You are on page 1of 17

Food Economics

ISSN: 2164-828X (Print) 2164-8298 (Online) Journal homepage: http://www.tandfonline.com/loi/sagc21

Households' food security status and its


determinants in the North-Central Nigeria

Abiodun Elijah Obayelu

To cite this article: Abiodun Elijah Obayelu (2012) Households' food security status
and its determinants in the North-Central Nigeria, Food Economics, 9:4, 241-256, DOI:
10.1080/2164828X.2013.845559

To link to this article: http://dx.doi.org/10.1080/2164828X.2013.845559

Published online: 25 Oct 2013.

Submit your article to this journal

Article views: 185

View related articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=sagc21

Download by: [University of Lethbridge] Date: 14 May 2016, At: 02:53


Food Economics, 2012
Vol. 9, No. 4, 241–256, http://dx.doi.org/10.1080/2164828X.2013.845559

RESEARCH ARTICLE

Households’ food security status and its determinants in the


North-Central Nigeria

ABIODUN ELIJAH OBAYELU*

Department of Agricultural Economics and Farm Management, Federal University of Agriculture, Abeokuta (FUNAAB),
Abeokuta, Ogun State, Nigeria
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Abstract
Study on household food security is very essential now that Nigeria is currently facing the problem of food crisis. This study
presents findings on household food security in the North Central Nigeria (NCN). The first objective was to review and
highlight the commonly used measures of households’ food security with their pros and cons; second, to determine the food
security status of households; and third, to analyse the determinants of household food security status in the study area. This
study employed a well-structured questionnaire to obtain data from a cross-sectional survey of 396 households selected from
Kwara and Kogi States of NCN with a multi-stage sampling technique. To analyse the data, descriptive statistics and
multivariate-ordered logit analyses were used. The descriptive results revealed that only 16% of the households were food
secure (FS), 36% food insecure without hunger, 28% FS with moderate hunger and 21% food insecure with severe hunger.
The results from the ordered logit revealed that geographical location, marital status, gender of the household head,
household size, food dietary diversity, total household expenditure, level of education, occupation of household head,
household dependency ratio, social capital and agricultural land-holding size significantly affect households food security
status. To transit from food insecure to FS, birth control, participation in agriculture, gender-sensitive food insecurity
alleviation policies that enhance endowments of female-headed households are imperative.

Keywords: Household; food security status; ordered logit; Rasch model; North-Central Nigeria.

1. Introduction life (USAID, 2008). This is one of the targets of the


Millennium Development Goals and widely consid-
Discussion on food security in different countries of
ered as a useful measure for evaluating the progress
the world has been an issue of considerable attention
of a country in terms of well-being (Vasco, 2006).
since 1970 (Ayalew, 1997) and still remains an
Attaining food security in its totality at household
ongoing issue. Household food security, although
level continues to be a challenge not only for the
complex, and a multi-dimensional phenomenon like developing nations but also for the developed world
poverty with an approximately 200 definitions and (FAO, 1997; IFPRI, 1996). For instance, the
450 indicators (Hoddinott, 1999); is most important Nigeria government has initiated a number of inter-
for the analyst since the household is the basic vention policies and programs aimed at ensuring
economic unit which determines the level of con- food security at household level in the country but
sumption by the individual. The importance of all these attempts have not yielded the much
households’ food security has become the concern acclaimed goals and objectives (Orewa & Iyangbe,
of national governments as well as the international 2009). Despite programs such as the River Basin
communities during the last few decades. A household Development Authorities (1977 to Date); Operation
is FS when all people at all times in a household have Feed the Nation (1976–1979); Green Revolution
sufficient physical and economic access to safe and (1979–1983); Directorate of Foods and Roads and
nutritious food to meet their dietary needs including Rural Infrastructure (1986–1993); National Agricul-
food preferences, in order to live a healthy and active tural Land Development Authority (1991–1999);

Correspondence: A. E. Obayelu, Department of Agricultural Economics and Farm Management, Federal University of Agriculture, Abeokuta (FUNAAB),
Abeokuta, P.M.B. 2240, Ogun State, Nigeria. Tel. +234-8034146503. Email: obayelu@yahoo.com

(Received 2 April 2013; accepted 8 August 2013)


# 2012 Taylor & Francis
242 A. E. Obayelu

Presidential Initiatives on Cocoa, Cassava, Cocoa, both developed and developing countries with
Rice, Livestock, Vegetables (1999–2007); FADAMA their pros and cons;
(irrigable land) I (1993–1999), FADAMA II (2004– . to empirically classify households in the North-
2007) and the ongoing FADAMA III projects Central Nigeria (NCN) into food security status
(2008–2013); the seven points agenda (2009– based on some available food security indica-
2012), the launch of commercial agriculture (2009– tors and
2014) and ongoing Agricultural Transformation . to analyse the basic demographic, economic and
Agenda which started in 2012 using the value chain social determinants of household food security.
approach on some specific agricultural products like
rice, cassava (Federal Ministry of Agriculture and
Water Resources, 2008; A. E. Obayelu & O. A. 2. Methodology
Obayelu, 2012), only 18% Nigerians were FS in This section gives a brief review of methods of
1996, 40% in 2005 (Sanusi et al., 2006) and 35% in measuring households’ food security status, the
2008 (Zoellick, 2008). One of the reasons for this is study area sampling procedure, methods of data
the fact that food production in Nigeria has not collection, methods and basis of models selection
increased at the rate that can meet up with the food and specifications for data analysis of the study.
demand of the increasing population of about 167
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

million (NPC, 2012). The apparent disparity


between the rate of food production and food 2.1. Review of methods of measurement of household food
demand has led to the increasing rate of food and security
livestock importation, as well as high rates in food Measurement of food security represents a challenge
prices thereby making it difficult for the poor to meet due to the complexity and wide array of factors
their food requirement (Adeoti & Egwudike, 2003; associated with this phenomenon (Abuelhaj, 2007).
Ogundari & Ojo, 2007). The inability to provide Literature had shown that modeling of food security
regular information on household food security and its determinants at household levels comprise
status in Nigeria has remained a source of concern many different methodologies and techniques. While
to both the government and the policy-makers alike. in developed countries, several methods have been
The policy-formulators and programme implemen- used by researchers to establish household’s food
ters are increasingly seeking measurement techni- security status, in developing countries, the use of
ques for food security that are simple to use, easy to some of the methods are prone to quite a number of
analyse and interpret as well as factors affecting food challenges (Table I).
security status of their people. Close examination
and analysis of household food security status and its
determinants is therefore very important not only 2.2. Study area and sampling procedure
from the economic point of view but also for human The NCN is one of the six geopolitical zones of
survival. Although, there are several factors affecting Nigeria. It comprises of Benue, Kogi, Kwara, Nasar-
food security that have been empirically investigated awa, Niger, Plateau states and the Federal Capital
in one part of the Nigeria or the other using different Territory Abuja. A multistage random sampling was
sets of target groups such as rural farm households in employed in the study to select the household heads.
just a state (e.g. Fakayode et al., 2009) or a Local The first stage was the random selection of Kogi and
Government Area (LGA) within a state (e.g. Idrisa Kwara States from NCN. The second stage was the
et al., 2008); we are not aware of any recent studies random selection of five LGAs per selected state
on households food security that have used micro- proportionate to the size of the LGAs in the state.
level data to analyse households food security status The third stage was the random selection of 5
and its determinants from a broader view like this Enumeration Areas (EAs) per LGA using the
study. The area-specific nature of the determinants Nigeria National Bureau of Statistic Core Welfare
of the households’ food security status in this study Indicator Questionnaire survey of 2006 categoriza-
will not only lead to the right formulation and tion of 10 EAs per LGA as a baseline. The strati-
implementation of food policy in the study areas fication of the study areas and selection of
but contribute to the existing empirical literature on respondents (household heads) were based on geo-
households’ food security in Nigeria and other graphy and random method, respectively, without
developing countries. any regard to the socioeconomic characteristics
This study has the following specific objectives: (such as education, income, age or gender distribu-
tion) of the area or those who were sampled. The
. to highlight the most commonly applied meth- fourth stage was the random selection of 400
ods of measuring household food security in household heads from the sampling frame of all the
Table I. Methods of measuring household food security/insecurity status.

Methods What is measured How does it measure Strengths Weaknesses References

United State Measure the Food security status of each household is i) It provides accurate and valid This method fails to capture food Bickel et al.
Household severity of assessed by their responses to 18 questions measurements of food security in the insecurity in elderly households as relevant (2000), Wolfe
Food Security household food (10 in households with no children) about general population, ii) the 18-item “core information to capture that are not et al. (2003),
Survey security food-related behaviors, experiences and module” of HFSSM is able to show both included in the module. That is, a Nord
Module (US- conditions that are known to characterize the prevalence and the severity of food situation where food maybe available and et al. (2008)
HFSSM) households having difficulty meeting their insecurity and hunger (Hamilton et al., accessible but may not be used due to
food needs. The questions cover a wide 1997), iii) the questions in the module are limited physical functioning by elderly
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

range of severity of food access problems able to measure four underlying people. This result to an underestimation

Households’ food security status and its determinants in the North-Central Nigeria
ranging from worrying about running out of conditions or behavior in the households: of the number of elderly who are food
food to children not eating for a whole day anxiety about the food budget or food insecure
supply; perceptions that food is inadequate
in either quantity or quality; reduced food
intake in adults and reduced food intake in
children
Household Measures HFIAS is an adaptation of the HFSSM It is a useful measure for comparing food The information arising from HFIAS can Deitchler
Food household food developed for developing or poor countries access across different population groups. be interpreted at community, district and et al. (2010)
Insecurity insecurity as by Food and Nutrition Technical Assistance It recall period used 4 weeks (30 days) national level. It is not appropriate for
Access Scale reported by the in 2007 to assess whether households have against 12 months often used by HFSSM. interpretation at individual or household
(HFIAS) households experienced problems with accessing food HFIAS provides information about the level, in the sense that it cannot be used
themselves access during the last 30 days. It contains 9 current state of food insecurity (rather for targeting of specific households/
items rather than 18 to capture experiences than past or future state of food individuals
associated with varying levels of food insecurity). This method classifies
insecurity severity. It reflects three domains households into three food security groups
perceived as central to the experience of food (FS, low food security, very low food
insecurity cross-culturally: i) anxiety about security)
household food supply; ii) insufficient
quality, which includes variety, preferences
and social acceptability and iii) insufficient
food supply and intake and the physical
consequences
Coping This is a relative In-depth interview to identify action people CSI is relatively easy to construct, using It is not an absolute measure of food Maxwell
Strategies measure of food take when they do not have enough food or household information with focus group insecurity that can provide an indication of (1996),
Index (CSI) insecurity money to buy food. A questionnaire (community) perceptions and is well the food gaps. It also does not have Maxwell
assessing the frequency of use of the strategy suited to assess the short-run impact of a information to distinguish between pre- et al. (1999)
is developed from which food security score food shock (transitory food insecurity). It crisis coping strategies (associated with
is derived by applying security weightings can also be used to monitor the impact of chronic poverty and food insecurity) and
various food aid interventions crisis-driven strategies; hence, posing
difficulties in distinguishing chronic food
insecurity from transitory food insecurity
Household Measures It provides a direct estimate of the food gap i) allows insights into the crisis-induced i) Information gathering is through key Boudreau
Economy household food by establishing a household “food balance” “change” in food security compared with informant and focus group interviews, (1999)
Approach security and early which matches “resources” (all income and baseline information, ii) it allows which may compromise representativeness
(HEA) food sources converted into kilocalorie or identification of chronic food insecurity, and requires high skills to avoid biased

243
244
Table I (Continued)

Methods What is measured How does it measure Strengths Weaknesses References

A. E. Obayelu
warning food cash equivalents) against “requirements” iii) has the capacity to estimate a food information and undertakes the internal
crises (food intake and other essential needs deficit taking into account both the impact cross-checking for consistency that is
converted in kilocalories or cash equivalents) of a shock on food availability, prices, food integral to the method; ii) it requires good
and non-food incomes and savings and the judgment and considerable experience to
households’ potential to cope quantify qualitative information, for
example on coping strategies
Household Measures Simple sum of the number of different foods Household-level dietary diversity is i) Household-level consumption patterns Ruel (2002)
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Dietary household food eaten by that person over the specified time strongly associated with per capita con- and changes have to be recorded, which
Diversity insecurity period or a weighted sum, where the weights sumption (a proxy for income) and can be time-consuming, ii) the method
Score reflect the frequency of consumption, and energy availability, suggesting that diet- does not shed light on causes (incomes,
(HDDS) not merely the number of different foods. In ary diversity could be a useful indic- prices, own-production) of a consumption
this method, the proportion of households ator of deterioration, iii) technique may result in
deviating from the score corresponding to a household food security estimates biased upwards because of the
minimum balanced and sufficient diet can be tendency to exaggerate shortfalls when
calculated according to given thresholds. A external aid is expected
calibration of the score is needed to provide
equivalences with kilocalorie (and, ideally,
macronutrient) intakes
Dietary Intake To determine The method estimates an average daily per i) It measures food consumption directly, i) It rely heavily upon respondents’ Jenson and
Assessment food security capita calories consumed, adjusting for the not the food availability; ii) it addresses memory that can lead to measurement Miller (2010),
(DIA) status access to culturally appropriate food. Cost of both dietary quality and calorie intakes at errors; ii) the assessment of adjusted recall Oluyole et al.
or COC minimum recommended energy level (Z) is individual levels estimations is a very difficult task that may (2009),
calculated as: Z = e(a+bL), where L denotes lead to high measurement errors; iii) the Babatunde
Recommended Daily Energy Level (Kcal); cost of applying recall methods in national et al. (2007)
“a” is the intercept term; b = coefficient of survey is high. iv) there is no consensus on
the calorie consumption. A household whose the thresholds needed for subsistence and
average cost of daily calorie consumption is any of the recommended thresholds would
equal to or more than Z is said to be food vary significantly across individuals, their
secure while a household with average cost of jobs and geographical locations
daily calorie consumption lower than Z is
considered food insecure
Food security To determine This method is based on the food security It has the similar strengths as the use Has similar weaknesses as COC Arene and
index food security line using the Recommended Daily Calorie of COC Anyaeji (2010),
status approach. Household whose food security Omotesho
index is greater or equal to the et al. (2010)
Recommended Daily Calorie Intake are
regarded as food secure and those whose
food security index is lower than the
recommended Daily Calorie Intake (2260
Kcal) are considered food insecure
Food Insecurity To determine The method is based on the perception or i) It captures psychosocial dimensions of i) It is a difficult task to establish cut-off Rose and
Experienced- food security experience reported by the respondents. The food security along with the physical points for classifying households into Charlton
based status extent of severity is based on a series of experience; ii) measures directly the different levels of food insecurity; ii) The (2001)
Table I (Continued)

Methods What is measured How does it measure Strengths Weaknesses References

Measurement questions, a number of which rely on phenomenon of food security based on the scales do not have questions on issues
Scales qualitative and subjective assessments experiences of individuals; iii) better related to water access, and food and
(FIEMS) understanding of causes and water safety hazards caused by microbial
consequences of food insecurity and and other environmental contaminants
hunger
Food Measures Food items are grouped into eight standard The method appears to be relatively un- i) It is slightly more time consuming and WFP/
Consumption household food food groups with a maximum value of seven biased ii) cut-offs are hard to justify FAO (2008)
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Score (FCS) insecurity days/week. The consumption frequency of

Households’ food security status and its determinants in the North-Central Nigeria
each food group is multiplied by an assigned
weight that is based on its nutrient content.
Those values are then summed obtaining
the FCS
Food poverty Measures food It is calculated as the ratio of food The method allows for determination of There is lack of knowledge of the type of Devereux
(purchasing security expenditures to the cost of the food basket. It food access gap (deviation of food foods on which the expenditures are made (2006)
power) involves converting food quantities into expenditures from the cost of a minimum when using this method
approach calorie (and other nutrient) equivalents using food basket)
nutrient conversion tables; deriving estimates
for apparent calorie consumption per person
per day using household composition data;
and comparing the apparent consumption
with the reference levels - the nutritional
norm (2100 kcals) and previous
consumption levels
Household Measures food It requires getting information about the i) It is used for the evaluation of national i) It measures the available amount of food Amaza et al.
Expenditure security quantity of food bought and expenditures on food and nutrition programs; ii) dietary and ignores the consumed amount at a (2006), Rose
Survey different food types (items) consumed within quality data can help understand the food given timeframe; ii) it does not account for and Charlton
Method and outside the house security dimensions the amount of food consumed outside the (2001), Bickel
(HESM) household, for example food consumed et al. (2000)
when visiting the relatives and friends; iii)
conversion of available food to the calorie
intakes involves major assumptions that
can cause measurement errors; iv)
difference of data collection methods
worldwide make it difficult to compare the
estimates across countries, regions and
different purchasing power parities
Anthropometry Indirect measure It involves in-depth interviews and i) It is highly standardized measurements The method is costly both monetarily and Wolfe and
of food security observation of individual characteristics like of weight and ii) it allows for mapping time wise; i). the indicators measure Frongillo
weight and height of infants, young children, nutritional security from the local to the nutritional status that results from the (2000), Mariara
youth and adults national levels interaction of food security and health et al. (2006)
status, hence is an indirect method to
measure food security

245
Source: Computed by the author from various sources.
246 A. E. Obayelu

houses within the selected EAs. Household heads statistics. Mean square residuals (MnSq) estimated
were selected because they were assumed as the as odd ratio were used to assess the goodness-of-fit
major decision makers. A total of 219 and 181 of each item compared with the assertions of the
household heads were selected from Kogi and Kwara Rasch model. MnSq are ratios of the observed versus
States, respectively, based on a higher population the expected scores. The expected MnSq values >
size in Kogi (3,314,043) compare to Kwara 1.3 are judged indicative of a poorly fitting or erratic
(2,365,353) as given by 2006 Nigeria Population item and MnSq values < 0.8 indicates that the item
Census (NPC, 2006). However, of the 400 house- is redundant or lacked variability with respect to the
hold heads sampled and interviewed, only 396 were information it shares with another item (Hamilton
found useful for the analysis. et al., 1997). That is, infit range of 0.8–1.3 are
generally considered to be quite good, The out-
comes of our data show that many questions of the
2.3. Types of data and methods of data collection 18 food security items were redundant and were
All the data on households’ demographic and socio- therefore removed from the scale because of their
economic characteristics (such as location, age, failure to demonstrate goodness-of-fit to the Rasch
educational level, marital status, gender, income, model. Only eight “basic” adult and household items
household size) as well as food security variables were therefore used (Annexure 2). Nearly all the
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

(like the types, the quantity and quality food con- items infit mean squares of our data from the eight
sumed in the previous seven days to the time of the items were within the fit criteria of 0.8–1.3 showing
interview, access to credit, total food expenditure that our data fit the model (Annexure 1). Other
share, access to public health services) were collected items were only omitted from measurement analysis
through a cross-sectional survey of selected house- but were used in the classification of households into
hold heads in Kogi and Kwara States, Nigeria food security status.
between 2006 and 2007. The main survey instru- Responses to the questions in the food security
ments used were structured questionnaire and per- survey module are combined into a scale using non-
sonal interview methods. linear statistical methods based on the Rasch meas-
A major limitation to this study is that due to lack urement model. We applied this model through
of physical records, data collected on households’ Conditional Maximum Likelihood Estimation from
food production, food purchases and food consump- winsteps scientific software implemented in SAS
tion were based largely on memory recalls which are with a STATA subcommand.
subject to bias on the part of the respondents. Rasch measurement model is a non-linear factor
analysis with binary variable such as “yes” or “no”
response to a survey item and it falls into the family
2.4. Empirical model selection and specification of Item Response Theory models (Hamilton et al.,
The model selection procedure employed in this 1997). The model is a one-parameter model, mean-
study involves two phases. The first phase was the ing that it models the “one” parameter difference
selection of appropriate model for classification of between person position and item difficulty. The
households into food security status and the second model assumes specifically that the log of the odds of
was selection of appropriate model to determine a household affirming an item is proportional to the
factors affecting food security status in the difference between the “true” severity level of the
study area. household and the “true” severity level of the item as
shown in Equation 1. A household’s responses
2.4.1. Rasch measurement model. We employed across the 18 food security questions are summed
Household Food Security Survey Module (HFSSM) (where yes = 1; no = 0) and this raw score value
methodology in this study. The questions (Annexure represents the household’s food security/insecurity
1) in the food security module measured four score. The model calculates a value on the scale for
underlying conditions/behaviors in the households: each household based on the number of questions
anxiety about the food budget or food supply; answered affirmatively adjusted for the number and
perceptions that food is inadequate in either quantity severity of the questions answered.
or quality; reduced food intake in adults and reduced
lnfPtðhÞ=½1  PtðhÞg ¼hbt, ð1Þ
food intake in children. Although HFSSM has
extensively been used in developed countries such where ln is logarithm, Pt(h) is the probability of an
as United States of America based on their ability to affirmative response to item “t” for a household that
keep record on foods they consume, before we has a food-insecurity score of h; 1 − P is the
adapted this technique to our data, we checked for probability that the household will deny the item.
the reliability as well as validity of our data with fit Because the logits are equal units of measurement,
Households’ food security status and its determinants in the North-Central Nigeria 247

they are additive. The index “t” runs over the 10 assumptions are more difficult to meet than those
adult-referenced items for childless households and of classical test theory. The benefits derived from its
over all 18 items for households with children use come at the cost. When data do not adequately
(Annexure 1). The food security scale provides a fit the model, the instrument construction process
continuous, graduated measure of the severity of must begin anew. Fit indices in the Rasch model are
food deprivation across the range of severity repre- being influenced by the testing context, making
sented by the items. Based on their food security definition of a standard for defining fit versus misfit
scale scores, households are also classified into food difficult. Varied standards exist for deciding fit to be
security status categories of the population. Follow- adequate or not, and for deciding on deletion of item
ing the recommended cut-points by Gulliford et al. or person data.
(2006), USDA (2000), households with all affirmat-
ive response to the 18 questions or 0–2 are classified 2.4.2. Ordered logit (Ologit) model. Literature has
as FS, those with 3–7 as food insecure without provided various models for determining factors
hunger (FIWH); 8–12 as food insecure with moder- influencing food security status of households with
ate hunger (FIWMH) and those between 13–18 as key among them that have been used by many
food insecure with severe hunger (FIWSH). For researcher in developing countries as Tobit model
adult-referenced items, households with 0–2 affir- (Etim & Solomon, 2010), Probit model (Oluyole
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

matives are classified as FS; those between 3 and 5 et al., 2009) and Logit (Babatunde et al., 2007)
affirmatives as FIWH; 6 and 8 affirmatives as based on the method they have employed to classify
FIWMH and between 9 and 10 affirmatives their respondents into food security status. This
responses as FIWSH. study however employed the ordered logit model
(the proportional odds model) to analyse factors
Key assumptions of Rasch model. Households with a affecting food security status of households. The
higher food-insecurity score have a higher probability selection of the model is in line with Greene (2000)
of answering “yes” to each item. The more FS an since our dependent variable (the household food
individual is, the more likely he or she will respond security status) is both categorical and ordinal. The
negatively to dietary quality items. The probability of coefficients show that for one unit increase in the
an affirmative response to each item is statistical predictor (independent variable), the response vari-
independent for all households that share the same able (dependent variable) is expected to change by
level of food insecurity. its respective regression coefficient while the other
variables in the model are held constant. In this
Strengths of the model. The model helps at converting study, a significant positive coefficient indicates that
series of questions in the HFSSM into a food a one unit increase in the independent variable
security scale and provides a concept of food security increases the likelihood that household will be FS.
that is well understood by policy makers. It provides On the other hand, a significant negative coefficient
a means of measuring both the prevalence of food indicates that a one unit increase in the independent
security and the severity of hunger by incorporating variable increases the likelihood that household will
the essential elements of perceptions of food insec- be food insecure.
urity and hunger by the people most affected. There In the case of a dummy, we cannot really take a
is no provision in the model for deliberate or derivative with respect to “X.” So we estimated the
unconscious deception, guessing or any other vari- effect on the predicted probability of switching the
able that might impinge on the responses provided dummy from zero to one, holding other x variables
because the left asymptote always approaches a zero at their means. The estimates are not dependent on
probability. With the model, we are able to divide the ancillary parameters (the cut-offs); the ancillary
the studied population into four distinct categories of parameters are used to differentiate the adjacent
food security (FS, FIWH, FIWMH and FIWSH) levels of the response variable.
based on the differing conditions, experiences and In the ordered logit model, there is a continuous,
behavior patterns that characterize each range of unmeasured latent variable Y*, whose values deter-
severity. mine what the observed ordinal variable Y (food
security status) equals. The model is explicitly stated
Weaknesses of Rasch model. As a limitation, Rasch in Equation 1:
model is overly restrictive or prescriptive because it XK
does not permit each item to have a different Yi ¼ b X þei ¼Zi þei .
k¼1 k ki
ð2Þ
discrimination. The model also requires larger sam-
ple sizes compared to classic testing procedures. The where Y* continuous latent variable, bk is the vector
model is termed a “strong” model since its of parameters or coefficients to be estimated by the
248 A. E. Obayelu

model, Xki represent vector P of the explanatory or explanatory variables was also based on empirical
independent variables, Zi ¼ Kk¼1 bk Xk , zi is the findings as presented in Table II.
random disturbance term reflecting that relevant
variables may be left out of the equation, or variables
may not be perfectly measured. The probabilities of 3. Empirical results and discussion
respondents being in any of the identified categories 3.1. Results of findings
are determined using the natural log of the cumu-
lative distribution following Booroah (2002). There are four household food security categories
The marginal effects of changes in the independ- classified by the Rasch measurement model in this
ent variables are computed as in Equation 2: study (Table III) which many of the previous studies
  (such as, Omotesho et al., 2010; Adenegan and
dprop y ¼ 0 x
Adewusi, 2007; Oluyole et al., 2009) in Nigeria and
¼ f ðl0  xb Þ.b,
dx other developing countries overlooked by just classi-
  fying households into FS or insecure through the use
dprop y ¼ 1 x
¼ ½f ðl1  xb Þ  f ðl0  xb Þ.b, of the notable Cost-of-calorie approach and Food
dx Security index approach. In Table III, only about
 
dprop y ¼ 2 x
16% households in the study area were FS, 36%
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

¼ ½f ðl2  xb Þ  f ðl1  xb Þ.b FIWH, 28% FIWMH and 21% FIWSH. This shows
dx that about 84% of people in the study area are still
 
dprop y ¼ 3 x food insecure at very levels of food insecurity.
¼ f ðl2  xb Þ.b. ð3Þ The results in Table IV are displayed as probab-
dx ility function. A total of 17 explanatory variables
Where 0, 1, 2, 3 are the various categories (FS, were included in the econometric model of which 11
FIWH, FIWMH and FIWSH); χ is the independent explanatory variables had significant influence on
variable, µ0, µ1, µ2 are the cut-off values for the household food security in the study area. The
ordered logit model and f is the cumulative probab- likelihood ratio chi-square of 418.28 with a p-value
ility function. The marginal effect for the dummy of 0.0004 suggests the fitness of the model. There
variable was calculated by taking the probabilities for was no intercept because it was not identified
each category at v = 0 and at v = 1, and take the independently of the cut-points and the STATA
difference. package sets the constant to zero and estimates the
cut points for separating the various levels of the
response variable. The pseudo R-square associated
2.5. Selection and definition of food security and with ordered logit model was observed as inappro-
explanatory variables with their a priori expectations priate measure of the predictive power of ordered
There are some variables that have been previously response models. Therefore, the chi-squared value
considered in Nigeria and other countries as affect- and the log-likelihood ratio criteria were used to
ing household food security status. The selection of evaluate the effectiveness of the model in line with
food security variables in this study relies on previ- Megan (2010). Also, geographical location, marital
ous empirical studies as well as economic theory like status, gender, household size, food dietary diversity
demand and consumption theory. For instance, we (Fddiv), total household expenditure, level of edu-
included the total household expenditure as a proxy cation, Occupation, Dependency ratio of household,
for total household income and food expenditure social capital, agricultural land holding had signific-
share following Adenegan and Adewusi (2007), ant influence on food security status of the house-
Babatunde et al. (2007), Omonona and Agoi holds (Table IV). The cut-off points shown at the
(2007), Arene and Anyaeji (2010), Oni et al. top of the output in Table IV indicate where the
(2011), Tshediso (2013) and a study (Umeh & latent variable is cut to make the four food security
Asogwa, 2012) and other analysis of food sector in groups.
Nigeria that revealed a demand shift for certain types If the latent variable Y* takes the value less than
of foods as income changes. The inclusion of the two −5.1811, the ordinal dependent variable will take the
variables was also based on the indirect relationship value of 0 (FS). If it is between −5.1811 and
between food security and food expenditure in the −1.3422, the ordinal dependent variable is 1
total household expenditure using the Engel’s Law, (FIWH). If the Y* is between −1.3422 and 1.4824,
which states, “the proportion of consumer’s budget the ordinal dependent variable will take the value of
spent on food declines with increase in the house- 3 (FIWMH) and if it is more than 1.4824, the
hold income” (Mundlak, 2000). The selection of our households will be FIWSH (Table IV).
Households’ food security status and its determinants in the North-Central Nigeria 249

Table II. Description of the variables specified in the model.

A priori
expectation
with respect
Variable acronym Variable meaning Types of measure to FS Source(s)

Dependent variables Whether household is FS, FIWH, Dummy (FS = 0, Nord (2000), Gulliford et al.
FIWMH, or FIWSH FIWH = 1, (2006), USDA (2000)
FIWMH = 2,
FIWSH = 3)
State of the household Whether respondent is from Kogi or Dummy (Kogi = ± Obamiro et al. (2003)
(State) Kwara state 1, otherwise 0)
Geographical Whether respondent is in rural* or Dummy (rural = ± Babatunde et al. (2007),
location (Gloc) urban area 1 not rural = 0) Ibrahim et al. (2009), Arene
and Anyaeji, (2010), Sultana
and Kiani (2011)
Marital status (Mstat) Whether respondent is married or not Dummy variable + Oni et al. (2011)
married (married = 1,
otherwise 0
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Age of household head Age of the household head - in years ± Oni et al. (2011), Sultana and
( Age) Kiani (2011)
Educational status (Edu) Educational level of household head Number of years + Oni et al. (2011), Sultana and
of formal Kiani (2011)
education
Household size (Hhsz) Number of adults and children who Number − Oni et al. (2011), Adebayo
are resident member (2012), Shaikh (2007)
Gender (Gend) Sex of the household head Dummy (male = ± Babatunde et al. (2007),
1, otherwise = 0 Paddy (2003)
Occupation (Occup) Occupation of the household head (1 Dummy (if into + Oni et al. (2011), Shiferaw
= farm 0= nonfarm) agriculture = 1, et al. (2003)
non-agriculture
= 0)
Total expenditure (Texp) Total expenses of households on food Amount in + Gebrehiwot (2009),
and non-food items (a proxy for naira (N) Omotesho et al. (2010),
household total income) Sultana and Kiani (2011)
Household food Proportion of household spent on food No unit because ± Hofferth (2003), Adenegan
expenditure share (Fes) expenditure from total households it is in ratio form and Adewusi (2007),
expenditure (food and non-food items) Olorunfemi et al. (2009)
Dependency ratio of This is the ratio of the number of non- Number − Hofferth (2003), Gebrehiwot,
household (Drh) working members of the household to (2009), Sultana and
total household size Kiani (2011)
Social capital (Soccap) Whether respondent belong to any Dummy variable + Babatunde et al. (2007),
local level institutions (member of social Bamire (2010), Sultana and
group = 1, Kiani (2011), Miller
otherwise 0) et al. (2011)
Remittances (Rem) Cash received from migrant members Amount in + Babatunde et al. (2007),
of family, friends and other groups naira (N) Babatunde and
Martinetti (2010)
Access to credit (Acrd) Privilege of getting credit for Dummy variable + Arene and Anyaeji (2010)
household food consumption (having access =
1, otherwise 0)
Fddiv How often household have consumed Number + Obayelu (2009), Hoddinott
different food groups over a referenced and Yohannes 2002)
period of seven days
Agricultural land Size of agricultural land held by ha + Pankomera et al. (2009),
holding (Land) household head Bamire (2010)
Access to public health Access to public health services by Dummy variable + Khan and Gill (2009),
services (Aphs) respondents (having access = Omotesho et al. (2010)
1, otherwise 0)

*Rural areas are areas lacking access to some basic social and infrastructural facilities like good road network system, telecommunication
system, water, electricity and so on with primary occupation of people living in this areas as farmers. Respondents were asked to identify
their location themselves.
250 A. E. Obayelu

Table III. Ordered dependent variable representing household food security status.

Ordered Cut-off rate based on affirmative response to the 18 Percentage


value Food security status food security questions Frequency (%) Probability

Y = 0 FS 1–2 63 15.9 0.1591


Y = 1 FIWH 3–7 141 35.6 0.3561
Y = 2 FIWMH 8–12 111 28.0 0.2803
Y = 3 FIWSH 13–18 85 20.5 0.2146
Total households sampled 396 100.0 1.0

The estimated coefficients and standard errors ordinal levels. The following sub-section discusses
reveal which factor influence respondents food the statistical significance of the coefficients and the
security status while a significant coefficient suggests marginal effects with particular reference to the FS
that the likelihood of food security will increase/ group (Y = 0).
decrease as the response on the explanatory variable
increase/decrease (Booroah 2002).
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

3.2. Discussion of findings


Table V shows that marginal effects of each
explanatory variable on the probability of food The results show that geographical location (rural =
security status. The marginal effects provide insights 1, otherwise = 0) had a significant, and negative
into how the explanatory variables shift the probab- relationship with the probability of being FS indic-
ility of household food security between the four ating that households who are in rural area, had
lower probability of being FS. Although this sound
Table IV. Results of the ordered logistic regression.
contrary to our expectation that people in rural areas
whose predominant occupation is farming are always
FS. This result is consistent with past studies such as
Number of
observation 396 Pseudo R2 0.3942 Arene and Anyaeji (2010), Ibrahim et al. (2009) who
LR chi2(17) 418.28 log-likelihood −321.42115 observed that urban households are more FS than
rural households but contrary to Sultana and Kiani
Prob> chi2 0.0004
(2011), Babatunde et al. (2007) who employed Cost
Variables Estimated β Standard P > ∣z∣
values error of Calorie (COC) function proposed by Greer and
Cut1 −5.1811 0.9885 Thorbecke (1986) to establish food security status
Cut2 −1.3422 0.9229 and logit model for its determinants and observed
Cut3 1.4824 0.9345 that rural households are more FS than urban
+
State of −0.0795 0.2595 0.759
households in Pakistan and Kwara State Nigeria,
household
+
Geographical −0.5697*** 0.2196 0.009 respectively. The results might be as a result of high
location quantity of food households consume in rural areas
+
Marital status −0.4840* 0.2518 0.055 rather than the quality. Households’ educational
Age of 0.0187 0.0124 0.133 status was significant (p < 0.1) and had a positive
household head
relationship with food security in line with our a
Educational status 0.0000158* 8.74e-06 0.070
Household size −0.4213*** 0.0641 0.000 priori expectation. This implied that increasing
+
Gender 1.8960*** 0.3845 0.000 education level is associated with an increasing
+
Occupation 1.3474*** 0.3201 0.000 probability of being FS (Table IV). The coefficient
Total expenditure 0.8672*** 0.2286 0.000 of the gender of the household head was also positive
Food expenditure 0.1511 0.4275 0.724
and significant at p < 0.01. This shows that respon-
share
Dependency ratio 1.4673*** 0.5173 0.005 dents who are male in the study area had higher
of household probability of being FS. This was in line with several
+
Social capital 0.3488*** 0.0952 0.000 other studies such as Oluyole et al. (2009), Omo-
Remittances 0.0000165 0.0000396 0.677 nona and Agoi (2007) used a household-based
+
Access to credit 0.0903 0.1568 0.564
survey, COC and logit model to both classify cocoa
Fddiv 1.8271*** 0.2831 0.000
Agricultural land 1.0660*** 0.2320 0.000 farming and urban households to food security status
holdings and factor influencing them respectively. This how-
Access to public −0.1859 0.3790 0.624 ever is in contrast to Paddy (2003) who observed
health services that female-headed households are more FS than the
(+) represent dummy variable (0 or 1). ***Significance at 1% male-headed. The coefficients of occupation and
level, **Significance at 5% level, *Significance at 10% level. social capital were positive and significant at p <
Households’ food security status and its determinants in the North-Central Nigeria 251

Table V. Results of the marginal effects on probability.

Marginal Standard Marginal Standard Marginal Standard Marginal Standard


Variables (Xs) effect (Y0) error effect (Y1) error effect (Y2) error effect (Y3) error
+
State of −0.002096* 0.00699 −0.017457 0.05665 0.016219* 0.05295 0.003335* 0.0107
household
+
Geographical −0.015238** 0.00679 −0.123879*** 0.04721 0.11508*** 0.04427 0.024040** 0.00999
location
+
Marital status −0.012847 0.00731 −0.105642* 0.05441 0.098029 0.05086 0.020461 0.01102
Age of 0.0004843 0.00033 0.0041357 0.00276 −0.003819 0.00256 −0.000801 0.00055
household
head
Educational 4.09e − 07* 0.00000 3.50e − 06* 0.00000 −3.23e − 06* 0.00000 −6.77e − 07* 0.00000
status
Household size −0.010888*** 0.00273 0.092973*** 0.0153 0.085862*** 0.01475 0.017998*** 0.00383
+
Gender 0.1032013*** 0.03907 0.267870*** 0.03467 −0.323363*** 0.04941 −0.047708*** 0.01032
+
Occupation 0.0525163*** 0.019 0.246130*** 0.04605 −0.255527*** 0.0534 −0.043119*** 0.01054
Total 0.0238758*** 0.00812 0.185742*** 0.04805 −0.173038*** 0.04541 −0.036580*** 0.01127
expenditure
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Food 0.0039046 0.01108 0.033342 0.09437 −0.030792*** 0.08719 −0.006454 0.01827


expenditure
share
Dependency 0.0379201*** 0.0148 0.323806* 0.11577 −0.299043*** 0.10748 −0.062683*** 0.02383
ratio of
household
+
Social capital 0.0090139*** 0.003 0.076971*** 0.02144 −0.071085*** 0.01999 −0.014900**** 0.00469
Remittances 4.27e − 07 0.00000 3.65e − 06 0.00001 −3.37e − 06 0.00001 −7.06e − 07 0.00000
+
Access to 0.0023342 0.00409 0.019932 0.03461 −0.018408 0.03198 −0.003859 0.00673
credit
Food dietary 0.0482961*** 0.01237 0.377444*** 0.05418 −0.332073*** 0.04851 −0.093667*** 0.02139
diversity
Agricultural 0.026088*** 0.00752 −0.233127*** 0.05038 0.207684*** 0.04451 0 .051531*** 0.01469
land
holdings
Access to 0.0044798 0.00853 0.0416651 0.08605 −0.037633 0.07599 −0.008512 0.01861
public health
services
Pr(y = 0) 0.02655 Pr(y = 1) 0.53240 Pr(y = 2) 0.39632 Pr(y = 3) 0.04472
(predict, (predict, (predict, (predict,
outcome (0)) outcome (1)) outcome (2)) outcome (3))

(+) dy/dx (marginal effects on probability) is for discrete change of dummy variable from 0 to 1, ***Significance at 1% level, **Significance
at 5% level, *Significance at 10% level.

0.01. This implies households who are into agricul- level of education of the household heads has
tural production (either primary or secondary occu- significant effect on the probability of households
pation) has probability of increasing their FS status being FS in Kwara State, Nigeria. Larger household
by 1.35 while those who are members of social sizes are associated with a negative food security
network groups have the probability of increasing status. Larger household sizes require increase food
their food security by 0.35 while holding other expenditure and competition for limited resources.
factors constant. A unit increase in household size will decrease the
The results of the marginal effects in Table V probability of being FS by 0.011. This is also in line
shows a unit decrease in household size is expected with other previous findings such as Babatunde
to lead to 0.01 increases in the probability of being et al. (2007).
FS. The findings agree with our a priori expectation Furthermore, by one unit increase in total house-
and previous studies carried out in recent study in hold expenditure (proxy of total household expend-
Nigeria. A unit increase in the level of education of iture), we expect 0.02 increase in the probability of
household head will increase the probability of being in FS with other factors held constant. This is
household to be FS by 4.09 when all other factors expected because, increased in income all things
are held constant. This is in conformity with being equal will results to increase access to food.
Babatunde et al. (2007), who also observed that the The result is also consistent with Kuwornu et al.
252 A. E. Obayelu

(2013), Arene and Anyaeji (2010), Babatunde et al. 4.2. Policy recommendations
(2007), Adenegan and Adewusi (2007) who
There are several implications of our findings. We
employed COC intake to classified their respondents
considered some of these by individually looking at
into food security status and logit models to deter-
each factor that has shown significant influence on
mined factors affecting the status with a positive and
household food security.
significant relationship between household income
The effect of geographical location captured by a
and food security. For household dependency ratio,
dummy (rural 1, otherwise 0) on household food
a one unit increase in households’ dependency ratio,
security status shows that there may not be uniform
is expected to lead to a 0.04 increase in the
factors affecting household food security across rural
probability of being in an FS, given that all other
and urban areas. So, there is need to put into
variables in the model are held constant. This
consideration the geographical location of house-
finding is contrary to our a priori expectation of a
holds while planning and designing policies for food
decrease in the probability of being in FS as increase
security.
in the number of non-working member of household
The study has provided evidence that gender of
or dependency ratio is expected to increase the food
head of a household play a key role in determining
insecurity level of household (Ojogho, 2010). A unit
food security status of households. Thus, gender-
increase in membership of social network (social
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

sensitive food insecurity alleviation policies that


capital) is expected to increase the probability of
enhance endowments of female-headed households
being FS by 0.009 in the study area. This findings is
should be a key ingredient of food insecurity reduc-
consistent with Oni et al. (2011) who observed that a
tion strategy.
unit increase in social capital (p < 0.01) increases the
Land-holding size was also found as one of the
probability of household in Ogun State, Nigeria to
significant factors in ensuring food security to the
be FS by 0.0991. A unit increase in household
households. Household head with greater farm land
diversity foods intake (Fddiv) is expected to lead to
size showed better food security status than the less
0.05 increases in the probability of being in an FS
endowed households. It is therefore very important
group, given that all other variables in the model are
for the government to reform the land use act in the
held constant. This is in accordance to our a priori
country so as to make agricultural land available to
expectation. In addition, a unit increases in the size
people in attempt at solving the problem of food
of agricultural land held by household head increase
insecurity in the country.
the probability of being FS by 0.026. This result is
To transit from food insecure to FS class as
also in line with our a priori expectation and
revealed by the other significant variables in this
Pankomera et al. (2009) and Bamire (2010 who
study, birth control through enlightenment pro-
observed that increase in the land holdings size of
gramme on number of children that household
farm households in the dry Savannas of Nigeria
head can cater for, participation in agricultural
improves probability of a household being FS by
production, increase in diversity of food intakes,
0.07 units.
membership of social network ( social capital)
become very imperative.
4. Conclusion and recommendations
References
4.1. Conclusion
Abuelhaj, T. (2007) Methodological concerns in the measurement
Results of the assessment of households’ food of undernourishment, dietary diversity and household food
insecurity. Accessed 1 October 2008, available at: http://km.
security status in the NCN, revealed that 16% of
fao.org/fileadmin/user_upload/fsn/docs/fao_tareq.pdf
households were FS, 36% FIWH, 28% FIWMH Adebayo, O. O. (2012) Effects of family size on household food
and 21% FIWSH. Household size, gender of security in Osun State, Nigeria. Asian Journal of Agriculture
household head, geographical location, occupation and Rural Development, 2(2), 136–141.
of household head, total expenditure, household Adenegan, K. O. & Adewusi, A. O. (2007) Determinants of food
security status of rural households living with HIV/AIDS in
dependency ratio, social capital, agricultural land- Southwestern Nigeria. African Journal of Biomedical Research,
holding size of the household head and Fddiv are 10(1), 9–18.
important determinants of household food security Adeoti, A. I. & Egwudike, O. I. (2003) Determinants of house-
status that have to be taken into consideration by hold food security among urban farmers. ASSET - Series A, 3
(3), 99–105.
governments and development agencies wishing to Amaza, P. S., Umeh, J.C., Helsen, J. & Adejobi, A. O. (2006)
promote the food security status of households in Determinants and measurement of food insecurity in
the study area. Nigeria: Some empirical policy guide. Paper presented at
Households’ food security status and its determinants in the North-Central Nigeria 253
International Association of Agricultural Economists Annual Greer, J. & Thorbecke, E. (1986) A methodology for measuring
Meeting, Queensland, Australia, August 12–18, pp. 15 food poverty. Applied at Kenya. Journal of Development
Arene, C. J. & Anyaeji, R. C. (2010) Determinants of food Economics, 24, 59–74. doi:10.1016/0304-3878(86)90144-6
security among households in Nsukka Metropolis of Enugu Gulliford, M. C., Nunes, C. & Rocke, B. (2006) The 18
State, Nigeria. Pakistan Journal of Social Sciences (PJSS), 30 Household Food Security Survey Items Provide Valid Food
(1), 9–16. Security Classifications for Adults and Children in the
Ayalew, M. (1997) What is food security and famine and hunger? Caribbean. BMC Public Health, 6, 26. doi:10.1186/1471-
Internet Journal of African Studies, 2, P.4. 2458-6-26
Babatunde, R. O. & Martinetti, E. C. (2010) Impact of remit- Hamilton, W. L., Cook, J., Thompson, W. W., Buron, L. F.,
tances on food security and nutrition in rural Nigeria. A Frongillo, E. A., Olson, C. M. & Wehler, C. A. (1997)
paper prepared and presented at Department of Public Household food security in the United States in 1995. Summary
Economics, Centre for International Cooperation and Devel- report of the food security measurement project. United
opment University of Pavia, March, 2010, Italy. States Department of Agriculture, Washington, DC.
Babatunde, R. O., Omotesho, O. A. & Sholotan, O. S. (2007) Hoddinott, J. (1999) Operationalizing Household Food Security in
Socio-economics characteristics and food security status of Development Projects: An Introduction. Technical Guide 1.
farming households in Kwara State, North-Central Nigeria. Washington, DC: International Food Policy Research Insti-
Pakistan Journal of Nutrition, 6(1), 49–58. doi:10.3923/ tute (IFPRI).
pjn.2007.49.58 Hoddinott, J. & Yohannes, Y. (2002) Dietary diversity as a food
Bamire, A. S. (2010) Effects of tenure and land use factors on security indicator. Food Consumption and Nutrition Division
food security among rural households in the dry Savannas of Discussion Paper No. 136. Washington, DC: International
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Nigeria. African Journal of Food, Agriculture, Nutrition and Food Policy Research Institute.
Development, 10(1), 1981–2000. doi:10.4314/ajfand. Hofferth, S. L. (2003) Persistence and change in the food security
v10i1.51470 of families with children, 1997–1999. Department of Family
Bickel, G., Nord, M., Price, C., Hamilton, W. & Cook, J. (2000) Studies, University of Maryland. Available at: http://www.
Guide to Measuring Household Food Security. Revised 2000 findarticles.com/p/articles/mi_m1309/is_3_40/ai_111027115
(Alexandria, VA: US Department of Agric, Food and Ibrahim, H., Uba-Eze, N. R., Oyewole, S. O. & Onuk, E. G.
Nutrition Service). (2009) Food security among urban households: A case study
Booroah, V. K. (2002) Logit and probit: Ordered and multinomial of Gwagwalada area council of the Federal Capital Territory
models. Quantitative Applications in the Social Sciences Series Abuja, Nigeria. Pakistan Journal of Nutrition, 8(6), 810–813.
No. 138 (Thousand Oaks, CA: Sage Publications). doi:10.3923/pjn.2009.810.813
Boudreau, T. (1999) The food economy approach: A framework for Idrisa, Y. L., Gwary, M. M. & Shehu, H. (2008) Analysis of food
understanding rural livelihoods. Relief and rehabilitation net- security status among farming households In Jere Local
work paper no. 26 (London: Overseas Development Government of Borno State, Nigeria. Journal of Tropical
Institute),. Agriculture, Food, Environment And Extension, 7(3), 99–205
Charlton, S. (2001) Low-income housing in South Africa: Practice & International Food Policy Research Institute (IFPRI) (1996)
policy reflections on the work of metro housing, Durban Metro- Feeding the World, Preventing Poverty and Protecting the Earth:
politan Council, 1997–2000. Unpublished report, Metro A 2020 VISION (Washington, DC: International Food
Housing Unit, eThekwini Municipality, Durban. Policy Research Institute).
Deitchler, M., Ballard, T., Swindale, A. & Coates, J. (2010) Jenson, R. T. & Miller, N. H. (2010) A revealed preference approach
Validation of a Measure of Household Hunger for Cross-Cultural to measuring hunger and under-nutrition. Working paper no.
Use (Washington, DC: Food and Nutrition Technical 16555, NBER Working Paper Series. Available at: http://
Assistance II Project (FANTA-2), AED, 2010). www.nber.org/papers/w16555.pdf
Devereux, S. (2006) Distinguishing between Chronic and Transitory Khan, R. E & Gill, A. R. (2009) Determinants of food security in
Food Insecurity in Emergency Needs Assessments (Rome: rural areas of Pakistan. A paper presented at National
SENAC, WFP). Conference on Socio-Economic Stability in Pakistan, Feb-
Etim, N. A. & Solomon, V. A. (2010) Determinants of rural ruary 16–17, 2009at Sarhad University of Science and
poverty among broiler farmers in Uyo, Nigeria: Implications Technology Peshawar, Pakistan.
for rural household food security. Journal of Agriculture and Kuwornu, J. K. M., Suleyman, D. M. & Amegashie, D. P. K.
Social Sciences, 6(1), 24–28. (2013) Analysis of food security status of farming households
Fakayode, S. B., Rahji, M. A. Y., Oni, O. A. & Adeyemi, M. O. in the forest belt of the Central Region of Ghana. Russian
(2009) An assessment of food security situations of farm Journal of Agricultural and Socio-Economic Sciences, 1(13),
households in Nigeria: A USDA Approach. The Social 26–42.
Sciences, 4 (1), 24–29. Mariara, J. K., Ndenge, G. K. & Kirii, D. M. (2006) Determinants
Federal Ministry of Agriculture and Water Resources (2008) of children’s nutritional status in Kenya: Evidence from demo-
National Food Security Programme, May 2008. graphic and health surveys. Presented at the Conference on
Food and Agriculture Organization of the United Nations (FAO) Reducing Poverty and Inequality: How can Africa be
(1996) Rome Declaration on World Food Security and included? (Oxford: Centre for the Study of African Econom-
World Food Summit Plan of Action. Accessed 3 January ies (CSAE)).
2012, available at: http://www.fao.org/DOCREP/003/ Maxwell, D. (1996) Measuring food insecurity: The frequency
W3613E/W3613E00.HTM and severity of “coping strategies.” Food Policy, 21(3), 291–
Gebrehiwot, T. (2009) Determinants of food security in rural house- 303. doi:10.1016/0306-9192(96)00005-X
holds of the Tigray. Unpublished MSc. in Applied Statistics, Maxwell, D., Ahiadeke, C., Levin, C., Armar-Klemesu, M.,
Thesis submitted to School of Graduate Studies of Addis Zakariah, S. & Lamptey, G. M. (1999) Alternative food
Ababa University, June 2009. security indicators: Revisiting the frequency and severity of
Greene, W. H. (2000) Models with discrete dependent variables. coping strategies. Food Policy, 24(4), 411–429. doi:10.1016/
Econometric Analysis, vol. 4 (NJ: Prentice-Hall), pp. 499–521. S0306-9192(99)00051-2
254 A. E. Obayelu
Megan, G. (2010) Three essays on buyer power theory, dominant Omotesho, O. A., Adewumi, M. O. & Fadimula, K. S. (2010)
HMOs and technology transfer. Unpublished PhD Dissertation Food security and poverty of the rural households in Kwara
submitted to Department of Economics, Northeastern Uni- State, Nigeria. Libyan Agriculture Research Center Journal
versity. Available at: http://hdl.handle.net/2047/d20000287 International, 1(1), 56–59.
Miller, C. M., Tsoka, M. & Reichert, K. (2011) The impact of the Orewa, S. I. & Iyangbe, C. O. (2009). The food insecurity profile
Social Cash Transfer Scheme on food security in Malawi. among the rural and low income urban dwellers in Nigeria.
Food Policy, 36(1), 230–238. doi:10.1016/j.foodpol.2010. American-Eurasian Journal of Scientific Research, 4(4),
11.020 302–307.
Mundlak, Y. (2000) Agriculture and Economic Growth: Theory and Paddy, F. (2003) Gender differentials in land ownership and their
Measurement (Cambridge, MA: Harvard University Press). impact on household food security: A case study of Masaka
National Population Commission (NPC) (2006) Population Cen- district. Master Thesis, Uganda. Available at: http://www.troz.
sus of Federal Republic of Nigeria: Analytical Report. uni-hohenheim.de/research/Thesis/MScAES/Paddy.pdf
Available at: http://www.population.gov.ng/index.php/state- Pankomera, P., Houssou, N. & Zeller, M. (2009) Household food
population security in Malawi: Measurements, determinants, and policy
National Population Commission (NPC) (2012) Nigeria’s popu- review. A paper presented at the Conference on International
lation stands at 167 million. Nigeria Punch Newspaper, August Research on Food Security, Natural Resource Management
5. Available at: http://www.punchng.com/news/nigerias- and Rural Development, October 6–8, 2009, University of
population-stands-at-167-million-npc/ Hamburg, Tropentag.
Nord, M. (2000) Does it cost less to live in rural areas? Evidence Rose, D. & Charlton, K. E. (2001) Prevalence of household food
from New Data on Food Security and Hunger. Rural Sociology, poverty in South Africa: Results from a large, nationally
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

65, 104–125. doi:10.1111/j.1549-0831.2000.tb00345.x representative survey. Public Health Nutrition, 5(3), 383–389.
Nord, M., Andrews, M. & Carlson, S.2008. Household Food Ruel, M. T. (2002) Is dietary diversity an indicator of food security or
Security in the United States, 2007. ERR-66, 2008 (Washing- dietary quality? A review of measurement issues and research
ton, DC: ERS/USDA). needs. FCND Discussion Paper No. 140. Washington, DC:
Obamiro, E. O., Dopper, W. & Kormawal, P. M. (2003) Pillars of IFPRI.
food security in Rural Areas of Nigeria. Internet Forum of food Sanusi, R. A., Badejo, C. A. & Yusuf, B. O. (2006) Measuring
Africa. March 31 to April 11. household food insecurity in selected Local Government
Obayelu, A. E. (2009) Households’ Food Consumption Patterns and Area of Lagos and Ibadan, Nigeria. Pakistan Journal of
Food Security Status in the North-Central Nigeria. Unpublished Nutrition, 5(1), 62–67. doi:10.3923/pjn.2006.62.67
PhD Thesis submitted to Department of Agricultural Eco- Shaikh, F. M. (2007) Determinants of Household Food Security and
nomics, University of Ibadan, Oyo State, Nigeria. Consumption Pattern in Rural Sindh: An Application of Non-
Obayelu, A. E. & Obayelu, O. A. (2012) Assessment of Food Separable Agricultural Household Model. Paper presented at
Insecurity Status and Government Interventions in Nigeria: the Mediterranean conference on agro-Food Social Scien-
The Challenges and Prospects in the Next Decades. In J. M tists, 103 EAAE Seminar ‘Adding Values to the Agro-Food
Jibrilet al. (eds.), Proceedings of the 46th Annual Conference Supply Chain in the future Euro-Mediterranean space, April
“KANO 2012” Agricultural Society of Nigeria (ASN)Confer- 23–25, Barcelona, Spain.
ence on “Agricultural Transformation in a Deregulated Eco- Shiferaw, F., Kilmer, R. L. & Gladwin, C. (2003) Determinants of
nomy: Prospects and Challenges” November 5–9, at the Bayero food security in Southern Ethiopia. A selected Paper Presented
University, Kano, Nigeria (Nigeria: Agricultural Society of at the 2003 American Agricultural Economics Association
Nigeria), pp. 48–53. Meetings, May 12, 2003, in Montreal, Canada.
Ogundari, K. & Ojo, S. O. (2007) Economic efficiency of small Sultana, A. & Kiani A. (2011) Determinants of food security at
scale food crop production in Nigeria. A Stochastic Frontier household level in Pakistan. African Journal of Business
Approach. Department of Agricultural Economics and Management, 5(34), 12972–12979.
Extension, Federal University of Technology Akure, Nigeria. Tshediso, J. S. (2013). Determinants of food security status of
Journal of Social Science, 14(2), 123–130. households receiving government grants in Kwakwatsi,
Ojogho, O. (2010) Determinants of food insecurity among Arable South Africa. Mediterranean Journal of Social Sciences, 4(1),
Framers in Edo State, Nigeria. Agricultural Journal, 5(3), 147–153.
151–156. doi:10.3923/aj.2010.151.156 Umeh, J. C. & Asogwa, B. C. (2012) Determinants of farm
Olorunfemi, S.Aderinola, E. A. & Ajibefun, A. I. (2009) Deter- household food expenditure: Implications for food security
minants of household food security in Southwestern Nigeria. in rural Nigeria. A paper presented at the International
International Journal of Business Research, 9(3). Available at: Conference on Ecology, Agriculture and Chemical Engin-
http://www.questia.com/library/1G1-208535099/determinants- eering, December 18–19, 2012, Phuket (Thailand), pp.
of-household-food-security-in-southwestern 212–217.
Oluyole, K. A., Oni, O. A., Omonoma, B. T., &Adenegan, K. O. United States Agency International Development (USAID)
(2009) Food security among cocoa farming households of (2008) United States Agency International Development.
Ondo State, Nigeria. Journal of Agricultural and Biological Madagascar Food Security Development programming
Sciences, 4, 7–13. Framework, From the American People, Madagascar.
Oni, O. A., Salam, K. K. & Idowu, B. O. (2011) Social capital United State Department of Agriculture (USDA) (2000) A Guide
dimensions and food security among farming households in to Measurement of Food Security, United State Department
Ogun State, Nigeria. Journal of American Science, 7(8), of Agriculture.
776–783. Vasco, M. (2006) Food Security in Vietnam during the 1990s: The
Omonona, B. T. & Agoi, A. G. (2007) An analysis of food security Empirical Evidence. WIDER Research Paper No. 2006/67.
situation among Nigerian urban households: Evidence from Helsinki, Finland: United Nations University (UNU)-
Lagos State, Nigeria. Journal of Central European Agriculture, World Institute for Development Economics Research
8, 397–406. (WIDER).
Households’ food security status and its determinants in the North-Central Nigeria 255
WFP/FAO (2008) Measures of Food Consumption - Harmonizing Wolfe, W. S., Frongillo, E. A. & Valois, P. (2003) Understanding the
Methodologies. Interagency Workshop Report WFP – FAO, experience of food insecurity by elders suggests ways to improve
April 9–10, 2008, Rome. its measurement. Journal of Nutrition, 133(9), 2762–2769.
Wolfe, W. S. & Frongillo, E. A. (2000) Building household food Zoellick, R. (2008) Food crisis: How prepared is Nigeria? THIS-
security measurement tools from the ground up. Background DAY Nigeria Newspaper, April 27, 2008. p. 25.
Paper, Food and Nutrition Technical Assistance Project,
Washington, DC. USA.
Downloaded by [University of Lethbridge] at 02:53 14 May 2016
256 A. E. Obayelu

Annexure 1. Eighteen (18) households’ food security items. Annexure 2. Logistic conditional analysis odds ratio estimates.

Number Questions 95%


Point confident
Q1 Adult cut size or skipped meals because there was not Parameter estimate level Infit
enough money for food*
Q2 Adult cut size or skipped meals three or more times in Worries whether food will run 0 0 0.8303
the last 30 days* out before buying
Q3 Adult does not eat whole day because there was no Food bought did not last and no 0.633 0.416 0.8307
enough food* money to buy new one
Q4 Adult does not eat whole day three or more times in Can not eat balanced meal 0.593 0.390 0.1658
the last 30 days* Cut/skip meals due to lack of 0.263 0.173 0.9618
Q5 Adult eat less than what they felt they should* money to buy food
Q6 Adult hungry but did not eat* Eat less than what we should 0.739 0.484 1.2696
Q7 Adult lost weight because there was not enough food* Hungry but did not eat due to 0.290 0.191 0.9756
Q8 Cut size of child’s meals** lack of food
Q9 Child skipped meal because there was not enough Lose weight due to lack of 0.189 0.124 0.9946
money for food** enough food to eat
Q10 Child skipped meal, three or more times in the last Eight not eaten for a whole day 0.093 0.060 1.0526
30 days** because of not having enough
Downloaded by [University of Lethbridge] at 02:53 14 May 2016

Q11 Child being hungry but did not eat because we could money for food
not afford more food**
Q12 Child not eating for a whole day because there was Generated from author data set.
not food**
Q13 Worried whether food would run out before getting
money to buy more*
Q14 Food bought did not last and no money to get more*
Q15 Adult could not afford to eat balanced meals*
Q16 Could not feed child with balanced meals because we
could not afford that**
Q17 Child not eating enough because we could not afford
enough food**
Q18 Adult feed child with low-cost foods because of
inability to buy food**

*Are the 10 adult-referenced food security items and **are the


child-referenced items. (Nord, 2000)

You might also like