submitted to UN Emergencies Unit for Ethiopiaby Laura Hammond

Executive summary and background to study

This report contains preliminary findings from the first month of a four month study of household level resource flows Ada Bai returnee settlement, Humera wereda, Ethiopia. This study is intended as part of a larger anthropological research project on the long-term process of reintegration which Tigrayan returnees to Ethiopia from Sudan are currently involved in. This is the fourth study in a series of reports submitted to the UN Emergencies Unit for Ethiopia and UN High Commissioner for Refugees (contact UN-EUE to request other reports).

The study is not intended to be an exhaustive economic survey. Rather, it is meant to provide profile information from ten household types selected to represent the economic, social, and religious diversity within the community and to give a preliminary sense of economic activity within the community. Data collection involves daily recall of economic transactions and agricultural production. The following data is solicited:

. household income levels. amounts and patterns of household expenditures. frequency of lending, borrowing and repayment. patterns of gift giving. levels and patterns of food consumption

Specifically, the study seeks to find answers to the following questions:

. What is the cost of living for returnees in the Humera area?. Is there significant economic stratification within the returnee community?. Do returnees have access to enough cash?. How do income levels relate to food consumption patterns?. What are the relative roles of farm and non-farm income to household economies?. What access do returnees have to credit and to what extent is it a part of overall household budgeting?

The following review of survey findings shows that returnee household economies can be characterised for the most part as semi-urban, with reliance on the market rather than on farm production in the provision of basic necessities. Returnees_ rate of borrowing is extremely high, supplemented by occasional waged labour and/or sale of farm products at the local market. Expenses are limited to basic food and nonfood items. Over time, it appears that household resources are being eroded; people are entering into increasing debt, selling expendable assets, and in some cases, limiting food intake. This may result in need for additional food aid in 1995.


Humera is located in the far northwestern corner of Ethiopia, bordered to the north by Eritrea and to the west by the Sudan. For the past thirty years it has been a major centre for commercial agriculture, with sorghum, sesame and cotton being the main crops. In June 1993, 12,000 Tigrayan refugees were repatriated from camps in the Eastern Sudan to the Humera area. They were joined in March 1994 by 2500 more returnees. The returnees are settled in three locations near Humera town. This research project is being carried out in Ada Bai, the largest of the returnee settlements with a population of approximately 7500.

Ada Bai is in many ways a large town rather than a rural village. Its population alone makes it one of the larger towns in the area. The town has a defined business district, with shops, tea and sewa houses,[15] tailors, weavers, metalworkers, and a video cinema. There is a church, mosque, health centre, and school for grades 1-5. There is an office for the baito[16] and a central meeting place for public meetings. There is electricity in the business district from 6 to 9 p.m. every night, provided on a rental basis to shop owners.

Residents of the three settlements were intended to become self-sufficient agricultural communities, both by providing labour to the large commercial farms in the area and by working on their own land. Each household was to receive between one and three hectares depending on number of members. For a variety of reasons, for each of the past two agricultural seasons there have been large numbers of households who have not received any land at all. For the 1994 agricultural season, 1947 of Ada Bai_s 2540 households were allocated farmland. Many of those who did not receive farmland entered into sharecropping arrangements with landholders from Humera town, providing all of the agricultural work needed in exchange for one half of the harvest. Households which lacked available labour or resources to become sharecroppers (primarily female-headed households) or who were allocated land too late in the season, did not farm at all in 1994 with the exception of their home gardens.

At the present time returnees in Ada Bai do not receive any external food aid. The last time a month_s grain ration was distributed to selected beneficiaries was in September 1994. Selected households are being assisted with low-interest loans from the baito. They also receive free health care, which is available from the local clinic and for referral cases from Humera hospital.

Demographic profile of the respondents

The ten respondent households were selected to represent the economic and social diversity of the community. Of the households sampled, two are female headed; two are Muslim; eight are Orthodox Christian, the head of one is an orthodox priest; nine are involved in farming; two had supplementary income (one a tailor and one a weaver). The average household size of the respondent group is 5.3 people.

Based on knowledge gained in the first year of this research project, an initial assessment was made of the relative economic strength of the households. Three of the households showed initial indications of being among the poorest in the community. Four households could be considered to represent the average household economic status, and three have a slightly higher economic standing in the community than the others. No very wealthy household was selected, as it was felt that such data would skew the averages and give the impression that people are better off than they really are. At any rate, such very wealthy households are very few in number in the Ada Bai community.

Four of the households had no livestock assets. These households also had very little expendable property. The other households own one donkey or goat with the exception of one household which has seven sheep and two goats. Those who do have expendable property own metal beds (valued at approximately 200 birr each), watches, tape recorders, and (in the case of some of the women) gold jewelry. Two households reported at the start of the study that they recently sold much of their furniture to cover their house expenses.


All of the participants in the study were known by the researcher prior to the beginning of the study. Indeed, the strong relationships developed in the first year of research in Ada Bai have helped to encourage honest reporting and to build trust in the research team on the part of the respondents. Confidentiality and anonymity was guaranteed to all participants in the study, as was the right of any household to withdraw from the study at any time, or to retract information already given. None of the households has availed itself of this opportunity, but it is believed that having the freedom to do so encourages people to be forthcoming with the data, much of which is considered to be extremely personal and secret. Returnees go to great lengths to hide their resource holdings from both their neighbors and local authorities, no matter how small their resource base, so as not to be excluded from distribution of relief aid. Returnees commonly feel that decisions about who should receive aid are unfair and that those in need are often screened out if, for instance, they own animals, have what local officials consider expendable assets, or are engaged in work that is deemed more profitable than farming. Assurances of confidentiality and awareness that the research team is not directly associated with any aid organization (as well as the fact that the research assistant is a returnee himself who is known and trusted by the community) allows for collection of data that is very likely to be more reliable than that obtainable by other aid actors present in the area.

The study was begun on 1 Hidar 1987 E.C. (11 November 1994).[17] At the outset, a baseline questionnaire was administered to all respondents. This recorded such information as age, education and training of each household member; number of chronically ill household members, estimated monthly income; property (food, cash, animals, other assets); amount of any remittances; resources shared with other households; landholdings; planted crops and yields; outstanding debts and schedules for repayment; and amount and date of last receipt of food aid.

Data is collected on a daily basis. Each of the ten households is visited and their economic transactions for the previous day are recorded. These include wage earnings, money obtained through sale, borrowing or gift as well as all expenses (itemized) and money lent or given. Meals eaten in the previous day are also recorded.

It should be noted that the start of the study coincided with the harvest time. This represents the period of greatest economic activity, and one would expect income, expenditure and food consumption levels to be at their highest point during the months of Hidar and Tahsas (November and December). The study will conclude in Megabit (March) which is the period of greatest economic hardship for most households. It is expected, therefore, that the final results of the study will reflect seasonal changes in economic activity and resource flow. Caution should be exercised in applying the data contained herein to non-harvest periods.

Expenditure profiles

While some returnees in the Humera area have found work as agricultural laborers, and many cultivate at least some of their own food crops, it is clear that the returnee settlements are not strictly agricultural settlements, but rather are semi-urban communities, heavily reliant on cash income and the market for daily household food supply rather than on subsistence farming. Cross-border trade plays a major role in the local village economy; most goods available in the market are brought to Ada Bai from the Sudan by traders. In addition, prices are very high due to the inaccessibility of the area from other Ethiopian trading centers (particularly Gondar and Mekele).

The study seeks to quantify the cost of living for returnees by examining daily expenses. Expenses are broken down by item into food and nonfood categories. As can be seen in Table 1, food expenses on average accounted for 47.9 % of the total household expenditures. Nonfood expenses accounted for 38.1% of total expenditures. It should be noted that nonfood expenditures are higher during the harvest period than at other times of the year. Most of the nonfood expenses are related to farming costs: purchase of quintal sacks and transport of grain from the fields to the farmers_ houses constitute major expenses at this time of year. Actual average expenditures are also given.Table 1. Types of Expenditures Expressed as Averages and Percentages of Total Expenditures

Type of Expenditure         Monthly                As Percent of Total    
                            Expenditures(Birr)18   Expenditures           

the ten households, the range of expenditures is very wide for both food and
nonfood categories. For monthly food expenditures, the amounts range from 44 to
305 birr per household, averaging 58 birr per person. Differences in household
size only partly explains this great degree of variation: per person food
expenditures also vary greatly, from 21.34 to 101.54 birr per month.  By way of
comparison, Webb and von Braun report 1988-89 spending levels of 33 to 37 birr
per person (US$ 16 to 18 at 1989 exchange rate).[19]   Other
factors that affect a household_s food expenditures include food stores prior
to the start of the study and the amount of available cash or credit.

Nonfood expenditures at the household level are even more variable, ranging from 5 to 311 birr. These figures are not computed on an individual basis as most of the expenses are intended for communal consumption or benefit. As noted above, these values tend to be higher during the harvest season than at any other time throughout the year. As is to be expected, the lowest expenditures are found among households with no farm income.

It has been charged by some visitors to the Humera area settlements that the returnees are wealthy compared to highland peasants in Tigray. As proof, they point out the existence in the homes of valuables such as beds, bicycles, and radios, plus the perceived high rate of market activity in the town. It is true that while in the Sudan living as refugees, many of Ada Bai_s residents were able to invest in such high value items, and they brought what they could back with them when they were repatriated. This "wealth", however, represents the asset base on which people are currently depending. As will be seen later in this report, there is a very real danger that such assets will have to be liquidated in the near future and that returnees will lose this important safety net.

Such misleading images of the "wealth" of returnees do not hold up when it comes to an examination of their current expenditure profiles. The expenditure profile given by all ten of the families shows "basic needs" expenses for household food and items such as kerosene, soap, flour milling, etc. With the exception of farm expenses (transport, purchase of quintal bags for harvesting), the data does not indicate that money is being spent by any of the households on valuable durable goods, animals or other items that could be considered a significant investment.

Total expenditures for the month, including food and nonfood expenditures, gifts and debt repayments, vary similarly. Household total resource outflows range from 50 to 578 birr.

Household incomes

The above information suggests that there is a significant variance in household incomes, with a vast discrepancy between what Webb and von Braun would call the "poor, poorer and very poor" households.[20] Consideration of the composition of household incomes shows that this is not the case. In many cases the poorest households may be those who borrow the most, because they have overextended themselves with regard to credit and are living beyond their means. While their expenditure profile is high, their effective economic status can be said to be among the lowest because they have no way of repaying their debts. Discrepancies in expenditure levels are correlated with differences both in access to, and utilization of credit.

Household income is considered here to be the sum of all earnings (including proceeds from sales), and loans and gifts received. Table 2 shows the relative weights of each of these components in each of the ten households. In five of the ten cases, borrowed money outweighs earnings. In the case of Household 3, in which 98% of the total household income was borrowed money, there is reason to believe that the household has sufficient assets and can be reasonably sure of earning wages during the sorghum harvesting season to repay the debt. Of more concern is Household 5, which has very few assets and very little prospect (because the household head suffers from chronic illness) of earning money through waged labour to repay the loans. The seriousness of the situation is not lost on the household, which at month_s end was trying to find a way to repay its debts and to reduce the amount of credit it seeks in the future. While mean statistics should be treated with caution due to the small sample size and the high degree of variation between household incomes, average earnings, received loans and received gifts are noted I Table 2 as 49.2%, 45.5%, and 5.4% respectively.

Contributions deemed "gifts" by the respondents do not significantly increase the household_s income base. The exceptions are Household 2, which hosted a funeral feast, and Household 6 which has relatives in Humera town who frequently visit and bring gifts. Given the high rate of borrowing and the relatively low rate of loan repayment (23% of debts were reportedly repaid within the month), it is likely that many of the smaller loans were extended without expectation of repayment and are in effect gifts.

Table 2. Earnings, loans received and gifts received as percentages of household income

Food         164.65House  47.9         Nonfood      34567*89130.7  
Purchases    hold #       %Earnings %  PurchasesOt  648.17343.58G  
                                       her          ifts           
                                       (gifts,      Received %     
                                       Rec_d %                     

Balance of payments

Only one of the respondent households showed a positive balance of payment at the end of the month. This was a household that had harvested two quintals of sesame, which it sold during the month. In the other nine households expenses exceeded incomes. Deficits ranged from a few to 150 birr. Those who experienced deficits had to draw from their meager cash reserves or sell off their assets to pay their debts.

The fact that household resources are not enough to cover basic needs even at the time of year (harvest time) that should theoretically be the most prosperous indicates a gradual decline in the ability of households to meet their consumption needs. The heavy reliance on credit and the inability of households to repay their debts suggests that people may be exhausting their available credit. Some will be able to improve their credit rating by earning money harvesting sorghum in the Sudan in January and February. Others will be forced to sell any assets they may have or to reduce consumption.

In deciding which assets to sell, it is clear that people would rather sell goods which can be relatively easily replaced, such as grain, than such high value property as beds or tape recorders, which are extremely difficult to replace. Three of the ten households reported selling grain during the month to meet their household expenses. They reasoned that if they sold their grain stores while the price of sorghum is still relatively high at 90 birr per quintal, they would be able to pay their debts or cover household expenses, and that with any luck they would be able to buy grain at a cheaper price in February and March. Such rationalization might prove sound if prices drop significantly. If they do not, however, further selling off of non-food assets will have to be resorted to in order to provide food for the household until the next rainy season. Thus, liquidation of household food stores may make the household more vulnerable to food shortages in the short term. Rather than having a store of grain in their homes, returnee farmers are reliant on beneficial market prices to meet their basic subsistence needs.

Food consumption patterns

A detailed analysis of the nutritional value of household food baskets reported in this study has not yet been done. Such analysis would seek to determine levels of calorie intake, and prevalence of the major necessary macro- and micro- nutrients.

Sorghum and sesame oil are the only locally grown items in the food basket for most of the year. During the rainy season, most households have home gardens in which they grow corn, okra, greens, and spices for cooking. The only vegetables available on the local market are onions, potatoes, and sometimes tomatoes. Market prices are reasonably stable.

Preliminary consideration of daily food consumption patterns, however, reveals a correlation between household income and nutritional quality of diets. Table 3 shows the various major meal types: injera with salt, injera with berbere, injera with pulses, injera with vegetable, and injera with meat. The other category is made up for the most part of bread, curdled milk, or other types of meals. Households 1, 2, and 7 have the lowest income of the group. Their reliance on sorghum injera with salt for a staple food diet is significantly higher than the other households. Indeed, it is quite common for these households to eat injera with salt and injera with berbere at all three meals for several days at a time. In comparison, the other households show a more balanced diet, reflecting their relatively better economic position.

Despite high variability in the quality of diet in different households, there did not appear to be a high incidence of missed meals. That is, while households may reduce or constrict their consumption by eating cheaper foods, they seldom go without food, at least in during the harvest season.

Table 3. Household food consumption, listed as by type as percentages of total consumption.

         1Househol   77Salt  18Berbe           9.07.01  3456789  5336369  
         d                   re       6Pulses  0.04.00  1.001.2  5.0      
                                               .511.01  56.251.  3.253.0  
                                               9.538.1  00-6.75  0.30.12  
                                               %14.0%1  1.75234  .08.0    
                                               00%Vege  5678     48  2    
                                               table    910Aver  31  8    
                                                        ageMeat  74 89    


For those who had farmland this year, yields averaged four quintals per hectare. Per person consumption is estimated by the World Food Programme and Relief and Rehabilitation Commission to be 150 kg per year.[21] A household of three people receives one hectare of farmland, and requires 4.5 quintals food for the year. If the household consumes all of the grain it produces, it will still have a 50 kg deficit for the year.

Such calculations are academic; in reality, most farmers consider the principal value of the grain they produce to be its market value. They will sell one or two quintals immediately after the harvest in order to pay back debts and cover the costs of the harvest. Additional quintals may be sold to pay for building materials for the house (which must be renewed each year), to pay for treatment of a sick household member, or to prepare the land for the next agricultural season. Farmers sell in the hopes that they will be able to buy sorghum when the price is low (in 1994 sorghum dropped to 30 birr per quintal after the harvest period). In this way, sorghum is treated as a monetizable asset by the household. The balance of the household food basket is made up from purchases at the local market with funds raised either through monetization, borrowing, or in relatively few cases, from wage earning.


Preliminary findings from this household economic study show that even at harvest time, returnees_ access to cash is severely limited. They meet their daily subsistence requirements through a combination of wage earning, asset monetization and borrowing. The high prevalence of these last two practices suggests that people are eroding their resource base even at this the most plentiful time of the year. Those who have come from the Sudan with large amounts of expendable assets will soon find themselves forced to sell, and those who have nothing will be hard pressed to make ends meet.

Credit is being utilized by the sample at an alarming rate. It is clear that people are borrowing beyond their means, with only the hope that a change in their circumstances will enable them to pay back their debts. A more likely scenario is that those who extend credit to these households will not be willing to continue such practices, and will start to demand repayment. This may also trigger sale of expendable assets.

The biggest problem that the returnees face is one of access to cash. There is not yet enough economic activity within the community to enable it to sustain itself without outside assistance. The market is not well developed; many shops are sitting empty because the owners do not have the money to stock them. Even in the harvest season, there are very few employment opportunities; workers must compete with each other for work and contracts are usually given for only a few days of work at a time. Nonfarm work is basically nonexistent. There is a definite need for increasing the income-generating opportunities available within the community.

There has been some discussion among those who are involved in assistance to the returnees of making funds available to farmers through revolving loan funds. Such funds would clearly be appreciated by members of the community, but care must be made to ensure that the money being loaned is being used as capital in such a way that the debtor has a realistic chance of being able to pay it back.

It is hoped that information from this survey can help provide a profile of returnee communities in northwest Ethiopia so that appropriate assistance can be designed and delivered not only to those returnees who are already living in the area, but for those who are to be repatriated in the future. In order for such planning to take place, a thorough understanding is needed of the ways that the returnee community economy differs from that of the highlands. If returnees are expected to remain in their new lowland homes, there must be more of an attempt to develop the economic infrastructure that can sustain them.

[15] Sewa is local beer, known also as tella in Amharic.

[16] Local governing council.

[17] E.C. is used to connote Ethiopian Calendar.

[18] US Dollar 1.00 = ETH Birr 6.30

[19] Patrick Webb and Joachim von Braun. 1994. Famine and Food Security in Ethiopia: Lessons for

Africa, John Wiley and Sons: Chichester, p.77.

[20] Ibid. p. 62.

[21] This figure is used to calculate food rations for relief distributions, and should be considered sufficient

only to provide a minimal standard of subsistence. A more appropriate figure for long-term consumption

is probably closer to 200 kg/person.

Back to document index

  • Disclaimer for EUE Materials

    Editor: Ali B. Dinar, (