Economic analysis of household food demand in Southwestern Nigeria.
The study examined the household food demand in south-west Nigeria. Data collected from 300 heads of household through a multistage random system techniques were analysed using a Working Leser demand function. In order to deal with zero consumption problem associated with those who did not demand the food items, various single equation models were applied. These include demand equation estimated by Heckman two-step sample selection model (Heckit.). Result showed that Ekiti State had the highest household daily per capita calorie intake and protein intake of 2,676.73 kcal and 70.44g respectively, followed by Ondo State(2,312kcal and 63.53g) and Osun State (2125.15kcal and 57.14g). Result of the Working Leser model showed that all the food commodities were normal goods while only gari and rice were elastic with regard to own-price.

Keywords: Economics Analysis, Food Demand, Nigeria

Article Type:
Food supply (Economic aspects)
Food (Supply and demand)
Econometric models (Analysis)
Olorunfemi, Sola
Pub Date:
Name: European Journal of Management Publisher: International Academy of Business and Economics Audience: Academic Format: Magazine/Journal Subject: Business, international Copyright: COPYRIGHT 2011 International Academy of Business and Economics ISSN: 1555-4015
Date: Summer, 2011 Source Volume: 11 Source Issue: 2
Event Code: 600 Market information - general
Product Code: 0101100 Food; 2000000 Food & Kindred Products; 8821000 Consumer Expend-Food & Tobacco NAICS Code: 111 Crop Production; 311 Food Manufacturing; 81411 Private Households
Geographic Scope: Nigeria Geographic Code: 6NIGR Nigeria
Accession Number:
Full Text:

The food supply of a nation is through domestic food production and import. In turn, ability of a nation to import food, depends on her export earnings; foreign exchange reserves; value of essential non-food import and debt service obligations (Olayemi, 1998). There is food insecurity and malnutrition among Nigerians. While the World Health Organization (WHO) recommendation is a minimum of 20 grams of animal protein per capita per day, only a total of 7.2, 7.6 and 4.9 grams were taken in Northern, Western and Eastern Nigeria respectively. This represents national average of 6.56 grams. Malnutrition for children aged 3-35 is 44 percent in Nigeria.

The move by government to import food massively to save the above situation only presents us with situation at the national level, which does not mean food security for various households in the country. Adequate food supply at the national level does not automatically lead to food security for all households. There may still be poor households who do not have the means or the purchasing power to procure the food they need.

Food prices continue to soar up day by day going out of the reach of the common man while the income of the households in the country are being debased by the staggering inflation rate. The inflation rate which was just 3.2% in 1972 rose to 39.6% in 1984, 40% in 1989, 72.8% in 1995 10% in 1998 and only 7.0% in 2000 (CBN 2000),. This underscores the fact that households' income can hardly cope with soaring food prices, which has necessitated large food spending out of households' income of between 60 percent and 80 percent coupled with poor income per head in Nigeria.

A look at the composite consumer price indices for food in Nigeria between 1970-2002 shows that it was 9.0% in 1970 but this progressively rose to 4567.07% in 2002 and 8,205.9% in 2006 (CBN, 2000, 2002, 2006). It was reported that the average income per month in Nigeria is N300 (Business Times 2000). Despite this humiliating income figure, notwithstanding our riches, households' income is not guaranteed because of the seemingly insuperable unemployment problem in the country. This picture portends a very gloomy household food security in the country. This picture portends a very gloomy household food security situation in Nigeria with a very bad implication for nutritional and health status of an average household. The economic development of Nigeria is directly linked with the levels of productivity of the average Nigerian. But the productivity of labour is dependent on the nutritional intake (and health status) of the labour itself (ILO, 1981). And in Nigeria, while the sectoral contribution of agriculture to total export was about 72.0% in 1970, it declined to 2.3% in 1985 and has largely continued to stagnate or decline thereafter (Okunmadewa, et al 1999). The contribution of total agricultural output to GDP was 42.7% in 1986; it dropped to 40.4% in 2001. This made foreign exchange to drop.

The costs of agricultural inputs continued to increase rapidly, also as a result of the low exchange rate of the national currency, the Naira. For instance fertilizer which was sold for an average of N141 per bag in 1985 and N200 in 1996 rose to N1,400 in 1997 (CBN, 1997, Okuneye, 2005). It is expected therefore that the cost of production of agricultural products would rise with an attendant increase in the cost of food. In the past three decades, the inflation rate has skyrocketed at a pace which has confounded policy makers. Both agricultural and manufactured commodities have been affected, though its effect on agricultural and food prices have been excruciating considering the fact that basic food need is topmost in the hierarchy of needs. This has affected the poverty level of different households. It is important to note however that the development of agriculture is highly necessary to ensure that more food is produced and made available to non-producers at reasonable prices so as to reduce poverty.

Although accurate information about the total food supply in the country is hard to come by, available evidences indicate a somewhat dismal picture in the production of the various major food crops.

Many things are unclear about the characteristics and causation of food demand in Nigeria. A great deal of probing investigation--analytical as well as empirical--is needed as background to understand food demand structure for appropriate public policy and action for eradicating famines and eliminating endemic under- nutrition. More evidence on this issue is necessary, particularly at the household level, as the general surveys may not be appropriate for bringing about possible solutions. Additional evidence is needed based on specific explanations. The present study attempts to fill this gap by providing further evidence on the understanding of food demand structure in South-western Nigeria.

As a result of this, it is pertinent, therefore, to analyze food demand in South-western Nigeria. Then the questions that arise are:

What is the food consumption pattern in south-western Nigeria?

What are the factors affecting household food demand in the region and

How can factors affecting household food demand be improved upon?

These are some of the questions this study attempts to answer.

The goal of this study is to analyze household food demand in South-western, Nigeria. This goal will be achieved through the pursuance of the following specific objectives:

(i) To examine the socio-economic characteristics of households food demand in South-western Nigeria;

(ii) To identify and measure the factors affecting the food demand among households in the study area;

(iii) To analyze food consumption patterns.


Food is a basic human need and the major source of nutrients needed for human existence. Food security indicates the availability of and access to food. The place of agriculture in an agrarian society cannot be overemphasized given its importance in the life of human's beings. Agriculture is expected to ensure adequate supply of food to the people. Globally, there is enough food for all, but more than 780 million people are chronically undernourished. Millions of people in the developing world simply cannot obtain the food they need for a healthy and productive life ( )

In the previous studies, there are several literatures on food demand in Nigeria, almost all of these contributed to the demand for individual food items. For instance Obi (2003) observed that production of animal protein has not been high enough to meet the demand of rapid population growth. Ganiyu (1982) observed that as consumer become more articulate and organised their demand for animal protein will have positive influence on production method. Several literatures are also available on the demand for rice as a food item. These include attempt by Odusina (2008) that looked at the urban demand for different rice types with a view to understanding the consumption pattern for local rice commodity and thereafter discovered the reasons for which local rice is not preferred to imported rice. To carry out these studies by various authors, they made use of the Almost Ideal Demand System (AIDS) and Linear Approximation of the strict AIDS model (LA-AIDS).

AIDS model is non linear model and is difficult to estimate. Some studies used the linear approximate of AIDS and these studies include works such as Akinleye (2007) Savadogo and Brandt (1988), Fulponi (1989) Mergos and Donatos (1989) and Soe et al (1994). The model hypothesis that the portion of total expenditure that accrued to a particular commodity is related to prices and income.

However, in this study we made use of Working Leser demand function as used by Chen et al (2003). In using this, an incidental truncation characteristic of the data was captured by using the Heckman's two step estimator.

Given that food consumption and expenditure decisions have long-term diet and health consequences and a direct effect on development; and the importance of per capita food intake on human welfare and productivity through its influence on the capability of man to perform work and the attitude of man towards the work. This study intends to derive the nutritional implications of the demand for different types of food items and the effect or otherwise of unit changes in prices and income on the households consumption in south-western Nigeria.


3.1 The Study Area

The data for this study were obtained from South-western Nigeria. South-western Nigeria is divided into six different states, namely: Lagos, Ogun, Oyo, Osun, Ondo and Ekiti states and fifty percent of these states was selected. The selected states were Ondo, Osun and Ekiti states. The study area is highly agrarian and highly endowed with mineral resources. The population is 5.94 million, which is 6.68 of the total population in Nigeria (CBN, 2003). According to information from the International Labour Office (ILO), the study area falls within three distinct vegetation zones--mangrove forest to the south, rain forest in the middle and guinea savannah to the north. The types of arable crops grown in the area include yam, cassava, maize, melon, cowpea, vegetables plantain and bananas.

Data were collected from 300 heads of households through a multi-stage random sampling technique and interviewed at intervals of two weeks for three months. The multi-stage random sampling employed involved 4 stages. At the first stage, Osun, Ondo and Ekiti states were randomly selected out of the 6 states in South-western Nigeria. At the second stage, the number of households from each state was selected using proportionality, such that the number of respondent households from each state is proportional to the number of local government areas in each state. That is

X = n/N x 100


n = the number of local government in each state

N = the number of local government in the three states

At the third stage, 5 local governments and 3 towns were randomly selected. At the fourth stage proportionality factor of 3 was introduced to derive the number of respondent households from each town. That is Y = K/3.


K = the number of households to be sample from a state

X = the number of respondents from each state

Y = the number of respondents from each town

Both primary and secondary data were used for this study. Questionnaire was administered to the respondents as enumerated above.

Data were collected on some household characteristics such as income, expenditure, quantities of food commodities consumed and the nature of household food security problems as envisaged by the household heads.

The secondary data were obtained from various issues of the Central Bank of Nigeria (CBN) publications.

3.2 Model Specification

The theory of consumer behaviour is the fundamental theory on which this research is base. The principal assumption on which the theory of consumer behaviour and demand is built is that, a consumer attempts to allocate his limited money income among available goods and services so as to maximize his utility. The usefulness of the theory lies in the fact that it can help us to understand how consumer demand responds to income. According to the neoclassical economic theory of consumer behaviour, each individual consumer is confronted with market determined prices of various commodities, with the consumer having only a known and fixed money income. It is these prices, according to Ferguson (1975) that helps the consumer to allocate his or her income to the various goods and services. Thus according to the theory, the amount of a commodity that a consumer would purchase therefore depends on the prices of the commodity and the money income of the consumer. Literature also asserts that commodities with negative income elasticities are said to be inferior, those with income elasticities between zero and one are said to be normal and a commodity with an income elasticity greater than one is said to be superior. Consumption pattern varies from one area to the other. This usually is the result of socio-cultural and economic differences while level of education is an important factor in consumption. However, income and price seem to be the dominant factors to reckon with. Income limits the extent to which one can expend on food, whether price of such commodity is high or low. Apart from prices and income being major factors in determining consumption pattern, household size could also be another important factor. Application of the theory of the household requires a specific model. In general, econometric studies of demand include both single equations and systems of demand equations. The demand functions can be generalized for a consumer or household buying n goods as:

[q.sub.i] = [q.sub.i] ([p.sub.1], [p.sub.2], ... [p.sub.j], ... [p.sub.n], 1), i = 1, 2, ... n ... (1)

Where [q.sub.i] = the quantity demanded; p = the price, the subscript i denotes the commodities; and I = income. These "n equations" can be estimated by single equations or by systems of equations. In this study, Equation 2 is estimated in a budget share form. Extending the demand function for individual consumers to that for a group of consumers in most empirical applications requires the inclusion of demographic variables besides prices and income.

Working lesser demand function had been used in literature to analyze food consumption. The original form of the Worker-Leser model was discussed by Working (1943) and Leser (1963). Intriligator, Bodkin and Hsiao (1996) and Deaton and Muellbauer (1980) provide a more detailed discussion of this functional form. In the Working-Leser model, each share of the food items is simply a linear function of the log of prices and of the total expenditure on all the food items under consideration. According to Chern et al, (2003) he specified


Where (i,j) = food items

[W.sub.i] = The expenditure share of food i among the food items;

[P.sub.j] = The price of food j;

x = The total expenditure of all food items included in the model.

[H.sub.k] = The demographic variables but not used in this study.

[[epsilon].sub.i]'s are random disturbances assumed with zero mean and constant variance. The parameters to be estimated included [alpha], [beta] and [gamma].

This model will be estimated for each food items by the ordinary least squares (OLS).

(a) Demand Elasticity and Income Elasticity for the Working-Leser Model

The expenditure elasticity (ei) can be expressed as:

[e.sub.i] = [1 + [a.sub.i]/[w.sub.i]] ... (3)

Taking a derivative of Equation (2) with respect to log ([p.sub.j]) yields, uncompensated own (j = i) and cross (j [not equal to] i) price elasticities ([e.sub.ij]) are as follows:


Where [[delta].sub.ij] is the Kronecker delta, which is unity if i = j; and zero, otherwise.

Income elasticity was derived from equation (2) by taking the logarithm of equation (2) while the expenditure elasticity, ei was estimated as presented by equation (5):

[e.sub.i] = [partial derivative][q.sub.i] / [partial derivative] x X / [q.sub.i] (5)

The Incidental Truncation Characteristics of the Data: When estimating elasticities, the use of household level micro data is a good way of avoiding the aggregation problem[Abowd et al, 2000]. However the use of household micro data is complicated by the econometric problem that arises when some heads of household made no demand for food items. Demand is one observed for those who consume food (demand are unobserved for those who don't consume). So information would not be taken from the entire population. Rather, the population would be limited and biased by including only individuals who consume. So the data would be non randomly selected or incidentally truncated. In this study, different estimators have been applied to estimate the parameters of the sample selection model. These are Heckman's two-step estimator (Heckit) and the Maximum Likelihood Estimator (MLE). In order to correct for the sample bias problem for household who does not demand food item, Heckman's twostep estimation (Heckit) procedure can be applied, as suggested by Chern et al (2003) and Heckman (1976). In the first stage, a probit regression is computed in order to estimate the probability that a given household actually make demand. This regression is used to estimate the inverse Mills ratio ( A) for each household, which is used as an instrument in the second regression. Recall that since we are not using demographic variables ([H.sub.k]) in equation 2 if [[alpha].sub.0] + [[alpha].sub.1] log x + [summation] [[beta].sub.ij] log [p.sub.j] = [[omega].sub.i],

then the first and second regression equations are given as equations 6 and 7:

[L.sub.i] = <' [[omega].sub.i] + [[psi].sub.i] .... 6

Where i = index for each survey household L = Boolean variable indicating membership into a plan < = Vector of variable coefficients to be estimated to = Vector of independent variables in equation 2, [psi] = Error term ~ N (0,1)

[K.sub.i] = [Real part]'[t.sub.i] + [[xi].sub.i] .... 7

Where k = satisfaction levels as measured by survey questions; [Real part] = Vector of variable coefficients to be estimated; t = Vector of independent variables used in the Probit model i.e. equation 6 plus the Inverse Mills Ratio; [xi] = error term ~ N (0, [sigma]2). The sample rule is that [k.sub.i] is observed only when [L.sub.i] is greater than zero


Where [[lambda].sub.[psi]] = - [[omega]'.sub.i] / [sigma][psi] and

[lambda] ([[lambda].sub.[psi]] = [phi] ([[omega]'.sub.i] / [sigma][psi]) .... 9

Equation 9 is the Inverse Mills Ratio for every household. For notational convenience this is put as

[psi] (< [??] + < [??] [[omega].sub.i] / [PSI] (< [??] + < [??] [[omega].sub.i] .... 10

where [phi], is the density probability function; and [PHI] is the cumulative probability function. [Real part] and [Real part.sub.[lambda]] can be estimated by the following equation

[K.sub.i] | [L.sup.*.sub.i] > 0 = E [[K.sub.i] | [L.sup.*.sub.i] > 0] + [[mu].sub.i]: = [t'.sub.i][Real part] + [[Real part].sub.[lambda]] [lambda] ([[lambda].sub.[psi]]) + [[mu].sub.i] .... 11

Where [[mu].sub.i] is heteroscedastic: var

[[mu]| [L.sub.i] = 1 [t.sub.i], [[omega].sup.2.sub.[xi]] ( 1 - [[Imaginary part].sup.2] [[sigma].sub.i]) .... 12

Least squares regression using incidentally truncated data produces inconsistent estimates of [Real part]. However, the least squares regression of k on t and [lambda] produces consistent estimators. Omitting [lambda] would

produce the specification error of an omitted variable. Unless [[Real part].sub. [lambda]] = p [[sigma].sub.[xi]] = 0. The hypothesis therefore is to test H0: P = 0 using t statistic on [[lambda].sub.i]. For maximum likelihood, recall from equation 6 and 7 that, for the sample selection model, there are two types of observation: Those where [k.sub.i] is observed and we know that [L.sub.i] > 0. For these observations, the likelihood function is the probability of the joint event [k.sub.i] and [L.sub.i] > 0. We can write this probability for the ith observation as the following (using Bayes Rule):


Thus the probability of an observation for which we see the data is the density function at the point [k.sub.i] multiplied by the conditional probability distribution for [L.sub.i] given the value of [k.sub.i] that was observed. Those where [k.sub.i] is not observed and we know that [L.sub.i] [less than or equal to] 0. For these observations, the likelihood function is just the marginal probability that [L.sub.i] [less than or equal to] 0. We have no independent information on [k.sub.i]. This probability is written as

Pr ([L.sub.i] [less than or equal to] 0) = Pr ([[psi].sub.i] [less than or equal to] - [[omega].sub.i] <) [PHI] (- [[omega].sub.i] <) = 1 - [PHI] ([[omega].sub.i] <) .... 14

Therefore the log likelihood for the complete sample of observations is the following:


Where there are No observations where we don't see ki and N1 observations where we do ([N.sub.0] + [N.sub.1] = N). The parameter estimates for the sample selection model can be obtained by maximizing this likelihood function with respect to its arguments. These estimates will be consistent and asymptotically efficient, under the assumption of normality and homoskedasticity of the uncensored disturbances.


Table 1 shows the structure of household income obtained for the study. The sample drawn from this study shows some degrees of diversity in household income distribution among the three states in southwestern Nigeria. Among the lowest income group, Ondo State had the highest percentage of 5% while Ekiti had the lowest 3.80%. For all the 3 states it was 3.07%. Among the highest income group, Ekiti State had the highest 72.15% followed by Osun State with 69.61% and Ondo State with 67.50%. The income difference among these three states appeared to be much. Table 2 presents a summary distribution of the shares of total food expenditure allocated to various food items in the three different states of the south--western Nigeria in 2004. The table shows that proportional income expenditure on rice is higher in all the three states, apart from this, the proportion of food expenditure for every other food differs. The second largest food expenditure for Ekiti State was on meat, which was 15.33 percent of the total expenditure followed by beans which was 14.54 percent of the total expenditure. The second largest expenditure made for food in both Osun and Ondo State was on yam and meat on which each of these states committed 18.09 and 19.41 percent respectively.

The level of composition of food and food nutrient intake derived from the quantities of various food commodities acquired and actually consumed by its members represents a measure of the household's food demand capacity and its degree of economic access to food.

We bring the analysis of food demand carried out in this study into focus by anlaysing the average food nutrient contents of household food consumption. It should be pointed out that these food nutrient contents were estimated from the data collected on the quantities of various food commodities, consumed by households and recorded during the household survey using the calorie and protein contents of foods provided by PCU (2002).

Table 3 presents the picture of the average energy intake per capita from different food items in various states considered. It is observed from the table that in Osun state, food intake sources are highly diversified, with beans constituting the single largest source of calorie intake accounting for only about 23.80 per cent, followed by rice with about 20.14 per cent. Meat and fish account for about 13.83 per cent while fruit and vegetable account for 8.80 per cent. In all, carbohydrate food--Yam, Rice and Gari provide about 44.48 percent of the average daily energy intake in the state.

On the average, each member of the household in Osun State had a daily energy intake of only 2125.15 kilocalories at the time of the study. This falls short of the FAO's recommended threshold for adequate energy. In Ondo State, carbohydrate foods accounted for 42.7 percent, followed by animal protein which accounted for 19.21 and fruit and vegetable which accounted for 13.18 percent. An average person in Ondo State had a daily energy intake of about 2312 kilocalories which, though slightly higher than that of Osun State was only a little bit higher than the FAO's recommendation. In Ekiti State, the food consumption revolves largely around Rice, Beans, Plantain, yam, fruit and vegetable. Each of these accounted for 16.49%, 17.45%, 15.16%, 14.73% and 12.07% respectively. A cursory look at Table 3 would confirm once again that carbohydrate food dominate the diets in the three states.

Carbohydrate, plant protein and animal protein dominate the diet in the Osun State as the three different food categories accounted for 44.48%, 23.80% and 20.06% respectively. Fruit and vegetables trailed behind as they accounted for 8.80%. The situation was different in Ondo State with the exception of carbohydrate that still lead the diet with 42.7%. However animal protein came second with 19.21% and plantain came third with 16.14%. The consumption of plant protein was just 10.75%. In Ekiti State carbohydrate alone account for about 41.67% of the average energy intake followed by animal protein and plant protein which were 22.48% and 17.45% respectively.

Going by the relative shares of calories derived from various food groups, it is evident that Ekiti state had the most diversified sources of energy intake at the time of survey, and with Ondo and Osun States having less diversified sources.

The average energy intake in all the three states was about 2371.32 kilocalories per capita per day, which was slightly above food energy adequacy.

Table 4 shows the average protein intake per capita derived from the consumption of different foods in the three states, and the relative percentage distribution of protein from different food items. It is observed from the table that beans, meat and rice constituted the three most important sources of protein in the three states, though not in the same order of importance in all the three states. In Osun state, these three sources contributed about 59.93 percent of the average protein intake in the state. Also in Ondo and Ekiti states, these three food items contributed 62.64% and 62.85% respectively. In all the three states as a whole, beans were the most important source of protein, followed by meat and then followed by rice. These three sources provided about 61.8 percent of the average protein intake. The average protein intake per capita per day was lowest in the Ondo state and highest in the Ekiti state. But while the average protein intake in Osun state fell short of an adequate intake threshold of 65 grams per capita per day, that of the Ekiti state was significantly

Higher than the threshold. However, the average intake for all studied states was only marginally higher than the threshold, as was the case with calories intake implying an unstable food security situation. In spite of the above discovery on food situation from each of the states, when the respondents were asked of the extent to which they are food-secured 57.8 percent in Osun State believed they were fully food-secured while it was just 27.5 and 35.4 percent for both Ondo and Ekiti states. This might be so because the respondents did not know the adequacy of both calories and protein intake per capita per day for them. To investigate the differences in the demand structure among income groups, the households were divided according to monthly income level. The mean budget share and elasticities were calculated to X-ray the food consumption pattern. The estimates of expenditure and income elasticities from the whole sample Working Leser model (OLS) are show in Table 5. The food items surveyed are rice, gari, beans, yam, milk, bournvita, meat, fruit and vegetable and plantain. The results indicate gari and rice to be normal goods in this estimation. The expenditure elasticity of both is above one. Other commodities are also relatively expenditure inelastic with yam having the highest expenditure elasticity followed by meat and plantain respectively. It is noteworthy that the own-price elasticities for gari and rice are very elastic. This indicates those South-western Nigerian consumers are sensitive to price changes in gari and rice. If this estimate represents south western consumer behaviour correctly, gari export--which should lead to an increase in price--might be much felt by consumers but it will also boost the effort of the farmers to produce more. From this estimation, gari is found to be a normal good. If this is so, gari consumption would increase as per capita GDP grows. If this is the case, it would be possible to project higher gari demand in the future as the income of

Nigerians increase. From this estimation, it could further be seem that the expenditure elasticities for all other foods are less than 1 and they are staple foods and that the consumption of each of these will decline as per capita income increases. This result is further complimented when the mean budget share is considered. The highest percentage of budget share for food in Osun, Ondo and Ekiti states went for rice at 30.90%, 31% and 26.57% respectively. The lowest budget share for food in Osun went for milk at 3.04%. It was 4.42 for fruit and vegetables in Ekiti state and 3% for plantain in Ondo state. These are shown in figures 8 to 10. However for all the states, the highest percentage of budget for food went for rice (29.17%) followed by meat (16.93%) yam (14.31%) and beans (11.66%) gari had 8.26% while plantain had 4.45%.

Table 3 shows expenditure elasticities by income bracket. Most estimates are relatively invariant with income level. Meat, gari, rice and plantain show that lower-income consumer demand tends to be more expenditure elastic, while higher-income consumers are less elastic. This means that the lower income classes are more sensitive to what they spent on these 3 foods, since the expenditure elasticities are respectively 0.567, 1.2165 and 1.481 respectively.

Table 7 shows own-price elasticities by income bracket. There are no significant variations of elasticity estimates by income level. The own-price elasticity for gari remains very high in these income submodels especially to a low income and middle-income earners. The estimates were -1.312 and -1.208 respectively but it was--1.12 for the high-income group. The coefficient of own-price elasticity for other staple foods to low income earners is higher than for every other income groups. The implication of these results is that, low income group are more sensitive to price change. Then we reject the Null hypothesis that the low- income group is not sensitive to price change.

Table 8 compares the own-price elasticity estimates from Working lesser (OLS) and Heckit sample selection model. It is surprising that the own-price elasticity for all staple foods except rice and gari are below 1 in absolute terms in both the Working lesser and Heckit model. The lowest estimates of ownprice elasticity for all the staple food are found in the Heckman's two-step where the inverse Mill ratio was used. The own price elasticity of gari and rice were -1.121 and -1.2017 under Working lesser model but -1.0059 and -1.012 under Heckit model. Gari and rice have higher own-price elasticity than every other staple in the two models. There are some possible reasons behind this finding. First reason is the importance of gari and rice in the Nigerian diet. Nigerians' eating culture has induced a wide range of variations in gari and rice prices and consumption level. Secondly, there are many varieties and uses of both. These varieties and uses offer many substitution opportunities, which tend to result in higher ownprice elasticity.


The analysis of household food demand has brought out the following issues:

The total percentage of household income spent on carbohydrate food is highest for all states, it is evident from the study finding that consumption of a variety of foods by the household may not have increased the probability of meeting the minimum recommended daily dietary requirement. Food staple, which are richer in one nutrient but deficient in others, are the main dietary sources. The consumption of carbohydrate food (Rice, Gari and Yam) is very high in all the three states. It was 44.48%, 42.70% and 41.67% in Osun, Ondo and Ekiti States respectively. The consumption of protein--rich food, is very low among the households in all the three states. This might be due to lack of nutrition--education to change consumption behaviour, low income is considered to be a major source of food insecurity. In this regard majority of those who gave reasons for not having three full meals daily which are 25.5% in all, out of this 12.17% identified unavailability of money to buy food, followed by inadequate quantity of food to eat (5.77) and lack of appetite (1.6%), average calories and protein intake was not adequate in Osun State because the average calories was just 2125.15 kcal/day and 57.14 protein gm/day. However for Ondo and Ekiti the average calorie was 2312 and 2676.73 while the average protein was 63.53 and 70.44 respectively. For the three states together, the average calorie and protein intake per capita is only at the threshold of adequacy. The expenditure elasticity results indicate that gari and rice are normal goods in south-western Nigeria; One set of results is related to the estimated low own-price elasticity for most of the staple foods. This means that the households are insensitive to price change. Moreover, the results show that lower-income households tend to be more expenditure--elastic, for gari, rice and plantain while higher-income consumers are less elastic. Since this elasticity has important implications for the impacts of south- western Nigeria agricultural and trade policies, it needs to be assessed carefully. The reliability of the estimates can be seen from the fact that the range is relatively robust among several model specifications.

In conclusion, from the study the available demand structure shows that food insecurity is still a serious problem among the households in south-western Nigeria especially in Osun State. An improvement in food security situation of household is advocated to put the nation on the right track of development.


Abowd, J., F Kramorz, and N. Margolia( 2000); High-wage workers and high-wage firms, Econometrical 67, 251-333

Business Times, Lagos, December 4 2000.

CBN (1997) statistical Bulletin Vol. 8 No. 1 Central Bank of Nigeria Abuja.

CBN (2000); Statistical Bulletin, vol. 11 No. 2 pp. 111-155.

CBN (2003); Statistical Bulletin, vol. 14.

CBN (2006); Statistical Bulletin Vol. 7 No. 2 Central Bank of Nigeria Abuja

Chern, W. S; K. Ishibashi, K. Taniguchi, and Y. Tokoyama, (2003); Analysis of the Food Consumption of Japanese Household FAO Economic and Social Development Paper 152-1-21.

Deatons, A. and J. Muellbauer (1980); An almost ideal demand system. America Economiv Review, 70.

Ferguson, C.E., and J.P. Gould (1975); Micreeconomics theory. Richard D. Irwin Inc. Georgeton Ontario, p.29-43.

Fulponi, L.,( 1989); The Almost Ideal Demand System An Application to Food and Meat Groups for France J. Agric. Econ., 40: 89-92.

Heckman, J.J. (1997); The common structure of statistical model of truncation sample selection and limited dependent variables and a simple estimator for such models. Annual of economic social measurement, 5:475-492.

ILO (1981); First Things First: Meeting the Basic Needs of the People of Nigeria United Printer LTD, ADDIS ABABA, ETHIOPIA PP. 1-58.

Itriligator et al ( 1996); Application to Household; Demand Analysis. In Econometric Models, Technique and Applications.

Lesser, C.E.(1963); Forms of Engel Functions. Econometrica, 31:694-703.

Mergos, G.J. and G.S. Donatos, (1989); Demand for Food in Greece: An Almost Ideal Demand System Analysis. J. Agric. Econo., 40: 178-184

Okunmadewa, F.A., Olomola and B. Adubi (1999); Trade Liberalization and Food Security: Situation Analysis, Nigeria Journal of Development Policy, Development Policy Centre Research Report No. 17 p. 2.

Okuneye, P.A. (2005); "Rising Cost of Prices and Food Insecurity in Nigeria and its implication for poverty reduction, Website: Vol 39-4-6, PDF

Obi, C.I.,( 2003); Game production. An alternative to beef cattle production in southern Nigeria. Nigeria Acad. Forum., 4: 36-40

Olayemi, J.K. (1998); "Food Security in Nigeria", Journal of Development Policy, Development Policy Centre Research Report No. 2 p. 1- 86.

PCU (2002); Fadama II: Development of Performance Monitoring and Evaluation Indicators, prepared by the Department of Agricultural Economics, University of Ibadan, Ibadan for Project coordinating unit of Federal ministry of Agriculture and rural development, Abuja.

Savadogo, K. and J.A Brandt, (1988); Household Food Demand in Burkina faso: Implications for Food Policy. Agric. Econo., 2:345-364

Soe, T., R.L. Betterham and R.G. Drynan,(1994); Demand for Food in Mynanmar (Burma). Agric. Econo., 11: 2-3, 207-217

Sonaiya, E.B., (1982); Beef quality and the Nigerian consumer. In: Beef production Nigeria Osinowo etal., (Ed), Proc. Nig. Nat. Conf. Beef Prod, pp. 555-564

Working, H. (1943); Statistical Laws of Family Expenditure Jurnal of American Statistical Association. 33:45-56

Sola Olorunfemi, Adekunle Ajasin University, Akungba Akoko, Ondo State Nigeria


Dr Olorunfemi Sola earned his Ph.D at The Federal University of Technology, Akure Ondo State, Nigeria. His area of research is development economics with emphasis on poverty and food security. He is an expert in computational general equilibrium modelling and a consultant in statistical information system. Presently, he lectures at the Department of Economics Adekunle Ajasin University, Ondo State, Nigeria.
Table 1: Household Income Distribution (grade level per month)

Income group     Osun state   Ondo state   Ekiti state   All States

< N7,500         3.92         5.00         3.80          3.07
N7,501-N15,000   26.47        27.50        24.05         23.37
> 15,000         69.61        67.50        72.15         73.56
Total            100          100          100           100

Source: Field Survey

Table 2: Percentage Distribution of Household Income
Expenditure by Food Items

Foods Items            Osun    Ondo    Ekiti   All

Rice                   30.90   30.90   26.57   29.17
Gari                   10.09   7.74    6.95    8.26
Beans                  8.29    12.15   14.54   11.66
Yam                    18.09   10.95   13.89   14.31
Milk                   3.04    4.90    7.72    5.22
Bournvita              3.62    6.10    5.64    5.12
Meat                   16.05   19.41   15.33   16.93
Fruits and vegetable   4.92    5.33    4.42    4.89
Plantain               5.00    3.41    4.94    4.45

Source: Field survey

Table 3: Percentage Per Capita Calorie Intake in Household by Food

Items        Osun State           Ondo State          Ekiti State

             Calories    % of     Calories   % of     Calories   % of
             kcal./day   Total    kcal/day   Total    kcal/day   Total

Plantain     128.5       6.05     379.5      16.14    405.8      15.16

Yam          265.5       12.50    346.86     15.00    394.33     14.73

Rice         428         20.14    447.54     19.36    441.5      16.49

Beans        505.7       23.80    248.60     10.75    467        17.45

Bournvita    7.50        3.53     5.81       2.51     26.25      9.81

Fruits and   187.1       8.80     304.61     13.18    323.2      12.07

Gari         231.58      11.84    192.8      8.34     279.6      10.45

Meat         293.93      13.83    296.98     12.84    255.3      9.54

Milk         57.34       2.70     89.3       3.86     83.75      3.13

Total        2125.15     100.00   2,312      100.00   2676.73    100.00

Items        All States

             Calories   % of
             kcal/day   Total

Plantain     304.6      12.54

Yam          335.6      14.07

Rice         439.01     18.66

Beans        407.1      17.33

Bournvita    13.19      5.28

Fruits and   271.63     11.35

Gari         241.32     10.21

Meat         282.07     12.07

Milk         76.80      3.23

Total        2371.32    100.00

Source: Field Survey

Table 4: Percentage Per Capita Protein Intake in Household by Food

Food         Osun State         Ondo State         Ekiti State
             Protein    % of    Protein    % of    Protein    % of
             (gm/day)   Total   (gm/day)   Total   (gm/day)   Total

Plantain     0.92       1.61    1.5        2.36    2.41       3.42

Yam          7.80       13.65   5.99       9.52    9.85       13.98

Rice         8.25       14.43   8.8        13.85   14         19.87

Beans        14.00      24.50   16.00      25.18   15         21.29

Bournvita    0.09       0.15    0.17       0.26    0.3        0.42

Fruits and   4.19       7.33    2.88       4.53    1.32       1.87

Gari         6.63       11.60   8.52       13.41   7.89       11.20

Meat         12         21.00   15         23.61   15.28      21.69

Milk         3.26       5.70    4.67       7.35    4.39       6.23

Total        57.14      100     63.53      100     70.44      100

Food         All States
             Protein      % of
             (gm/day)     Total

Plantain     1.61         2.46

Yam          7.88         12.38

Rice         10.35        16.05

Beans        15           23.65

Bournvita    0.18         0.27

Fruits and   2.79         4.57

Gari         7.68         12.07

Meat         14.09        22.1

Milk         4.10         6.42

Total        63.18        100

Source: Field Survey

Table 5: Pooled Sample Elasticities for Major Food Consumption (OLS)

Food items             Mean budget   Own-price          Expenditure
                       share         elasticity         elasticity

Rice                   29.17%        -1.2017(0.016)     1.16 (0.10)
Gari                   8.26%         -1.121 (0.031)     1.012 (0.003)
Beans                  11.66%        -0.0137 (0.017)    0.014 (0.003)
Yam                    14.31%        -0.001 (0.003)     0.10 (0.09)
Milk                   5.22%         -0.0028 (0.004)    0.02 (0.004)
Bournvita              5.12%         -0.0059 (0.003)    0.018 (0.005)
Meat                   16.93%        -0.00138 (0.003)   0.054 (0.007)
Fruits and vegetable   4.89%         -0.0063 (0.006)    0.0072 (0.003)
Plantain               4.45%         -0.0028 (0.003)    0.032 (0.005)

Note: The numbers in parentheses following the elasticities estimates
are standard errors. All estimates are statistically significant at
the 5 percent level. Source: Data analysis

Table 6: Expenditure Elasticities by Income Bracket (OLS)

            Income level

Food        Income 1              Income 2
items       [less than or equal   N7,501 - N15,000
            to] N7,500
            Mean     Elasticity   Mean     Elasticity
            budget   estimate     Budget   estimate
            share                 Share

Rice        30.74%   1.481        29.80%   1.134
                     (0.021)               (0.032)
Gari        6.03%    1.2165       9.20%    1.076
                     (0.02)                (0.031)
Beans       9.37%    0.013        13.12%   0.0116
                     (0.005)               (0.022)
Yam         16.18%   0.297        13.83%   0.360
                     (0.011)               (0.046)
Milk        5.44%    0.0156       4.05%    0.029
                     (0.012)               (0.012)
Bournvita   4.61%    0.0014       3.57%    0.036
                     (0.002)               (0.008)
Meat        14.85%   0.567        18.33%   0.116
                     (0.014)               (0.31)
Fruit and   7.20%    0.029        4.94%    0.046
Vegetable            (0.006)               (0.038)
Plantain    5.58%    0.117        3.16%    0.047
                     (0.015)               (0.012)

            Income level

Food        Income 3
items       > N15,500
            Mean     Elasticity
            Budget   estimate

Rice        28.81%   1.060
Gari        7.76%    1.0116
Beans       11.28%   0.0136
Yam         14.29%   0.067
Milk        5.59%    0.0197
Bournvita   5.70%    0.0194
Meat        17.62%   0.0476
Fruit and   4.64%    0.0029
Vegetable            (0.002)
Plantain    4.30%    0.0153

Note: The numbers in parentheses underneath the
elasticities are standard errors. All estimates are
statistically significant at the 5 percent level.
Source: Data analysis.

Table 7: Own--Price Elasticities by Income Bracket (OLS)

             Income level

             Income Level 1        Income Level 2
             [less than or equal   N7,501 - N15,000
             to]  N7,500
Food Items   Mean     Elasticity   Mean     Elasticity
             budget   estimate     Budget   estimate
             share                 Share

Rice         30.74%   -1.062       29.80%   -1.005
                      (0.02)                (0.031)
Gari         6.03%    -1.312       9.20%    -1.208
                      (0.008)               (0.044)
Beans        9.37%    -0.987       13.12%   -0.017
                      (0.033)               (0.023)
Yam          16.18%   -0.609       13.83%   -0.011
                      (0.012)               (0.006)
Milk         5.44%    -0.5165      4.05%    -0.004
                      (0.010)               (0.006)
Bournvita    4.61%    -0.745       3.57%    -0.003
                      (0.012)               (0.005)
Meat         14.85%   -0.0611      18.33%   -0.0053
                      (0.005)               (0.005)
Fruit and    7.20%    -0.012       4.94%    -0.0034
Vegetable             (0.008)               (0.009)

Plantain     5.58%    -0.5195      3.16%    -0.0092
                      (0.008)               (0.06)

             Income level

             Income Level 3
             > N15,500

Food Items   Mean     Elasticity
             Budget   estimate

Rice         28.81%   -1.031
Gari         7.76%    -1.121
Beans        11.28%   -0.014
Yam          14.29%   -0.0008
Milk         5.59%    -0.003
Bournvita    5.70%    -0.006
Meat         17.62%   -0.0014
Fruit and    4.64%    -0.0063
Vegetable             (0,006)

Plantain     4.30%    -0.0028

Note: The numbers in parenthesis represent the standard
errors of the estimate.
Source: Data analysis

Table 8: Comparison of Own-Price Elasticities

                    Mean                   Working
Food item           budget   % of zero     lesser     Heckit
                    share    consumption   (OLS)

Rice                29.17%   0.00%         -1.2017    -1.012
                                           (0.016)    (0.016)
Gari                8.26%    0.00%         -1.121     -1.0059
                                           (0.031)    (0.009)
Beans               11.66%   4.20%         -0.0137    -0.0003
                                           (0.017)    (0.013)
Yam                 14.31%   8.40%         -0.001     -0.0001
                                           (0.003)    (0.003)
Milk                5.22%    21.80%        -0.0028    -0.0015
                                           (0.004)    (0.003)
Bournvita           5.12%    26.40%        -0.0059    -0.0045
                                           (0,003)    (0.002)
Meat                16.93%   5.40%         -0.00138   -0.0011
                                           (0.003)    (0.002)
Fruit & vegetable   4.89%    16.90%        -0.0063    -0.0013
                                           (0.006)    (0.004)
Plantain            4.45%    44.10%        -0.0028    -0.0011
                                           (0.003)    (0.03)

Source: Data analysis
Gale Copyright:
Copyright 2011 Gale, Cengage Learning. All rights reserved.