key: cord-0972625-wtj0dk9c
authors: Kh’ng, Kerry; Chang, Ching-Cheng; Hsu, Shih-Hsun
title: Implications of food and nutrition security on household food expenditure: the case of Malaysia
date: 2022-04-07
journal: Agric Food Secur
DOI: 10.1186/s40066-022-00367-4
sha: 13590b09e6c719333e8d41f3a708a7ed9e0d8e54
doc_id: 972625
cord_uid: wtj0dk9c

INTRODUCTION: Food security is attracting more attention in Malaysia not only at the national level that concern toward the enhancement of food self-sufficiency but also at the individual level which concerns more on nutrition and health. The economic recession triggered by the COVID-19 pandemic has brought the food and nutrition security challenge to the higher priority. In this study, we assessed the feasibility of encouraging a healthy eating plan by taking into account two important elements, food cost and nutrient intake, to help tackle the food and nutrition insecurity challenges at the individual level. METHOD AND MATERIALS: This study used a goal programming model with dietary intake data from Malaysian Adult Nutrition Survey reports to develop food plans that can improve nutrition quality without increasing food cost. Missing data, such as nutrient compositions and food prices, were collected separately from existing governmental and non-governmental sources. Benchmark nutrient intakes were derived from Malaysian Dietary Guidelines and Malaysian Recommended Nutrient Intakes reports, whereas benchmark costs were estimated by mapping food prices to dietary intake. The cost of healthier diets was also assessed to examine the acceptability of dietary changes for the low-income population. RESULTS: The results showed that healthier diets following national dietary guidelines are achievable with reasonable food choices shift without changing the cost of meal plan. Greater intake of milk and vegetables (for more calcium) and smaller intake of seafood and egg products (for less protein) will contribute to raise diet quality and achieve more adequate nutrition. However, the cost attached to healthier food plan is still likely to be burdensome for the food-insecure and low-income population. CONCLUSIONS: Our results suggest that policymakers should implement income-relevant laws to cut poverty and improve the population’s dietary intake. Income growth as a result of better skills and education is needed to ensure that the real incomes of Malaysian are well sustained, and increased to help low-income population make better and healthier food choices.

The definition of food security was first introduced in the 1970s from the perspective of food supply to ensure that all people everywhere have sufficient food to eat [1] . This definition was expanded in 1996 to incorporate nutrition and cultural dimensions [2] . Unlike food security,

Agriculture & Food Security *Correspondence: emily@econ.sinica.edu.tw 1 Department of Agricultural Economics, National Taiwan University, Taipei, Taiwan Full list of author information is available at the end of the article nutrition security was interpreted from the viewpoint of food demand as "a person is considered nutrition secure when she or he has a nutritionally adequate diet and the food consumed is biologically utilized such that adequate performance is maintained in growth, resisting or recovering from disease, pregnancy, lactation and physical work" [3] . With the widespread recognition of the importance to include nutritional aspects into food security by international organizations such as FAO, UNICEF, and IFPRI, analysis has also expanded to consider both global and individual levels, as well as multiple forms of malnutrition [4] . Following the 2007-08 food price spikes, growing attention to the food self-sufficiency and greater awareness of the public health implications of malnutrition has in turn influenced food security policy in important ways [5] .

For decades, food insecurity has always been correlated with poverty and health issues [6] [7] [8] [9] . Although the incidence of poverty in Malaysia had significantly reduced [10, 11] , it does not mean that Malaysian households are free from food insecurity situation. Many demographic indicators, such as sex, age, source of income, household type, homeownership, marital status, immigrant status, and aboriginal status, can be relevant to household food insecurity. Among them, household income was found to be a factor that driving significant impact to food insecurity [12] . Researchers also found that 13.4% of Malaysian adults tend to reduce the size of meals and skip main meals due to financial constraints [13] . Intervention measures, such as the National Plan of Action for Nutrition (NPAN), has been launched by the Malaysian government to ensure food security and nutrition security for all households and to prevent diet-related noncommunicable diseases [14] . Strategies and activities have also been identified for implementation to ensure the availability of quality and safe food to all households at affordable prices. However, meeting healthy diet recommendations may involve substantial adjustment in dietary pattern which will also influence food purchase cost. In fact, about 17.3% of Malaysian household income are devoted to food consumption expenditures [15] . It accounts as the second highest percentage to overall basic necessities' expenditure but as the highest for families in the rural areas. The high percentage of income committed to food consumption may cause low-income families especially vulnerable to food insecurity, poor nutrition, and leading to various health issues.

In March 2019, the Employees Provident Fund (EPF) of Malaysia released an expenditure guide for Malaysian individuals and families. The guideline indicated that an individual who is single and a public transport user required a budget of RM 1870 per month for his or her living [16] . The suggested budget on food was RM 550, equivalent to 29% of the overall budget and 50% of the minimum wage. The imbalance ratio of food budget to the minimum wage has stimulated our curiosity, for which we find that it is necessary to provide a better understanding of how healthy eating habits will impact the low-income group especially those who had been attached or is still attaching to the minimum wage. In addition, the recent COVID-19 outbreak has also brought the income issue to a higher level of concern, unemployment from either job losses or reduction in working hours has significantly affected the livelihoods of Malaysians [17] , and hence consolidated our motive of investigation.

This study aims to evaluate the feasibility of encouraging a healthy eating habit that follows national recommended dietary guidelines to the low-income group in Malaysia through modeling the realistic diets by using a food-based goal programming optimization model approach. It is an approach extended from linear programming and was popularly adopted to develop economically feasible food plans while promoting healthy dietary patterns simultaneously [18] [19] [20] . It was also empirically proven as an efficient and effective method to solve dietary problems [21] . To the best of our knowledge, such application can be barely found in the case studies of Malaysia and researches that gauge the balance in between food and nutrition security at the individual or household levels are also limited in Malaysia. Hence, this study serves as a contribution to identify the diet plans at individual level by taking into considerations of the important linkage of economic accessibility, food availability, and food utilization to help realize local government initiatives as well as global initiatives for tackling food insecurity challenge for the low-income population.

This study adopted the goal programming approach using dietary intake data from Malaysian Adult Nutrition Survey (MANS) of the year 2003 and 2014 to design food-based dietary recommendations for Malaysia. Goal programming is a tool for solving multiple-goal problems with an objective function to minimize the sum of absolute values of deviations from various goals [22] . The approach was applied [21, 23, 24] in formulating the optimal food plan which aims to improve nutritional intakes via more prudent food group and subgroups choices under cultural, habitual dietary patterns, and economic cost considerations.

The general structure of the model is as follows:

where subscript k denotes m food subgroups, i the n nutrients, and j the seven major food groups; Y denotes the objective variable to be minimized; X opt k : denotes the optimal daily intake quantity of food subgroup k; X obs k : denotes the observed daily intake quantity of subgroup k; a ik : denotes the amount of nutrient i in unit of each subgroup k; N The objective function aims to minimize the gap between the quantity of optimal food intake and the observed quantity of intake by the study population. To standardize the gap across different food groups, it is divided by the observed quantity consumed.

The objective function can be transformed into a linear function [21] with two sets of non-negative decision variables representing, respectively, the positive deviation (D + k ) and negative deviation (D − k ) from the observed food intake, as follows:

The linearized objective function (Y*) is then defined as follows:

Three sets of constraints are included. First, Eq. (2) ensures that the daily total nutrient intake meets the desired level for Malaysian adult population. Second, Eq. (3) ensures that the serving amount of each major food group is within the range from the 5th percentile (as a lower limit) to the 95th percentile (as an upper limit) of

Minimize Y * = k=n k=1 D + k + D − k , k = 1, . . . , n, and

daily consumption. Third, Eq. (4) ensures that daily per capita food cost is within the boundary of observed levels to prevent food plan from incurring unreasonable cost.

In essence, the model is seeking an intervention that can encourage people to make better food plans with least possible changes in dietary habit and cost. The finding of the optimal food plan may not satisfy all the nutrient requirements. Nevertheless, more realistic, affordable, and healthier recommendations are made to the target population for nutrition promotion purpose.

The complete list of data sources for this study is provided at Appendix 1. The main data source on food intake was extracted from the published reports of two Malaysian Adult Nutrition Surveys in 2003 (MANS 2003) [25] and in 2014 (MANS 2014) [26] . MANS 2003 was a nationwide survey with a total of 6886 households in Peninsular Malaysia, Sabah, and Sarawak. It contains a total of 126 food items. MANS 2014 was the second survey with a total of 2973 households and 165 food items. Both surveys conducted with a stratified random sampling method with proportional allocation. Specifically, data retrieved were the estimated mean per capita food intake (in gram) by food items per day of the surveyed households.

Several steps were taken to convert the food intake data into nutrient forms before they are incorporated into the goal programming model. First, the mean food intake of 126 and 165 food items in MANS 2003 and MANS 2014 were, respectively, grouped into 77 and 76 subgroups at per capita basis following Malaysian Dietary Guidelines (MyDG) [27] . Second, the nutrient compositions of the mean food intake by 77 and 76 subgroups were estimated using Malaysian Food Composition Database (MyFCD) [28] . Other sources, such as Food and Nutrient Database from Taiwan [29] , Nutritionix website, and many more, were also used for nutrient estimations for subgroups that were not available in MyFCD. Third, because the food intake constraints in the model were specified in serving units by 7 major food groups, the food intake data were further converted into food groups based on one single nutrient indicator of each food group [30] . The mapping of 77 and 76 subgroups with 7 food groups is shown in Appendix 2.

For instance, recommended intake for cereals and cereal products was based on 30 g of carbohydrates per serving per person. Hence, the mean servings for bijirin (cereals) subgroup are equal to its intake volume in gram divided by 30 g of carbohydrates and by summing up all the relevant food subgroup servings to form the food group serving. The overall conversion process is illustrated in Fig. 1 .

The constraints of the goal programming model require the lower and upper limit of food intake in serving units and nutrient forms. For the serving units, data from Malaysian Dietary Guidelines (MyDG) of seven major food groups were used as the limits and they are listed in Table 1 . Note that the lower limit is based on 2000 kcal per day instead of the original 1500 kcal available in MyDG taking into account the standard benchmark of 2100 kcal per day for adults as suggested by the United States Department of Agriculture [31] . Food group A was based on 30 g of carbohydrates per serving; B and C were based on 14 g of protein per serving; D and E were based on 7 g of protein per serving; F was based on 80 g per serving size [32] ; G was based on 15 g of carbohydrates per serving 

For the constraints in nutrient forms, the upper and lower limits are listed in Table 2 . They are averaged from the units derived from 2017 Recommended Nutrient Intakes (RNI) report of Malaysia [33] . Note that iron nutrient is excluded due to the difficulty in conversion to a single representation of total population.

The food intake data of MANS did not include cost data. Therefore, the food cost data by subgroups are collected separately from various public and private online sources over three-year period of 2016, 2018, and 2019. For consistency purpose, the cost data are traced back using the Consumer Price Index (CPI) of Malaysia with the latest available figures derived from the Department of Statistics Malaysia (DOSM).

To match with the MANS food intake data, all food cost data are converted from kilograms to grams, whereas for liquidized food items they are converted from milliliters to grams. After that, all cost data are averaged and converted to Ringgit Malaysia (RM) per gram, eventually multiplied by the daily mean intake from MANS to get the total cost estimates. The food cost conversion steps are illustrated in Fig. 2 , whereas the estimated upper and lower limits for food cost are listed in Table 3 . 

The optimized results on per capita and daily basis are shown in Table 4 for MANS 2003 and Table 5 for MANS 2014. Overall, the serving and nutrient goals that were not achieved in the observed dietary intake are met in the optimized diets. The most noticeable result from nutrition perspective is the lack of consumption in calcium nutrient. Malaysian adults are found to have consumed less calcium than the minimum recommendations in both observed years. This is consistent to the lesser quantity intake for milk and vegetables food groups listed in Table 6 . Hence, the intake of milk and vegetable products are suggested to increase. Previous research [36] found that income, gender, health condition, and location of residence are highly correlated to low intake of vegetable products. Among them, low income is an expected factor due to consumers are likely to allocate their limited budget to more essential food (such as rice) and nonfood items (such as clothing and housing). Also, domestic milk production is slow in its growth in Malaysia and thus of high reliance on imported products. Milk products are not exactly available or affordable by all, which could trigger its low intake [37] . In addition, some studies also found that Malaysians could have high prevalence rate of lactose intolerance [38, 39] and they are advised to consider soybeans and vegetables as alternatives to supplement the intake of calcium. Next, although Malaysian adults consumed lots of cereal-based foods, their consumptions are still within the recommended range. This is consistent with the findings from MANS reports, in which rice is the top food consumed by the surveyed groups [40, 41] . It is likely attributable to social and cultural norms, whereby rice can be transformed into different types of cuisine and is exceptionally easy to blend into every meal that On the other hand, seafood and eggs that contribute to major source of protein in the observed diet are encouraged to reduce to an appropriate level as suggested in the optimized diets. The higher intake of seafood and eggs may be due to several reasons, not only the prices but also the conveniences of readily available, easily reachable, and generally acceptable by all ethnicities. Unlike poultry products, pork and beef are prohibited for Muslims and Hindus, in addition to some Chinese that adopt some Buddhism beliefs and do not consume beef. Nevertheless, a meal plan that is inclined to a specific selection may not conform to a balanced diet that is beneficial to health. Thus, a cut in consumption of this food group is suggested.

To be in line with the concept of affordability, we define a target diet that is able to achieve food group servings recommended by MyDG and nutrient content recommended by RNI with reasonable costs. The target diet emerged in which food choices shift is required to commensurate the definition from food and nutrient perspectives. Based on the findings of this study which had referenced the clinical suggestions from guidelines released by the governmental portals, calcium was the nutrient found to be consistently less consumed. For improvement, an increase in the consumption of milk and vegetable products is considered adequate to meet most people's nutrient need, whereas the consumption of seafood and egg products is suggested to reduce. Nevertheless, it is noted that the consumption of these food items must be based on individual's health condition and activity levels. Maintaining balanced dietary intake and avoid falling into chronic disease predicaments are the ultimate goals to achieve. Although there could be many factors attached to the chosen diet plan, but from health maintaining perspective, a consistent one is still the income and food cost that are affordable.

The Minimum Wage Order of Malaysia was established in the year 2012. Before the commencement of this law, Poverty Line Income (PLI) was adopted as a measure As aforementioned, we defined a target diet as "a diet that is most closely represent a realistic diet as be able to achieve the recommended servings by MyDG and nutrient recommendations by RNI with reasonable cost. " Pertaining to this definition, it is achievable from nutrient perspective with reasonable food choices shift. However, cost on food is rising over time due to inflation. Although the ratio of food expenditure to income is showing a decreasing trend, it is still higher than the figure proposed in the expenditure guideline [16] released by the Employees Provident Fund of Malaysia in 2019. The guideline indicates that a total budget of RM 1870 per month is required for the living of an individual who is single and a public transport user, whereby RM 550 should be allocated to food. The recommended food budget is equivalent to only 29% of the total budget, yet it occupies 50% of the minimum wage, and with our simulated optimized cost, there is still a gap of 14% (43%-29%) even though the optimized cost had been controlled at the observed cost level. Hence, if one's earning is near or at the minimum wage level, the higher food cost attached to a healthier diet following the nutritional guidelines proposed by the government could raise public concerns on the affordability of healthy diet for low-income group.

In fact, according to the statistics from the Central Bank of Malaysia, the starting salaries for workers with no prior working experience in the non-executive level were all just close to the minimum wage [46] . It was reported that Malaysian's salaries were found to have misaligned with their productivity levels. With the same level of output produced, Malaysians received lower pay than the workers in the benchmark economies. Researchers [47] also found that Malaysians tended to have prevalence of food insecurity. There was a total of 6 parameters used to measure the prevalence, and 4 of them are believed to be highly correlated to the income level: (1) could not afford to feed children with various food (20.8% prevalence), (2) only rely on cheap food and affordable food to feed children (23.7% prevalence), (3) skipped main meal (15.2% prevalence), and (4) reduced size of meal (21.9% prevalence).

Since COVID-19 pandemic started to outbreak in Malaysia, the gradually revealing income matter could have worsened, the unemployment rate had increased drastically within a short period of time from about 500,000 people to more than 700,000 people unemployed in just 4-month period from January to April 2020 [48] . There were also more than 30,000 businesses closed down as of November 2020 [49] . With the extension of Movement Control Order and lockdown measures enforced by the government to bring down the infectious cases, many Malaysians were suffered from job lost and hence income shrinkage. The tough battle with the disease is going on in 2022 and is expected to continue [50, 51] . While there are proposals [52, 53] to increase the minimum wage level to RM 1,500 per month, the estimated optimized cost per month is equally going to increase as the core inflation rate has set to rise. Hence, wages-relevant laws, such as the Minimum Wage Order, shall position into a more careful assessment that is always up to the par of inflation rate as rising food cost amid the post-pandemic economic recovery could aggravate food insecurity in low-income households.

We demonstrate that improving dietary quality is possible without increasing the existing cost by using goal programming models. Although acceptability is achievable with food choices shift based on each individual consumption preference, the problem of affordability still remains challenging for low-income households. Income is an indispensable element when we strive to make improvement toward food and nutrition security [54] [55] [56] . Research [47] found that education also plays an important role in combating food insecurity. We believed that not only education on nutrition is required, but the most closely linked element to income generation is also education which could aid lowincome households to stay out of the poverty cycle. Hence, policies aiming at income growth are likely to benefit the low-income groups in terms of nutrient availability and the quality of their diet [57, 58] . However, we believe our approach could provide valuable information for food and nutrition security program planners in the disadvantaged environments by identifying key problem nutrients and related food costs in the available local diet. There are three limitations of this study that we would like to highlight. First, the data from MANS have the prevalence of underreporting as admitted in the original MANS 2014 report. Hence, the actual nutrient intakes are believed to be higher than reported, which means the food cost in this study could be underestimated. Second, food intake constraints are specified in food groups instead of food subgroups because the referenced data from MyDG is in food groups form. Hence, the constraints are believed to be more stringent. Third, the cost data were estimated based on a mixture of raw and processed food prices without further breaking into condiment granularities consideration. For food, like roti canai, the collected data were purely based on available price that could be found online, such as in a forum or travel tips website. Besides that, some prices were mapped by food type similarity, for example, rice and rice porridge were collected from the same source. Hence, the estimated costs are believed to be lower than the actual ones.

Besides the aforementioned limitations, more knowledge is needed on how nutrition information on food products influences consumer choices. In model formulation, consumers' sensitivity to health information from different food products should also be further investigated. Future research can be designed to assess the impact of alternative promotion strategies to achieve more balanced diets for a more prosperous population in Malaysia. [44] , and 2020 [45] , effective 1st Jan, 2013 and 2019, and 1st Feb, 2020, respectively

The optimized food cost in 2019 and 2021 is estimated by assuming that the optimal servings in both years are equivalent to the optimized servings in 2013 and food cost inflation rates are equivalent to 18 

The Four Dimensions of Food and Nutrition Security: Definitions and Concepts. Food and Agriculture Organization of the United Nations (FAO)

The state of food insecurity in the world

A definition of nutrition security. CARE Mimeo

Food and nutrition security indicators: a review

Global Food Price Volatility and Spikes: An Overview of Costs, Causes, and Solutions

Poverty, obesity, and malnutrition: an international perspective recognizing the paradox

Poverty, malnutrition, underdevelopment and cardiovascular disease: a South African perspective

Food security and nutrition in the context of the global nutrition transition. Rome: Food and Agriculture Organization of the United Nations

Urban agriculture, poverty, and food security: empirical evidence from a sample of developing countries

The role of government in poverty reduction. Proceeding of the National Seminar on Poverty Eradication through Empowerment

php?r= column/ cthem eByCa t& cat= 120& bul_ id= TU00T mRhQ1 N5TUx HVWN0 T2Vjb XJYZz 09& menu_ id= amVoW U54UT l0a21 NWmdh MjFMM WcyZz 09

Food insecurity in Canadian households

Household food insecurity in Malaysia: findings from Malaysian Adults Nutrition Survey

Changing trends in dietary pattern and implications to food and nutrition security in Association of Southeast Asian Nations (ASEAN)

php?r= column/ cthem eByCa t& cat= 323& bul_ id= c3JpR zRqeT NPamM xL1Fp TkNBN UVBQT 09& menu_ id= amVoW U54UT l0a21 NWmdh MjFMM WcyZz 09

Expenditure Guide for Malaysian Individuals and Families

Covid-19 pandemic and unemployment in Malaysia: a case study from Sabah

Food-based dietary guidelines can be developed and tested using linear programming analysis

Diet optimization methods can help translate dietary guidelines into a cancer prevention food plan

convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year submit your research ?

Designing optimal food intake patterns to achieve nutritional goals for Japanese adults through the use of linear programming optimization models

Diet planning in the third world by linear and goal programming

Goal programming for decision analysis

Design of optimal food-based complementary feeding recommendations and identification of key ' "problem nutrients"' using goal programming

To meet nutrient recommendations, most French adults need to expand their habitual food repertoire

Nutrition Section, Family Health Development Division. Malaysian Adult Nutrition Survey

The National Health and Morbidity Survey 2014: Malaysian Adult Nutrition Survey. Kuala Lumpur: Ministry of Health

National Coordinating Committee on Food and Nutrition (NCCFN)

Malaysian Food Composition Database (MyFCD) 1997

Food and Drug Administration (FDA)

Needs-based food and nutrient security indices to monitor and modify the food supply and intakes

GFA-28. United States Department of Agriculture

Food and Agriculture Organization (FAO)

Recommended Nutrient Intakes (RNI) for Malaysia. Putrajaya: Ministry of Health

An introduction to chemistry-atoms first

FAO/INFOODS Databases: Density Database Version 2.0

Who are eating and not eating fruits and vegetables in Malaysia?

Food Value Chain in ASEAN: Case Studies Focusing on Local Producers

Hypolactasia & lactose intolerance among three ethnic groups in Malaysia

Lactose tolerance and intolerance in Malaysians

Food consumption patterns: findings from the Malaysian adult nutrition survey(MANS)

Food consumption patterns: findings from the Malaysian adults nutrition survey (MANS) 2014

Putrajaya: office of the prime Minister of Malaysia; 2003. 43. Minister of Human Resources. Minimum Wages Order 2012. Attorney General's Chambers of Malaysia

Minister of Human Resources. Minimum Wages Order (Amendment) 2018. Attorney General's Chambers of Malaysia

Minister of Human Resources. Minimum Wages Order 2020. Attorney General' s Chambers of Malaysia

Are Malaysian Workers Paid Fairly?: An Assessment of Productivity and Equity)

Food insecurity situation in Malaysia: findings from malaysian adult nutrition survey (MANS) 2014

my/ v1/ index. php?r= column/ cthem eByCa t& cat= 124& bul_ id= RDllT kpHej RFNGl RdlRL WWJzM i91QT 09& menu_ id= Tm8zc nRjdV RNWWl pWjRl bmtla Dk1UT 09. Accessed

The Straits Times. 30,000 Malaysian businesses have folded up since movement curbs in March

COVID-19 in 2022: controlling the pandemic is within our grasp

Pandemic or endemic: where is COVID heading next?

Minimum wage report expected in 2022

Cola to RM300

Systems, food security and human health

Public policies for improving food and nutrition security at different scales. Food Security

Social grants, remittances, and food security: does the source of income matter? Food Security

Health and hunger: nutrient response to income depending on caloric availability in Nepal

Does a nutritious diet cost more in food deserts?

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

The authors gratefully acknowledge the financial support from Academia Sinica (AS-107-SS-A02) and the Ministry of Science and Technology (MOST 105-2923-H-002-002-MY3, MOST 109-2621-M-001-007) of Taiwan.

"-" denotes food subgroup was excluded from modeling due to low intake, that is, intake value was zeroed after rounded off to two decimal places Authors' contributions KK contributed to research proposal writing, data collection, data analysis, data interpretation, article writing, and supervision. CCC contributed to data interpretation, article writing, and supervision. SHH contributed to data interpretation and supervision. All the authors read and approved the final manuscript. 

The datasets used and/or analyzed during the current study are available from the corresponding author on request.

Ethics approval and consent to participate Not applicable.

Not applicable.

The authors declare that we do not have competing interests or personal relationships that could have influenced the work reported in this paper.Author details 1 Department of Agricultural Economics, National Taiwan University, Taipei, Taiwan. 2 Institute of Economics, Academia Sinica, Taipei, Taiwan.Received: 2 August 2021 Accepted: 3 March 2022