Baseline socioeconomic survey report: agriculture in Borno State, Nigeria by International Institute of Tropical Agriculture

\\TA Baseline socioeconomic survey report: agriculture in Borno State, Nigeria P.S. Amaza , J.K. Olayemi, A.O. Adejobi Y. Bila , and A. Iheanacho

",'cr.II)

of \ t.:lldugun

I +I

Coo"',," Inlemallofl."

lk\dormcnt A!clll.-Y

Intemationallnstitute of Tropical Agriculture - Institut international d'agriculture tropicale - www.iita.org

Baseline socioeconomic survey report: agriculture in Borno State, Nigeria

P.S. Amaza , J.K. Olayemi, A.O. Adejobi Y. Bila, and A. Iheanacho

ÂŠ International Institute of Tropical Agriculture (1ITAl, 2007 Ibadan, Nigeria To Headquarters from outside Nigeria: IITA, do Lamboum (UK) Ltd, Carolyn House 26 Dingwall Road, Croydon CR9 3EE, UK Within Nigeria: PMB 5320. Ojo Road Ibadan, Ojo State ISBN 978 131 315 3 Printed in Nigeria by IITA Correct citation: Amaza. P.5., J.K. Olayemi, A.O. Adejobi, Y. Bila, and I. Iheanacho. 2007. Baseline socioeconomic survey report: agriculture in Borno State, Nigeria. International Institute of Tropical Agriculture, Ibadan, Nigeria. 84 pages.

Contents 1. Introduction ... ................. ........... ..... ._._ ... _......................._ ... _........ . The agricultural sector in Nigeria .. ............. ...................... _...... ..... . The PROSAB Project ............ ........... .............................................. . The socioeconomic baseline study ................. .. ........................ ..... . Scope of the study ....... ..... .................... ............................. .......... ..

3 4

2. Methodology ............. ... .............. ......... ............. ...... ...... ..... ... ....... ...... . Description of the survey area ...................................................... .. Sampling technique and data collection procedure ...... ................ .. Analytical processes adopted in accomplishing the Terms of Reference ......... .............................................................. ..

3. Description and analysis of household characteristics ...................... Distribution of household heads by gender ..................... ............... Distribution of household heads by age .. ........................................ Distribution of household heads by years of formal education ....... Distribution of household heads by marital status ...... .......... .......... Distribution of household heads by occupations ........... .................. Household asset ownership structure ............................................. Number offarm plots owned by household heads and their spouses.. Household farm sizes .......... ........ .. ................. ..... ... .... ...... ...... ...... .. Farming experience of household heads .. ..... ............. .... ................ Distribution of households by types of land tenure ......................... Distribution of sources and amounts of credit obtained for farming .... Distribution of the most important crops grown by households ...... Continuous cropping of farm plots .......................... ................ ........ Types of crops no longer grown ..... ............ ...... .. ...... ....................... Estimated gross margins in crop production .. ..................... ........... Factors determining profitability in crop farming ............... .. ............ Composition of household livestock................................................ Profitability of livestock farm ing.. ............ .. . ..................................... Household non路farming activities ........ .. ............................... Household food consumption patterns .... ...... ..................... ............ Consumption of nonfood commodities by households ...................

22 22 22 23 24 25 26 27 28 28 29 30 30 31 32 34 35 36 40 41 42 47

4. Household poverty and food insecurHy analysis .............. ................. Household food insecurity statistics .......... ...... ...... ....... ........ .......... Detenminants of household food insecurity..................................... Classification of households by poverty status .......... .. ........ ........... Detenminants of household poverty intensity ............ ...................... Elasticities of household poverty intensity......................................

50 50 50 54 55 61

7 7

5. Major findings and recommendations .................. ..... ....................... .. 63 Summary of major findings.... ..... .................. ............... ................... 63 Policy recommendations ................ ................................................. 70

iii

Tables 1. Percentage distribution of household heads by gender ...... ...... ....... 22

2. Percentage distribution of household heads by age category and gender .................. ...... .......................................................

3 . Percentage distribution of household heads by years of

formal education ............. ...................................................... .

4 . Percentage distribution of household heads by marrtal status

5. Percentage distribution of household heads by their main occupations ..... ..... ......... .......... .......................................... ..... 6. Percentage distribution of household asset ownership ............ . 7. Distribution of number of farm plots owned by households........................................... .......... ............ ..... .. 8. Average household farm sizes (ha) .......... . 9. Average farming experience of household heads (years).... ........... 10. Percentage distribution of households by types of land tenure .. 11. Percentage distribution of households growing various crops. 12. Percentage distribution of reasons given by households for dropping some crops from cultivation.... ............ ............... .... . 13. Estimated gross margins (N/ha) ............ .. .... ..... ...................... . 14. Percentage distribution of major factors determining profitability in crop farming .................... ................................... ........ 15. Average number of various animal species owned by men in households ...... ............................. ........................................ 16. Average number of various animal species owned by women in households .................... .......................... ................ .. 17. Average number of various poultry species owned by men in households .. .......... ...... ...... .. .. ............................................ 18. Average number of various poultry species owned by women in households................ ..... .. ................... ............ ................ 19. Percentage distribution of households identifying factors determining the profitability of livestock fa rming ................ 20 . Average quantities of cereals consumed by households (kg/month) .... 21 . Average quantities of legumes consumed by households (kg/month)... 22. Average quantities of roots and tubers consumed by households (kg/month) ................ .... ............... ....... .......................... 23. Percentage distribution of households by types of nonfood items consumed ........ ........ .......... .. ........ ........ ...... .............. .............

24 24 2S 26

27 28 29 29 31 33

34 3S

37 38 39 39 40

43 44 4S 47

24. Average household expenses on nonfood consumer items (Nlmonth) ... 48 25 . Percentage shares of major nonfood consumer expenses in total household nonfood expenditure ................ ............. ......

26. Summary statistics on food insecurity among households in the project area ... ..... ....... .............. .... ..... ... ... ....... ....... ..........

27. Results of the Logit Regression Analysis of household food insecurity status .... .. ................ .. ...... ...... ....... .....

.... .... ....... ... . 51

28 . Results of Tobit Regression Analysis of household poverty intensity .... ....... .. .. ...... ...... ...... .... ..... .... ...... ................. ...... .... 56 29. Elasticity estimates of household poverty intensity. .. .......... .... ..... 61

References ............ ..... ..... ....... ......... ..

. .... .. .. .. ... ............... ............. .... 73

Annexes 1. Conversion factors for calorie requirements for different ag e groups .... 76 2. Barno State showing the project area .... ... ... .......... .. ... .... ... .... ....... ... 77 3. Communities covered by the socioeconomic baseline survey....

Introduction

The agricultural sector in Nigeria Nigeria is the most populous country in Africa, with an estimated population of about 130 million (2004) currently growing at an annual rate of about 2.8%. It lies wholly within the tropics, along the Gulf of Guinea in West Africa. It covers a geographical area of 923 768 km' and spans four broad agroecological belts or zones: the mangrove rainforest along the southern coast, the rainforest further inland, the southern and northern Guinea savannas in the central region and the Sudan savanna and Sahel in the northernmost region of the country. An estimated 65% of the population reside in the rural areas where agriculture is the predominant occupation. It is estimated that about 70% of the rural population are engaged in agriculture (FOS 1999). Generally, the agricultural sector is the single largest sector of the economy, contributing about 41 % to the country's gross domestic product. It also contributes significantly to national employment, w ith about 60% of the country's total workforce engaged in agriculture. The sector accounts for most of Uhe country's food supply and it is also an important contributor to the nation 's foreign exchange earnings as well as Uhe supply of industrial raw materials (Olayemi et al. 2004). Nigeria has a total land area of about 98.3 million ha. But although about 71.2 million ha are cultivable, only about 34 .2 million ha or 48% of the total cultivable land are actually cultivated. Due to its high agroecological diversity, the country produces a very wide range of agricultural products, consisting of staple food crops, cash and industrial crops, livestock, fish, and forest products. But overall , cereals and roots and tuber crops constitute the largest categOl)' of agricultural products. Nigeria is a nation of small farmers who account for over 90% of the country's total agricultural production. These farmers cultivate small land holdings that are often less than 2 ha in area and in fragmented plots.

The traditional system of agricultural production still predominates , with its characteristically low technological base, high reliance on manual labor, and hence low resource productivity. Agricultural production still depends heavily on the vagaries of nature; as a result, seasonal and annual fiuctuations in agricultural outputs are common. Women in Nigeria play important roles in the agricultural economy and especially in arable food crop production, processing , and marketing .

It is estimated that women produce as much as 60-80% of the food in Nigeria (Akinyele et al. 1991). However, the rate at which women participate in famning varies widely across the country. In the main, cultural and religious factors are important determinants of their rates of participation. Generally, women's participation is highest in the southeastern and middle belt areas of the country and lowest in the Muslim areas of northern Nigeria (Akinyele et al. 1991). Anambra and lme States have the highest rates of women's participation in the southeast; Cross River, Rivers , Akwa Ibom, and Delta States have relatively high rates in the south-south area with an average of about 75% . In the middle belt area of the country, rates of women's participation in farming are relatively high in Benue, Plateau , Adamawa, and Taraba States. For cultural reasons, the rate of women's participation in farming in the Yoruba southwest is much lower, being less than 40% . The rate exceeds 40% only in Ogun State (FOS 1999). But women in the southwest dominate food processing and marketing activities. As men produce food crops, their wives take charge of the processing and marketing activities. In the Muslim areas of northern Nigeria, the rate of women's participation in farming is the lowest in the country, due to the fact that Muslim women are supposed to be kept at home. Hence, in Bauchi, Barno, Kaduna, Kano, and Niger States, the rate is less than 20% (FOS 1999). Women are largely involved in subsistence food crop production on small plots of land often less than 0.5 ha in size. The capacity of women to participate in larger-scale, commercial crop production is constrained by difficulties associated with inadequate access to land and credit, as well as the laborious nature of traditional farming methods. According to Daramola (2004 ), manual farm operations impose severe limits on famners ' capacity to increase their farm sizes and productivity:

The traditional system of agricu~ural production still predom inates, with its characteristically low technological base, high reliance on manual labor, and hence low resource productivity. Agricultural production still depends heavily on the vagaries of nature: as a result, seasonal and annual fluctuations in agricultural outputs are common. Women in Nigeria play important roles in the agricultural economy and especially in arable food crop production , processing, and marketing. It is estimated that women produce as much as 60-80% of the food in Nigeria (Akinyele et al. 1991). However, the rate at which women participate in farming varies widely across the country. In the main , cultural and religious factors are important determinants of their rates of participation. Generally, women's participation is highest in the southeastem and middle belt areas of the country and lowest in the Muslim areas of northem Nigeria (Akinyele et al. 1991). Anambra and Imo States have the highest rates of women's participation in the southeast; Cross River, Rivers, Akwa lbom, and Delta States have relatively high rates in the south-south area with an average of aDout 75% . In the middle belt area of the country, rates of women 's participation in farming are relatively high in Benue, Plateau, Adamawa, and Taraba States. For cultural reasons, the rate of women's participation in farming in the Yoruba southwest is much lower, being less than 40%. The rate exceeds 40% only in Ogun State (FOS 1999). But women in the southwest dominate food processing and marketing activities . As men produce food crops, their wives take charge of the processing and marketing activities. In the Muslim areas of northem Nigeria, the rate of women's participation in farming is the lowest in the country, due to the fact that Muslim women are supposed to be kept at home. Hence, in Bauchi , Barno, Kaduna , Kana, and Niger States, the rate is less than 20% (FOS 1999). Women ane largely involved in subsistence food crop production on small plots of land often less than 0.5 ha In size. The capacity of women to participate in larger-scale, commercial crop production is constrained by difficulties associated with inadequate access to land and credit, as well as the laborious nature of traditional farming methods . According to Daramola (2004), manual farm operations impose severe limits on farmers' capacity to increase the ir farm sizes and productivity:

the manual system of farming is technically inefficient, labor-intensive, and costly to sustain. Although the use of animal power for land cultivation has been adopted to a limited extent in the savanna/Sahel belt of the country, its use in other areas is almost nonexistent, for ecological and cultural reasons. Mechanization of agriculture in the country has generally made little headway due to technical , ecological , and socioeconomic factors (Olayemi et aL 2004). It may, therefore , be inferred from the foregoing that Nigeria's agrarian economy is still largely in the pre-modern phase , although some indications of modernization are in evidence here amI there . The challenge is how to accelerate this modernization process on a sustainable basis in the face of the foregoing constraints.

The PROSAB project Livelihood strategies for most men and women in Barno State are based on agriculture. Farming is characterized by a variety of cropand livestOCk-based production systems. Most crops are grown for home consumption as well as for local markets. In the southern part of the State, maize, sorghum , cowpea, and, to a lesser extent, rice and soybean are the major crops. Towards the north, sorghum and millet become the dominant crops. Livestock species (i.e., small and large ruminants as well as poultry) are an integral part of the farm ing system and provide income as well as household food security. However, due to adverse biophysical conditions such as erratic rainfall, marginal soil fertility, and a non-conducive policy environment, the agricultural sector is no longer able to cater for the growing popu lation of the State, and it is becoming less capable of coping with unexpected shocks. Hence, in many parts of the State, farmers have been obliged to diversify their livelihood sources with incomes from outside the agricultural sector. Cognizant of the foregoing conditions, which are also preva lent in many other parts of the country, the Federal Government of Nigeria prepared and adopted a new national Rural Development Strategy in 2001 . Its aims are to improve livelihoods and food security through a process of community-based

agricu~ure

and rural development programs.

The strategy calls for a community-driven development approach that ensures the active participation of beneficiaries and local governments at

all levels of decision-making. It is within this development framework that the Canadian International Development Agency signified, in September 2003, its assistance to the agricultural and rural development sector of Nigeria by funding a project for Promoting Sustainable Agriculture in Barno State (PROSAB) which was proposed by the International Institute of Tropical Agriculture (UTA). The project is being implemented in three agroecological zones of Barno State , the southern Guinea savanna (SGS), the northern Guinea savanna (NGS), and the Sudan savanna (SS). (A map of the project area is presented in Annex 1 of this report.) The goal is to contribute to improved rural household livelihoods in the project areas. The specific objectives are to contribute to (1) improved food security, (2) reduced environmental degradation, (3 ) improved sustainable agricultural production through the transfer of improved agricultural technologies and management practices to female as well as male farmers , (4 ) improved market access, (5) a more enabling policy environment, and (6) enhanced capacity of project partners. The project was designed in a participatory way and took into consideration the experience gained from past projects in northern Nigeria especially in the project area . The project was to be implemented by IITA in collaboration with local and regional development partners , including non-governmental and community-based organizations. Strong linkages were also to be forged with other donor-supported projects in the State as well as across States.

The socioeconomic baseline study In pursuance of the goals of the project, a baseline survey was carried out in the project area in 2004 to provide sex-disaggregated baseline da1a on socioeconomics, resource use patterns, market opportunities, and their effects on land degradation and agricultural productivity in potentially targeted project communities. The major baseline indicators or criteria designed for measuring and monitoring farmers' economic status and progress included the following:

â&#x20AC;˘

Agricultural production characteristics, such as farm size distribution , distribution of the number of farm plots owned , distribution of important crops grown.

â&#x20AC;˘

Livelihood diversification indices, such as households' engagements in non-farming activities and enterprises.

•

Farm households' access to improved farm inputs and gender-based differentials in access.

•

Farm households' food and nonfood expenditure and consumption patterns.

•

Farm households' food secu rity and poverty status.

Socioeconomic measurements as well as food security and poverty status analyses were carried out to help support and explain the various baseline indicators. These were to provide information on the capabilities and constraints under which the farme rs, their spouses, and other family members operated to achieve the desired goal of improved household welfare . These socioeconomic frameworks were taken into consideration in making recom mendation s for efficient operations under the PROSAB project.

Scope of the study The report will cover the following . 1. Household socioeconomic characteristics Distribution of household heads by gender. Di stribution of household heads by age. Distribution of household heads by years of formal education . Distribution of household heads by types of occupation. Household asset ownership structure . Marital status of household heads. 2. Household crop production activities Number of farm plots owned by household heads. Number of farm plots owned by spouses of household heads . Sizes of household farm plots. Types of land tenure fo r households. Types of crops grown by households Distribution of household plots by years of continuous cropping . Crops that households have stopped growing , and the reasons. Quantities of crops produced by hou seholds. Percentage of produced crops sold by households. Estimate of gross margins from crops produced. Factors that determine profitability in crop farming. Constraints inhibiting profitability in crop farm ing .

3. Crop production inputs Sources and quantities of inputs used by households. Types of farm power used by households. Sources offarm credit. Modes of transportation used by households.

Livestock farming activities of households Distribution of households by animal stock composition. Distribution of households by poultry stock composition . Factors that determine profitability in livestock farming . Constraints to profitability in livestock farming.

Non-farming income sources available to households Distribution of households by types of non-farming employment. Distribution of households by the types of non-farming enterprises in which they engage. Estimate of household incomes from non-farming activities.

6. Household food and nonfood expenditure and consumption Consumption of own-produced food by households . Consumption of purchased food by households. Distribution of food quantities consumed by households. Distribution of nonfood expenditure by households.

7. Food security level and poverty status of households The food insecurity level of households . The poverty status of households .

Methodology

Description of the survey area The baseline field survey was carried out in the communities covered by the PROSAB project in Borno State. Borno State in northeast Nigeria covers an area of 69 435 km'. The State is demarcated into four agroecological zones: the NGS and SGS in the south , the SS in the southem and central parts , and the Sahel in the north. The annual precipitation ranges from less than 600 mm in the north to 1500 mm in the south. Rainfall, however, varies from year to year but has tended to decrease over the last two decades. The population of Borno State is estimated to be 3.35 mil lion (2001 estimate) . More than one million people live in Maiduguri, the State capital. The population of the project area in the State is estimated to be about 1.67 million . The project area covers three agroecolog ical zones located between latitude 10' and 12' north of the equator and longitude 11 ' 30 ' and 14' east (Annex 2) and comprises four Local Government Areas (LGAs): Damboa in the SS, Hawul in the SGS , Biu and Kwaya Kusar in the NGS. Numerous ethnic groups and cultures characterize the area, with approximately 80% of the population being small-scale farmers. Agriculture and trading constitute the major economic activities of the area (BOSADP 1998). The baseline study area is characterized by a relatively wet and humid climate as compared with the drier northern part of the State. Hence, agriculture is more widely practiced in the study area. The annual rainfall ranges from 700 to 1200 mm with about 140 to 150 days of rainfall. The rainy season starts in April and lasts until October. The humid ity is about

49% with an annual evaporation rate of 2034 mm. The temperature varies from 15 째C to 20 째C during the harmattan period (i.e., between December and February) and also from 32 째C to 38 째C during the hot, dry season (i.e ., between March and May).

The vegetation of the study area is typical of the savannas , consisting of shrubs interspersed with trees and woodland . Most parts of the area are mountainous with abundant rivers that are, however, seasonal in nature. The agricultural activities in the project area can be categorized into cropping activities and animal husbandry. The cropping pattern is almost uniform throughout the area, probably due to sim ilarities in vegetation . There are two cropping seasons . One is the wet cropping season that starts with the onset of rain and lasts to the end of the rainy season when most crops are harvested. The other is the dry cropping season, which starts soon after this . Crops may be grown sole, multiple , mixed, or in relay. Crops may also be grown in rotation, depending on preference. Major crops grown in the area are ma ize, sorghum, ccwpea, groundnut, and vegetables . T he animals reared are cattle , sheep, goats, pigs, and pou ltry.

Sampling technique and data collection procedure The baseline data were obtained through a household survey conducted between June and August 2004. The main instnuments for data collection were well-structured Questionnaires administered on households by trained enumerators under the supervision of PROSAB's agricultural economist and research collaborators from the Department of Agricultural Economics, University of Maiduguri. Four LGAs were covered for the purpose of data collection . Data were collected from 39 communities and settlements spread across the project area. (Annex 3 provides a list of the settlements and communities .) First, the 30 communities that had been already identified and selected during the livelihood analysis carried out in preparat ion for project implementation activities were purposively included in the sampled communities. Then, in addition, nine other communities where project activities were expected to scale-out were also selected for the data collection exercise. In each selected community, a random selection of households was carried out, using a sampling proportion to size procedure , with 40 respondents selected in relatively small communities , 80 in med iumsized communities and 120 in large-sized communities. To ensure gender balance, the survey was so designed that 25% of all respondents were female. This impl ied that 25% of the respondents were female

household members who were not necessarily household heads. But in the remaining households, only household heads were interviewed. The survey instrument was also modified in some areas to incorporate gender-specific issues. A total of 2250 questionnaires were administered out of which 2205 were retrieved and used in descriptive analysis. However, only 1998 questionna ires were used in household food security and poverty status analyses, because of incomplete information in some questionnaires.

Analytical processes adopted in accomplishing the Terms of Reference In executing the Terms of Reference for the study, the following analy1ical procedures were adopted .

Descriptive statistical and budgetary analysis For household socioeconomic analysis, descriptive statistics, such as mean, standard deviation, frequency distributions, and so on, were computed and used. However, for the purpose of evaluating the profrtabi lity of farm production activities, budgetary analyses were conducted involving the computation of gross marg in (GM) and retums per naira ouUay (RPN). These were carried out separately for crop and livestock production activities. The monetized values of variable inputs and incidental production costs were subtracted from gross revenue (GR) to arrive at GM estimates for crop and livestock enterprises. The RPN was calculated by finding the ratio of the GM to the total variable cost (TVe) in each case . That is :

OM = OR - TV C

(I )

and RPN = O~VC

(2)

From the above , it was possible to carry out a sensitivity analysis by increasing the costs and decreasing the revenues by 10% and then recalculating the GM to see whether crop and livestock production still gave a positive GM in the face of changing cost and revenue scenarios, as suggested by Gittinger (1972) .

Measurement of household food insecurity status For household food insecurity analysis, a combination of analytical tools was used . To analyze the food security status of households, a cost-ofcalorie index was constructed , followed by the use of a Logit regression model to identify the major determinants of household food insecurity. Cost-of-Calorie Index The cost-of-calorie method proposed by Greer and Thorbecke (1984) was used in this study to determine a threshold food security line. The method yields a threshold value that is usually close to the minimum calorie requirement for human survival. Two steps-identification and aggregation-are involved in constructing the index . Identification is the process of defining a minimum level of nutrition necessary to maintain a healthy living. This minimum level is referred to as the " food insecurity line". In the context of this study, the food insecurity line is the calorie level below which people are classified as being food insecure or are subsisting on inadequate nutrition in the study area. Calorie adequacy was estimated by dividing the estimated calorie supply for each household by the household size, adjusted for adult equ ivalent, and using the consumption factors for various age-sex configurations . This method has been applied in several studies with a main focus on food security (Hasan and Babu 1991 ; Makinde 2001). Following the method , the food insecurity line is given as: LnX = a + bC

(3)

where X is the adult equivalent food expenditure and C is the actual calorie consumption per adult equivalent in a household. The recommended minimum daily calorie requirement per adult equivalent is 2250 kcal (see FAO 1982; Food Basket 1995) and this was used to determine the food insecurity line, using the equation:

(4 )

e (II+tt..)

where, S = the cost of buying the minimum calorie intake requirement (i.e., the food insecurity line ); a and b = parameter estimates in equation 3, and L = recommended minimum daily calorie intake level.

Based on the calculated S, households were classified as being food secure or food insecure, depending on which side of the line they fell. But due to differences in household composition in terms of age and sex, there was a need to calculate the levels of expenditure required in households with different age-sex compositions, The approach used to achieve this was to divide household expend iture by household size to get the per capita expenditure as used by the World Bank (1996) and in several other stud ies. The household expenditure was then decomposed on per adu lt equiva lent basis, using the conversion fa ctor developed by Institut Panafricain pour Ie Developpement (1981 ) and as adapted by Storck et al. (1991). The factors are presented in Annex 1. Determinants of food insecumy

The determinants of food insecurity of households were also analyzed. For this , a Logit regression model was used. The model is presented as fol lows : (5)

I I = b0

+ ÂŁ... ~bX J JI

(6)

j=1

Where Y, is the observed response for the ith household . It is a binary variable in which Yi = 0 for a food secure household and Y, = 1 for a food insecure household . I, is an underlying and unobserved stimulus index for the ith household. Conceptually, there is a critical threshold, (, for each household. If I, < ( , the household is observed to be food secure, but if I., ~ [", the household is observed to be food insecure. Equation (6) states the functional relationship between the response (Y,) and the stimulus index (I,) that determines the probability of being food insecure . Equation (7 ) states that I, is, in turn, detemnined by a vector of explanatory variables (X,),

1- P;

j ='

I , =In--=b o + Lbi X J; (7)

The Logit model assumes that the underlying stimu lus index (I,) is a random variable that pred icts the probability of being food insecure. Therefore , for the Ith household , the relative effect of each explanatory variable (X,,) on the probability of being food insecure is measured by

differentiating P; with respect to X" using the quotient rule (Green and Ng'ong'ola 1993). That is:

(8)

where, P; =the probability of an ith household being food insecure . X,. is the vector of explanatory variables in equations (6-8) and is defined as follows. AGE FARMINC FARMSZ HHSZ FAMEX COOP

= = = = = =

Age of head of household (in years). Farm income of a household (in naira per annum). Farm size of a household (in ha l . Farmer's household size. Famning experience (in years). Cooperative society membership (D

= 1 if farmer

is a member; D = a otherwise). EDUC DIST GEND

= = =

Level of farmer'S education (in years). Distance to input source (in km) . Gender of head of household (D = 1 for male; D=

DIVER ASSETS FARMEN

= = =

a for female).

Diversifi cation index (using the Herfindal index). Total value of household's disposable assets (in naira). Household production enterprise portfolio

= 1, if household is engaged in farming enterprises alone ; D = 0 otherwise). (D

CREDIT

Household head's access to credit facilities (D = 1 if there is access; D = 0 otherwise).

CDR

EXTAG

Ch ild dependency ratio. Household head's access to agricultural extension services (D = 1 if there is access; D

=0 otherwise).

EXCOM

Extent of famn production commercialization

REMIT

Total annual value of remittances received per

(Le., proportion of farm produce sold). adult eq uivalent by a household (in naira ). HLAB FLAB D

= = =

Hired labor (in standard days). Family labor (in standard days). 0-1 dummy variable .

The diversification variable (DIVER) was measured by using the Herfindal index defined as:

DIVER

" : = I,R

(9)

j= 1

where ,

( 10)

Where A, = share of farm revenue from enterprise i operated by the farm household and n

= number of farm enterprises owned by

household.

Measurement of household poverty status In the context of this study, poverty is defined as the inab ility of household to satisfy its basic needs for food, clothing and shelter, its inability to meet its social and economic obligations, its lack of gainful employment, its deprived access to basic facilities such as education, health, potable water and sanitation, and, hence , its restricted welfare status (Obadan 1997; Englama and Bamidele 1997). To determine the poverty status of households in the study area, a poverty line was constructed, using two-thirds of the mean per adult equivalent expenditure, below which a household was classified as being poor and above which a household was classified as being non-poor. The use of monetary income or consumption to identify and measure poverty has a long tradition , right from the study of Rowntree (1901) up to the recent World Bank's (1996) study on global income poverty. One interesting thing , however, is that most of these studies shared common approaches and methods. These studies were based on household income and expenditure surveys and this has made the approach to become the standard for quantitative poverty analysis (World Bank 2001). In his early study, Rowntree (1901) defined poverty as a level of total eamings that is insufficient to obtain the minimum neceSSities of life (i nclud ing food, house rent, and other basic needs) and for the maintenance of physical efficiency. He generated different poverty lines for different families, depend ing on their sizes, and compared these

with their earnings to arrive at their poverty status . The World Bank, on the other hand , has been assessing global income poverty by using expenditure data collected through household surveys. This is because consumption level, which is reflected in consumption expenditure, has been conventionally viewed as a preferred welfare indicator. Also, for practical reasons of reliability, consumption expenditure levels are thought to better capture long-run welfare levels than current income levels (World Bank 2001) . However, the

I ~erature

is expl icit on th e fact

that consumption expend iture may not fully capture a household's or an ind ividual's command over goods and services; but in the absence of more practical approaches, consumption expenditure has become the most w idely used variable for determining the poverty line (Omonona 2001; World Bank 2001).

Empirical model for determinants of household poverty status For the determinants of household poverty status, a Tobit regression model was conceptualized . The full model, which was developed by Tobin (1958) , is expressed in equation 11, following McDonald and Moffit (1980) and as adapted by Omonona (2001) . V',

Vi Vi

= = = =

~TX, +e, Q ~V;' $ Q

V'ifV'>Q , , 1, 2, ..... ...n

(11 )

=Lim ited dependent variable depicting the depth of household poverty, Xi =Vector of explanatory variables, ~T =Vector of unknown parameters, ei =Independently distributed error term.

where Vi'

The limited dependent variable V is defined as: (12)

(Z - Y,)fZ where Z

= Poverty line, and Yi =

Mean household food expenditure per

adult equivalent. The vector of explanatory variables is as defined earlier for equations

(6-8).

The empirical model in equation 11 was used to draw inferences on the causal factors for household poverty" The probabilities of being poor and the depth or intensity of poverty in the context of household characteristics (as captured by the >\s) were obta ined from the Tobit regression estimates" Following the Tobit

decompos~ion

framework suggested by McDonald

and Moffit (1980) and as adapted by Omonona (2001), the Tobit model can be further disaggregated to determine the effect of a change in the value of the ith variable on change in the probability of a household being in poverty and the expected depth of the poverty. For it can be shown that: E(V ) = F(Z) E (V,"),

(13)

where E(V,") is the expected value of V, for those households that are already poor, and F is the cumulative normal distribution function at Z, where Z is

X~ I15 .

The effect of a change in the level of any of the

household characteri stics (represented by the independent variable X,l, on the poverty level of a household can be decomposed into two, by differentiating equation (13) with respect to the specific household characteristic (>\) That is:

Multiplying by XIE(V). the relationship in (14) above can be converted into elasticity forms : aE(V,lIa>\x/ E(V,l

=F(Z){aE(V,")laX)X,I E(V,l +E(V,"XaF(Z)laX)

(15)

Rearranging equation (15) using equation (14), we have: (aE(vyaX}X,IE(V,) = (aE(vyaX) >\1 E(V: ) +(aF(Z)l aX,lX,IF(Z)

(16)

Therefore, the total elasticity of a change in the level of any explanatory variable (X,l consists of two effects: i.

change in the elasticity of household poverty intensity, and

change in the elasticity of the probability of being poor.

A prior; expectations for the explanatory variables in food insecurity and poverty status models The a priori expectations with respect to the explanatory variables in both household food insecurity and household poverty models are presented as follows, based on a classification of the variables into those relating to household demographic characteristics, household economic characteristics or activities, institutional factors affecting household poverty status , and household vulnerabil ity factors .

Household demographic characteristics These include the marital status and gender of head of household , household size, age of household head, and level of education of household head . Others include child dependency ratio and adult dependency ratio. Marital Status: The relationship between the marital status of a

household head and the food insecurity and poverty levels of the household has been found to be ambiguous in the sense that it can

be positive or negative. For while some studies show that households whose heads are married tend to be poorer than those with unmarried heads (Omonona, 2001), other findings show a contrary tendency (FOS, 1999). Similarly, studies have not been able to consistently associate the level of household food insecurity with the marital status of the head of the household. Gender: The relationship between the poverty status of a household and

the gender of the head of the household cannot be determined a priori. This is because, while several studies have revealed that female-headed households are likely to be more food secure and less poor than maleheaded households (World Bank, 1992; Makinde, 2001; Omonona 2001 ), others have found that female-headed households are likely to be more food insecure and poorer than male-headed households (Bennett, 1992). Household size: Household size could have a negative or positive

correlation with household food insecurity status and the probability and intensity of poverty, depending on the dependency ratio, which is usually positively correlated with the intensity of poverty because the larger the number of less active adults (e.g. , the old or the unemployed) and children in a household, the heavier the burden of the active members in meeting the cost of min imum household nutrition and, hence , the

higher the level of food insecurity and the probability or intensity of poverty, and vice versa (Hassan and Babu 1991; World Bank 1996; FOS 1999; Omonona 2001). But, on the other hand, household size may be negatively correlated with household food insecurity and poverty status if the household dependency ratio is low. Age: The relationship between the age of the head of a household and household food security status and poverty level may be difficult to determine a priori; for while the age of the head of household has been found to be negatively related to the probability and intensity of poverty in many studies, the World Bank (1996), Dercon and Krishman (1996), and Omonona (2001) have found the age of the head of household not to be a significant determinant of the poverty level among farming households. Also, no consistent pattern of association has been found between the age of the head of a household and the food security status of the household. Formal education: The level of the formal education of a household head would tend to be a positive factor in the adoption of improved farm production and management techniques. Hence, it is hypothesized that the educational status of the head of household is positively correlated with household income-earning capacity and , therefore, negatively correlated with the food insecur~y and poverty status of the household . Several studies have revealed that the incidence of food insecurity and poverty is higher among people with little or no education (World Bank 1996; FOS 1999; Omonona 2001).

Household economic characteristics (household productionl consumption activities) Household economic characteristics are described in terms of the employment status of the head of household, the level of food selfsufficiency of the household, and the ratio of food expenditure to total expenditure in the household . Other characteristics include the size of family labor, size of hired labor, household production enterprise portfolio (i.e., whether or not the household engages in other occupations outside farming), total household farm size, access to agricultural extension services, degree of diversification in farm production , and period of farm cultivation.

Employment: The employment status of the head of a household is crucial to the well being of the household . Employment status is expected to be negatively correlated with poverty intensity, that is, if a household head is employed , the probability and intensity of poverty would tend to be lower. Empirical studies have confirmed this negative correlation between the employment status of a head of a household and the poverty status of the household (Hassan and Babu 1999 ; FOS 1999; Omonona 2001 ). Similarly, the employment status of the head of a household is expected to be negatively correlated with the food insecurity status of the household . Level offood self-sufficiency: Level of food self-sufficiency is defined as the ratio of the quantity of own-produced food consumed to the total quantity consumed by a household. This variable is expected to be inversely related to the household food insecurity level. That is, a household that produces a large share of its food consumed would tend to be more food secure than a household that is more dependent on food purchases, as it is likely to be more able to cater for itself in times of general food shortages. Hasan and Babu (1991) have also found that there exists a negative relationship between the food self-sufficiency ratio and poverty level of households in rural Sudan . Ratio of food expenditure to total household expenditure : Th is is hypothesized to have a positive correlation with the household food insecurity level and poverty status, based on Engel's law which says the higher the income, the lower the proportion of such income that is spent on food, and vice versa. Studies have shown that the higher the ratio of food expenditure is to total household expenditure (or income), the higher are the probability and intensity of poverty and food insecurity (Hasan and Babu 1991 ). Household labor size: The num ber of persons available for work in a household determines the total income of the household. Therefore , the number available for family labor may be inversely related to the food insecurity level and poverty status of the household, ceteris paribus, as higher income will leave the household with higher disposable income to acquire more food and other household goods and services that will enhance the welfare of the household. However, because large

household labor size is often associated with la rge household size, and vice versa, there may also be a positive correlation between household labor size and household food insecu rity level and poverty status (World Bank 1996).

Hired labor size: The amount of hired labor that a household employs has a negative effect on the disposable income ava ilable to the household and, hence, has a positive relationship with the food insecurity and poverty status of the household . Literature has shown that the greater the amount of hired labor employed by a household , the lower the disposable income available to the household and, hence, the more intense poverty in the household is likely to be, and vice versa (Dereen and Krishman 1996). However, the amount of hired labor may also exhibit a negative correlation w ith food insecurity and poverty status as it may tend to increase household productivity and ineeme in the absence of adequate family labor to work during the peak farming season (Reardon 1997; Leavy and White 2000).

Household production activities or occupations: Among the rural people, those who engage in agriculture as a single and sole source of ineeme tend to be poorer and more prone to food insecurity than those who combine agricultural and nonagricultural activities. That is, rural households, which are engaged in other occupations, in addition to farming , are often less poor and less prone to food insecurity than households that are engaged in farm ing alone (Omonona 2001). Hence the literature has reported an increased engagement of rural households in nonagricultural ineeme-earning activities in recent years in Nigeria (Meludu 1993; Jambiya 1998; Yunusa 1999).

Total household farm size: The total size of a household farm is inversely related to the food insecurity level and poverty status of the household. It has been found that, given other factors, the larger the farm size of a household, the lower the probability of such a household being poor, and vice versa (FOS 1999; Omonona 2001). Similarly, the larger the size of a household's farm, the lower the probab ility of the household being food insecure, and vice versa.

Degree of diversification of farm production: Generally, agricultural intensification and diversification have often been recommended to promote income growth and stability, Agricultural enterprise diversification ensures that the farmer derives income from a wide range of sources, thereby redu cing income instability. Diversification has, therefore, been found to reduce household food insecurity and poverty level among farming households in Nigeria (Omonona 2001). Period of annual farm ing activities: The length of the period of the year in which a farm household ca n engage in farming activities is another important factor that affects the household's poverty status. A household that is able to produce or cultivate throughout the year is expected to be better off than the household whose production period is dependent on the rains and restricted to the rainy season alone , The length of period of farm ing activity also affects the amount and stability of off-farm income available to the household and , therefore, influences the food security and poverty status of the household (Hasan and Babu 1991). Degree of commercialization of farm production: The degree of commercialization of farm production, which is measured as the percentage of total farm produce marketed, influences the food insecurity status of a household, It is hypothesized that the extent of agricultural output commercialization would tend to be positively related to household food insecurity, since most households in the project area might not be able to generate agricultural products in adequate quantities to provide marketable surpluses after making allowance for home consumption , Hence, such households might often be forced to sell food output meant for home consumption to meet household needs, such as the education of children and medical expenses. This may tend to reduce the food security status of the household. The literature has shown that the poverty status of a household and the extent of commercialization of farm production may be positively or negatively related, depend ing on types of enterprise combination. But Omonona (2001 ) found a negative relationsh ip, whereby the higher the extent of commercialization, the higher the income accruable to the household and, hence, the lower the probability of poverty, given other factors.

Institutional factors Institutional factors, which influen ce or refiecl the levels of household food insecurity and poverty, include membership of cooperative societies and access to farm services (extension services , credit facilities, fertilizer supply, etc. ). Membership of cooperative societies: Membership of one or more

cooperative societies is beneficial in many ways. It improves the access of members to many facilities , which can enhance their farm productivity and income . Studies have shown that membersh ip of cooperative societies or farmers' associations exhibits a negative correlation with the food insecu rity and poverty status of a household. It has been shown that the probabi lity of households whose heads are members of cooperative societies being food insecure or being poor is lower than that of households whose heads who are not members (FOS 1999; Omonona 2001). Access to farm services: Farm services , such as agricultural extension

services, credit facilities, efficient supply of improved inputs (e .g., fertilizers, herbicides and planting materials), are very important in enhanci ng farm productivity. Hence, it has been established by previous studies that improved access to all or some of these services would tend to reduce the incidence and severity of household food insecu rity and poverty (FOS 1999; Omonona 2001). Vulnerability factor

The level of valuable, disposable assets owned by a household is regarded as a factor determining the degree of vulnerability of the household to food insecurity, poverty incidence, and poverty intensity. This is because these disposable assets can be easily sold in times of need and , to that extent, constitute a determ inant of the probabil ity of a household being food insecure or poor, These assets could also serve as a measure of household wealth and, hence, the level of their possessions is expected to have a negative relationship with poverty status , That is , given other factors, the higher the value of disposable household assets possessed , the lower the probability of household food insecurity and poverty, and vice versa (Bender 1991 ; Barney and William 1994),

Descriptive analysis of household characteristics

In this section, the major socioeconomic characteristics of households covered in the survey are described. In the main, these characteri stics relate to the relative frequen cy distribution of heads of households by gender, age, years of formal education and marital status. Also included are household asset ownership structures, size distribution of household farms, types of land tenure, sources of farm credit, types of crops grown, composition of household livestock (an imal and poultry stock). household farm income distribution, household non-farmi ng employment and income distribution, and household food and nonfood consumption patterns.

Distribution of household heads by gender The distribution of household heads by gender is shown in Table 1. The pattern of gender distribution of household heads was similar across the agroecological zones surveyed. But. in relative terms , the percentage of male-headed households wa s higher in the 55 than in the 5G5 and NG5. The percentage was lowest in the NG5 . On the other hand, the percentage of female-headed households was highest in the 5G5 and lowest in the 55 . But. on average, about 76% of the households covered in the survey were male-headed while about 24% were fema le-headed .

Distribution of household heads by age Age has been found to determine how active and productive the head of the household would be. Age has also been found to affect the rate of household adoption of innovations that, in turn . affect household productivity and livelihood improvement strategies (Dercon and Krishman 1996). Table 1. Percentage distribution of household heads by gender. Agroecological zones

Gender Female Male Total

SGS

NGS

Average (all areas)

10.2 89.8 100 .0

28.2 71 .8 100.0

24 .6 75.4 100.0

24 .3

Source: Field survey, 2004.

75.7 100.0

Table 2. Percentage distribution of household heads by age category and gender. Agroecological zones

Age category

30 or below 31-40 41-50 51-BO Above 60

SGS

NGS

Average (all areas)

Female

Male

Female

Male

Female

Male

Female

Male

21 .2 18.4 2.7 4.7 0

78.8 81.1 97.3 95.3 0

36.8 31 .4 34.4 21 .4 13.1

63.2 68.8 65.5 78.6 86.9

38.7 22.6 23.9 21 .6 14.0

61 .3 77.4 76.1 78.4 86.0

36.2 25.2 24.7 18.8 127

63.8 74.8 75.3 81 .2 87.3

Source: Field survey, 2004.

Table 2 shows the distribution of household heads by their age ranges. The distribution of the age of household heads was fairly similar across the ecological zones surveyed . But on average, about 54% of the household heads were between 31 and SO years of age . The mean age of household heads was 46 years (with a standard deviation of 13.68) . In the whole population , about 70% of household heads were in the active and productive age range of under 50 years . The predominance of active and productive heads of households in the project area has direct bearing on (1) increased availability of able-bodied labor for primary production; (2) ease of adoption of innovations; and (3) reduction in the degree of risk aversion. All these have great potentials for increasing agricultural productivity and production and , hence, for improving household livelihood and reducing poverty.

Distribution of household heads by years of formal education The level of education determines the level of opportun ities available to improve livelihood strategies, enhance food security, and reduce the level of poverty. It affects the level of exposure to new ideas and managerial capacity in production as well as the perception of the household members on how to adopt and integrate innovations into the household's survival strategies. Table 3 shows the distribution of the levels of fomnal education among household heads .

Table 3. Percentage distribution of household heads by years of formal education. Agroecological zones SGS

SS Education (years)

NGS

Average

(all areas) Female Male

Female

Male

Female

Male

Female

Male

10.1

89.9

38.0

62.0

35.8

64.2

32.1

67.9

8.8 14.3 7.1

91.2 85.7 92.9

19.8 19.2 19.4

80.2 80.8 90.6

14.4 14.1 12.3

85.6 85.9 87.7

16.2 16.4 16.4

83.8 83.6 85.8

No forma l education Up to 6 7-12 Above 12

Source: Fteld survey, 2004.

Table 4. Percentage distribution of household heads by marital status.

Agroecological zones SGS

NGS

1.2 42.7 54.6 1.5 0.0 100.0

0.7 53.3 44.0 1. 5 0.5 100.0

0.9 39.1 58.5 1.2 0.3 100.0

Marital slatus Single Monogamous Polygamous Widowed Oil/orced Total

Average (all areas) 0.8 45.5 51.6 1.8 0.3 100.0

Source: Field survey, 2004.

The majority (51%) of household heads had no fo rmal education . In all the agroecolog ical zones in the project area , the pattern of the distribution of the levels of formal education of household heads was similar. However, the highest level of illiteracy was found in the SS with about 73% of respondents having no formal education at all, while the SGS and NGS had relatively well-educated household heads.

Distribution of household heads by marital status The distribution of the marital status of household heads in the study area is presented in Table 4 . There was a high level of homogeneity in the distribution of household heads' marital status in the project area due to similarities in cultural and religious practices. The majority of household heads in the study area were polygamous, except in the SGS where about 53% were monogamous. However, substantial percentages of household heads were also monogamous in the SS and NGS. On average , about 45% of all household heads in the project areas were monogamous .

On the other hand , on ly a negligible percentage of household heads were single, widowed , or divorced . Furthermore , all the single household heads were male, except in the SGS where a small nu mber (about 14) were women. But all the widowed and divorced household heads were female .

Distribution of household heads by occupations In the study, different farming and non-farming occupations of household heads were identified. As practiced in many rural econom ies in Nigeria, the nural farm ing households in the project area had hig hly diversified income-generating activities. The major types of these incomegenerating activities o r livelihood strategies that were identified in the course of the study are presented in Table 5. The distribution of occupations was similar across the agroecological zones. It is also evident tha t crop farm ing was the most important occupation of the heads of households in the project area, while a civil service job was the second most important occupation . Combined crop and livestock fanmi ng was also rel atively important. It could, however, be observed that, on the whole , hunting , artisan work, trading, and non-farming wage labor were relatively minor occupations in the project area. Table 5. Percentage distribution of household heads by their main occupations.

Agroecological zones 5GS

NG5

96.8 0.9

88.5 7.7

91.1 1.0

Average (all areas) 90.1 3.4

8.6 50.0

21 .1 41.3 4.3

7.7 16.7

8.1 3.8

14.7 40.0 3.5 9.7 6.3 11 .1 to .5

19.9 38.8 2.6 3.7 1.9 9.5 13.7

Occupation Crop farming Livestock farming Combined crop and livestock farming Civil service job Trading Arti san work Hunting Non-fa rm ing labor Others'

Note : 路Other occupations include farm wage labor and casual non-farming labor. Percentages do not add up Lo 100% due to multiple occupations. Source: Field survey, 2004.

It is interesting to note that livestock farm ing alone was not an important occupation of household heads. On the other hand , combined crop and livestock farming was more favored. This was because of the complementarity between crop and livestock enterprises. On the whole. therefore, efforts to improve household livelihood strategies in the project area wou ld have to focus on improving combined crop and livestock farming . But sight shou ld also not be lost of the increasing importance of non-farming employment opportunities as alternative rural household livelihood im provement strategies in the area.

Household asset ownership structure Hassan and Babu (1991) have found that the level of asset ownership in a household is an indication of its endowment and provides a good measure of household resilience in times of food crisis . resulting from fam ine, crop failures , or natural disasters. This is because a household can easily fa ll back on its assets in times of need by selling or leasing them . The asset ind ication of the households covered in the study is presented in Table 6 . Land was the asset predominantly owned by households . Th is is not surprising as land is the basic resource required for fanming . But the sma ll proportion of households that owned ox plows and work bulls in the project area was indicative of the fact that most farming households did not practice mechanized or semi-mechanized fa rming . Instead , they still relied on hand implements in their farming activities. Table 6. Percentage distribution of household asset ownership. Agroecological zones SS

SGS

NGS

Average (all areas)

78 .8 61 .3 8.8 19.0 87.2 31 .0 29.2 8B.7 12.5 5.0

93.1 70.0 5.2 16.6 32 .1 7.2 6.3 86.0 39.3 14.6

88.5 79.6 6.5 19.7 58 .1 17.9 7.9 55.6 12.7 20.0

86.3 65.1 6.3 18.7 49.4 12.6 10.3 78.8 25.0 12.4

Assets La nd Livestock Car Motor cycle Bicycle Ox plow Work bull House Bank accounl Sewing machine Source: Field survey, 2004.

Number of farm plots owned by household heads and their spouses Because of some practices associated with traditional land tenure systems, many fa rm households in Nigeria still operate fragmented plots. Hence, the number of farm plots owned by households in the project area was investigated and reported (Table 7). The pattern of distribution of farm plots owned by households is similar in the three agroecological lanes. An average household head had an average of three plots and provided labor for an average of three plots , which means that the household, on average , was able to provide sufficient labor for most of its own plots and this confirms the fact that most of the households in the project area were able to supply the bulk of their own farm labor. The average total numbe r of farm plots owned by a household in the study area was about six, if the number of farm plots owned by spouses are added to those owned by the household head . This validates the fact that women often own a sizeable amount of farm assets in the area. Table 7. Distribution of number of farm plots owned by households.

Agroecological

Number owned by househotd head

Number owned by head's spouse(s)

Mean Standard deviation Minimum Maximum

2.7 0.8 1.0 4.0

2.6 2.2 1.0 15.0

Mean Standard deviation Minimum Maximum

3.5 1.6 1.0 12.0

2.9 1.8 1.0 12.0

NGS

Mean Standard deviation Minimum Maximum

3.0 1.0 1.0 6.0

2.7 1.5 1.0 8.0

Average (all areas)

Mean Standard deviation Minimum

3.1 1.4 1.0 12.0

2.7 1.7 1.0 15.0

Statistics

zones SS

SGS

Maximum Source: Field survey, 2004 .

Table 8. Average household fann sizes (ha). Agroecological zones 5S 5GS NG5 All areas

Average area (ha)

Standard devia1ion

Minimum size

Maximum size

4.2 3.7 4.3 3.7

1.7 1.2 0 .9 1.6

0.2 0 .01 0.1 0 .02

20.0 25 .0 B.O 41.0

Source: Field survey, 2004.

Household farm sizes The average sizes of household farms in the project area are shown in Table B. Nigeria is known to be a nation of small farmers who often operate on fragmented farmlands. As shown , the average farm size in the project area ranged from 3.7 ha in the SGS to 4.3 ha in the NGS , thus showing large disparities in farm sizes among the farming households in the project area. However, there were many farm households that operated small and fragmented plots in the project area . A striking finding from Table B is that the disparity between the minimum and the maximum famn sizes was large, with a minimum size of only 0.1 ha and a maximum size of 41 ha for all areas covered in the study. In sum, Table 8 reveals that the average farmer operated small fragmented plots that added up to an average of about 3.7 ha per household .

Farming experience of household heads (years) Farming experience is an important factor determining both productivity and production level in farming . But the effect of farming experience on productivity and production may be positive or negative. Generally, it would appear that up to a certain number of years, farm ing experience would have a positive effect while, after that, the effect may become negative . The negative effect may be derived from aging or reluctance to change from old and accustomed famn practices and techniques to modern and improved ones.

Table 9. Average farming experience of household heads (years). Agroecological zones

Mean

SS SS NGS All areas

26.5 26 .7 2B .l 25 .2

Standard deviation

M inimum experience

Maximum experience

12.8 15.2 30.3 16.9

1.0 1.0 1.0 1.0

60.0 80.0 70 .0 80.0

Source: Field survey, 2004.

Table 10. Percentage distribution of households by types of land tenure. Agroecological zones

Tenure Inherited Rented Pledged Borrowed Gift Purchased Sharecropping (rent in kind) Others Total

SGS

NGS

77.4 2.6 6.3 4 .1 2.6 7.0 0.0 0.0 100.0

76.1 2.2 1.0 1.8 5.3 10.9 0.2 2.5 100.0

76.8 4.9 4.9

1.9

2.7 6.8 0.0 0.0 100.0

Average (all areas) 69.9 3.8 2.3 2.6 4 .1 15.2 0.6 1.4 100.0

Source : Field survey, 2004.

The farming experience of household heads in the project area varied widely, with a minimum of only one year and a maximum of 80 years. The average farming experience, however, did not vary widely among the agroecological zones as the variation was between 27 years in the 55 to about 28 years in the NG5. The average for all LGAs was about 25 years . This implies that the average farm household head had considerable experience and this might have a negative impact on the attitude to change in farming practices.

Distribution of households by types of land tenure Land tenure systems in Nigeria vary widely among various parts of the country. But the predominant systems revolve around community and family ownership, with members having due access on a usufruct basis. The typology of land tenures practiced in the project area is presented in Table 10.

The typology of land tenure system in the project area was simila r across all the agroecolog ical zones . The commonest type of land tenure was ownership by inheritance from family or community. This tenure system often involved inheritance to use rather than to own in fee simple . It is also an important facto r in land fragmentation as land is subdivided and shared among the children of deceased parents . The next most important tenure system was by land purchase. The sharecropping system that is common in other parts of the country (most especially in the southwest) was not widely practiced in the project area.

Distribution of sources and amounts of credit obtained for farming The availability of credit to farm households is vital to agricultural production . In this study, the distribution of households that obtained credit, the sources , and amounts obtained were analyzed . The analysis showed that the pattern of distribution of households that obtained credit for fanning purposes was similar across the project areas. It was, however, revealed that most households in the project area did not obtain credit in the 2003 planting season . Overall. very few farm households (less than 1%) were able to obtain credi t in that year. The reason for th is might be the usual problems, either lack of access to credit institutions experienced by most fanners or difficu lties associated with loan applications , especially the problem of bureaucratic delays. This might have had negative implications for agricultural production in the project area in the year. Generally, credit sources in the project area were commercial banks , specialized banks fo r agricultural loans, Borno State Agricultural Development Project (BOSADP), cooperative societies , individual moneylenders, and relatives and friends. But virtually all these credit sources failed to make any significant impact on the credit needs of the farmers in the project area.

Distribution of the most important crops grown by households The distribution of the most important crops cultivated by households in the project area is presented in Table 11 .

Table 11. Percentage distribution of households growing various crops . Agroecological zones SS

SGS

NGS

Average (all areas)

31 .6 0.7 23.4 31 .2

85.5

71 .9 0.7 15.2 10.7 0.7

72.9 0.2 9.8 7.4 0.3 0.0 0.6 0.2 1.8 0.0 0. 1 0.1 0.0

Crops Maize Millet Sorghum Groundnut Cotton Soybean Cowpea Pepper Rice Tomalo Bambara nut Cocoyam Garden egg

2.6 3.4 0.1 0.1 0.5 1.8

0.4

1.5 0.4

3.3

0.4

0.1

0.4 0.4 0.1

Source: Field survey, 2004.

T he pattern of distribution of the important crops grown by households was similar among all the agroecologica l zones in the project area. The most widely cultivated crops were the cereals. Of these, maize was the most important and most widely cultivated . Legumes were next to the cereals in terms of relative importance , as reflected in the percentage of households growing them, especi ally cowpea and groundnut. On the other hand , vegetables and roots and tubers were not w idely grown . In sum, it cou ld be con cluded that the most important and most widely cultivated crops were the cereals and this may be connected with the food or dietary habits of the population in the area as well as the local climate, which favored the cultivation of cerea l crops.

Continuous cropping of farm plots Due to increasing population pressure on farmland and some traditional land tenure practices , continuous cropping of farmland is becoming an important feature of farming in many parts of the country. This is in contrast to the age-old practice of shifting cultivation. In the project area , this change was also being observed. Hence, it was found that the average number of years of continuous cropping of farmland ranged from about 9 years to about 19 years across the project LGAs. This average was lowest in the NGS (about 9 years) and highest in the SS (about 16 years).

The increasing incidence of continuous cropping and increasing intensijy of land use could be expected to affect the fertility of farmland. This raises the issue of the need for the introduction of new, yield-enhancing technologies for sustainable growth in farm productivijy and production in the project area . It is , therefore , not surprising that one of the most important strategies adopted by PROSAB to promote agricultural growth in the project area involves the continued introduction and promotion of improved farm technologies and farm management practices.

Types of crops no longer grown The types of crops that farm households bring into cultivation or drop from cultivation may be an indication of the perception of the value or importance of various crops to household livelihood strategies. Hence, in this study, households were asked to indicate the crops that they had hitherto cultivated and that had been dropped from cu ltivation in recent years. An analysis of the information provided showed that cotton , sorghum, rice, and millet were the crops most w idely abandoned in the project area. However, while the discontinuation of the cultivation of cotton, rice , and sorghum was fairly equally spread in all the four LGAs covered in the study, millet was more often dropped from cultivation in the SS agroecological zone. On the whole, about 22% of all farm households had dropped cotton from production in recent years ; about 17% had dropped sorghum; about 15% had dropped rice, while about 10% had dropped millet. Normally, cotton and rice are considered to be commercial crops in the project area while millet and sorghum are stap le crops. The dropping of cotton and rice may have someth ing to do with the fact that they are relatively unprofitable to grow while that of millet and sorghum may be due to households' changing food security strategies (e .g., increased reliance on the market for their supply for household consumption compared with reliance on households' own-production). Reasons for dropping various crops from production were many and sometimes interrelated, but the reasons and the percentages of households citing each reason are summarized in Table 12.

Table 12. Percentage distribution of reasons given by households for dropping some crops from cultivation. Agroecological zones

SG5

NGS

Average (all areas)

2.6 62.0 7.9 1.5 1.1 2.4 8.3 1.6 10.9 2.0 100.0

0.3 63.2 8.0 3.5 2.2 2.2 8.2 0.1 11 .1 1.1

1.5 49.6 7.3 4.4 8.6 3.6 7.5 1.7 13.0 2.6 100.0

1.3 58.3 7.6 3.1 4.0 2.8 8.0 1.1 11 .6 1.9 100.0

Reason Poor food security value Poor soil fertility and low crop yield Disease , pest, and animal predation

Low market prices Poor market demand for produce Land constraints Labor constraint

High production cost Inadequate credil to purchase inputs Others Total

100.0

Source: Reid survey, 2004.

The reason most mentioned for abandoning the

cu~ivation

of some

crops was poor soil fertility, leading to low crop yields. On average, about 54% of households abandoned crops for th is reason in all the LGAs. Next, in relative order of importance, was inadequate access to credit facilities to purchase required inputs, cited by an average of 12% of all households . Labor constraints, high incidence of disease, pest and animal predation, and poor market demand for produce were some of the other important reasons for dropping crops from cultivation . These reasons point to a number of important policy issues in the agricultural development of the State . First, the problem of inadequate access to agricultural credit points to the need for efficient credit institutions for susta inable agricultural development. Secondly, the interrelated problems of poor soil fertility and low crop yield in the project area call for measures to make fertilizers readily available to farmers and to put in place appropriate soil fert ility enhancing/conserving farm management practices, such as minimum tillage and the use of legumes and organic fertilizers. Thirdly, the problem of animal predation is of particular interest in the sense that most of the communal crises in the project area in particular, and in many parts of northern Nigeria in general, could be traced to disagreements over grazing rights between pastoralists and crop farmers. These have often led to the loss of an imal and human lives as well as crops and the problem thus deserves

special attention. Finally, the problem of poor market demand is partly a reflection of inefficient marketing systems and poor price mechanisms that call for the evolvement of better and more efficient marketing and pricing systems/practices.

Estimated gross margins in crop production A summary of the gross margin analysis carried out in the study is presented in Table 13. The NGS earned the highest gross revenue/ha from crop production while the SS earned the lowest. But generally, the variation in gross revenue across agroecological zones was marginal. The average gross revenueslha from crop production was about H98 500, implying that an average household earned almost .. 100 OOO/ha from its combination of crop enterprises. Wha t a household earned in gross revenue, however, depended on the types of crops cultivated, the yields of the various crops , and their producer prices. The variable costs of crop production did not vary widely across agroecological zones as shown in Table 13. This was not surprising, since most households could be expected to use similar variable input combinations purchased at fairly similar prices (especially fertilizer prices), and given that most parts of the project area practiced similar crop production systems . The most important item of variable cost was for fertilizers that varied from about"9 OOO/ha in the SS to about 1'0116 OOO/ha in the NGS. Fertilizer cost alone accounted for between 39% of the total variable cosUha in the SS and about 69% in the SGS. Table 13. Estimated gross margins IN/ha).

Agroecological zones Items Gross revenue Variable cast Gross margin Return per naira outlay (RPN)

SGS

NGS

Average (all areas)

1t 7.427.9 23,512.5 93,915.4

91,025.0

92,374.25 20,779.3 86,594.95 4.12

98.488.9 20,061. 1 78.427.8

3.99

18,105.7

72,919.3 4.03

Source: Computed from survey data .

3.91

The gross margin/ha from crop production is the difference between gross revenue and lotal variable cost. Hence, variations in gross margins are a reflection of variations in gross revenues and/or total variable costs. In this regard, it may be observed from Table 13 thai gross margins/ha varied marginally across agroecological zones from about N72 200/ha in the SGS to about N93 915/h a in the SS . But the average for all agroecological zones was N78 428/ha . Variations in RPN are a reflection of variations in both variable cost and gross margin as the RPN outlay is defined here as gross margin earned per naira of variable cost incurred. As shown in Table 13, RPN outlay was highest in the NGS and lowest in the SS. This implies that surplus eamingslha available 10 cover both fixed cost outlay and net farm revenue from crop production were highest in the NGS and lowest in the SS . Given that average fi xed costlha does not normally vary widely where crop production systems are similar (as in the project area), it could be inferred that the profitability of crop farming was also highest in the NGS, with relative profitability in the SGS and SS following in that order.

Factors determining profitability in crop farming Major factors identified by respondents as those determining profitability in crop farming in the project area are summarized in Table 14. The availability of a market outlet for produce was the single most important factor determining profitability in crop farming in the project area as a whole. It was , in particular, the single most important factor in the SS and NGS. High market prices for produce came second as the most important single factor in the project area as a whole . II was, in fact, Ihe most important faclor in Ihe SGS and the second mosl important in the NGS. Low cost of crop production was the third most important Table 14. Percentage distribution of major factors determining profitability in crop farming. Agroecolog ical zones Factors Availability of market for produce High price for produce Ava ilability of cheap labor Low cost of production Other factors Total

Average

SGS

NGS

(ali areas)

25.4 13.4 16.3 6.4 36.5 100.0

16.5 24.3 9.0 11 .5 36 .7 100.0

45.6 17.4 5.6 2.4 26.6 100.0

22.3 19.7 8.6 13.7 35.7 100.0

Source: Field survey. 2004.

factor in the project area as a whole as well as in SGS. The availability of cheap labor as a factor in the profitability of crop farm ing ranked high

in the SS. It should be pointed out that cheap labor itseij must have been a contributory variable to a low cost of crop production. As such, the availability of cheap labor and a low cost of production as factors determining profitability in crop production were closely related. Other determining factors which, together, were very important were the availability of credit and improved farm inputs (especially fertilizers), high crop yields, the availability of processing facilities for produce and of transportation. The policy implications of the forego ing include the need for policies to ensure adequate access by farmers to improved production inputs (especially fertilizers) which would increase crop yield and thereby reduce the unit cost of crop production, and policies to ensure adequate access to markets for farmers' produce. Constraints to profitability in crop farming, as identified by respondents, were, on the whole , the obverse of factors promoting profitability. Hence, the major constraining factors identified included poor market demand for produce, low market prices for produce, high cost of labor and of fertilizers. It is, however, important to point out that the high cost of fertilizers was the single most constraining factor identified by respondents in every LGA covered in the study. This suggests that the maintenance of soil fertility was a priority that should be accorded a special 路attention in programs designed to improve farm productivity and household livelihoods in the project area.

Composition of household livestock Livestock are an important livelihood asset in the project area hence the distribution of the different livestock species kept by households was studied. Household livestock covered consisted of animal stock species, made up of cattle, sheep, goats and pigs, and poultry stock species made up of chickens, guinea fowl, and ducks.

Animal stock species The average number of various animal stock species (i.e., cattle, sheep, goats, and pigs) kept by men in households is presented (Table 15), while the average number kept by women is in Table 16. Taken together, the two tables reveal the following situation.

Table 15. Average number of various animal species owned by men in households. Animal stock species

1. Cattle Adults Weaners Sucklings Totat 2 . Sheep Ad u ~s

Weaners Sucklings Total 3 . Goats

Average

(all areas) 8

5 3 16 10 5 3 18

Ad u ~s

Weaners Sucklings Total 4 . Pigs Adults Weaners Sucklings Total

8 3 4 15

6 3 3 12

7 3 3 13

6 3

7 3 3 13

2 11

6 3

4 16

2 1 8

2 18

Source: Field sU/vey, 2004.

Cat/Ie: On the ave rage, men in households in the SS and SGS areas

kept more cattle than those in the NGS. For women , the distribution of cattle across agroecological zones was more uniform. On the whole , men kept more cattle than women in the project area. Sheep: The average number of sheep kept by men was highest in the

SS and lowest in the SGS . But the ownership of sheep by women was fairly uniform across the agroecological zones . However, there was no significant gender difference in the number of sheep owned by men and women in the project area . Goats: Goat ownership among men was highest in the SS and lowest

in the SGS. But, on the whole, there was little disparity in the ownership patterns among men and women across the LGAs. Also, the average number of goats owned by men and women in all the LGAs was about the same. That is, there was no significant gender difference between men and women in the average number of goats owned in the project area as a whole . Pigs: Both men and women kept pigs in the SGS (Tables 15 and

16). The difference in distribution might be connected with religious prohibitions, as Muslims are normally not permitted to keep pigs.

Table 16. Average number of various animal species owned by wOmen in households. Animal stock species

Agroecological zones NGS SGS

Average

(aU areas)

1. Cattle Adults

1 1 7

4 2 2 8

Weaners

2 2 9

7 3 2 12

7 2 3 12

8 3 3 14

7 3 3 13

6 3 3 12

7 3 3 13

8 3 3 14

7 3 3 13

Sucklings Total 2. Sheep Adults Weaners

Sucklings Total 3. Goats Adults Weaners

Sucklings Total 4 . Pigs Adults Weaners Sucklings Total

5 5 4 14

1 1 1 3

Source: Field survey, 2004

Generally, both men and women in the project area had only small numbers of animal stock and most of these were kept under the freerange system . Th is is indicative of a poor state of development of the livestock industry in the project area.

Poultry species Household poultry stock species kept by households in the project area were chickens, guinea fowl, and ducks. The average numbers of these kept by men and women in households are pnesented in Tables 17 and 18. Taken together, the two tables reveal the following about the state of the poultry industry in the project area. Chickens: The average number of birds kept by men was about the

same in all the LGAs except in the NGS where the average was much lower. Similarly, the average number of birds kept by women was about the same in the SS and NGS. Thus, there was, however, no significant gender difference in the average number of birds kept by men and women in households across the agroecological zones as a whole.

Table 17. Average number of various poultry species owned by men in households.

Poullry stock species 1 . Chickens Layers Broilers Cockerels Chicks Local chickens Tolal 2. Guinea fowl 3. Ducks Total

Agroecological zones SGS NGS

Average

(all areas)

5 5 5 5 6

4 6 5

2 7 24

26 7 14

4 5 4

3 7

7 6 26 19 18 29

16 30 21 30

Source: Field survey, 2004 .

Table 18. Average number of poulbry species owned by women in households. Poultry stock species

Agroecological zones SG5 NGS

Average (ali areas)

1. Chickens Layers Broilers Cockerels Chicks Local chickens Total 2. Guinea fowl 3. Ducks

3 7 7

5 S

6 6 27 11

6 4

4 2 2 8 6 22 10 20

3 5 4

6 6

24 10 19

Source: Field survey, 2004.

Guinea fowl: Most of the guinea fowl kept by men were found in the SGS and NGS wh ile most of those kept by women were found in the SGS. On the whole, the distribution of guinea fowl was very unequal for both men and women across the agroecological zones, but on average, men kept more gu inea fowl than women . Ducks: The average number of ducks kept by men ranged between

13 in the SS to 30 in the NGS, with an average of 29 birds for all agroecological zones . Similarly, the range for women was between 12 in the NGS and 30 in the SGS, but with an average of 23 for all agroecological zones. On the whole , therefore, men kept more ducks than women in the project area.

Overall evidence from analysis indicated that the ownership of large-size poultry farms in the project area was rare. Much more common were small-scale poultry holdings reared mostly on a free-range basis.

Profitability of livestock farming Although quantitative data were not available for the estimation of gross margins in livestock production in the project area , respondents were asked for their perception of the profitability of livestock farming. Their responses indicated that the majority perceived livestock farming as being profitable. Specifically, about 52% of respondents in the NGS , 56% in the SS, and 52% in the SGS perceived livestock farming as being profitable. Across all the three zones about 54% of respondents perceived livestock farming as being profitable. The percentage distribution of households identifying various factors determining the profitability of livestock farming in the project area is presented in Table 19. The availability of a market for livestock products was the single most important factor identified in the NGS. But the single most important factor identified in the SGS was the high prices of products , while low cost of production was the most important factor in the SS. Overall, the availability of a market for products was the single most important factor in all agroecological zones, followed by high product prices, low cost of production, and availability of livestock feeds, in that order. Other factors, such as easy access to credit. availability of processing facilities, access to good transport facilities , and so on, were also identified. Table 19. Percentage distribution of households identifying factors determining the profitability of livestock farming. Agroecological zones

SGS

NGS

10.6 11.6 21 .0 30.3 26.3 100.0

19.4 27.7 14.4

35.7 12.3 6.8

Factors Availability of market for products High prices of products Availability of livestock feeds Low cost of production Other factors Total Source: Field survey, 2004.

9.0 29.5 100.0

Average (all areas) 21.1

6.5

19.5 13.7 17.2

36.7 100.0

26.5 100.0

Major factors constraining the profitability of livestock farming, in relative order of importance, were the problems of livestock diseases, high cost and/or scarcity of livestock feeds, drugs, and vaccines, lack of a market for produce, and low product prices . The foregoing point to the need for policies and programs designed to enhance producers' access to product markets and to control livestock pests and diseases. Also important is a policy to make livestock feeds, drugs, and vaccines easily available to farmers.

Household non-farming activities It is now common to find some levels of livelihood diversification in many rural households in the country. Hence, it was not surprising that the survey found many farming households in the project area were also engaged in diversified income-generating activities, involving both farming and non-farming activities. Their level of involvement in nonfarming economic activities was , however. not very high and the level also varied across the LGAs. Analysis revealed that on ly a mi nority of households was engaged in off-farm economic activities, meaning that the majority of households concentrated on fanming activities alone. On average, only about 18% of households in the SS, 32% in the SGS, and 43% in the NGS were engaged in non-farming economic activities. The average for all agroecologica l zones was about 38% of households . Off-farm economic activities of households covered activities both agricuijure-related and those not related to agriculture, although only a relatively small percentage of households were engaged in off-farm agriculture-related activities. In all, only about 18% of all households were engaged in agricu lture-related off-farm economic activities in the SGS . The corresponding percentages of households engaged in agriculture-related non-farming activities in other agroecological zones were 33 in the SS and 37 in the NGS. The average for all the zones was about 29% . Major non-farming economic activities of households included petty trading , civil service employment, artisanship, processing of agricultural products, marketing of agricuHural products, and so on . But the most widely pursued non-farming activity by households was the marketing of agricultural products . This activity was most prevalent in the SGS and

55 but least prevalent in the NG5. Thus, the marketing of agricultural products constituted the most important livelihood option besides farming. This should not be unexpected, as agricultural product marketing is very close and complementary to agricultural production activity

Household food consumption patterns The types of food items consumed by households in the project area were categorized into cereals, legumes, and roots and tubers. The average quantities of the various food items consumed by households in each of the three categories are presented in Tables 20-22. Cereals: The major cereals consumed by households were green maize,

maize grain, and maize flour, local and imported rice, millet grain and flour, and sorghum grain and flour. But in relative terms, maize was the most important cereal food consumed by households, especially in the 5S. Millet and sorghum were next to maize in retative importance and , again, relatively more of these were consumed per household in the SS than in other areas. Rice was the least consumed of all the cereals and virtually all the quantities consumed were locally produced . But in relative tenms , more rice was consumed per household in the SS than in any other part of the project area. Most of the cereals consumed were own-produoed by the households. In all, over 90% of the quantities of various cereals consumed were own-produced, except that households produced 48% of the local rice consumed and 78% of the millet flour. But there were some variations across agroecological zones in respect of the percentage shares of cereals that were own-produced. In this regard, it might be observed that only 69% of the green maize consumed by households was own-produced in the NG5 while 80% of maize grain and 76% of maize flour were own-produced in the SG5. All the millet flour was purchased in both NG5 and SGS. Generally, therefore, there was a very high degree of subsistence and self-sufficiency in cereal crop production in the project area. Legumes: As shown in Table 21, major legume crops consumed by

households were cowpea, ground nut and soybean but only in modest quantities. Generally, the average monthly consumption of any of these legumes per household was less than 10 kg. In relative terms , groundnut was the most quantitatively consumed, followed by cowpea. Very little soybean grain and flour were consumed.

Table 20. Average quantities of cereals consumed by households (kg! month). Agroecological zone SS SGS NGS 26 .9 0.0 26.9 100.0

16.6 0.6 17.2 96.5

10.28 0.6 14.7 69.4

Average (all areas) 16.6 0.9 17.5 94 .8

31 .1 0.1 31.2 99.7

20.2 5.1 25.3 79.8

25.5 1.1 26.6 959

25.2 1.2 26.4 95.4

33.1 0.9 97.4

17.2 5.3 22.5 76.4

25.7 2.6 28.3 90.8

27.2 2.8 30.0 90.7

12.5 13.7 26.2 47.7

6.3 5.9 12.2 51.6

8.3 4.7 13.0 63.8

8.3 8.8 17.1 48 .5

0.0 0.5 0.5 0.0

0.0 1.5 1.5 0.0

0.0 1.6 1.6 0.0

0 .0 1.2 1.2 0 .0

22.5 0.0 22.5 100.0

20.0 0.0 20.0 100.0

23.0 0.0 23.0 100.0

20.0 0.0 20 .0 100.0

13.0 0.9 13.9 93.5

0.0 0.7 0.7 0.0

0.0 2.3 2.3 0.0

3.2 0.9 4.1 78.0

30.2 0.1 30.3 99.7

10.5 0. 1 10.6 99.1

10.6 0.1 10.7 99.1

16.6 O. t 16.7 99 .4

33.6 0.0 33.6 100.0

10.7 0.0 10.7 100.0

10.2 0.0 10.2 100.0

17.8 0.0 17.8 100.0

Items Green Maize: own-produced

purchased total Percentage share

of own-produced Maize grain : own-produ ced purchased

total Percentage share of own-produced

Maize flour: own-produced purc hased 10tal Percentage share

34.0

of own-produced

Local rice : own-produced purchased total Percentage share

of own-produced Imported rice: own-produced purchased totat Percentage share of own-produced Millet grain : own路produced purchased 10tal Percentage share of own-produced Millet flour: own-produced purchased tota l Percentage share of own-produced Sorghum grain : own -produ ced purchased total Percentage share of own-produced Sorghum own-produced fl ou r: purchased total Percentage share of own-produced

Table 21. Average quantities of legumes consumed by households (kg/month). Agroecolog ical zones NGS SS SGS Items Cowpea grain : own-produced purchased total Percentage share of

Average (all areas)

6.8 0.6 7.4 91 .9

4.1 0.3 4.4 93. 2

4.5 0.6 5.1 88.2

5.2 0.5 5.7 91 .2

8.3 0.5 8.8 94.3

4.6 0.2 4.8 95.8

5.2 0.3 5.5 94.5

5.8 0.3 6.1 94.4

5.6 0.1 5.7 98.2

3.9 0.2 4.1 95.1

5.4 0.0 5.4 100.0

5.3 0.1 5.4 98.1

0.0 0.0 0.0

12 0.0 1.2 100.0

0.0 0.0 0.0

7.5 0.0 7.5 100.0

0.0 0.0 0.0

2.3 0.0 2.3 100.0

own路produced

Groundnut (shelled): own路produced

purchased total Percentage share of own-produced Groundnut (unshelled): own-produced purchased total Percentage share of own-produced Soybean grain : own-produced purchased total Percentage share of own-produced Soybean flour: own-produced purchased total Percentage share of own-produced Source: Field survey, 2004.

Over 90% of the legumes consumed were own-produced by households. In fact, soybean was produced only in the SGS and the soybean consumed was own-produced. Roots and tubers: Major roots and tubers consumed by households

in the project area were cassava tubers and cassava products , yam, sweetpotato, Irish potato , and cocoyam. Quantitatively, however, cassava and cassava products were the most consumed , followed by sweetpotato , Irish potato, yam, and cocoyam, in that order. In relative temns, little yam and very little cocoyam were consumed by households in the project area, an evident reflection of the local taste and lack of preference for these commodities.

Table 22. Average quantities of roots and tubers consumed by households (kg/month).

Agroecological zones SGS NGS SS Items

Average (all areas) 4.5 O.B 5.3 84.9 3.3 0.1 3.4 97.1

Cassava roats: own-produced purchased Total Percentage share of own-produced own-produced Gari: purchased Total Percentage share of own-produced

6.5 0.4 6.9 94.2 2.4 0.0 2.4 100.0

3.1 0.1 3.2 96.9 4.7 0.1 4.8 97.9

3.1 2.6 5.7 54.4 4.0 0.2 4.2 95.2

Cassava flour: own-produced purchased Tolal Percentage share of own-produced Cassava chips: own-produced purchased Tolal Percentage shane of own-produced

3.0 0.0 3.0 100.0 5.0 0.0 5.0 100.0

3.8 0.0 3.B 100.0 2.8 0.1 2.9 96.6

2.2 0.2 2.4 91.7 2.8 0.0 2.8 100.0

3.6 0.1 3.7 97.3 4.6 0.0 4.6 100.0

Yam lubefS: own-produced purchased Tolal Percentage shane of own-produced Sweetpotato: own-produced purchased Tolal Percentage shane of own-produced

3.5 0.9 4.4 79.5 12.7 1.4 14.1 90.1

2.7 0.4 3.1 87.1 5.B 0.1 5.9 98.3

2.4 4.6 7.0 34 .3 3.0 0.5 3.5 85.7

2.7 1.0 3.7 73.0 B.8 0.4 9.2 95.6

Insh potalo: own-produced purchased Total Percentage share of own-produced Cocoyam : own-produced purchased Total Percentage share of own-produced

2.0 0.3 2.3 87.0 0.0 0.0 0.0

7.0 0.2 7.2 97.2 0.0 0.1 0.1 0.0

5.2 1.4 6.6 78.8 0.0 0.0 0.0

5.8 0.7 6.5 89.2 1.6 0.1 1.7 94.1

Source: Field survey, 2004.

There were also some discernible disparities among agroecological zones in the relative quantities of roots and tubers consumed by households. In this regard, the average consumption of cassava roots was relatively low in the SGS but almost equal in the SS and NGS. Similarly, the average consumption of gari was highest in the two Guinea savannas. The average consumption of cassava flour was highest in the SGS and of cassava chips in the SS . The average consumption of yam was generally low in the project area, but higher in the NGS than in the other agroecological zones. The average consumption of sweetpotato was relatively high in the SS while that of Irish potato was highest in the SGS and lowest in the SS. Cocoyam was, however, consumed in a significant quantity only in the SGS. In general, households own-produced 90% or more of their consumed roots and tubers . The exceptions were in respect of cassava roots in the SGS (where only about 54% of the total quantity consumed by households was own-produced), gari, cassava flour, cassava chips , and sweetpotato, also in the NGS, and Irish potato in the SS and NGS. Generally, yam had the lowest own-produced share of the total quantity consumed by households , with a share of about 34% in the NGS . In addition to the regular consumption of home-prepared food , members of households also sometimes resorted to the consumption of food prepared outside the household, in local restaurants and other eatingplaces. This was often done for convenience . However, this practice was not widespread in the project area . Information collected revealed that only about 11 % of households in the NGS, 6% in the SGS, and 3% in the SS ever consumed food away from home and only very occasionally too . The amounts spent monthly per household on such foods were very modest, being less than ...500 on average. Two strong inferences which could be drawn from the analysis of household food consumption patterns are that, first, households in the project area had weak taste and preference for most root and tuber foods, except cassava and its derivatives and, secondly, the degree of household self-sufficiency in root and tuber production and supply was high . This implied a low degree of dependency on markets by houfeholds for root and tuber food supply, and thus reflected a low level of r.ommercialization of root and tuber production .

Consumption of nonfood commodities by households Nonfood commodities consumed by households included goods and services periodically procured for household consumption but excluded durable capital assets, such as motor vehicles and houses. They constituted those goods and services on which households incurred recurrent expenditure. However, some classifications might sometimes appear to be ambiguous. Table 23 presents the percentage distribution of households consuming various types of nonfood items in the project area. The items that enjoyed the most widespread consumption by households (with 70% or more of households consuming them) were kerosene fuel , shoes, clothing , education, and health services in all the agroecological zones of the project area. Evidently, kerosene was much more popular than fuelwood as a household energy source. The widespread consumption of education and health services was also noteworthy as it serves as an indicator of the priority accorded to children's education and family healthcare by the households . Table 23. Percentage distribution of households by types of nonfood Items consumed. Agroecological zone SS

SGS

NGS

Average (ali areas)

87 .5 40.1 79 .4 17.8 91.2 95.2 67.3 13.3 59.9 55.7 87.5 34.9 38.5 27.2 28.2 82.4 3.0 38.8

42.4 40.9 94.0 15.9 96.4 83.5 24.0 10.9 40.4 79.0 84.7 45.5 43.2 25.1 2t .0 65.1 4.9 45.7 24.0 38.5 7.9

60.8

57 . t 49 .6 91 .9 18.0 95 .8 88.4 40.0 12.3 47.0 75.6 86.4 49.2 37.6 26 .9 26.9 71.6 8.5 44.1 28.3 30.0 6.2

Commodities Beverages and cigarettes Fuelwood Kerosene Petrol (for vehicles) Shoes Clothing Repairs and other expenses on vehicles Vehicle maintenance Home repairs Educational expenses Health expenses (modern) Health expenses (traditional) Kitchen eqUipment Furniture Taxes and levies Donations Group contributions Wedd ing expenses Festival expenses Funeral expenses Other nonfood expenses

63.0 23.8

3.0

Source: Field survey 2004.

74.1 96.0 21.2 97.4 93.7 46.6 14.7 62.7 85.7 92.7 73.6 41 .9

30.4 25.0 88.3 22.7 54.4 25.8 26.9 15.5

Table 24. Average household expenses on nonfood consumer items (Hlmonth). Items Clothing Shoes Fuelwood Kerosene Beverages and tobacco

Furniture

Vehicle fuel Vehicle maintenance

Detergents Kitchen equipment Home repa irs Other repairs Group contributions Donations Taxes and levies Ceremonies and festivals Educational expenses Modern health expenses Traditional health expenses Other nonfood expenses Total

Agroecological zones SS SGS NGS 11235. 1 2865.0 306.6 202.6 232.1 26602.0 985.4 5305.7 246.0 5357.2 12472.0 3773.4 5400.0 11706.9 5251 .5 13050.9 6862.1 4723.9 2908.2 30573.1 149859.7

6705.7 1840.8 474 .3 197.8 123.3 9489.0 1079.6 5281.0 198.3 1990.1 12791. 1 3590.6 7471.4 2036.8 2234.9 6812.1 5877.1 4806.9 1881 .1 9771 .0 84652.9

Average

(all areas)

5056.0 1703.6 461.2 200.4 111.8 9644.5 927 .7 3534.7 281.5 2369.2 7956.5 3206.1 4890.7 2999.2 1301 .6 8966.7 5837.9 5535.1 1572.1 4158.9 70715.4

7172 .5 1934.2 389.2 200 .5 139.7 11617.3 1240.7 6973.3 232 .3 2506 .2 12083.6 3668 .5 8975.0 4118 .7 3486 .5 10635.1 63 18.9 4737.4 1791.0 12836.9 101057.5

Source: Field survey, 2004.

Table 24 shows the average monthly expend iture by households on various nonfood items consumed. Analysis shows that only eight major household items accounted for a disproportionately high share of all expenditure. These were clothing, fumiture, vehicle maintenance, home repairs, group contributions (to societies. community associations, religious organizations, etc). ceremonies and social events. education. and healthcare . Their respective percentage contributions are summarized in Table 25. The listed eight major household nonfood items accounted for more than 70% of the total household expenditure on nonfood

~ems

in the NGS

and SGS. The case of the SS was different in that household donations and miscellaneous expenses accounted for a relatively high share of 28.2% of total nonfood expenditure.

Table 25. Percentage shares of major nonfood consumer expenses in total household nonfood expenditure.

Agroecological zones SGS NGS

Items

Average (a ll areas)

4.4

15.1 11.2 8 .0 8 .8 7 .9 6 .2 7 .9 6 .9

112 13.6 12.7 6 .9 7.1 5.0 10.0 8.2

12.0 11.5 10.5 8.9 7.1 6.9 6.5 6.2

Subtotal

58.8

72.0

74.7

69.6

Other nonfood items

41.2

28.0

25.3

30.4

Total

100.0

Home repairs Furniture Ceremonies Group contributions Clothing Vehicle maintenance Healthcare Education

8.3 17.7 8.7 3.6 7.5 3.5 5.1

Source: Field survey. 2004.

Generally, local conditions and circumstances often dictated the relative distribution of total household expenditure on various consumer items. For example, the cost of repairing homes might be high due to the use of nondurable building materials in many localities. This might also be true of expenditure on furniture. Also, most communities took periodic religious and secular/cultural ceremonies very seriously hence relatively high shares of household expenditure were often devoted to them. Group contributions , in the context of this study, involved some household income savings, especially contributions to cooperative societies or mutual assistance groups . Vehicle maintenance expenditure included expenses on maintaining motor vehicles, motorcycles , bicycles, and other vehicular devices used locally. The general inference from the foregoing is that most households in the project area placed much of their expenditure priorities on maintaining their home environment and on taking care of household members. This was evident from the high expenditure shares allocated to home repairs , household furniture, clothing, vehicle maintenance, and educational and healthcare services.

Household poverty and food insecurity analysis

This section presents the results of the household food insecurity and poverty analysis carried out in the study, based on the models developed in section three. The results are presented as follows.

Household food insecurity statistics The summary statistics on the food insecurity status of households are presented in Table 26. Based on the recommended daily energy level (L) of

2250 kilocalories, the food insecurity threshold or line (S) for the households in the project area was found to be N63Jl daily or H1975.01 monthly per adult equ ivalent. On an annual basis, this was equivalent to H23 700.12 per adull equivalent. Using this defined insecurity line, it was found that 58% of all households sampled were food insecure by headcount (H). Furthermore, the estimated aggregate income gap (G) of -375.74 indicated the amount

(H375.74) by which food insecure households were below the minimum expenditure level required to meet their basic food needs.

Determinants of household food insecurity The various variables included as the determinants of household food insecurity in the Logit regression model used were as defined earlier. The results of the Logit regression analysis are presented in Table 27. From the Logit analysis, 11 regression coefficients were found to be statistically significant at

0.05 (i.e., at 5% level or below). However,

some variables which were earlier included in the model were dropped from the analysis because of the problem of multicollinearity. The result shows that the significant determinants of household food insecurity in the project area were household size, gender of household head, level of education of household head, household farm size, structure of household farm enterprises, share of household's own-produced food in the total quantity consumed, extent of agricultural production commercialization, household's expenditure on education, household's access to agricultural extension services and

agricu~ural

credit, household

head's membership of cooperative societies or farmers' organizations, and the total value of household assets.

Table 26. Summary statistics on food insecurity among households in the project area.

Measures Cost-of-calories equalion FAO recommended daily energy level (L) Food insecurity line Z (cost of Ihe minimum energy requirement per adult equivalenl). Head count (H ) poverty index: Aggregate income gap (G)

Values Can slant = 4.154 (0 .534) Slope coefficient = 0.0019 (0.0004) 2250 Kcal H63 .71 daily H1975.01 monthly H23 700.12 yearly 0.58---375.74

Table 27. Result of the Logit Regression Analysis of household food insecurity status. Variable Constant HHSZ GENO EOUC CDR RFETE FARMSZ CREDIT FARMEN FLAB HLAB PERCUL RQPQC DIVER EXCOM EDUCEX EXTAG COOP ASSETS REMIT

Parameter estimate

2.388 -0.014" 0.946'

--O.895r" --0.003 1.317 -0.1184' --0 .009 1.025' --0.471 0.0 18 0.651 -0.220" --0 .234 0.261" 0.034" --0.1308"

-0.034 .... -0.3E-04" --O.5E-04

Source: Computer printout of data analysis Note: 路SigniflCant at 5% level; Significant at 1% H

t-value 1.373 - 2.031 2.097 -3.226 - 0.054 1.367 -1.899 - 0.403 1.743 -0.345 0.088 1.56 -3.766 -1.396 2.946 3.860 -2.623 - 3.928 -4.396 -0.086

~vel

The pattem of the behavior of these variables is explained as follows. Household size (HHSZ): The coefficient of the variable is statistically significant at the 1% level and its sign is negative. This shows that households with large sizes had higher probabilities of being food insecure than those with smaller sizes, and vice versa. That is, household size is a negative factor determining the food security status of a household in the project area.

Gender of the household head (GEND): This is a variable with a

.coefficient that is statistically significant at the 5% level and has a positive relationship with household's food insecurity status. In particular, the resul t shows that households headed by males had higher probabilities of being food insecure than those headed by fema les. Educational level of household head (EDUC): The coefficient of

this variable is statistically significa nt at the 1% level and carries a negative sign, thus suggesting that the higher the educational level of the household head, the more food secure (or less food insecure) the household tended to be , and vice versa. This is as expected , since the level of education should positively affect the income earning capacity and level of efficiency in managing the household's food resou rces. Farm size (FARMSZ): The total farm size of a household is another

significant determinant of food insecurity status. The coefficient of the variable is negative in sign and statistically significant at the 5% level , meaning that farm size exhibits a negative relationship with the food insecurity status of a household. That is, households with larger farm sizes tend to be more food secure than those with smaller sizes , and vice versa. As a household's farm size increases, the probability of food insecurity tends to decline. Structure of household enterprises (FARMEN): The coefficient of

this variable is negative and statistically significant at the 5% level. Th is implies that households engaged in farming alone have higher probabilities of food insecurity than those that have diversified enterprise structures, involving farm and non-farming enterprises. This is plausible because households that have other sources of income in addition to farming alone tend to be more resilient in times of food crisis than those engaged in farming alone. Alternative income sources outside farming provide enhanced security for household livelihood. Share of own-produced food in the total quantity consumed (RQPQC):

The coefficient of the share of own-produced food in the total quantity of food consu med by a household is statistically significant at the 1% level and exhibits a negative relationship with food insecurity. This shows that the higher the share or ratio, the lower the probability of food insecurity tends to be , and vice versa .

Extent of agricultural production commercialization (EXCOM): The extent of agricultural production commercialization is an income-determining factor that is expected to affect food insecurrty. The coefficient of the variable is statistically significant at the 1% level and exhibits a positive relationship with food insecurity status , suggesting that the higher the extent of commercialization , the higher the probability of food insecurity tends to be, and vice versa. This is , however, contrary to a priori expectations. The reason for this resuH is probably because most of the households produced at a scale primarily meant for home consumption, although they might be forced to sell when the need to earn more money income arose. Thus they might deplete the stock meant for home consumption and thereby expose the household to higher risks of food insecurity. Expenditure on education (EDUCEX): This variable is another important determinant of food insecurity. The coefficient of the variable is statistically significant at the 1% level and carries a positive sign , suggesting that the higher a household's expenditure on education. the higher the probability of food insecurity tends to be, and vice versa. This is plausible, as the education of children is a priority area for which a household could deny itself some food and other necessities. Household's access to agricultural extension services (EXTAG) : The coefficient of the variable is statistically significant at the 1% level and has a negative relationship with the food insecurity status of a household. This implies that households with access to agricultural extension services tended to have higher probabil ities of being food secure than those that did not have such access , and vice versa . This is because contact with extension services tends to enhance the chances of a household having access to better crop production techniques, improved inputs, as well as other production incentives that pos itively affect farm productivity and production and thus household food security status. Household heads ' membership of cooperative societies (COOP): Whether or not household heads are members of cooperative societies or farmers' associations is another important variable that affects the food security status of households in the project area. The coefficient of the variable is statistically significant at the1 % level and carries a negative sign . This implies that households whose heads were members

of cooperative societies or other farmers' organizations had higher probabilities of being food secure than those households whose heads were not members . This can be dosely linked to the beneficial effects of their membership , in terms of production and other welfare-enhancing services that these societies and organizations often offer. Value of household assets (ASSETS): Ownership of household assets is considered to be one of the strategies for enhancing households' resi lience in the face of eoonomic crisis and adverse Circumstances, such as crop fa ilure, drought, and so on . It is believed that some of the assets could be disposed of to cushion the effects of a transitory economic crisis on the households. The variable is an important factor affecting food security in the study area . The coefficient of the variable is statistical ly sig nificant at the 1% level and carries a negative sign , thus suggesting that the higher the value of household assets, the lower the probability of food insecurity tended to be , and vice versa.

Classification of households by poverty status From a mean monthly expenditure of H3670 .00 per adult equivalent, a pcverty line of H2446.67 monthly was derived (i.e., H611 .67 weekly or

H81 .56 daily) . This line was defined as two-thirds of the mean household expenditure per adult equivalent. Households with a mean monthly expenditure per adult equivalent below this poverty line were classified as being poor, while those with a higher mean monthly expend iture were dassified as being non-poor. This analysis revealed that about 67 % of the households in the project area were poor, while about 33% were non-poor. This poverty line is comparable to that of the FOS (1999), which was derived from a similar approach of two-thirds of the mean per capita household expenditure. The FOS estimate also put the percentage of the poor in the northeastern zone of the country (to which Barno State belongs) at 67%. It should be painted out, however, that the poverty threshold used in the current study is below the internationally recognized level of one US $1 day. At the time of this study, the average exchange rate was $1 = N135 .

Determinants of household poverty intensity In this section, the factors that affect household poverty status and the elasticities that show the degree and direction of the responses of poverty level to changes in these variables are presented. Two types of elasticities were generated from the Tobit regression model used: the elasticity of the probability of a household being poor and the elasticity of the intensity of poverty of a household that is already poor. As already discussed, the analysis of household monthly expenditure revealed that the mean monthly per adult equ ivalent expenditure (MMPAEHE) was H3670.00, out of which 72% went on food consumption alone. From this, a poverty line of N2446.67 monthly was obta ined (+-1611 .67 weekly or N81 .56 daily). For the study area, 67% of the households were classified as being poor (i .e., with a mean monthly expenditure per adult equivalent that was below the poverty line). In the Tobit regression analysis used. only poor households were considered . Hence. the dependent variable (as defined

ea~ ie r

section three) measured the intensity of poverty among households in the project area . The values of this dependent va riable ranged between 0 and 1; the farther away the value is from 0, the worse the poverty situation . The result of the regression analysis is presented in this section. A multicollinearity test was first carried out on the variables included in the Tobit model and , as a resu lt, some of the explanatory variables initially proposed for inclusion were dropped from the analysis. The results of the Tobit regression analysis are presented in Table 28 and show the va rious paramete r estimates from the Tobit regression analysis. Table 28 reveals that 15 out of the 23 explanatory variables related to household livelihoods included in the model had statistically sign ificant coefficients at between 1% (P < 0.01) and 10% (P<O.I), representing about 63% of all the explanatory variables. Also, the sigma (a ) value was 0.36. with a t-value of 19.28 . This was statistically significant at the P<0.01 level. thus indicating that the model had a good fit to the data. Furthermore, the value of the intercept was 0.44, meaning that the autonomous poverty intensity was 0.44 in the study area. The 15 explanatory va riables . which were found to significantly affect household poverty intensity. are discussed as fo llows.

Table 28. Result of Tobit Regression Analysis of household poverty Intensity. Variable Constant Demographic HHSZ GEND EDUC CDR Eoonomic (Production/Consumption RFETE FARMSZ CREDIT FARMEN FLAB HLAB PERCUL RQPQC DIVER EXCOM

Estimate

t-value

0.44'"

2.145

0.211'" 0 .375""" 0.22E-04 0.293"

3.64 3.08 1.41 2.71

0.045 -0 .103'" -0.196" -0.227'" -0.657" 0.851E-03 -0.342 0.471

0.50 -0.83 -2.34 -2.06 - 1.825 0.56 -2.96 1.22 3.06 - 1.77

- 0.112"· -0.031'

Health/Hygiene MEDOP DlSWAT DISMED HELTEX

-0.086"" 0.751E-03 -0.002 0.673E-06

-3.20 0.56 -0.98 0.3 1

Institutional Influence EDUCEX EXTAG COOP

0.962· ..• - 0.013' -0.084 ....

2.03 -1.98 -2.77

Vulnerability and Resilience ASSETS REMIT

- 0.162' -0.205·.....

-1 .71 - 3.47

Source: Computer printout of Tobit analysis .... = Stgnificant at P < 0 01 ; -=Significant at P < 0.05 : • = Significant al P < 0.1 0. a:::; 0.36: Log likel ihood function = 64.91

Household's size (HHSZ): Households with large sizes had a higher intens~y

of poverty than those with smaller sizes. The household size

variable has a regression coefficient of 0.211, meaning that a unit increase in household size would bring about an increase of 0.211 in the probability of household poverty, and vice versa . The coefficient is positive and statistically significant at 1%.

Gender of the household head (GEND): The coefficient of the variable is statistically significant at the 1% level and shows a positive relationship with the intensity of poverty. This result shows that households headed by males had a higher poverty intensity than those headed by females in the study area . Because GEND is a dummy variable, its coefficient of 0.375 implies that, given all other factors, the probability of the poverty intensity of male-headed households was autonomously higher than that of female-headed households by 0.375.

Child dependency ratio (CDR) : The degree of child dependency of a household is believed to affect the welfare of such a household. In this study, the child dependency ratio was found to have negatively affected the poverty status of the households in the project area. That is, a high dependency ratio was found to be inimical to households' poverty status. The regression coefficient of 0.293 for the child dependency variable implies that a unit increase in the child dependency ratio would increase the probabil ity of poverty intensity by 0.239 in an average household in the project area , and vice versa.

Household farm size (FARMSZ): Household farm size was one of the highly sign ificant factors affecting the intensity of poverty among households in the project area. Households with larger farm sizes were, on average, less poor than those that cultivated smaller farm sizes. This was because households with larger farm holdings were expected to generate more income , which would enhance their consumption level and subsequently improve their household poverty status. The regression coefficient is -0.103, meaning that a unit increase in the size of farm holding would lead to a reduction in the probability of household poverty by 0.103 , and vice versa .

Household head's access to credit (CRED/T): Households whose heads had access to credit facilities had a lower level of poverty intenSity than those whose heads did not have such access. This might be due to the fact that those households w ith access to credit were able to acquire more productive resources for their household enterprises. This would subsequently enhance the household's income-generating ability and household welfare. This variable has an intercept dummy of -0.196, meaning that the autonomous poverty intensity of households whose heads had access to credit facilities was, on average , lower by 0.196 than that of households without access.

Household production enterprise structure (FARMEN) : Households whose enterprise structure was not restricted to farm production alone had a lower intensity of poverty than those that depended solely on farm production. An intercept dummy of -0.224 implies that the probability of poverty intensity was autonomously reduced by 0.224 among households whose enterprise structure was not restricted to farm production alone in the study area , compared w ith households having only farm production enterprises. Family labor (FLAB) : Households with higher levels of family labor supply had a lower intensity of poverty than those with lower levels of supply. That is, given other factors, the higher the level of household labor available, the lower the intensity of household poverty. This might be because the abundance of family labor would tend to reduce the need for expenses on paid (hired) labor, thereby leaving some extra money to take care of other household needs . The regression coefficient also indicates that, given other factors, a unit increase in the quantity of family labor would reduce the probability of household poverty intensity by 0.657, and vice versa. Extent of household production diversification (DIVER) : This is another variable that sign ificantly affected the poverty status of households in the project area. The coefficient of the variable is statistically significant at the 1% level and carries a negative sign. This shows that households with relatively more diverse farm enterprises and , hence, more diverse sources of farm income tended to have lower probabil ities of poverty intensity than households with relatively less diverse enterprises and income sources, given other factors. The regression coefficient shows that, given other factors , a unit increase in the diversification index would tend to reduce the probability of poverty intensity by 0.112, and vice versa. Production diversification is a well-known strategy to minimize risk. In this case, it could serve as a strategy for minimizing the risk of losses in farm income and, hence, the risk of a more intense househo ld poverty level. Extent of agricultural output commercialization (EXCOM): A negative relationship between the extent of production commercialization and household poverty intensity implies that the higher the extent of commercialization, the lower the intensity of poverty, and vice versa.

This is plausible, because the sale of output is expected to generate income for households to meet their needs , such as expenses for healthcare and the education of their children and this tends to reduce the household's poverty level , given other fa ctors. The regression coefficient shows that an increase of one unit in commercialization by a household would reduce the probability of household poverty intensity by 0 .031, and vice versa. Health/hygiene-related variable (MEDOP) : The statistical sign ificance of the MEDOP coefficient shows that households that combined traditional methods (herbs) with modern methods of healthcare wene less poor than those that did not. Since the cost of receiving treatment in the hospitals was often perceived to be too high in the project area , households that resorted to using less expensive tradrtional healthcare methods might save some money. This could be spent on other household items that might enhance the welfare of such households . The intercept dummy of -0.086 suggests that, on average, the autonomous poverty intensity of households that combined traditional methods (herbs ) with modern methods of healthcare would decrease by 0.086 , given other factors. Household expenditure on education (EDUCEX) : Households with higher expenditure on education wene on , the average, poorer. This may be due to the fact that expenditure on education as a n item of priority expenditure would deprive a household of some other basic needs . This could have a negative impact on household welfare and increase the intensity of household poverty. The regreSSion coefficient of 0.962 implies that a unit increase in the expenditure of a household on education would increase the probability of household poverty intensity by 0.962, and vice versa . Household's access to agricultural extension services (EXTAG) : The coefficient of this variable is statistically significant at the 10% level and shows a negative relationship with the poverty status of households. This implies that households that had access to extension services had lower probabilities of being poor than those that did not have such access, and vice versa. This might be because contact with extension services provided more access to improved crop production techniques, improved inputs, and other production incentives. These would positively affect

farmers' outputs and their income-generating abil ity, thereby reducing their poverty level. An intercept dummy of -0 .013 suggests that, on average, the autonomous poverty intensity of households that had access to extension services was lower by 0.013 in the project area. Membership of cooperative societies or other farmers' associations (COOP): This variable exhibits a negative relationship with household

poverty intensity. This implies that the intensity of poverty was lower in a household whose head was a member of a cooperative society or any other farmers' association than in one whose head did not belong to such an organization . This might be as a result of various benefits accruable to members of cooperative societies, such as credit facilities, access to improved production inputs, and access to information that could enhance their productive capacity. The intercept dummy of the variable suggests that membership of cooperative societies would autonomously reduce poverty intensity by 0.084, and vice versa. Total value of household assets (ASSETS) : A negative relationsh ip

between the value of household assets owned and the intensity of household poverty implies that the higher the value of household assets, the lower the household poverty intensity, and vice versa. The value of household assets measures the ability of the household to withstand economic shocks and income shortfalls to finan ce the purchase of household needs; ownership of assets serves as a surety and a fallback strategy for the household against transitory poverty because some of these assets could be sold to procure basic household needs in periods of temporary financial distress. The regression coefficient of -0.162 implies that a unit increase in the value of household assets would reduce the probability of household poverty by 0.162, and vice versa . Total value of remittances received by household (REMIT) : The negative

relationship between total value of remittan ces received by a household and poverty intensity implies that households with higher values of received remittances tended to have lower probabilities of poverty intensity. Also, a unit increase in the value of household remittances received would reduce the probability of poverty of an average household by 0.21 , given other factors.

Table 29. Elasticity estimates of household poverty intensity.

Variab le

Elasticity of probability of poverty (a)

Elasticity intensity of poverty (b)

Total elasticity (a+b)

0.276 0.031 -{)062

0.277 0.031 - 0.060

0 .555 0062 -0 .122

-1.23 -1 .57 0.162 -{).029 -4 .594 -0.382

-1.03 -1 .09 0.180 -{).028 -{).0004

-2.26 -2.66 0.362 -0.057 -4.594 -{).785

Household size Expenditure on education (N) Household assets (Nj Extent of agricuftural production diversifICation Farm size Child dependency ratio Extent of output commercialization

Total value of remittances received Family labor (mandays)

-0.403

Source: Computed from Tobit regression resu lts .

Elasticities of household poverty intensity Elasticity coefficients were computed for only nine of the variables included in the model because other variables with statistically sig nificant coefficients were dummies. Elasticity coefficients computed were those of household size , total household expenditure on education, total value of household assets, child dependency ratio, extent of household agricultural output commercialization , extent of household agricultural production diversification , farm size, total value of remittances received , and quantity of fam ily labor available. As shown in Table 29, only the coefficients of farm size, extent of agricultural production diversification , and total value of rem ittances received by household were elastic (i.e., >1) out of the nine computed. The important factors that reduced household poverty intensrty, in order of importance, were the total value of remittances received by households, farm size, the extent of agricultural production diversification, the quantrty of family labor available, and the total value of household assets. A 1% increase in three variables would reduce the intensity of household poverty as follows: by 4.46% with an increase in the total value of remittances received by households, by 2.26% with an increase in farm size , by 2.26% with an increase in household agricultural product

diversification , and vice versa. On the other hand , a 1% increase in the amount of fami ly labor available would lead to a decrease of only about 0.80% in household poverty intensity and in the total value of household assets to a decrease of only about 0.12%, and vice versa. Some variables were, however, found to increase household poverty intensity. These, in order of importance , were the household size , child dependency ratio, and household expenditure on education. From Table 29, it can be deduced that a 1% increase in household size would increase household poverty intensity by about 0.60%; the same increase in the ch ild dependency ratio would increase household poverty by 0.04%, in household expenditure on education by 0.10% , and vice vensa.

Major findings and recommendations

This section presents the summary of major findings from the study, conclusions from the findings, and policy recommendations for improving rural household livelihoods and for reducing household food insecurity and poverty intensity in the project area.

Summary of major findings Household socioeconomic characteristics On average, about 76% of households covered in the survey were maleheaded while about 24% were female-headed . Most households had between 6 and 11 members in adult equivalents. About 70% of household heads were in the active and productive age range under 50 years. The average age was 46 years . The majority (51 %) of household heads had had no formal education at all . Five was the average number of years of forma l education received . About 52% of households were polygamous while about 45% were monogamous.

Economic activities of farm households Crop farming was the most important occupation of households in the project area, and civil service employment was the second most important. Combined (mixed) crop and livestock farming was the third most important farming activity, although livestock farming alone was of minor importance. The analysis of the structure of household asset ownership showed that land was the most w idely owned asset, foll owed by livestock, and then by bicycles. The average household cultivated small, fragmented plots, which added up to about 3.7 ha. The average number of fragmented farm plots cultivated by male and female members of households was six.

The average farming experience of household heads covered about 25 years. The commonest type of land tenure was ownership by inheritance from family or community. Credit sources in the project area included commercial banks, specialized banks for agricultural loans, Bomo State Agricultural Development Project, cooperative societies , individual moneylenders, and relatives and friends. But virtually all these failed to make any significant impact on the credit needs of fanmers. Hence, in 2003 for example , less than 1% of fanm households were able to obtain cred it. The most important and most widely cultivated crops in the project area were cereals, probably due to local tastes and preferences and local agro-climatic conditions . The average number of years of continuous cropping of farmlands ranged between 9 and 19 across the LGAs. Cotton, sorghum, rice , and millet were the crops most widely abandoned by farmers. The major reasons given for abandoning these crops, in relative order of importance, were poor soil fertility and low crop yields, inadequate access to credit to purchase requ ired inputs , labor supply constraints, high incidence of diseases, pests, and animal predation , and poor market demand for produce. The average gross revenuelha from crop production was about H98 500, implying that an average household earned almost ...100 000 gross revenuelha from its combination of crop enterprises. The most important item of variable cost was fertilizer. This cost between N9000lha in the SS (i.e., Damboa LGA) and about N16 Ooolha in the NGS (i.e., Kwaya Kusar LGA). Fertilizer cost alone accounted for between 39% of total variable costlha in the SS (i.e., Damboa LGA) and about 69% in the SGS (i.e., Hawul LGA). The gross marginsl ha varied widely across LGAs, from about N68 OOOlha in Biu LGA to about N105 OOOlha in Kwaya Kusar LGA, both in the NGS. But the average for al l the agroecologica l zones was about N85 4001ha .

Variations in RPN were a reflection of variations in both variable cost and gross margin earned per naira of variable cost incurred. The highest RPN was in the NG5, (i.e., Kwaya Kusar LGA). It could be inferred that the profitability of crop production was also highest in the NG5, i. e ., (Kwaya Kusar LGA) with relative profitability in the 5G5 and 55. The availability of a market outlet for produce was the single most important factor determining profitabil ity in crop farm ing in the project area as a whole. It was, in particular, the single most important factor in the NGS (Biu and Kwaya Kusar LGAs) and the 55 (Damboa LGA). A high market price for produce came second but was the most important single factor in the project area as a whole . Constraints to profitability in crop farming, as identified by respondents, were, on the whole, the obverse of factors promoting profitability. Hence , the constraining fa ctors identified included poor market demand for produce, low market price for produce, high cost of labor, and high cost of fertil izer. It is, however, important to point out that high cost of fertil izer was the single most constraining factor identified by respondents in every LGA covered in the study. This suggests that the maintenance of soil fertility was a priority in the project area and so should be accorded a special place in programs designed to improve farm productivity and improve household livelihoods in the project area. Livestock was also an important livelihood asset in the project area . The types of livestock kept by households consisted of animal and poultry stock species . Major animal stock species were cattle, sheep, goats, and pigs; major poultry stock species were chickens , guinea fowl , and ducks. Generally, both men and wom en in households kept only small numbers of animal stock species. On average, men in households kept 13 head of cattle, 13 sheep, 13 goats, and 8 pigs; women kept 9 head of cattle, 13 sheep, 13 goats, and 3 pigs. On the whole, men kept more cattle than women but there was no gender difference in the number of sheep, goats, and pigs kept by men and women. On average, men kept 26 chickens, 29 gu inea fowl, and 29 ducks; women kept 24 chickens 14 guinea fowl , and 23 ducks.

Overall evidence indicated that large-scale poultry ownership was uncommon in the project area . Much more common were small-scale poultry hold ings, with the birds reared mostly on a free-range basis. In assessing the profitability of livestock farming, about 54% of respondents perceived livestock farming as being profitable. Overall , availability of a market for products was the single most important factor determining the profitability of livestock farmi ng in al l the LGAs, fo llowed by high product prices , low cost of production, and the availability of livestock feeds , in that order. Other factors, such as easy access to credit, availability of processing facilities, access to good transport facilities, and so on, were also identified. Major factors constraining the profitability of livestock farming , in relative order of importance, were the problems of livestock diseases , high cost andlor scarcijy of livestock feeds , drugs, vaccines, etc., lack of a market for produce , and low product prices .

Household non-farming activities Analysis revealed that only a minority of households was engaged in off-farm economic activities, meaning that the majority of households concentrated on farm ing activities alone. On average, only about 18% of households in the SS (i.e., Damboa LGA), 32% in the SGS (Le .. Hawul LGA), and 57% in the NGS were engaged in non-farming economic activities. The average for all LGAs was about 38% . Off-farm economic activities of households covered both agriculturerelated activities and those not related to agriculture, although only a relatively small percentage of households were engaged in off-farm agriculture-related activities . In all , only about 18% of all households were engaged in agriculture-related off-farm economic activities. Major non-farming economic activities of households induded petty trading, civil service employment, artisanship, processing of agricu ltural products, marketing of agricultural products, and so on . But the nonfarming activity most widely pursued by households was the marketing of agricultural products. This activity was mosl prevalent in the SGS and SS (Le. , Hawul and Damboa LGAs) but least prevalent in the NGS (Le., Kwaya Kusar LGA). Thus, the marketing of agricultural products constituted the most important livelihood option besides farming . This should not be unexpected as agricultural product marketing is complementary to activities in agricultural production .

Household food consumption pattern The types of food items consumed by households in the project area were categorized into cereals, legumes, and roots and tubers. The major cereals were green maize, maize grain, and maize flour, local and imported rice, mil let grain and flour, and sorghum grain and flour. But in relative terms, maize was the most important cereal food consumed by households, especially in the 55 (i .e., Damboa LGA) . Millet and sorghum were next to maize in relative importance and, again , households consumed relatively more of these in the 55 (Damboa LGA) than in other areas. Rice was the least consumed of all the cereals and virtually all the quantities consumed were locally produced. Most of the cereals consumed were own-produced by the households. In all, over 90% of the quantities of various cereals consumed were own-produced , except for local rice (only 48% own-produced by the households ) and millet flour (78% own-produced). Generally, there was a very high degree of subsistence and self-sufficiency in cereal crop production in the project area. The major legume foods consumed by households consisted of cowpea, groundnut, and soybean . But it would appear that households consumed only modest quantities of these. Generally, the average monthly consumption of any of these legumes by a household was less than 10 kg. But in relative terms, groundnut was the most quantitatively consumed, followed by cowpea . Very small quantities of soybean grain and flour were consumed . Over 90% of the legumes consumed were own-produced by households. Major roots and tubers consumed by households in the project area were cassava tubers and cassava products, yam , sweetpotato, Irish potato , and cocoyam . Quantitatively, however, cassava and cassava products were the most consumed, followed by sweetpotato, Irish potato, yam, and cocoyam, in that order. In relative terms, households consumed little yam and very little cocoyam in the project area, an evident reflection of the weak local taste and lack of preference for these commodities. In general, households produced 90% or more of their consumed roots and tubers.

Consumption of non-food commodities by households Non-food commodities consumed by households included goods and services periodically procured for household consumption but excluded durable capital assets such as motor vehicles and houses . They consisted of those goods and services on which households incurred recurrent expenditure. The items that enjoyed the most widespread consumption by households (with 70% or more of households consum ing them) were kerosene fuel, shoes, clothing, education and health services in all the four LGAs of the project area . Only eight major household

expend~ure

items accou nted for

a disproportionately high share of all expenditure . These were expenditure on clothing, furni ture, vehicle maintenance, home repairs, group contributions (to societies, community associations, religious organizations, etc. ), ceremonies and social events, education, and healthcare . The eight items everywhere accounted for more than 70% of the total household expenditure on nonfood items except in Damboa LGA (i.e., the SS). Generally, local conditions and circumstances often dictated the relative distribution of total household expenditure to various consumer items .

Household food insecurity analysis Summary statistics computed on the food insecurity status of households showed that the daily food insecurity threshold or food insecurity line per adult equivalent was Using this food

~3 . 7 1 ,

insecur~

i.e., H975.01 month ly or H23 700.12 yearly.

line per adult equivalent, 58% of all households

in the project area were found to be food insecure. Furthermore, the estimated aggregate income gap was 1'01375.74, which was the amount by which an average food insecure household was below the minimum monthly expenditure required to meet its basic food needs. From the Logit analysis of determ inants of household food insecurity,

11 regression coefficients were found to be statistically significant at P S 0.05. The factors (variables) which tended to reduce household food insecurity (given other factors) were (1 ) female gender of household head , (2 ) educational level of household head , (3) household farm size,

(4) structure of household enterprises, (5) degree (extent) of household agricultural enterprise diversification, (6) share of household 's ownproduced food in total food consumed, (7) household's access to agricultural extension services, (8) household head 's membership of a cooperative society or fanmers' association, and (9) total value of household assets . The factors (variables) which tended to increase (enhance) household food insecurity (given other factors) were (1) household size, (2) male gender of household head , (3) extent of agricultural production commercialization, and (4) amount of household expenditure on education.

Household poverty intensity analysis The analysis of household monthly expenditure revealed tha t the mean monthly per adult equivalent household expenditure was 1'013670.00 out of which 72% went to food consumption alone. From this, a month ly poverty line of 1'012446.67 (N611 .67 weekly, or NB1 .56 daily) was derived . Based on this line, 67% of the households in the project area were classified as being poor. The Tobit regression analysis was then used to measure the intensity of poverty in households, or to analyze the detenminants of poverty intensity. The analysis revealed that 15 out of the 23 explanatory variables included in the Tobit regression model had statistically significant regression coefficients at between 1 and 10% levels of significance. These represent about 63% of all the explanatory variables. Factors (variables) which tended to reduce household poverty intensity were (1 ) female gender of household head , (2) household farm size, (3) household head's access to credit, (4) household enterprise diversification, (5) fam ily labor size, (6) household enterprise structure , (7) extent of household production commercialization, (8) household healthcare indicator, (9) degree of household's access to extension services , (10) household head's membership of cooperative society or farmers' organization, (11) total value of household assets , and (12) total value of remittances received by house hold. Factors (variables) which tended to enhance (increase) the intensity of household poverty were (1) male gender of household head, (2) household size (3) household child dependency ratio, and (4) household expenditure on education.

Elasticities of household poverty intensity Elasticity coefficients were computed for only nine non-dummy variables with statistically significant coefficients. The result showed that household poverty intensity was elastic with respect to (1) the extent of agricultural production diversification, (2) household's farm size, and (3) total value of remittances received by household . All these elasticity coefficients were negative but greater than unity in absolute values . This impl ies that a 1% increase in the value or level of each of these variables , given the values of all others, would decrease the intensity of household poverty by more than 1% and vice versa. Other variables with significantly negative elasticity coefficients but which were less than unity in absolute values were (1) value of household assets, (2) extent of agricultural production commercialization, and (3 ) quantity of family labor available. This implies that an increase of 1% in the value of each of the se variables, given the values of all other variables, would decrease poverty intenSity, but by less than 1%, and vice versa. Three variables had statistically significant but positive elastiCity coefficients that were also less than unity in value. These were (1) household size , (2) child dependency ratio, and (3) household expenditure on education . This implies that a 1% increase in the values of each of these variables, given the values of all other variables, would also increase the level of poverty intensity, but by less than 1% and vice versa.

Policy recommendations Broadly, policy recommendations are grouped into six categories covering policy issues relating to (1) gender mainstreaming , (2) production input and credit support, (3) household enterprise organization, (4) technical services support, (5) input and output market infrastructure , and (6 ) social capital formation.

Gender The analysis of both household food insecurity and poverty status showed that female-headed households were more food secure and less poor than male-headed households. This finding points to the need for gender mainstreaming with regard to the promotion of entrepreneurship among females in households to enhance their income~arning capacity through training for technical and managerial skill acquisition, credit support, resource supply support, and social capital format ion .

Women's comparative advantage in rearing small stock (sheep, goats, and poultry), small-scale agroprocessing, and produce marketing should be fully exploited to promote household food security and reduce household poverty.

Production input and credit support As findings from this study have revealed, inadequate proeluction input anel creelit supply constituted some of the most bineling constraints to householel crop anellivestock production, especially in respect of the supply of fertilizer, improveel seeds . pesticides , anel veterinary proelucts. There is, therefore, a need for measures to ensure an aelequate supply of these inputs as well as creelit through both public anel private sector initiatives .

Organization and structure of household enterprises Findings from this stuely have revealed that household enterprise diversification is a strong factor in household food security and poverty reduction . Farm enterprise diversification , involving mixed (crop and livestock) farming and also promoting crop anel livestock interaction, as well as farm/non-farming enterprise diversification can, therefore, serve as an effective strategy for promoting householel food security and for reelucing householel poverty in the project area. Commercialization of agricultural proeluction was also found to be a strong factor in household poverty reduction . It follows, therefore, that measures that promote both household enterprise diversification and agricultural production commercialization are highly elesirable. Such measures woulel include the aelequate supply of improved farm inputs, the provision of technical services, including technical training and agricultural extension services, and the supply of aelequate credit to farmers. Householel food consumption analysis revealed that households had strong tastes and preferences for cereals and cereal products. This finding points to the need for special emphasis on the production of cereals and cereal products in the project area.

Technical services support As found in this study, there was an increasing incidence of continuous cropping and increasing intensity of land use , resulting in soil infertility, low crop yields, anel increasing incidence of disease and pest infestation. Policy measures to control these problems or ameliorate their effects should indude those related to increased technical support for farmers in the project area. Four types of measures are proposed as follows.

First, there should be measures to address the issue of soil fertility maintenance and improvement through the introduction of appropriate tillage practices and the use of both organ ic and inorganic fertilizers. Secondly, there should be adequate technical training of farmers in appropriate soil management techniques and improved farm management practices for crops and livestock. Third ly, there should be adequate strategies for promoting appropriate farm mechan ization, including the promotion of animal traction technology, with the use of draft animal equ ipment and work bu lls. And, fourthly, there should be an accelerated promotion of adaptive agricultural research and technology to develop and distribute improved farm inputs, especially organic and inorgan ic fertilizers, improved seeds, and livestock, pesticides, and veterinary drugs and vaccines.

Market infrastructure Inadequate market access for farm produce (or inadequate market demand for produce) was widely identified by farmers as one of the major constraints inhibiting the profitability of crop and livestock farming in the project area. Measures to address this problem should include the provision of improved market infrastructures , such as market stalls, community storage facil ities , rural access roads , rural transportation faci lities , and agricultural price information systems. Furthermore, the private sector should be encouraged to invest in agricultural input and output market infrastructure and facilities.

Social capital formation In this study, membership of cooperative societies and farmers' organizations was identified as one of the sign ificant factors enhancing household food security and reducing househOld poverty status. As SUCh, measures should be put in place to encourage the formation of effective farmers' cooperatives and other farmers' organizations for the purpose of knowledge transfer, input and output marketing and distribution, savings mobilization, and farm credit sourcing and supply.

References Adejobi, A.O. 2004 . Rural poverty, food production and demand in Kebbi State, Nigeria. PhD thesis, Department of Agricultural Economics, University of Ibadan, Nigeria. Adejobi , A.a ., P.M. Kormawa , and A.L. Kehinde. 2004 . Profitability and determinants of manure use among crop farmers in Kaduna State, Nigeria. Bowen Journal of Agriculture 1(2): 160-172, October. Akinyele , L. , J. Omueti, and T. Ekpenyong 1991 . Economic and democratic reforms in Nigeria's development. Society for International Development, Lagos , Nigeria. Ben nett, L. 1992. Women , poverty and productivity in India. Economic Development Institute Seminar Paper No. 43, Worl d Ban k, Washington DC , USA. Bamey, C. and J.H. William .1994. Demographical behaviour and poverty: micro-level evidence from Southern Sudan . World Development 22 (7): 1031-1 044 . Bender, W. and S. Hunt, 1991. Poverty and food insecurity in Luanda. London: Food Studies Group Working Paper No. 1, Un iversity of Oxford. UK. Burra, N. 1998. Child labour, poverty and development: an agenda for UNDP in the next millennium. http://rrmeet.undp.org. in/_disc6/00000003.htm . Dercon , S. and P. Krishman. 1996. Income portfolios in rural Ethiopia and Tanzania: choices and constraints. Jou rnal of Development Studies 32(6): 850-875. Englama, D. and A. Bamidele. 1997. Measurement issues in poverty. Pages 141-156 in Selected papers from the Nigerian Economic Society's Annual Conference . FOS. (Federal Office of Statistics). 1999. Poverty and agricultural sector in Nigeria. Federal Office of Statistics Publication, Lagos , Nigeria. FAa (Food and Agriculture Organization). 1982. Food consumption tables for the Near East. Food and Nutrition Paper No. 20 , FAa. Rome , Italy.

Food Basket Foundation International. 1995. Nutrient composition of commonly eaten foods in Nigeria-raw, processed , and prepared , Unpublished paper, Food Basket Foundation International, 46 Ondo St., Old Bodija, Ibadan, Nigeria.131 pp. Gittinger, J.P 1972. Economic analysis of agricultural projects , The Johns Hopkins University Press, Baltimore , USA Green, D.A.G. and D,H. Ng'ong '0Ia.1993. Factors affecting fertilizer adoption in less developed countries: an application of multivariate logistic analysis in Malawi. Joumal of Agricultural Economics 44(1): 611-625. Hassan, R.M . and S .C. Babu. 1991 . Measurement and detemninants of rural poverty: household consumption patterns and food poverty in rural Sudan . Food Policy 16 (6): 451--460. Jambiya, G . 1998, The dynamics of population, land scarcity, agriculture and non-agricultural activities: West Usambara mountains, Lushoto District, Tanzania . ASU Working Paper No. 28, African Studies Center, University of Leiden , The Netherlands. Leavy, J. and H. WMe. 2000. Rural labour markets and poverty in sub-Saharan Africa. http://www.ids.ac.uklidsJpvty/pdf-files/ ru rallabou ri nssa ,pdf. McDonald , J .F. and R.A. Moffit. 1980. The uses of Tobit analysis. Review of Economics and Statistics 62:318-321 . Meludu, N.F. 1998. Changing lifestyles in famning societies of Sukumaland : Kwimba District, Tanzania. Dar es Salaam Institute of Resource Assessment and Leiden African Studies Center Working Paper No. 27. University of Leiden, The Nethertands, 40 pp. Obadan , M.a. 1997. Analytical framework for poverty reduction: issues of econom ic growth versus other strategies. Pages 1-18

in Proceedings of the Nigerian Economic Society's Annual Conference. Olayemi , J.K" A. Fadahunsi, and A.a . Adejobi. 2004. The challenge of building democratic, sustainable peasants' societies for a developed Nigeria. Technical report presented at the Third World Forum workshop, CESAG, 2-5 November 2004 , Dakar, Senegal.

Omonona , B.T. 2001 . Poverty and its correlates among rura l farming households in Kogi State. PhD thesis , Department of Agricultural Economics, University of Ibadan, Nigeria. Reardon , T. 1997. Using evidence of household income diversification to inform study of the rural non-farm labour market in Africa. World Development 25(5):735-747. Storck, H., B. Emana, B. Adenew, A. Borowiecki, and W. Hawariat. 1991 . Farming systems and farm management practices of smallholders in the Hararghe High lands. In Farming systems and resource economics Vol. 2. Wissench ails verlag vauk Kie!, Germany. Tobin, J. 1958. Estimation of relationship for limited dependent variables. Econometrica 26: 26-36. World Ban k. 1992. Nigeria: income distribution and poverty profile. Prepared by Mustapha Rouis in connection with the Basic Economic Mission to Nigeria. January-February World Bank, Washington DC, USA. World Bank . 1996. Nigeria: poverty in the midst of plenty. The challenge of growlh with inclusion . World Bank Poverty Assessment. Population and Human Resources Division, West Africa Department, Africa Region . Report No. 14733 UNI. World Bank, Washington DC, USA. World Bank. 2001 . World Development Report. World Bank, Wash ington DC , USA. Yunusa, M.B. 1999. Not farms alone : a study of rural livelihoods in the Middle Belt of Nigeria . ASC Working Paper No. 38. Afrika Studie Centrum. Leiden, The Netherlands/Center for Research and Documentation (CRD ), Kano, Nigeria.

Annex 1. Conversion factors for calorie requirements for different 8ge groups Age group

Male

Female

< 10 years

0. 6

0.6

10--13

0.9

0.8

14-16

1.0

0.75

17-50

1.0

0.75

> 50

1.0

0.75

Source: Culled from Storck at 01. (1991).

Annex 2.

CI ~_I

'' ~..o....L'~~'~'

PROS~ CcrnmlSltlE!s

Figure 1. Bomo State showing the project area.

Annex 3. Communities covered by the socioeconomic baseline survey SINo.

1 2 3 4

5 6 7

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

35 36 37 38 39

Agroecological zone

LGA

Village/Community

NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS NGS SGS SGS SGS SGS SGS SGS SGS SGS SGS NGS NGS SGS SGS SGS SS SS SS SS SS SS SS

Siu Siu Siu Siu Siu Siu Siu Siu Siu Siu Siu Siu Siu Kwaya Kusar Kwaya Kusar Kwaya Kusar Hawul Hawul Hawul Hawul Hawul Hawul Hawul Hawul Hawu l Hawul Hawul Hawul Hawul Hawul Hawul Hawul

Filin Jirgi Galdimare Kabura Mirnga Buratai Mbulamile Wakama Tum Sabon Layi Mandafuma Yamarkumi Maina Hari Mandagirau Wandal i Guwal Kwaya Kusa r Vina Dam Tilla Kukurpu Kwajafa Tashan Alade Shaffa Azare Yimirshika Ngwa Sakwa

Biu

Kimba Sa bon GaM Kumboa Azir Wawa Damboa Gumsuri

Damboa Damboa Damboa Damboa Damboa Damboa

NGS = northern Guinea savanna SGS = southern Guinea savanna 58 = Sudan savanna

Mararna

Hema Tangaramta Kidang Mbulatawiwi Gusi