Food Demand Projection and Consumption Patterns of Urban Households in Nigeria’s Kano State

ABSTRACT


INTRODUCTION
In most industrialized nations, household spending contributes more than half of GDP, making it a key factor in economic growth (Sadiq et al., 2020b).Households tend to alter their purchasing habits as their wealth grows quickly and a wide range of new products enter the consumption basket.According to the United Nations Industrial Development Organization (UNIDO, 2018), this phenomenon is seen as a welfare-improving aspect of contemporary economic development.The changing nature of purchasing habits has significant repercussions for the expansion of industries and the overall economy.The ability of industries to achieve growing returns on scale has the greatest immediate impact.In addition, changes in household spending due to income present both opportunities and difficulties for both developed businesses facing a slowdown in demand growth and for emerging sectors of the economy providing luxury products (UNIDO, 2018).The ability of mature sectors to reach economies of scale may be hampered by a decline in demand, which could further spur inventive activity as business owners try to prolong a slowdown in economic growth by introducing new products (Sadiq et al., 2021).Demand expansion for emerging businesses may result in rising profits, allowing for further investment in R&D (Research and Development) operations (UNIDO, 2018).The variability in demand that increases at high levels of income hinders the achievement of scale economies, and this is a problem that new industries must concurrently deal with.According to recent research, income-induced changes in household consumption have an effect on trade patterns, labor supply, and wage disparities between skilled and unskilled employees in the larger economy (UNIDO, 2018).
The analysis of how people utilize their money to meet their needs and wants in accordance with their choices and the resources available could be referred to as the study of consuming behaviour (Katsaiti et al., 2017).A key element of economic health and, thus, a key indication of living standards is the utilization of goods and services.Both today and in the future, wealth and income are available to sustain consumption (by means of the savings that create income).According to the Organization for Economic Co-operation and Development (OECD, 2013), "Income, consumption, and wealth are three elements of the broader idea of economic wellbeing, and it is crucial to understand the links between them" (Sadiq et al., 2020a).The most common methods for determining living standards are income and consumption.According to the Nigeria Bureau of Statistics (National Bureau of Statistics, 2020), income is defined as revenues from current transfers and productive activities.Given that it has a continuous flow or is steady, evaluating expenditure over one week or a month might give an idea of a household's consumption patterns throughout the year.But the amount of income varies significantly from month to month and vice versa between weeks.
Economic theory must address the issue of unsustainable consumption habits in a developing country like Nigeria (Ibbih & Siyan, 2018).According to this literature, current consuming habits that deplete economic resources faster than the environment can replenish them could render progress unsustainable (Sadiq, 2023).Of course, other factors affect household and children's consumption in addition to income and pricing.Additional factors like age or parental education will also influence choices about consumption.The underlying empirical realities connected with the determination of expenditure shares, namely budget allocation towards diverse consumer commodity groups to incomes and prices with a focus on income and price elasticities, had not been the focus of expenditure data analyses in the research area.Previous research has employed single-equation approaches to assess consumer demand.Consumer theory is not given much consideration in single equation specifications, which are mostly focused on estimating elasticities.In addition to being inappropriate for constrained budget sharing and having empirical findings that are more likely to be inaccurate across a larger range of data, using the parameter estimates for welfare concerns won't reveal anything about household wellbeing.Besides, demand analysis has long since moved away from simple equation estimation and toward the most complex methods.System-wide strategies guarantee that the demand systems are consistent with consumer theory.
Furthermore, consumption patterns are dynamic and not static.Despite the ongoing widespread inflation that regularly distorts consumers' spending due to income fluctuations, the literature reveals little to no research efforts that apply the sophisticated methodologies utilized to solve consumer theory problems in the studied area.It is crucial to analyze the pattern of food consumption to forecast the future demand for agricultural goods, food security, and nutrition in the research region, as well as how it responds to changes in price and income.
In order to forecast future food demand under various price and income scenarios, it is helpful to have a deeper comprehension of demand elasticities.This knowledge could be valuable to policy planners when making crucial decisions about the future of policy.Accordingly, it was in light of those above that this research was planned in order to get insight into the food consumption patterns of the urban households in the study area.This research contributes to becoming a basis for policymakers to increase food productivity.The research's conclusions will offer accurate information on budget shares of household food spending, household income, and household food spending's price elasticity in Kano's metropolis.This information will be helpful to government parastatals, non-governmental organizations (NGOs), research institutes, etc.The broad objective of the study was to determine the food demand projection and consumption patterns of urban households in Nigeria's Kano State, while the specific objectives were to assess the households' food budget share to determine the income and price elasticities of households' food consumption; and, to determine the challenges to households' consumption pattern.

Research Methods
Kano State, known as the commercial hub of Nigeria, is located there (Sadiq et al., 2022).It is situated in a typical tropical region between latitudes 10°30'N and 12°38'N and longitudes 7°45'E and 9°29'E (Sadiq et al., 2022).This place has a Koppen's Aw climate, which is a tropical continental climate with dry and wet seasons.The dry season runs from mid-October to mid-May, while the rainy season lasts from mid-May to mid-October (Sadiq et al., 2022).The state's total annual precipitation is approximately 800 mm in the north and 1100 mm in the south (Sadiq et al., 2022), with a mean annual temperature of about 26 °C (Sadiq et al., 2022).The area has Sudan Savannah vegetation because of the rainfall quantities.Not only is Kano state the most populated state in northern Nigeria, but also in the entire nation.The predicted population as of 2022, according to the National Population Commission (NPC) and National Bureau of Statistics (NBS), was 15,462,200, premised on a 3.2% national growth rate (National Bureau of Statistics, 2020).Given this population and the state's 20,131 square kilometer territory, the population density in 2006 was calculated to be around 466 people per square kilometer.Because of this, it has the northern Nigerian states' greatest population density.The success of the study area's crop and animal production, both at the subsistence and commercial levels, is widely known.In the state's outlying areas, where crop production is mostly dependent on irrigation during the off season and on rainfall, these agricultural systems are most prevalent.Rain affects animal productivity as well, but somewhat indirectly.The main food crops grown for domestic consumption are maize, millet, sorghum, and rice, while groundnuts and cotton are grown for industrial and export use.Nigeria's Kano state leads the nation in the production of animal skins and hides chili peppers, gum Arabic, cotton, sesame, garlic, and soybeans.
In drawing a representative sample size, a multi-stage sampling technique was adopted.Firstly, out of the eight local government areas (LGAs) situated in the state metropolis (National Bureau of Statistics, 2020), six LGAs, namely, Dala, Fagge, Gwale, Nasarawa, Tarauni and Ungoggo were randomly selected.Subsequently, using a freelance survey, from each of the selected LGAs, twenty-four (24) households were randomly selected.To ensure uniformity, the composition of the twenty-four households encapsulated low, medium, and high incomes vis-à-vis 8 households each.Thus, a total of 144 households were randomly selected.Using an easy cost-route approach, the households cross-sectional survey data were collected using a well-structured questionnaire complemented with an easy cost-route approach in the year 2022.In order of arrangement of the specific objectives, the first two objectives were achieved using a descriptive statistics and Linear Approximate/Almost Ideal Demand System (LA/AIDS) model respectively; while the third and fourth objectives respectively were achieved using food demand projection model and exploratory factor analysis.

Model Spesification 2.2.1. Almost Ideal Demand System (AIDS)
Due to its many advantageous characteristics, the "Almost Ideal Demand System (AIDS)" is frequently used in applied demand analysis: (a) it approximates any demand system arbitrarily well with a first-order function; (b) it complies with the choice axioms; (c) it allows non-linear Engel curves while still aggregating precisely over customers; (d) simple parameter constraints can be used to test and enforce the homogeneous and symmetry property; and, (e) if the translog price index is approximated, for example, by the Stone index, the demand equations become linear (Henningsen, 2017).
The AIDS model is based on the microeconomic theory of the family and is as follows (Sadiq et al., 2020a): Descriptions:   = budget share of the i th commodity (i.e.  =     /)   = is the price of the j th commodity x = total household expenditure on all the food items considered for the study P = price index for the group of commodities   = stochastic term, and it is assumed to be zero and has constant variance   = intercept   = price coefficient   = expenditure coefficient.
P is a translog price index defined as: The price index from equation (2) makes equation (1) a non-system of equations, raising estimation difficulties.To avoid non-estimation, many empirical studies used price ( * ) instead of P (Deaton & Muellbauer, 1980).
Since the translog price index Pt is the only nonlinear component of the Marshallian demand equations, the Stone index was used to simulate the translog price index (Deaton & Muellbauer, 1980;Henningsen, 2017).This approach makes the parameters of the demand equations linear, rendering the method of share equations considerably simpler to estimate.The term "linear approximation of the AIDS" (LA-AIDS) is used to describe this modification of the original AIDS model: For any anticipated demand system to be considered theoretically consistent, it must adhere to four general limits that are implied by consumer theory.The AIDS inherently satisfies the adding-up criteria, but parameter limitations can impose global homogeneity and symmetry restrictions.Yet parameter limitations typically cannot impose monotonicity and concavity characteristics.The coefficients are subject to some limitations imposed by microeconomic household theory, including the premise of utility maximization.
The "adding-up" requirement guarantees that the expenditure portions will always equal one (∑    = 1).The following conditions are met: The "homogeneity" requirement ensures there is no "money illusion", meaning that if all prices & income fluctuate at a uniform rate, the amounts consumed remain the same.This suggests that customers are not concerned with absolute prices or levels of money or income, but just with relative prices and income.Hence, demand theory does away with the illusion of money.It is completed if: The "symmetric", when the spending function of AIDS is subjected to Shepard's Lemma, condition follows.Only if all prices are equal can the symmetry limitations guarantee symmetry.It is accomplished if: Its "monotonicity", the expenditure function must increase monotonically with prices, which means that its first-order derivatives with reference to prices must always be positive.
The concavity condition ensures that the utility maximization issue has a singular solution.If the expenditure function's Hessian matrix, also known as the "Slutsky substitution matrix" in this case, is concave -is a definite negative preposition (Henningsen, 2017).The substitution matrix's i th and j th elements are as follows: Descriptions: The estimated parameters are used to calculate the demand elasticities, which have conventional ramifications.Following are the details of the expenditure elasticity (∈  ), a measure of how responsively demand changes with changes in consumption expenditure: MBS and ABS means marginal budget share and average budget share, respectively.Price elasticity can be calculated in two different ways: uncompensated (Marshallian) elasticity, which includes impacts on both prices and income, and compensated (Hicksian) elasticity, which only includes effects on prices (Sadiq et al., 2020a).When the total expenditure and other prices are maintained constant, or ceteris paribus, the uncompensated own-price elasticity (∈  ) and the cross-price elasticity (∈  ) measure how a change in the price of a product affects the demand for that product and that of other items, respectively.
Following (Sadiq et al., 2020a), the Marshallian own and cross-price elasticities are presented below: Shown below (Sadiq et al., 2020b), the Hicksian own & cross-price elasticities (∈  *  ∈  * ) quantify the price impacts on demand under the assumption that the real expenditure (   * ⁄ ) is constant.
Moreover, using ∈  , ∈  and ∈  , one can estimate the compensated price elasticity, with the following permutation: The sign of the estimated ∈  * shows the substitutability or complementarity of the destinations under consideration, according to (Sadiq et al., 2021).If the compensated crossprice elasticity of a commodity pair is negative or positive, respectively, then that commodity pair is referred to be a complement or substitute.A food item is categorized as a Giffen or inferior commodity (∈  < 0), a luxury commodity (∈  > 1), or a necessity/requirement commodity (0 <∈  < 1) based on the value of spending elasticity.If the elasticity value of a given product's own price is greater than unity, the demand for that commodity is considered to be price elastic (inelastic) in absolute terms (less than unity).The Hicksian elasticity describes the change in demand for a good produced by a change in price when the actual expenditure change brought on by the aforementioned price change is offset by a change in spending, keeping satisfaction or utility constant.

Food demand projection model
The growth formula as adopted by Hina & Abbas (2021) was used to estimate the future demand for food items.Hence, the three most significant variables affecting future demand are changes in consumption patterns, per capita real income (y) & population growth (N).
Where,   = the household demand for i th commodity in period t  0 = the per capita consumption of the i th commodity in the base year (2022)   = the projected population in the year t for the area  = growth in per capita income (GDP -3.06%)   = the expenditure/income elasticity of demand for i th commodity  = year (1, 2, 3 ….. n): for the base year, t is 0 NPC and NBS database were used to project the future population of metropolitan Kano from 2022 to 2040 with the aid of a simple compound formula as presented below: Where,   = future population   = present population  = population growth rate  = number of year (s).

Households' Food Composition
Two factors lead researchers to consider households' consumption expenses as a standin for income.Firstly, according to Friedman (1957), who bases his argument on the permanent-income hypothesis, expenditures are more likely to depict permanent income and are, hence, a more reliable indicator of consumption patterns (Ahmed & Shams, 1994).Second, spending information is typically more trustworthy than income information.On the aggregate, a cursory review of the households' expenditure budget shares in the studied area showed cereals to account for 60.73%: coarse and non-coarse being 37.61and 23.12% respectively; protein accounts for 17.39%: animal and non-animal proteins respectively being 12.45 and 4.94%; while root and tubers, and oil (lubricants) had share budget expenditures of 9.97 and 11.92% respectively (Table 1).Besides, individual-wise, in descending order, local rice and foreign rice had the most appreciable shares in the budgetary expenditure, though didn't exceed 20% each, then distantly followed by meat, groundnut oil and spaghetti whose individual shares is a single-digit but greater than 5% while the remaining ten (10) food commodities had share budget expenditures each of less than 5%.Therefore, in a nutshell, with the share contribution of carbohydrates and starchy foods in the consumers' food basket being 70.70%, it can be concluded that the dietary diversity of the households in the study area is very low.Thus, this indicates a poor balance in the dietary nutrition of the households in the study area and this might be associated with inflationary effect that erodes their purchasing power, thus forced the studied households to strike a balance of survival between the body and soul.
Inspite of the cosmopolitan characterization of the study area-elites dominated, the expectation is that dietary consciousness will be the driving force/shape most of the households' budgetary expenditure but high costs of living standard that marred their purchasing powers adversely led to poor balance diet in the households' food baskets in the study area.Furthermore, the marginal propensity to consume (MPC) of the households was to some extent moderate for indomie (26.72%); low for local rice (16.75%) and groundnut oil (14.47%); and, very low for five commodities viz.Irish potatoes (7.48%), meat (6.62%), fish (6.24%), foreign rice (6.11%) and Semovita (5.27%).Besides, the MPC of four commoditiesyam (4.64%), spaghetti (4.32%), beans (3.34%) and millet (2.45%) were marginal while the households MPC for maize, garri and palm oil commodities were negative.
The gentle slope in the households MPC might be attributed to their purchasing power in relation to the general price level given that the MPC of poor people towards food commodities tend to be high compared to the rich people who have low MPC for food commodities.Nevertheless, the negative MPC associated with the three food commoditiesmaize, garri and palm oil might be connected to cultural food dietary behavior/ attitude in the study area-the north-western part of Nigeria unlike in the southern part of the country where these food commodities have strong ties/affinities to cultural dietary.

Food Expenditure Estimates
Empirically, the LA-AIDS model was found to be best fit for the households' expenditure prediction as it fulfilled its theoretical consistency of adding-up, homogeneity, symmetry conditions and 100% monotonicity fulfillment at 144 out of 144 observations.In a nutshell, the model keeps to the hypothesis of the consumer theory that underpins the fulfillment of four general restrictions that must be satisfied by any estimated demand system for theoretical consistency.Because of the use of budget shares, the adding-up condition was satisfied automatically.The condition of homogeneity implies that that the prices and income are homogenous of degree zero, i.e.only real prices and income matter to the households and not nominal prices and money-income level, thus the neglect/ discard of money illusion.The fulfillment of the symmetry condition implies that cross-price derivatives are identical.The monotonicity condition implies that the households' expenditures monotonically increases in prices, i.e. the first derivatives of the expenditures with respect to prices are non-negative.
Furthermore, for the coefficient of multiple determination (R 2 ), the values of the fifteen food items fitted into the equations ranged between 0.0114 for local rice to 0.308 for yam (Table 2).These R 2 values are not uncommonly low in consumption analysis when dealing with cross-sectional data as evidently pointed out by (Haq et al., 2011).In a related research, Hina & Abbas (2021) in their study on consumption pattern in Pakistan reported low R 2 values.The positivity and significant of the intercepts of six commodities (local rice, foreign rice, yam, spaghetti, indomie and groundnut oil) at equal or less than 10% probability level, indicate an exogenous/external increase in the demand for these commodities that is independent of the changes in prices and income.In the LA-AIDS matrix, out of the 224 estimated slope coefficients, 69 parameter estimates were different from zero at 10 percent error gap (Table 2).

Expenditure/Income Elasticity
Except maize, garri and palm oil, empirically the expenditure elasticity showed all the remaining food commodities to be normal goods as evident from their respective elasticity estimates that were positive signed (Table 1).Further, seven commodities (millet, yam, Irish potatoes, semovita, indomie, fish and groundnut oil); five commodities (local rice, foreign rice, beans, spaghetti and meat); and, three commodities (maize, garri and palm oil) were found to be luxury, necessity and inferior commodities respectively, vis-à-vis their respective coefficients that were greater than unit, less than unity and negative.
Necessities are the commodities that are common in the food baskets of households, i.e. everyday goods or goods purchased at all the times by households and easy to afford.Therefore, local rice, foreign rice, beans, spaghetti and meat are more affordable compared to millet, yam, Irish potatoes, semovita, indomie, fish and groundnut oil.Except foreign rice, all the normal goods have relative high expenditure elasticities, thus implying that most of the households, especially the poor, face tight budgetary constraints and they deemed all these food commodities to be very important as they fulfill their fundamental needs.106 For the commodities identified to be luxury, an increase in income will lead to an increase in their budget shares while for the necessary goods, any increase in the households' income will lead to a decrease in the budgetary expenditure of any given household in the study area.In other words, for the luxury foods, their demands will increase if the households' real income increases in tandem with the overall economic growth of the study area.However, if the households' real income depletes, in relative terms, less expenditures will be apportioned to these goods.For the necessary goods, if the relative income of households increases, less expenditure will be budgeted for these commodities while if the reverse is the case (decrease in households' real income), more expenditures will be budgeted for these commodities.Nonetheless, for the inferior commodities, neither an increase nor a decrease in income will affects the demand for these commodities.It is worth to mention that necessary/necessity goods are income inelastic while luxury goods are income elastic.It can be inferred that the growth for millet, yam, Irish potatoes, semovita, indomie, fish and groundnut oil in the study area will surpass the growth in the income.Nevertheless, given a fixed supply for millet, an upward shift of the demand curves will imply that the market equilibrium prices will increase.Since the own-price elasticity of millet is greater than unity, it is anticipated that the increase in the price due to the shift in the demand curve will lead to an increase in the demand for millet by more than the proportionate price change.Consequently, as households' expenditures increases, likewise diversifying their diets, they tend to increase their consumption of non-staple commodities rather than the staple commodities.

Price Elasticity 3.5.1. Own-price elasticity
Uncompensated price elasticity is the elasticity that is not adjusted for income while compensated price elasticity is adjusted for real income changes so as to maintaining consumers' utility.Besides, the former and latter respectively are composed of two effects (income and substitution effects) and one effect (substitution effect).For the own price elasticity, it shows the responsiveness of demand for a particular commodity to a change in the real prices; while for the cross price elasticity, it reflects the responsiveness of demand for a particular good to a change in the real price after real income compensation of the consumers.The uncompensated own price elasticity shows the responsiveness of a good to a change in its own price while the compensated own price elasticity shows the responsiveness of a commodity to a change in the price of other commodity(s).
The uncompensated own price elasticity of local rice, foreign rice, semovita and indomie noodles being negative, signifies they are everyday food commodities in the food baskets of most households in the study area (Table 2).However, the remaining eleven food commodities can be categorized as Giffen commodity owing to their positive own price elasticity coefficients.A Giffen good in economics is referred to as a low income, non-luxury good that defies standard economic and consumer demand theory.The demand for a Giffen commodity increases with the rise in its price and vice-versa.This is quite possible in the study area because most of the households are low income earners, thus makes these food commodities to be low income goods with very little or no-substitutes.The expenditure elasticity of those food commodities that are negative and greater than unity, thus consistent with the demand theory, implies that households are highly sensitive to a price change in these commodities.In other words, local rice, foreign rice and semovita showed a very elastic behavior with respect to their elasticity magnitudes of -3.39, -1.29 and -1.44 respectively; while indomie noodle was inelastic.Therefore, a change in the price of the former will lead to a more than proportionate change in their respective demand; while it will lead to a decrease in the case of the latter.
Furthermore, based on the uncompensated own price elasticity estimate, the empirical evidence showed that if the prices of the local rice, foreign rice and semovita plummet by 1 percent, their demands would surge by 3.39, 1.29 and 1.44 percent respectively.As defined for complemented own price elasticity, out of this demand increase, price effect, i.e. substitution effect is purely responsible for 3.23, 1.23 and 1.39 percent respectively for the aforementioned commodities while the income effect due to the declined price accounts for 0.16 (3.39-3.23),0.06 and 0.05 percent rise in the demand for the respective commodities as a result of increase in the real/relative income.Though, the absolute/nominal money income remains unchanged.However, due to the small budget shares of these commodities in the households' food basket, income effect is relatively small on their respective demands.In other words, changes in their respective prices had minimal effect on the households' relative/real income.Further, if a 1 percent increase in a per capita income is accompanied a percent decline in the price of any of these foregoing commodities, the demand for the local rice, foreign rice and semovita would rise by 4.35 (3.39 + 0.96), 1.67 and 2.78% respectively.Though, the rise in the per capita income represents a shift in the commodity (local rice/foreign rice/semovita) demand curve that normally leads to a surge in the commodity price.A compensation for a particular commodity is worthless if there is no distinction between the uncompensated and compensated own price elasticities.Consequently, based on the empirical evidences, uncompensated vis-àvis compensated own price elasticities of all the food items were approximately similar, and this might be attributed to their proportions in the total households' budgetary expenditures being marginal-to-small.However, for food items viz.local rice, indomie and groundnut oil that showed to some extent a reasonable margin, it is suggested that government should indemnify the households when it observes a surge in the prices of these food items to enable the consumers to maintain a status quo ante in their welfare level in case of price change in the study area.

Cross price elasticity
The degree of the responsiveness of demand for a good to a change in the price of another commodity is called cross elasticity.Positive and negative cross price elasticity respectively, imply that two goods are substitute and complements.The uncompensated cross price elasticity gives the 'gross cross' effect that includes both the substitution and income effects while the compensated cross price effect, the 'net effect' of price change on demand, represents only the substitution effect, i.e., the pure price effect.The empirical evidence showed that the uncompensated cross price elasticity has a total of 103 gross substitute and 122 complement commodities (Table 3a) while the compensated cross price elasticity has a total of 107 and 118 commodities that are net substitutes and complements respectively (Table 3b).The positive sign associated with the cross-price elasticity of demand for local rice to the respective prices of foreign rice, beans, yam and semovita implies that the local rice and the commodities in pair are substitutes.In a nutshell, for example, the positive cross price elasticity of demand for local rice due to change in the price of foreign rice indicates that the duo commodities are substitute.The cross-price elasticity of local rice-to-foreign rice, local riceto-beans, local rice-to-yam, and local rice-to-semovita are positive because the price of the former in relation to the demand for the latter moves in the same direction.The pure price or substitution effect of 10% fall in local rice price will lead to a decrease in the demand of beans and yam by 7.69 and 2.26 percent respectively.The increase in the relative/real income, i.e. income effect, due to the fall in the local rice price will induce the consumers to plummet their beans and yam demands by 0.48 (i.e., 7.69-7.21)and 0.43 percent respectively.Furthermore, the cross-price elasticity of demand for local rice due to a change in the price of meat being negative sign means that both commodities are complement.The pure price effect of a 10% fall in the price of local rice will lead to a decrease in the demand for meat by 8.40%.The increase in the relative income due to the decrease in the local rice price will make the consumers to increase the demand for meat by 0.86% (i.e., 9.26-8.40).
Between the uncompensated and compensated cross elasticities, the signs associated with some elasticity estimates differ.For example, millet-to-yam cross elasticity, the empirical evidence showed the uncompensated cross price elasticity to be negative, indicating the duo goods to be gross complement whereas the compensated cross price elasticity was positive, implying the two commodities to be net substitute.Given that yam has a relative high income elasticity of demand, an increase in the relative income due to a decrease in the price of millet will lead to a rise in the demand for yam.Since the price effect is lower than the income effect, thus it can be inferred that a fall in the pure price of millet will lead to an increase in the demand for yam.More uncertainty/ambiguity exists with respect to uncompensated cross-price elasticity.The significant cost impacts, however, undoubtedly have an impact.When looking for information on possible substitutions, compensated cross price elasticity is the most suitable.

Households' Demand Projection
The empirical evidence of households' projected demand of food commodities from 2022 to 2042 (four years interval) for the total demand (million metric ton) per year showed that local rice, Irish potatoes and indomie noddles will witness a steep rise in their future demand while the demand for foreign rice, yam, semovita, spaghetti and groundnut oil will be marked by a gentle rise in the future (Table 4 and Figures 2-3).Besides, the rise in the future demand for beans, millet, meat and fish will be marginal; though plato, the future demand for palm oil will witness a slight marginal increase.However, the future demand for garri will be on the decrease and this might be attributed to the status (inferior) of the commodity in the study area.Generally, the total demand for local rice, foreign rice, beans, maize, millet, yam Irish potatoes, garri, semovita, spaghetti, indomie noodles, meat, fish, palm oil and groundnut oil will increase respectively from 13.83, 12.02, 3.79, 1.46, 1.76, 3.39, 2.77, 1.42, 3.19, 5.03, 10.23, 6.71, 2.83, 3.32 and 6.05 million metric (MMT)/year in 2022 to 43.47, 27.32, 10.16, 1.81, 6.10, 11.15, 17.01, 1.35, 12.34, 13.41, 59.97, 18.68, 13.67, 5.57 and 32.84 MMT/year in the year 2042.Furthermore, on a per capita basis (kg/year), local rice, Irish potatoes and indomie noodles will witness a steep increase in their future demand.In contrast, yam, semovita, groundnut oil and meat will witness a gentle rise in their respective future demand.Besides, the future per capita demand of foreign rice, beans, millet, spaghetti and meat will be marked by a marginal rise.However, the future per capita demand of maize, garri and palm oil will plummet, i.e. will witness a marginal decline, and the possible reason may be attributed to the cultural consumption behaviour of the study area towards these commodities.Generally, it can be inferred that the changes in the households' consumption expenditures in the study area owed to a rise in income, population growth and cultural consumption behaviour.

Consumption Constraints
To determine the common factors affecting consumption, the varimax rotation applied to the fifteen variables showed only five factors to be interpretable as evidenced by their respective Eigenvalues that were greater than unity (Table 5).From the total variance, the combined variance of these factors is 74.80%, wherein factors 1, 2, 3, 4 and 5 variances are 24.96, 17.87, 14.25, 10.14 and 7.59% respectively. Besides, Hair et al. (1998) sentenced a contributory variance value of 70% or above to be satisfactory in social sciences.The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy value being 0.51, a mediocre level, indicates that the R-matrix has a common factor; the correlation pattern is relatively compact and the factor analysis yielded distinct and reliable factors for prediction with certainty, efficiency and accuracy.The KMO value is within the threshold recommended to be acceptable by Field (2005).Further, Bartlett's test of Sphericity (BTS) being different from zero at a 1% probability level implies that the R-matrix is not an identity matrix.Besides, each factor has an internal consistency in its factor loadings, as evidenced by their respective Cronbach's Alpha test of reliability that is not less than 0.70, thus ensuring satisfactory reliability.A value less than 0.70 indicates unsatisfactory internal consistency reliability (Youssef et al., 2012).The acceptable reliability estimates is within the range of 0.70 and above for social sciences.For the extracted factors, factor loadings whose absolute values were less than 0.40 were outrightly excluded (Bagheri & Fami, 2016;Sadiq et al., 2019;Sadiq et al., 2018;Sadiq et al., 2017).In labelling the factor that was loaded from two factors, only the higher factor score was considered.The extracted factors affecting consumption were labelled inflationary, population explosion, occupational, debt and economic challenges.The dimension labelled "inflationary challenge" showed households' concern about the effect of a high inflation rate, such as low monthly income, high cost of non-food expenditure, saving for the future, high food prices and misery instincts that inhibited their consumption.The dimension labelled "population explosion challenge" points to the households' concern about the social menace, such as the vulnerability of households to population explosion-overdependence ratio and insecurity/theft that affect their consumption.The dimension labelled "occupational challenge", showed households' concern about occupational hazards such as loss of employment and continuous salary cuts that affected their consumption.The dimension labelled "debt challenge" showed households' concern about how debt prevented them from accessing good health care, thus affecting their consumption.Lastly, the dimension labelled "economic challenge", showed households' concern on economic woes, such as the need to bequeath fortune and lack of steady income that affected their consumption.

CONCLUSIONS
Based on the empirical evidence, it was established that food dietary diversity was very low as cereals dominate, and this is owed to food inflation that inhibited the purchasing power of the households.Besides, rice, beans, spaghetti and meat were necessary commodities; millet, yam, Irish potatoes, semovita, fish and groundnut oil were luxury goods, while maize, garri and palm oil were inferior goods.In addition, rice, semovita and indomie noodles were everyday goods as evidenced from the uncompensated own price elasticity; likewise, demands for rice and semovita that were highly price sensitive as established by compensated own price elasticity.The uncompensated own price elasticity showed that the government should indemnify the households so as to maintain welfare if it observes a surge in the food prices of local rice, indomie noodles and groundnut oil.Furthermore, it was established that the future per capita demand for local rice, Irish potatoes and indomie noodles will surged.Therefore, in order to bridge the increasing future demand of these commodities in particular, government is advised to increase their productivity.Nevertheless, the determined obstacles to households' food consumption were inflation, households' population explosion, occupational hazards, debt and economic woes.Thus, the study advice policy makers to soften the macro-economic policies, i.e., implements policies that have human-face and are economically friendly.
Figure 3. Per capita demand projection

Table 1
Note: ABS = Average budget share; MBS = Marginal budget share; G/Oil = Groundnut oil

Table 2 .
LA/AIDS food expenditure parameter estimates

Table 3 .
a. Uncompensated own and cross price elasticities of demand for food items

Table 4 .
Projected total and per capita food demand from 2022 to 2042