Analysis of Household Food Demand in South Kalimantan Province Application Quadratic Model Almost Ideal Demand System (QUAIDS)

Food consumption patterns are influenced by households’ socioeconomic characteristics, such as the number of household members, education of household members, household income, and area of residence. This study aims to analyze the households food consumption pattern in South Kalimantan Province based on socio-economic characteristics. Quadratic Almost Ideal Demand System (QUAIDS) model is used to analyze income elasticity and price elasticity which represent the relationship between food demand and price and income in households in South Kalimantan. The data used are data from the National Socio-Economic Survey (Susenas) KOR and the South Kalimantan Consumption Module that was collected in March 2022. The findings indicated that the food and beverage commodity group accounted for the largest share of food expenditure in South Kalimantan, reaching 32 percent in each socio-economic variable. The Quadratic Almost Ideal Demand System (QUAIDS) model’s independent variables price, income (expenditure approach), and socio-economic characteristics can be used to estimate the share of food expenditure at the household level. The income elasticity of food demand as a whole shows a positive value, indicating that all commodity groups fall into the category of normal goods. Meanwhile, the price elasticity itself shows a negative sign, indicating that an increase in commodity prices results in a decrease in demand for commodity consumption. The cereals and tubers group is the most inelastic of the five commodity groups, with a cross elasticity of food commodity groups showing negative (complementary) and positive (substitution) values.

Introduction

Humans have a very basic requirement that must be met: food. Economic instability may result from a shortage of food compared to demand. If food security is jeopardized, there may also be a number of social and political upheavals. Even national and economic stability may be at risk due to this dire food situation. Thus, the availability of food is essential.

Economics researchers have utilized food as a parameter to characterize the degree of economic well-being in the community. This is predicated on Engel’s theory, which holds that a rise in per capita income will cause the percentage of money spent on food to decline. When comparing food expenditures between residential area, urban and rural households generally exhibit distinct patterns. Urban households consume more non-food commodities than rural households. Susenas data from 2022 indicates that this is the exact situation of South Kalimantan Province [1].

Households’ income level determines how much and what kinds of food they eat. The likelihood that a household will select healthy food in terms of both quantity and type increases with the household’s purchasing capacity. The tendency for people to consume higher-quality food at a higher cost per unit of nutrients will increase with income.

As money rises, there will be a general increase in the intake of foods with higher nutritional content and a diversification of food consumption habits. In addition to income, household tastes and price levels have an impact on the demand for food commodities. The number of household members, their level of education, their habits and cultural norms, and their tastes all have an impact on the food consumption preferences of the household. Different income levels will undoubtedly result in varying degrees of preference and educational attainment among household members, which will impact lifestyles, particularly with regard to consumption. So that, depending on their unique circumstances, each household will react differently to changes in income and prices [2].

In addition to being impacted by dietary practices that vary from one group to another, budgetary restrictions also have an impact on the choice of food products that households will consume. Research on households consumption habits is therefore required. Information on the demand for food items and the circumstances surrounding price fluctuations will be provided by this study. A household’s desire for different commodities linked to food security and insecurity issues can also be predicted using research on consumption patterns. The data will also be used to forecast future demand, balance community welfare, and assess the effects of governmental policy changes.

The government would be able to create policies with the help of studies on food demand elasticity and income elasticity. Furthermore, food safety and the study of food demand patterns are closely intertwined, which aids in understanding the demand-side factors that affect consumption and how it changes in response to regional price and income influences. Given this context, research on food demand studies in South Kalimantan is imperative.

This study aims to analyze household food consumption patterns in South Kalimantan Province based on socioeconomic characteristics using the QUAIDS model and analyze income elasticity and price elasticity of food demand in households in South Kalimantan Province.

It is anticipated that the findings of this study will serve as a resource for future researchers and as a study guide for anyone with an interest in learning more about the trends in food demand and how South Kalimantan households react to variations in income and price. Additionally, when interventions or food assistance programs are implemented to improve welfare, particularly for low-income population groups, the findings of the study of food demand and consumption based on income groups and/or the nutritional status of the population are helpful input for decision makers.

Research Methods

Place and Time of Research

Beginning with the proposal writing phase in February 2023 and continuing with data processing until the research report was finished in June 2024, the study was carried out in the province of South Kalimantan.

Types and Sources of Data

Susenas data from March 2022 Kor and KP South Kalimantan Province, as well as other data from the Central Statistics Agency’s data release, are used in this study as secondary data. The Susenas Consumption Module’s data is cross-sectional and includes household sample units; in 2022, there were 8,300 households split over 13 districts and cities, with 8180 households eligible for data. Ten homes were chosen from each chosen BS using a systematic sampling technique that included implicit stratification according to the household head’s educational attainment [3].

Analysis Method

The categorization of Individual Consumption According to Purpose (COICOP) code, a common categorization of household consumption to acquire goods and services organized according to their usage, served as the basis for the formation of food commodity groups in this study. The Indonesian Household Expenditure/Consumption Standard Classification 2003 is the source of the COICOP code that is utilized [4]. Food commodities are categorized as follows:

1. Cereals and tubers group;

2. Group of fish, meat, eggs, milk, and their products;

3. Groups of vegetables, nuts, fruits, condiments, fats and oils;

4. Prepared food and beverages group;

5. Other food groups.

This study uses a semilog-shaped QUAIDS model analysis as follows:

$$\begin{array}{rcl}{w}_i& =& {\alpha }_i+{\sum }_{j=1}^n{\gamma }_{ij}lnln{p}_j+{\beta }_{i}lnln\left\{\frac{x}{a\left(p\right)}\right\}+\frac{{\lambda }_i}{b\left(p\right)}\\ & & {\left[lnln\left\{\frac{x}{a\left(p\right)}\right\}\right]}^2+{\alpha }_{i1}educ+{\alpha }_{i2}dloc+{\alpha }_{i3}dgen\\ & & +{\alpha }_{i4}hsize+{\alpha }_{i5}inc2+{\alpha }_{i6}inc3+{\alpha }_{i7}IMR+ui\end{array}$$

where

w_i—budged share of ith food group

p_j—aggregate price of the jth food commodity group

x—household expenditure on food consumption

a(p)—price index

b(p)—aggregate price

dloc—location dummy

educ—household head education dummy

dgen—household head gender dummy

hsize—household size

Inc2—middle income group

Inc3—high income group

IMR—Inverse Mills Ratio

i—1,2,3,4,5

α_i, α_i1, α_i2, α_i3, α_i4, α_i5, α_i6, α_i7, γ_ij, β_i, λ—estimated parameters

The AIDS model is a demand model derived from an indirect utility function that is linear in the logarithm of total income. However, the AIDS model cannot capture the effect of Engel curve nonlinearity as often found in empirical demand studies [5]. In addition, the AIDS model cannot capture information on income class differences and regional differences. To maintain the positive properties of the AIDS model as well as to maintain consistency with the Engel curve and the influence of relative prices in utility maximisation, the quadratic form of the logarithm of income is added to the AIDS model so that the model becomes Quadratic AIDS (QUAIDS) [6].

Other studies have shown that in addition to food price factors, socioeconomic factors also influence the rate of change in households’ opportunities to consume a food group, namely total monthly food expenditure, area of residence, household size, years of schooling of the household head, age of the household head, gender and marital status of the household head, floor area per capita, poverty status and the household’s main source of income [7].

The drawback of using household survey data is the absence of prices of the commodities consumed. There are several approaches to overcome this problem, such as the use of unit values. The commodity price or unit value in the Susenas data is listed in the value spent by the household to buy the commodity. This value is obtained from the purchase price multiplied by the amount purchased by the household.

To overcome this problem, the price variable in this study uses a unit value corrected by the price differential method [8]. The unit value is corrected by adding the mean value and the residuals of the regression estimating the difference in the mean value of the unit value with socio-demographic factors, namely regional classification, average years of schooling of the family head, household size, gender of the household head and income group. Mathematically, it can be written as follows:

$$\begin{array}{rcl}{v}_i-v_{median}=& & {\alpha }_i+{\alpha }_{1}dloc+{\alpha }_{2}hsize+{\alpha }_{3}dgen\\ & & +{\alpha }_{4}educ+{\alpha }_{5}inc2+{\alpha }_{6}inc3+{\epsilon }_i\end{array}$$

The correction price is formed from the addition of the average per unit value of the i-th commodity group at the commodity level and the residuals:

$$\begin{array}{r}{\left(p_i\right)}_{median}={\left(v_i\right)}_{median}+{\left({\hat{e}}_i\right)}_{median}\end{array}$$

The resulting price values indicate that each household in a region is assumed to face the same market price for each good. The price values are free from endogeneity problems caused by different quality factors among households in a group [8].

Based on several models, as a basis for estimation, the equation used in this study is a model developed by Poi [9], a QUAIDS model that accommodates the variable classification of household size, income class. To overcome the problem of empty data, the variable Inverse Mills Ratio (IMR) is added, thus, the QUAIDS equation becomes:

$$\begin{array}{rcl}{w}_i& =& {\alpha }_i+{\sum }_{j=1}^k{\gamma }_{ij}lnln{p}_j+{\beta }_{i}lnln\left\{\frac{m}{P(p)}\right\}\\ & & +\frac{{\lambda }_i}{b(p)}{\left[lnln\left\{\frac{m}{P(p)}\right\}\right]}^2+{\sum }_{s=1}^k{\vartheta }_s{D}_{s1}^k+u_s^k\end{array}$$

where

w_i – share of expenditure for five groups

p – price

m – total expenditure and

D – household characteristics

Variable Operationalisation

The variables used (according to data availability) are:

1. The value of the proportion of expenditure on selected food consumption consisting of quantity and rupiah value (consumption module data) either from purchase or from own production and provision for each commodity.

2. The commodity price is calculated as unit value, which is the quotient between the rupiah value of food expenditure and the quantity of food consumed.

3. Total household expenditure per month as an approximation of monthly income. Household expenditure here includes purchases for food and non-food (goods and services) at home and abroad.

4. Household size is the number of household members, which is all the people who usually reside in a household (krt, husband/wife, children, children-in-law, grandchildren, parents/parents-in-law, other relatives, domestic helpers or other art).

5. Income groups are divided into three: the lowest-income 40% of the population, the middle-income 40% and the highest-income 20%.

6. A dummy variable indicating:

- Region of residence: urban/rural.

- Gender of household head.

- Education of household head.

Results and Discussion

Consumption Patterns

People’s food and non-food consumption patterns generally depend on the socio-economic characteristics of each region. South Kalimantan Province, which consists of two regions, namely 11 districts and two cities, has an average monthly household expenditure in general in 2022 of IDR 1,407,719. This is higher than the national average of IDR 1,327,782. [2]

In general, food expenditures account for 56.10% of household budgets in rural areas and 47.27% of overall household budgets in urban areas. This demonstrates that the majority of spending in rural areas goes toward food, whereas the majority of spending in urban areas goes into non-food groupings (Fig. 1).

Fig. 1. Proportion of household expenditure.

The budget share by food commodity group serves as the dependent variable in the study that employs the QUAIDS model. Table I shows the budget proportion for each food commodity group in both urban and rural areas.

Table I. Average Budget Share of Food Commodity Groups

Budget share	Urban	Rural	Total
Cereals and tubers	0.11	0.14	0.13
Fish, meat, eggs, milk, and their derivatives	0.22	0.21	0.21
Vegetables, legumes, fruits, oil, coconut, beverage dan spices	0.18	0.19	0.19
Prepared food and beverages	0.36	0.30	0.32
Miscellaneous food items	0.13	0.16	0.15

Metropolitan communities spend 36% of their overall food expenditures on processed foods, compared to 30% in rural regions (Table I). Given that there is a higher range of variety of processed foods available in urban regions compared to rural ones, it is evident that people in urban areas consume processed foods more frequently than those in rural communities. These findings are consistent with study by Molina et al which found that people in urban regions earn more than those in rural areas and that the location of residence has a positive relation to the part of expenditure [10].

Zero Consumption

Zero consumption of a food group is something that may occur when using susenas micro data. In this study, zero consumption also occurred (Table II).

Table II. Proportion Zero Consumption of Food

Commodity group	Urban	Rural	Total
Cereals and tubers	1.49	0.33	0.79
Fish, meat, eggs, milk, and their derivatives	1.34	0.61	0.90
Vegetables, legumes, fruits, oil, coconut, beverage dan spices	0.97	0.37	0.61
Prepared food and beverages	0.03	0.16	0.11
Miscellaneous food items	0.79	0.25	0.46

Parameter Estimation

In the data processing results, parameter estimation shows that of all the coefficients, 78% of them have significance values at the 1%–10% level. This indicates that price changes affect the proportion of expenditure on the five food commodity groups analysed. A positive price effect means that if there is an increase in price, it will increase the budget share of the commodity group; if it is negative, the effect of price increase will decrease the budget share of the commodity group. In other words, changes in quantity consumed are strongly influenced by price changes and price changes will quickly drive an increase or decrease in demand for a commodity (Table III).

Table III. Parameter Estimation Coefficient of QUAIDS Model

Variable	w_group1	w_group 2	w_group 3	w_group 4	w_group 5
ln (price of group 1)	0.131***	−0.062***	0.006	−0.031***	−0.043***
	(0.013)	(0.008)	(0.009)	(0.009)	(0.013)
ln (price of group 2)	−0.062***	0.010	0.033***	0.024**	−0.005
	(0.008)	(0.011)	(0.008)	(0.012)	(0.018)
ln (price of group 3)	0.006	0.033***	0.006	0.061***	−0.105***
	(0.009)	(0.008)	(0.011)	(0.011)	(0.015)
ln (price of group 4)	−0.031***	0.024**	0.061***	−0.027	−0.027
	(0.009)	(0.012)	(0.011)	(0.030)	(0.029)
ln (price of group 5)	−0.043***	−0.005	−0.105***	−0.027	0.180***
	(0.013)	(0.018)	(0.015)	(0.029)	(0.044)
ln × (income)	−0.020	0.066***	0.059***	0.258***	−0.364***
	(0.015)	(0.020)	(0.018)	(0.030)	(0.030)
ln ×2 (income squared)	0.009***	0.007***	0.006***	0.018***	−0.039***
	(0.002)	(0.002)	(0.002)	(0.003)	(0.005)
educ (highest education)	0.002***	−0.003***	0.000	−0.002***	0.003***
	(0.000)	(0.000)	(0.000)	(0.001)	(0.000)
dloc (region classification)	−0.009***	−0.004***	−0.011***	0.002	0.022***
	(0.002)	(0.001)	(0.002)	(0.002)	(0.003)
dgen (gender of head of household)	−0.001***	0.002***	0.001***	0.003***	−0.005***
	(0.000)	(0.000)	(0.000)	(0.001)	(0.001)
hsize (household size)	−0.002***	0.001***	0.001***	−0.002***	0.001***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Middle income group	0.001**	−0.001**	0.001	−0.001*	0.001**
	(0.000)	(0.000)	(0.000)	(0.001)	(0.001)
High income group	0.000	−0.005***	0.000	−0.001	0.005***
	(0.001)	(0.001)	(0.001)	(0.002)	(0.001)
ivexp (instrument variable)	0.003***	−0.002***	−0.005***	−0.005***	0.009***
	(0.000)	(0.001)	(0.001)	(0.001)	(0.001)
Mills group 1	0.001	0.024	0.031**	0.009	−0.065***
	(0.011)	(0.015)	(0.014)	(0.025)	(0.018)
Mills group 2	0.040***	0.015	0.031***	−0.055***	−0.031**
	(0.009)	(0.011)	(0.011)	(0.020)	(0.014)
Mills group 3	0.011	0.037***	0.037***	−0.065***	−0.020
	(0.010)	(0.011)	(0.012)	(0.019)	(0.019)
Mills group 4	0.017***	−0.033***	−0.036***	0.104***	−0.051***
	(0.006)	(0.010)	(0.009)	(0.018)	(0.011)
Mills group 5	−0.026**	0.004	−0.019	−0.098***	0.139***
	(0.011)	(0.011)	(0.012)	(0.019)	(0.020)
constanta	−0.060*	0.227***	−0.251***	0.822***	0.261***
	(0.032)	(0.041)	(0.039)	(0.068)	(0.053)

The price variable itself has a positive and significant effect on each food commodity group’s expenditure share. For instance, group 1’s cereals and tubers have a coefficient of 0.131, meaning that, under the assumption of ceteris paribus, a 1% price increase in the commodity group of cereals and tubers will result in a 0.131% increase in the group’s share.

The price coefficient shows that, if there is a 1% increase in price, the highest increase in the proportion of expenditure (share) is the group of beverages, cigarettes, and other consumption materials 0.180%, followed by the group of cereals and tubers 0.131%, the group of food and processed drinks 0.027%, the group of fish, meat, eggs and milk 0.010%, and the lowest is the group of vegetables, fruits, nuts, seasonings, fats and oils which is 0.006% with the assumption of ceteris paribus. From the description of own price and all food prices, it can be concluded that price has significance on food demand.

The expenditure share is positively impacted by the income variable. In other words, an increase in revenue will be followed by an increase in the percentage of expenses. This illustrates that consumption patterns are still dominantly quantity-oriented, or it can be said that consumers do not feel sufficient so they need to increase consumption if income increases. Of the five commodity groups, all income variables are significant at the 1 per cent real level. The coefficient for fish, meat, eggs and milk commodity groups has a positive effect, meaning that if income increases, the share of the group will also increase.

This is not the case, however, for the commodity groups of cereals and tubers, which have negative coefficients. This means that if income rises, households will likely shift their demand toward more expensive and high-quality foods, like processed foods, which will lower the proportion of expenditure for these two groups.

Seventy-two percent of the coefficients exhibit significance in the Inverse Mills Ratio (IMR). This suggests that the issue of sample selectivity bias is present in these coefficients and that the estimated parameters in the equation become unbiased when IMR factors are added. Additionally, the IMR variable’s negligible impact on 28% of the coefficients suggests that selectivity bias is not an issue for this commodity group.

Income Elasticity

The background factor influencing food demand is income. Demand and income generally have a positive relationship; the higher the household income, the higher the demand. The response of changes in demand brought on by changes in income is ascertained using the income elasticity value. Income elasticity will explain the nature of the commodity group whether it is inferior goods, normal goods, or luxury goods.

Based on the results of this study, the income elasticity of household food demand in South Kalimantan Province as a whole shows positive results (greater than zero). This means that all commodity groups fall into the category of normal goods (Table IV).

Table IV. Income Elasticity

Commodity group	Income elasticity
Cereals and tubers	0.009
Fish, meat, eggs, milk, and their derivatives	0.009
Vegetables, legumes, fruits, oil, coconut, beverage dan spices	0.008
Prepared food and beverages	0.009
Miscellaneous food items	0.016

The processed food commodity group has the largest income elasticity value of any group, and it is positive. This suggests that, given ceteris paribus, any rise in household income will result in a higher allocation of expenditure on this commodity. On the other hand, the processed food group’s high elasticity value may be due to its comparatively higher price when compared to other food groups, which makes it the primary food choice as income rises.

This result is in accordance with food demand research in Tanah Laut Regency with LA AIDS, the results of income elasticity show that eggs, milk and its derivatives, fruit vegetables, seasonings, nuts, fats and oils, fish and meat, and processed food and beverages are luxury goods. An increase in the income of farming households will have a large effect on the increase in consumption of these four food groups [11].

Price Elasticity

The amount of own-price elasticity for food items is displayed in Table IV both overall and by socioeconomic characteristics, including income group, household head’s educational attainment, and region. All of them exhibit negative elasticity magnitudes, with the exception of the group of cereals and tubers, suggesting that, ceteris paribus, a rise in commodity prices results in a decrease in demand for commodity consumption. This is in line with the law of demand’s negative trend [12]. Furthermore, research shows that price increases have little impact on the demand for cereals and tubers because they are a staple food that is always consumed (Table V).

Table V. Price Elasticity

Comodity group	Price elasticity
	1	2	3	4	5
Cereals and tubers	0.283	−0.771	0.156	−0.502	0.242
Fish, meat, eggs, milk, and their derivatives	−0.518	−0.827	0.070	0.147	0.111
Vegetables, legumes, fruits, oil, coconut, beverage dan spices	0.081	0.126	−1.074	0.414	−0.342
Prepared food and beverages	−0.276	0.060	0.167	−1.043	−0.102
Miscellaneous food items	0.129	0.125	−0.495	−0.203	−0.720

Conversely, it is deemed inelastic if its own price elasticity is less than one. Accordingly, a 1% increase in commodity prices, ceteris paribus, causes the quantity of commodities in the group of cereals and tubers to decline by 0.283%, the group of fish, meat, eggs, and milk to decrease by 0.827%, and the group of beverages to decrease by 0.720%.

The group of cereals and tubers has the lowest own price elasticity score (0.283), making it the least elastic of the five commodity categories. This issue occurs because rice is the main food source that South Kalimantan Province residents use to achieve their nutritional needs. in order to reduce the likelihood that households will respond to price rises. The category with the highest own-price elasticity is that of relatively more costly fruits, vegetables, nuts, seasonings, fats, and oils. The commodity group comprising vegetables, fruits, nuts, seasonings, fats, and oils has a price elasticity of 1.074, assuming ceteris paribus. This means that the decrease in family food consumption is almost equal to the increase in household food prices.

Urban households had higher own-price elasticity values than rural ones for cereals and tubers, fish, meat, eggs, and milk, and for vegetables, fruits, nuts, seasonings, fats, and oils. This suggests that urban inhabitants are more sensitive to fluctuations in commodity prices than their rural counterparts. The food and beverage commodities group’s price rises will be more swiftly absorbed by rural areas.

Price increases for a product can affect not just the commodity itself but also shifts in demand for other commodities. Cross-price elasticity illustrates how changes in the prices of other commodity groups affect the demand for the commodity group. Cross-price elasticity will establish whether a commodity or collection of goods is complementary or substitutable.

Negative cross-price elasticities show that linked food commodities are typically complimentary, while the majority of commodity groups have very tiny elasticity values near 0 or below 1, indicating that food commodities are inelastic. Because their absolute values are less than one, food commodities can complement or replace one another, but their relationships are still weaker.

Commodity groupings can be substituted for one another, as in the case of cereals and tubers and vegetables, fruits, nuts, spices, fats, and oils. Assuming ceteris paribus, the demand for fruits, vegetables, nuts, seasonings, fats, and oils will rise by 0.081% for every 1% increase in the price of cereals and tubers.

In the meantime, fish, meat, eggs, and milk are complementary to other commodities like cereals and tubers; therefore, if the price of cereals and tubers rises, the demand for these commodities will fall.

Conclusions and Suggestions

Conclusion

Based on the results of the analysis and discussion described earlier, it can be concluded as follows:

1. The consumption pattern of South Kalimantan Province’s households are influenced by socio-economic. Income level affects food consumption except for the consumption of cereals and tubers. Area of residence affects food consumption except for the consumption of processed food and beverages. Education of the household head affects food consumption except for the consumption of vegetables, nuts, fruits, seasonings, fats and oils. Household size and gender of household head affect food consumption in all food groups. The food budget share in South Kalimantan is highest in the commodity group of processed food and beverages, which reaches 32 per cent.

2. Income elasticity as a whole shows positive results. This means that all commodity groups fall into the category of normal goods. The income elasticity value for the processed food commodity group is the highest among other groups and has a positive sign which means that any increase in household income will cause the allocation of expenditure on this commodity to increase.

3. Price elasticity alone demonstrates that all of them are negative, with the exception of the cereals and tubers group. This indicates that, with the exception of the cereals group, which is South Kalimantan’s primary food source, rising commodity prices lead to falling demand for commodity consumption. In contrast, negative cross price elasticity indicates that related food commodities are often complimentary, and the cross elasticity of all food commodity groups displays a value between 0 and 1, showing that food commodities are inelastic. There is a substitution between the groupings of vegetables, fruits, nuts, seasonings, fats, and oils and cereals and tubers. Complementary relationships occur in the group of cereals and tubers are complementary to the group of fish, meat, eggs and milk.

Suggestions

Several things can be suggested from the findings of this study, namely:

1. People’s consumption patterns depend on their socioeconomic circumstances, including education. There is a need for adequate education so that people can choose quality consumption.

2. It is necessary to maintain price stability at the consumer level, this can be done for example by market operations, monitoring price distribution and providing subsidies if there is a significant price increase.

3. The consumption pattern of the people of South Kalimantan, which is dominated by the consumption of food and beverages, can be an opportunity for the government to attract investors in the culinary field so as to be able to open wider employment opportunities.