Weredas
, during the survey year

 

Legeambo

Mekete

Households having access to non-farm income sources

180 (46%)

103 (41%)

Mean non-farm income (Birr/annum)

for households with access                                                                                                 

624

475

for all sampled households

284.40

194.90

Estimated mean gross farm income (Birr/annum)

1897.7

2069.0

Share of non- and off-farm income in total household income

for households with access                                                                                                

24.7%

18.7%                                                                                           

for all sampled households

13%

8.6%

 

5.1.3        Low level of productivity

 

As mentioned earlier, the low productivity of land and labor could explain partly why people in the area are very poor and food insecure. Data collected from sampled households indicate that about 20% of the farmers produce only 3 or less quintals of wheat-equivalent grain on a hectare of land during a year they consider normal. More than 60% of the sampled farmers reported that they produce 9 quintals or less per hectare. This level of productivity is very low even compared to the national average. 

 

While land degradation and declining soil fertility are the direct causes of the low return to labor and land, extreme poverty which induces cultivation of marginal and degraded land plays a part indirectly. It contributes both for low land productivity and unsustainable farming system practiced by peasants in the study areas. This situation coupled with high population pressure creates a big challenge to government and non-governmental organizations working to improve the level of food security in the study areas. Any development intervention demands an intergraded program that encompasses environmental rehabilitation, voluntary migration, family planning, resettlement and the development of the non-farm economy. Moreover, institutional support in terms of agricultural marketing and extension should also be strengthened to improve labor productivity and farm income.  

Table 3. Land productivity in cereals production during normal year
              (quintal of wheat equivalent[7] per hectare of cultivated land)
____________________________________________________

Productivity|      ________Households_______________

 (qt./ha)   |       Freq.      Percent        Cum.

------------+---------------------------------------

< 3         |         128       19.97        19.97

3.00–6      |         150       23.40        43.37

6.01–9      |         109       17.00        60.37

9.01-12     |          87       13.57        73.94

12.01-16    |          62        9.67        83.61

16.01-20    |          39        6.08        89.69

>20         |          66       10.31       100.00

------------+----------------------------------------

   Total    |          641      100.00_______________                     

5.1.4 Low income and food consumption

 

Poverty is mainly manifested in terms of hunger, malnutrition and poor access to social services like health and education. All of these manifestations of poverty, however, should be measured in terms of a single indicator to compare the level of poverty in different communities and to monitor the progress made to eradicate poverty. This indicator is called the poverty line which is used as a threshold level of per capita income or consumption, below which an individual is considered to be poor and unable to satisfy the minimum food (energy) and non-food consumption requirements. This minimum income level is usually computed based on an income level that is agreed on (by the government of the FDRE) as sufficient for minimum food and non-food consumption expenditure for an adult person.

 

Real per capita consumption (food and non-food consumptions) expenditure for rural people was Birr 995 in 1995/96, which was taken by the government as the poverty line for rural Ethiopia (MOFED, 2002). As price varies from time to time and affects the basket of commodities required to maintain the minimum level of consumption, this level of income should be adjusted for price inflation. The level of poverty line is, therefore, adjusted to account for the effect of price inflation between 1995 and 2001 and reflects the situation in 2001, the latest year for which data on a moving average consumer price index for rural Ethiopia was available. Based on this inflation rate, rural poverty line for the year 2001 is calculated as Birr 1022.9[8].  There is a big disparity between this level of poverty line and farm households’ actual income. The average household, for instance, needs an additional income of Birr 1140 and 1057 in Legeambo and Meket Weredas, respectively, to fulfill the minimum food and non-food consumption requirement of its members. Households’ income is in general very low and it may have a negative effect on labor productivity by reducing resistance to diseases and capacity for work.       

 

Table 4: Estimated level of poverty in south and north Wello, Ethiopia[9]

 

Good (Normal) year

Legeambo

Meket

Average HH size in AEa

3.58

3.52

Average consumption expenditure[10]

Per capita food and non-food consumption expenditure

(for year 1995/96 ) (Br.)

995

995

12 months moving average general price index in 2001 (1996=100)

102.8

102.8

Per capita food and non-food expenditure after adjusting inflation in rural prices (for the year 2001)

1022.9

1022.9

HH minimum food and non-food consumption expenditure

In 2001

3662

3601

Average Production
 and incomea

Income

from food crops production

Production (qt.)

9.03

10.64

Weighted average price (Br./qt. food crop)[11]

190

175

Estimated annual income from
crop production (Birr)

1715.7

1862.0

Average estimated income from off/non-farm activities (Br./annum)[12]

624.0

475.0

Average estimated income from livestock sales (Br./annum)

182.0

207.0

Total estimated income (Br./annum)

2521.7

2544.0

Additional income required to meet minimum expenditure (Birr/annum/household)

1140.3

1057.0

Average household capacity to meet minimum consumption requirement (%)

68.9

70.6

Estimated percent of population living below poverty line

62.1%

58.7%

a Weighted average for beneficiary and non-beneficiary households.

 

The proportion of the population living below the poverty line under normal agricultural year is estimated at 62.1% in Legeambo and 58.7% in Meket Weredas. This is very high even compared to the national average.  When compared to the 1995/96 national data, the percentage of population living below the poverty line is over 17% in Legeambo and 13% in Meket Weredas.   

 

5.2 Livestock as indictor of poverty trend  

 

Poverty is a major economic and social problem in Ethiopia and affects more than 40 percent of the population. Even though it is a rural and urban phenomenon, it is closely related to the performance of the rural and agricultural sectors. The performance of Ethiopia’s rain-fed agriculture has not been encouraging. This, coupled with low employment opportunity in the non-farm sector, high population pressure and the erratic nature of rainfall put rural livelihood at risk in the study areas as in most parts of the country.  The production environment has led farmers to look for assets that could play a buffer role in time of drought and towards the end of every crop year when most households face food shortage problems.

In the study areas, livestock has served as buffer against hard times. Households who have opportunity to save usually keep their money in the form of livestock. They produce or buy livestock (particularly small ruminants) to sell and buy food grains during years of drought or to fill the gap in food requirements towards the end of the agricultural year when they fall short of food. Therefore, changes in the size of livestock of an average household could be taken as a proxy to indicate the dynamics of poverty and vulnerability to drought[13].

Table 5: Change in the size of livestock in the past seven years

 

Legeambo

Meket

HHs reported change in their livestock size

72.8%

61.5%

Reported change in
livestock  size during the
past seven years

Increased (TLU/valid case)

0.27

0.11

Decreased (TLU/valid case)

3.06

2.30

Estimated net change
during the past 7 years

TLU/valid cases

-2.79

-2.19

TLU/sampled HHs

-1.95

-1.35

Ox-equivalent/sampled HHs

-2.40

-1.69

Estimated net change
per year

TLU/valid cases

-0.28

-0.24

TLU/sampled HHs

-0.20

-0.15

Ox-equivalent/sampled HHs

-0.25

-0.19

Livestock size of an average HH in 2003 as % of what it owned

five years ago

54.6%  

(average of the two Weredas)

N (Number of sampled households)

 395

     251

 

Accordingly, farmers were asked to indicate the size and composition of their livestock during the survey year and to compare it with what they owned some seven years ago. Table 6 shows the change in the size of livestock during this period. Data collected from farmers in Legeambo Wereda indicate that 395 sampled households lost on average about 79 TLU (in terms of oxen-equivalent, about 99 oxen[14]) of different kinds of livestock every year during 1996 to 2003.  In other words, during the survey year, on average four farmers own one ox (or one ox-equivalent livestock) less than what they owned some seven years ago. In Meket Wereda, the loss is a bit lower. About 251 farmers lost various livestock that is equivalent to 38 TLU (i.e. equivalent to 47 oxen) every year in the past seven years. This is equivalent to a loss of 1 ox among every five farmers. In general, about 68.5% of households residing in the two Weredas reported change in the size of their livestock. And when compared to what they owned five years ago (in 1998), the average size of livestock of a household during the survey year (2003) is only 54.6%. In conclusion, the survey result indicates the worsening trend of poverty in the study area. Other studies also indicate a similar trend in the size of livestock of households owned in the study areas[15].

                                              

                                                                                                                   

 

 

 

6. Agricultural production and food security

As the non-farm economy is not developed, poverty and food security are closely related to the performance of the agricultural sector. Any analysis of poverty and food security, therefore, necessitates a closer look at the performance of the farm both at household and local economy (community) levels.  

 

6.1 Analysis at household level

 

Different types of food crops, notably cereals and pulses, are produced in the study areas. Compared to farmers in Meket Weredea, crop production is, however, less diversified in Legeambo Wereda where crops like Teff, Maize and Sorghum are uncommon. Grain production does not vary much between beneficiary and non-beneficiary households in Legeambo Wereda where an average household produces about 9.3 quintals of various food crops. However, in Meket Wereda, CfR beneficiary households produce on average 5 quintals less than what non-beneficiary households produce (Table 6). Data collected from farmers also indicate that in drought years, production could decline on average by as much as 62% in Legeambo Wereda and by about 58% in Meket Wereda.

 

The level of food security at household level was estimated considering this production data, income from non-farm activities[16] or remittances, and households’     minimum food and non-food consumption requirement. The sampled households were categorized into four groups based on their location and participation in the relief (CfR) program. The food balance sheet (i.e. the level of food security) was computed for average households from every group.

The average household that participated in the relief program in both Weredas, and non-beneficiary household of Legeambo Wereda could not satisfy their food requirement even in normal (good) agricultural year. They suffer from food shortage for a period of 1 to 2 months in a year that is considered normal by them. Only the average non-beneficiary household from Meket Wereda could feed his/her household and have some surplus that could supply for 4 to 5 extra months.

 

On the other hand, none of the average households could feed themselves in drought years. They face food shortage for a period of 6 to 8 months. If one assumes that households could get cash for their non-food requirement by working outside their farm or totally abandon their expenditure on non-food consumption during drought period, the period of food shortage could narrow marginally to 5 to 6 months (in both Weredas).

 


Table 6: Production and consumption (Food balance sheet) of an average household in the study area        

Wereda

 

Normal agricultural year

Drought year

CfR participant

Non-participant

CfR Participant

Non-participant

Legambo

Annual food (energy) production in wheat equivalent (qt.)

8.94

9.32

3.5

3.6

Annual food (energy) requirement in wheat equivalent (qt.)

8.76

10.21

8.76

10.21

Annual food (energy) balance from own production 
(before expenditures for non-food purposes) 

Shortage/surplus in wheat equivalent

0.18 qt.

-0.95 qt.

-5.26 qt.

-6.67 qt.

Shortage/surplus in months

0.25

-1.11

-7.17

-7.76

Annual cash expenditures (food and non-food purposes)[17]

Birr/annum

931

931

372.4

372.4

In Wheat equivalent (qt./annum)[18]

5.6

5.6

2.3

2.3

Estimated cash income from off-farm/non-farm business (Birr/annum)[19]

591

(3.58 qt. wheat)

734

(4.49 qt. wheat)

354.6
(2.15 qt. wheat)

440.4
(2.7 qt. wheat)

Estimated income from livestock sales (Quintal of wheat equivalent/HH) [20]

0.40

1.3

0.40

1.3

Annual food (energy) balance from own production
and other incomes (including other expenditures)

Shortage/surplus in wheat equivalent (qt.)

-1.44

-0.7

-5.01

-6.21

Shortage/surplus in months

-1.96

-0.81

-6.9

-5.83

N

303

92

303

92

Meket

Annual food (energy) production in wheat equivalent (qt.)

9.62

14.04

3.8

4.16

Annual food (energy) requirement in wheat equivalent (qt.)

7.79

10.15

7.79

10.15

Annual food (energy) balance from own production
(before expenditures for non-food purposes) 

Shortage/surplus in wheat equivalent (qt.)

1.83 qt.

3.89 qt.

-3.99 qt.

-6.01 qt.

Shortage/surplus in months

2.8

4.6

-6.23

-7.01

Annual cash expenditures (for food & non-food purposes)

Birr/annum

931

931

372.4

372.4

In Wheat equivalent (qt./annum))

5.6

5.6

2.3

2.3

Estimated cash income from off-farm/non-farm business (Birr/annum)

421

(2.6 qt. wheat)

654
(3.96 qt. wheat)

252.6

(1.53 qt. wheat)

392.4

(2.38 qt. wheat)

Estimated income from livestock sales (Quintal of wheat equivalent/HH)

0.48

1.6

0.48

1.6

Annual food (energy) balance from own production
and other incomes (including other expenditures)

Shortage/surplus in wheat equivalent (qt.)

-0.69

3.85

-4.28

-4.31

Shortage/surplus in months

-1.06

4.62

-6.6

-5.1

N

193

58

193

58


6.2 Analysis at community level

 

The level of food security at community level (i.e. taking into account each and every household who participated in the survey which constitutes 5% of households in the community) is estimated based on information obtained from sampled farmers. Information on farm production and income including remittance and non-farm and off-farm income, and secondary data on food and cash requirement for non-food purposes are collected. Similar to the analysis made above for the average household, some adjustment is made on the data to account for the reduction of non-food expenditures and cash income during drought years. Accordingly, non-farm income and cash expenditures are assumed to decline by 40% during drought year (when compared to their level during normal years).

Table 7: Food security situation during normal and drought years
                                             (percent of sampled farmers)

 

Normal (good) year

Drought (Bad) year

All
sample

Legeambo

Meket

All
sample

 Legeambo

Meket

Food secured HHs[21]

54.66

53.16

57.03

19.72

19.75

19.68

Food insecured HHs

45.34

46.84

42.97

      80.28

80.25

80.32

                N

644

395

249

644

395

249

 

As shown in Tables 7 and 8 lack of access to adequate food which is the worst manifestation of poverty is a widespread phenomenon in the community.  During a year considered normal and drought, about 10% and 39% of the sampled households, respectively, satisfy only 25% or less of their food requirement. This implies that they require food aid for 9 or more months. The percentage of chronic food insecure people (or households which could not meet part of their family food requirement from own source during a normal year) is 45%, while during drought year about 80% of the people depend on food aid for a long period of time (of which, more than 60% need assistance for a period of 6 to 12 months). In addition to the 45% chronically food insecure households, 35% of the households will join this group during drought years.

 

In conclusion, both analyses made at household and community levels indicate that food insecurity is very rampant in the community that is hit by drought too frequently[22]. The study also indicates the need for long-term comprehensive development interventions. Government and non-government organizations working in the area should not limit themselves to food aid and relief-related development efforts which usually lack long-term financial and non-financial commitments.  In this regard, the resettlement program initiated recently by the government is a step in the right direction as it may ease the population pressure in the areas. However, settlement, if well planned and implemented, should be considered only as one of the ingredients that could help to achieve sustainable livelihood. It should also not be considered as an end by itself.

Table 8: The degree of food security (aid requirement) in the study areas

     Food security situation
(aid requirement)

Normal year

Drought year

Food security level

in percent

Food aid

requirement
in months

Number

Percent

Number

Percent

25% or less 

 9 or more months

61

9.47

253

39.29

26 – 50%

 6 up to 9 months

85

13.20

135

20.96

51 – 75%

 3 up to 6 months

88

13.66

82

12.73

76 – 100

 3 or less months

58

9.01

47

7.30

Satisfy 100%+ from
own sources

 (Food secured HHs)

 No need

352

54.6

127

19.7

Chronically food insecure households

45.4%

Households vulnerable to drought

34.9%

N

644

 

6.3 Resettlement program – a speedy pathway to food security   

 

Resettlement program of moving people who suffered from drought and shortage of productive land to relatively land and water (rain) abundant parts of the country is one of the recent initiatives taken by the government to relax the problem of shortage of farmland and fix chronic food insecurity problems in a short period of time. This is in addition to the effort that has been carried out since the early 1990s to improve the productivity of existing cultivated land using modern farm inputs like inorganic fertilizers and improved seeds. The resettlement program has been pursued both by the present (EPRDF-led) and the previous (Derg) governments. There are some differences and similarities between the two programs. The present resettlement program is said to be voluntary (i.e. based on the willingness of the settlers)[23].  The other difference between the previous and the present programs is that the latter is carried out within a given region which may minimize problems that may arise due to cultural and language differences between the recipient and sender communities. Moreover, the present program seems better in terms of government support and commitment as it has been carried out during a period of peace and increased international financial support.

 

There are also similarities between the present and the previous resettlement programs.  Like the 1984/85 resettlement program, the current program has been initiated following a big drought that threatened the lives of 14 million Ethiopians. This fact may indicate that the current program, as its predecessor, is a spontaneous attempt that emerged when the problem reached a critical and desperate level which left no option for other alternatives that could bear fruit gradually but on a sustainable manner.  On the other hand, many observers in aid/donor organizations and civic societies worry about the pace of the current program which plans to relocate about 2 million people within 3 to 5 years.  They fear that this condition, coupled with shortage of social and economic infrastructures in an environment tough for human settlement, could lead to social problems in the short-run and challenge the success of the program in the long-run.

 

The issue of sustainable development that maintains the balance between environment, agriculture and population is the other point that worries some observers of the resettlement program. They fear that relocating peasants with low knowledge of and experience in sustainable agriculture, low technologies and high fertility rate to environmentally fragile areas is a difficult task unless supported differently.  Still others say that Ethiopia’s potential to feed itself should primarily rely on increasing yield on existing farmlands and providing employment in the non-farm sectors that could strengthen the performance of the rural sector and its interface with the urban sectors which will ultimately lead to the commercialization of Ethiopian subsistence agriculture (EEA, 2004).  In conclusion, the efficacy of resettlement program to relax the constraint of nonrenewable resources like farmland is short-lived, unless it is effectively supported by other long-term interventions that enhance labor productivity in the farming sector and improved employment opportunities in the non-farm sectors.

 

 

7. Determinants of grain production and food insecurity

 

7.1  Determinants of grain production

 

A regression model is estimated to verify some of the results obtained through the descriptive analysis and establish whether basic economic relationships assumed to exist in a production process is supported by empirical evidence. Accordingly, a non-linear (Cobb-Douglas) production function was estimated using Tobit regression model to identify the determinants of farm output. Estimates derived from regression models and inferences made based on those estimates are valid under certain conditions – conditions that amount to the regression model being “well-specified”. One of the major factors that determine the specification of the model is the type of the production function adopted for regression. In this regard it is important to consider the peculiar production feature of the study area where many farm households produce positive amount, while there are also other households who produce nothing because of lack of productive resources, mainly fertile land[24].

 

Tobit model is an appropriate technique to run a regression of dependent variable that is essentially continuous over a range of values but also takes on zero (the threshold value) with positive probability over a number of explanatory variables.  The model fits  dependent variable on independent variables where the censoring values are fixed (Hog and Lunde, 2002). The shorter version of the functional form adopted and estimated is:

 

 

 

 

Y=f(X,Z) and in log form

LnYi = ∑βjlnXij + ∑αjlnZij + γ + µi

 

Where: Y    is quantity of output (cereal production)

             X   is a vector of physical inputs including land, labor and oxen

             Z   is a vector of other factors that affect the operation of a farmer like age, sex,
                  engagement in land rental market, off farm activities, etc.

γ and µ are constant and error terms, respectively.

 

The independent variables that are supposed to affect the level of production are broadly classified into four groups:

 

i.         Physical inputs – land, labor and ox. Capital has little contribution in subsistence mode of production. Measuring any capital stock used in the production process is also difficult. However, cultivated land and ox/oxen used in the production process could be used as a proxy measure for capital stock. Due to lack of data, fertilizer was not incorporated in the regression model.

 

The size of land cultivated and owned by sampled households was also entered into the
model. Labor (excluding children) was measured and entered into the model in terms of man-equivalent labor to reflect variations attributed to age and sex. The number of oxen owned by a household was considered to estimate the impact of available drought power on production.

 

ii.       The characteristics of farm manager (household head) – age, sex and level of education of household head were considered. While age was a continuous variable, sex and education level (being able to read and write) were entered into the model as dummy variables.

 

iii.      Factor (land) market – this refers to the existence and farmers’ financial ability to command scarce resource through the market. Specifically, the amount of land rented-in and land shared-in were considered.

 

iv.     Non-farm income – this indicates remittance and participation in off-farm activities (a dummy variable). The result could indicate the degree of linkage between farm and non-farm sectors.

 

Land productivity was not considered in the model due to lack of data and some irregularities in the collected data.

 

 

 

 

 

Table 9: Summary Statistics of variables estimated in the regression model

-------------------------------------------------------------------------

       Variable    | Obs   unit            Mean    Std. Dev.  Min     Max

-------------------+-----------------------------------------------------

Land               | 646   ha               0.68     0.30    0.13    2.00

Family labor       | 646   AE               1.84     1.30    0.40    6.10

HH members         | 646   No.              3.66     2.11    1.00   11.00

Children < 7 years | 646   No.              1.19     1.03    0.00    5.00

Children b/n 7&14  | 646   No.              1.08     1.04    0.00    6.00

Children b/n 14&50 | 646   No.              1.32     1.07    0.00    5.00

HH members > 50    | 646   No.              0.08     0.32    0.00    2.00

OX ownership       | 443   No.              0.47     0.65    0.00    4.00

Livestock own.     | 443   TLU              1.33     1.41    0.00   10.53

HH head Age        | 646   year            56.00    23.00   17.00   71.00

HH head Sex        | 646   percent male    74.00    43.98    0     100.00

HH head education  | 646   percent able

                           to read&write   50.50    39.12    0     100.00

HHs’ share-in land | 620   percent of HH   11.94    32.44    0     100.00    HHs’ rent-in land  | 632   percent of HH   18.98    39.25    0     100.00

HHs’ engage in off

farm activities    | 645   percent of HH   45.74    49.85    0     100.00

Amount of income

from off farm act. | 285   Birr           515.00  1141.14    0    6384.49        

HHs having 

remittance income  | 629   percent of HH   12.3     22.33    0     100.00

-------------------------------------------------------------------------

 

The result of the regression model of the Cobb-Douglas (CD) production function is reported in Table 10. Land and oxen which could also be used as proxies for capital stock are found important to explain existing variation in the level of production among sampled households. The coefficient for land is statistically significant at 1%.  However, the coefficient for oxen is relatively high but significant only at 5% level.  The contribution of labor is statistically insignificant to explain observed differences in output among sampled households.

 

Table 10: Determinants of Tobit estimates: Dependent variable: Output

 

Tobit estimates: Dependent variable - Output

Coefficient

t-value

Coefficient

t-value

Constant

15.82

3.87

12.82

3.05

ln (Land)

2.18

1.77*

1.93

1.55

ln (labor)

2.67

1.88

3.26

2.22**

ln (Oxen)

5.27

2.47**

3.63

1.56

Age

0.08

1.23

0.01

1.34

Sex

1.64

0.59

1.54

0.56

Head education

-0.79

1.99**

-0.52

1.29

Land rent-in

 

 

0.03

0.02

Land share-in

 

 

4.17

2.06**

Remittance

 

 

7.41

2.27**

Participation in off farm-activities

 

 

0.77

0.57

 

 

 

Number of obs  168                                                  LR chi2(6) = 18.12                                                  Prob > chi2= 0.0059

Log likelihood= -582.6723

Number of obs = 155                                                  LR chi2(11) = 22.95                                                  Prob > chi2= 0.0180

Log likelihood = -533.3385

*, **, *** indicate significance level at 10%, 5% and 1% respectively.

Age and sex of household head are also found statistically insignificant to explain observed variations in output. The neutrality of being male-headed or female-headed household on grain production reinforces the abovementioned result on labor and indicates the existence of surplus male labor in the study areas. The coefficient for household heads’ education level (i.e. by and large refer to the ability to read and write) is found negative and significant at 5% level. Although difficult to interpret from theoretical point of view, this is because of the negative correlation between literacy level and ownership of productive resources which are more important than the level of education to explain existing variation in output (see Annex 4).          

 

The impact of improved access to land on grain production is analyzed by incorporating the amount of land rented-in and shared-in into the model. These two forms of acquiring land, however, have different implications in terms of efficiency in resource utilization and production risk which is very common in the farming system of the study areas. Land sharing which involves cultivating someone’s land for sharing some portion of the final output also implies the sharing of production risk like crop failure. This type of production arrangement may lead to suboptimal utilization of resources when compared to renting land in which all cost and benefit associated to the use of someone’s land is exclusively related to a person that rented the land.

 

The regression model reveals the positive impact of improved access to land through land sharing arrangement on grain production (at community level). The coefficient is high and statistically significant at 5% level. Moreover, the coefficient for labor becomes significant when access to farmland is enhanced. This is not a surprising result as land and labor have a negative correlation (Annex 5). Land sharing compensates some of the negative impacts of mismatched ownership of productive resources on grain production. In other words, it substitutes to some degree the weakness of factor market in the study areas. Policy makers should deal with the problems that lead to disproportionate ownership of productive resources among households in the study areas and find ways that facilitate the functioning of factor market. Contrary to what was expected, land rented-in is found statistically insignificant to affect grain production.

 

The fact that land sharing becomes less preferable but more important in terms of improving grain production could also provide some empirical evidence on the behavior of peasants working in a risky production environment.  It signifies the existence of production risks that provide the economic rationale for land sharing.   However, the popularity of land rental market vis-ŕ-vis land sharing makes this argument a delicate issue. Why does the less popular alternative (land sharing) become more rewarding?  Possible explanations could include lack of financial capacity to rent land and purchase other resources simultaneously or those who engage in land sharing arrangement could be wealthier but more risk averse than those who rented land.  This question, however, should be explicitly answered by future researches.

 

Nevertheless the result indicates the necessity to provide institutional support to overcome existing problems of production risks and/or shortage of financial resources.   Whenever they design development projects, local government and non-government institutions should, therefore, be able to provide institutional support, for example, in the form of insurance against potential production risks.  The support could take different forms. For instance, rainfall insurance or fertilizer subsidy could encourage farmers to use fertilizers at the level recommended by extension institutions. Such intervention will make farmers risk takers and allow them to push their existing production function to a higher level or to the point where the law of diminishing returns setoff.    

 

Compared to households that did not get remittance, households that have external income in the form of remittance have a higher probability of producing more cereals.

The result, therefore, indicates the existence of linkage between farm and non-farm economy especially if remittances are originated from non-farm activities. Moreover, it could be considered as the positive contribution of labor that migrated outside their village.

 

7.2 Determinants of food security (household level analysis)

 

Despite high correlation between grain production and food security, they could not be affected by the same factors. A logit regression model was formulated, therefore, to know the determinants of food security and whether the two were affected by the same variables or not.

 

Accordingly, different variables were hypothesized to determine food security at household level. As independent variable household size and its composition, non-agricultural income in the form of remittance and off-farm income, age, sex and level of education of the household head and size of livestock owned by a household were entered into the model.  The level of own food production was not considered due to the problem of endogenity. Due to lack of appropriate variable, instrumental variable was also not used to substitute it. The dependent variable is the level of food security of a household which is expressed as a dummy variable where 0 represents households that could not fulfill the food requirement of their members and 1 otherwise. 

 

Regression result indicates that household size which is measured in terms of a standard consumption unit increases the chance of falling into food poverty. The result is positive and significant even after family members were grouped into different age groups to see the impact of dependency related to age differences[25].    

 

The probability of becoming food secure is not affected by households’ participation in off-farm activities and the amount of remittances they got. On the other hand, as the age of household head increases the probability of falling into food poverty increases. Even though the coefficient is very low, it is significant at 10% level. The result implies that the probability of falling into food poverty is higher for households headed by seniors (old age) than those headed by youngsters. Sex and level of education (being able to read and write) of the household head were found statistically insignificant to explain variations in the probability of being food insecure.     

 

Livestock size which is measured in terms of tropical livestock unit (TLU)[26] has no effect on the probability of being food secure. The coefficient is very small and also statistically insignificant. Even though this is contrary to what was expected, it has indicated that variation in the size of livestock (i.e. asset accumulation) during the survey year is too little to cause differences in food security among different households. 

 

Table 11: Logit estimates: Dependent variable: Food insecurity during normal year

 

 

 

 

Logit estimates

Coefficient

z-value

Constant

-0.851

2.37

Number of children less than 7 years old

0.540

3.55***

Number of persons between 7 - 14 years

0.428

2.93***

Number of persons between 15 – 50 years

0.521

3.26***

Number of persons above 50 years

0.417

0.80

Remittance

-0.310

-0.41

Off-farm income

-0.000

-0.19

Age of hh head

0.003

1.86*

Sex of hh head

-0.216

-0.59

Education of hh head

-0.000

-0.72

Size of livestock

-0.001

-0.78

 

Number of obs= 277                                                  LR chi2(11)= 74.88                                                  Prob > chi2=0.0000

Log likelihood = 

         - 150.889                  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

*, **, *** indicate significance level at 10%, 5% and 1% respectively

 

8. Summary and recommendations

 

The level of poverty in Ethiopia is not only becoming a major challenge to its future growth but also a threat to its stability and peace. Moreover, Ethiopia became synonymous with aid, particularly, food relief that has exceeded what is normally considered as a “relief” because food aid has become part and parcel of the rural culture and way of life in some parts of Ethiopia, leading to dependency syndrome. This is partly reinforced by the state and non-state institutions that have emerged since the early 1970 to deal with relief and relief related development programs.   

 

Result from the regression model shows that access to productive farm resources principally to land and ox significantly affects the level of grain production. On the other hand, labor is not a constraint given the current farm size.

 

The probability of becoming food insecure varies indirectly with household size. In addition to production enhancing interventions, the new food security strategy of the government should, therefore, incorporate family planning as one of its priority areas in the fight against food insecurity in the study areas.

 

Due to problems related to uneven distribution of factors of production and weak factor market, development programs should be sufficiently flexible to deal with problems of different farmers differently. For instance, farmers who lack sufficient land should be supported differently from farmers who encounter labor shortage like female-headed households. Land shortage could be compensated to some extent by providing incentive and support to farmers to grow high-value crops using small scale irrigation or through the use of land saving technologies like chemical fertilizers that will increase the return per unit area. On the other hand, labor saving technologies or alternative livelihoods are more important for labor scarce families.

 

Existing labor in the area is surplus given current farm size. Any additional labor, therefore, does not have a positive value added to output. On the other hand, model statistics indicate that households who received remittance have a higher probability of high grain production. As most of these remittances have originated from non-farm activities, there may be a positive return from policy interventions that encourage migration and alternative livelihood. Government should, therefore, provide various incentives for farmers in the area to encourage migration in search of alternative livelihoods. This could be implemented by providing long-term tenure security to farmers migrating to other areas in search of employment[27]. For farmers who own uneconomical holding or cultivate very fragile and degraded land, the government could go to the extent of even paying some money to encourage voluntary abandoning of farming occupation on such lands[28].  In parallel, concerted effort should be made to create labor-intensive employments in the non-farm sector (either in rural or urban areas) to reduce existing population pressure in rural areas and allow farming not to be a reservoir of unproductive labor that damages the sustainability of the system.           

 

 

 

 

 

 

 

 

 

 


Annex 1: Distribution of farmland among sampled farmers ______________________________________________________________

Land in ha.   | Freq.         Percent        Cum.

--------------+------------------------------------

     <=0.13   |         12        1.86        1.86

       0.19   |          1        0.16        2.02

       0.25   |         70       10.87       12.89

       0.38   |         13        2.02       14.91

       0.50   |        192       29.81       44.72

       0.63   |          7        1.09       45.81

       0.68   |         40        6.21       52.02

       0.75   |        150       23.29       75.31

       0.88   |          1        0.16       75.47

       1.00   |        119       18.48       93.94

       1.13   |          1        0.16       94.10

       1.25   |         20        3.11       97.20

       1.38   |          1        0.16       97.36

       1.50   |         14        2.17       99.53

       1.75   |          2        0.31       99.84

    >= 2.00   |          1        0.16      100.00

--------------+--------------------------------------

      Total   |         644      ___100________________        

Annex 2: Drought affected population

Year

Disaster/drought affected population (million)

Proportion affected

1980/81

2.82

7.7

1981/82

3.70

9.8

1982/83

3.30

8.5

1983/84

4.21

10.5

1984/85

6.99

17.0

1985/86

6.14

14.5

1986/87

2.53

5.8

1987/88

4.16

9.3

1988/89

5.35

11.6

1989/90

3.21

6.8

1990/91

7.22

14.8

1991/92

7.85

15.6

1992/93

4.97

9.6

1993/94

6.70

12.6

1994/95

3.99

7.3

1995/96

2.78

4.9

1996/97

3.36

5.8

1997/98

4.10

6.8

1998/99

7.19

11.7

1999/00

10.56

16.6

2000/01

6.24

9.6

Average

5.37

10.3

2002/2003*

14.5

21.0

* Estimated.

Source: Mulat Demeke (2003)


Annex 3: Crop production in the study areas (qt./ average household)

Wereda

Crops

Normal agricultural year

Drought year

Beneficiary

Non-beneficiary

Beneficiary

Non-beneficiary

Legeambo

 

Cereals

All cereals

7.03

7.01

2.83

2.62

Barley

6.29

6.06

2.47

2.27

Wheat

0.74

0.86

0.36

0.34

Sorghum

0

0.09

0

0

Maize

0

0

0

0.01

Pulses

All pulses

1.92

2.32

0.68

0.91

Lentil

0.89

1.28

0.31

0.51

Peas

0.69

0.79

0.25

0.30

Beans

0.31

0.21

0.11

0.08

Vetch

0.03

0.04

0.01

0.02

Oil crops

Flax

0.31

0.05

0.001

0.03

All Food crops

9.26

9.38

3.51

3.56

Share of Belg in total

production (%)

51.8%

56.5%

45.8%

49.7%

N

303

92

303

92

Meket

Cereals

All cereals

6.71

10.91

3.20

4.06

Barley

2.44

4.06

0.81

1.38

Wheat

1.26

2.30

0.50

0.70

Teff

1.55

2.26

1.05

1.40

Sorghum

1.31

2.27

0.79

0.50

Maize

0.14

0.02

0.05

0.08

Oat (Aja)

0.01

0

0.005

0

Pulses

All pulses

1.78

2.56

0.79

0.76

Lentil

0.20

0.49

0.04

0.21

Peas

0.52

0.77

0.16

0.18

Beans

0.93

1.08

0.28

0.28

Vetch

0.13

0.22

0.31

0.09

Oil crops

Flax

0.01

0

0.08

0.01

All Food crops

8.50

13.47

4.07

4.83

Share of Belg in total

production (%)

17.5%

16.3%

13.2%

10.8%

N

193

58

193

58

 


Annex 4: Partial correlation of level of education of HH head and ownership of productive
                 resources (Land and oxen)  (N=443)
-------------------------------------------------

             |  HH Head Edu.    Land   Oxen

-------------+-----------------------------------

HH Head edu. |    1.0000

Land         |   -0.0135      1.0000

Oxen         |   -0.0481     -0.0719   1.0000

-------------------------------------------------

 

Annex 5:Partial correlation of land, labor and oxen owned by sampled
        households (N=443)

-------------------------------------------

             |   Land     Labor     Oxen

-------------+-----------------------------

Land         |   1.0000

Labor        |  -0.1327   1.0000

Oxen         |  -0.0719   0.2828   1.0000

-------------------------------------------


Bibliography  

 

Addis Fortune (2004). A weekly Newspaper, Volume 5, No. 221, Addis Abeba.  

Amhara National Regional State (2002). Three Year (1996 – 1998 EC) Food Security Strategic Plan. Bahir Dar. Amharic Version.

Austrian Development Co-operation (2003): Ethiopia: Subprogram Food Security, Support of Agriculture and Rural Development with Emphasis on Natural Resources Management. Minoritenplatz 9, 1014 Wein.

Bekele Shiferaw (1998). Peasant Agriculture and Sustainable Land Use in Ethiopia: Economic Analysis of Constraints and Incentives for soil conservation. Published Ph.D Thesis. Dissertation no. 1998.1. Agricultural University of Norway.

Daniel Assefa (2003).Family Planning Services: an Important Front in the Battle against Poverty in Ethiopia. EEA/EEPRI.

De Graaff, Jan (1996). The Price of Soil Erosion: An economic evaluation of soil conservation and watershed development. Wageningen Agricultural University, Published Ph.D Thesis. Wageningen, The Netherlands.

DPPC (2000). National Food Aid Targeting Guidelines. Addis Ababa, Ethiopia

EDRI and IFPRI (2004). Ethiopia Strategy Support Program, Program Description. September 2004, Addis Abeba.

Ethiopian Economic Association (1999). The First Annual Report on the Ethiopian Economy. Vol. I. 1999/2000. Addis Abeba, Ethiopia.

_________(2002). Land Tenure and Agricultural Development in Ethiopia. Addis Abeba, Ethiopia.

________ (2004). Report on the Ethiopian Economy. Volume III 2003/04. Addis Abeba, Ethiopia.

FAO (2001). The State of Food and Agriculture. Rome, 2001.

Federal Democratic Republic of Ethiopia (2002): Ethiopia: Sustainable Development and Poverty Reduction: Strategy Paper for Promoting Development and Poverty Reduction. Ministry of Finance and Economic Development, May 2002.

Francois Bourguignon (2004). The Poverty-Growth-Inequality Triangle. A Paper Presented at the Indian Council for Research on International Economic Relations, New Delhi, Februrary 4, 2004.

Getahun Tafesse (2003).The Roles and Externalities of Agricultural Growth to Poverty Reduction in Ethiopia. Unpublished Report made for FAO-ROA Project.

Holden, S.T. and Bekele Shiferaw (1999). Incentives for Sustainable Land Management in Peasant Agriculture in the Ethiopian Highlands. In: Sanders, D.W., Huszar, P.C., Sombatpanit, S. and Enters, T. (eds). Incentives in Soil Conservation From Theory to Practice. World Association of Soil and Water Conservation. Oxford and IBH Publishing Co. Pvt. Ltd. New Delhi.

Hog and Lunde (2002). Application of Tobit Model. Department of Econometrics, NYU, New York, NY 10012.

IDS and SC-UK (2002). Destitution in the Northeastern Highlands (Amhara Region). Interim Report for discussion at a Policy Consultation Workshops in Bahir Dar and Addis Abeba.

Keddeman, W. (1992). An Economic Analysis of Soil Conservation Projects in   Ethiopia. In: Kebede Tato and Hans Hurni (eds). Soil Conservation for Survival. The soil and Water Conservation Society.

Julius Holt and Dessalegn Rahmato (1999). Study Report: Sustainable Livelihoods in North Wollo and Wag Hamra Zone. Save the Children UK.

 Julius Holt and Mark Lawrence (1993). Making Ends Meet: A Survey of the Food Economy of the Ethiopian North-east Highlands. Mary  datchelor House, 17 Grove Lane, London SE5 8RD. Save the Children UK

Leisinger, K.M., Schmitt, K. and ISNAR (eds). (1995). Survival in the Sahel: An ecological and developmental challenge. ISNAR, The Hague.

Masefield, A. (1996). The Great Grain or Cash Debate: Food for work versus cash for work
                     in the context of employment based safety net policy in
Ethiopia.

MOFED (2002). Ethiopia: Sustainable Development and Poverty Reduction: Startegy Paper for
                      Promoting Development and Poverty Reduction.
Addis Abeba, Ethiopia.

Mulat Demeke (2003): An Overview of Agricultural Production and Land Resource Management in Ethiopia. Paper Presented at the Founding Workshop of Farmers Competence Consortium. 14-17 July 2003, Addis Abeba.

NBE (2004). Quarterly Bulletin, Volume 19, No. 4, Second Quarter, 2003/2004. Addis Abeba.

Save the Children UK (2001). Cash-for-Relief Piloting Review Report.

Save the Children UK (2002)2: Cash-for-Relief Project 2002.

Save the Children UK (2002). Monitoring Report: Cash-for-Relief – North and South Wollo.

Save the Children UK (2003). Making Ends Meet: A Survey of the Food Economy of the Ethiopian North-East Highlands. Mary Datchelor House, 17 Grove Lane, London SE5 8RD.

 



[1] This paper is a modified version of the impact assessment study carried out to evaluate the Cash-for-Relief project of SC-UK.

[2] This estimate is based on the assumption that all energy requirement comes from grain (cereals, pulses and oil crops) consumption which largely reflects the reality especially in most parts of central and northern Ethiopia. The minimum daily calorie requirement per person used to compute the food balance sheet is 2100 kcal.

[3] Some Weredas in western Haraghie and Arsi became vulnerable to drought in 2002/03 for the first time, while about 35% of the Wollo population received food aid annually between 1997 and 2001.

[4] Interested readers could get a copy of the main report of the CfR impact assessment study from the EEA/EEPRI.

[5] However, recent data indicate that population growth has started to decline gradually.

[6] Normally an index should be developed to encompass the various factors involved in determining the level of food security of a given household. Compared to other factors, however, the size of farmland is believed to be the single most important factor that determines the level of food security. 

[7] The various food crops produced in the area were expressed in terms of their wheat energy equivalent.
   The major food crops of the area are barley, wheat and maize, sorghum, lentil, peas, beans, vetch and
   flax.

[8]  Still, there is a two year gap between our data on the poverty line which is computed for 2001 and the actual
   income collected from framers in 2003. However, this has little effect to change the analysis or the conclusion
   derived from it. 

[9] The level of poverty is calculated based on minimum consumption level for rural Ethiopia based on the 1995/96 HICE which is also used by the government in its 1999/00’s exercises to determine the level of poverty line. It is remembered that the basket of consumption goods considered in the calculation of the poverty line was identical in the 1995/96 and 1999/00 calculations. However, attempt is made to update the poverty line using the rural general price index using 1995/96 as the base year. However, data is available only for 2001. 

[10] Data on per capita and average household minimum expenditure for food and non-food consumption was taken from a document prepared for the SDPRP by the Ministry of Finance and Economic Development of the FDRE; and on price index from EEA database.

[11] Average weighted price calculated based on price and production data collected from the household survey. The share of the various crops in total production was taken into account in the calculation of the average weighted price.

[12] Data on income from off-farm and non-farm activities for normal/good agricultural year was obtained from the survey
    result. But for drought year and income from these activities is assumed to decline by 40% as the general economic activities
    in the area (except for activities related to food/cash aid operations) could weaken.

[13] Due to lack of baseline data, the study chooses an indirect way of measuring the dynamics of poverty
    in the study areas.

[14] 1 TLU=1 camel, or 0.7 ox or cow, or 10 sheep or goats, or 0.5 Donkey or bull, or 0.45 heifer or calf,
    or  0.7 mule, or 0.8 horse or 100 chicken (Hans E.Jahnke, 1982, quoted from EEA, 2002).

[15] Although the data is  a bit old,  a study by SC-UK reported that the size of livestock in 1992 increased for
    14% of the farmers, remained stable for 15%, decreased for 42% and much decreased for 29% when
    compared to the trend level (Julius Holt and Mark Lawrence,1993).

[16] Income from livestock sales and off/non-farm activities were also considered.

[17] Average household cash expenditures for food and non-food purpose were calculated based on data collected from baseline study conducted in South Wollo Highland Belg FEZ by SC-UK. The report classified the community according to their wealth status into better off, middle, poor and  very poor households who constitute 20%, 35%, 25% and 20% of the population respectively. Their annual cash expenditure is 1550 Br., 1000 Br., 725 Br., and 450 Br. In bad (drought) periods, cash expenditure was assumed to decline to 60% as households are forced to reduce their expenditures and consumption patterns.

[18] Based on various survey conducted in the study area, the price for one quintal of wheat is taken as 165 Br.

[19] Farmers’ income from off-farm and non-farm activities was recorded for an average (normal) year.  Income from these activities during bad or drought year is assumed to be 60% of the average year.

[20] According to a study made in the study areas, livestock sales could produce an average income that could purchase 1.3 and 1.6 quintal of grain in S.Wollo and N.Wollo respectively. But income from livestock sales is concentrated in a few hands. Roughly 60% of income from the sale of oxen and cattle and 90% of income from sales of shoats goes to the 20% of households with the largest holdings of these animals (SC-UK, 1993). Based on this information and survey results from this study, assumptions made that income from livestock sale of beneficiary households is only 30% of non-beneficiary households’ income from livestock sale.

[21] Food secured households are households who can produce sufficient food from their own farm to feed their family. Similarly food insecured households are households who could not meet part of their family food requirement from own production.

[22] Farmers covered by the survey reported that drought occur every 2 to 3 years.

[23] However, many question whether this massive relocation program is voluntary or not. For instance, Benjamin Joffe-Walt (cited by the weekly Addis Fortune, Volume 5, No. 221) quoted some farmers who said that they were moved by force.

[24] Even though environmental degradation and drought are the principal causes of food insecurity, lack of productive resources has increasingly led many households into chronic food insecurity problems. There are many households in the study areas that depend on food aid even in normal year.

[25] Regression results on household size measured in terms of adult equivalent was not reported.

[26] TLU is an index number that aggregates the different types of livestock a household owned to a single
    number. Even though different livestock which implies different degree of liquidity are aggregated into an
    index, the total livestock size still could broadly indicate existing differences in households’ vulnerability
    to drought. Households in the area normally keep any of their savings in livestock which could be
    liquidated any time the household face food shortage.

 

[27] In this regard, the recent attempt by the government to provide farmers with land (user) certificate could relax the problem. But this depends on the type and clarity of the rights the certificate could guarantee farmers and the efficiency of the law enforcing agencies to enforce the law in case of potential disputes.

[28] This is not a hypothetical recommendation. Some farmers in the study area have already considered their farmland as something that could not support their current and future livelihood. This makes them too careless to conserve or develop their farmlands.