Cost of the Diet – novel approach to estimate affordability of a nutritious diet
A woman cooking at home in Tanzania
By Abigail Perry, Save the Children UK
Abigail Perry currently works for Save the Children UK as a nutrition adviser based in London, providing technical support for Asia and Africa programmes. Previously she has worked on a range of nutrition programmes and research projects in Darfur, South Sudan, Pakistan, Zambia and the UK.
The author would like to acknowledge the invaluable contributions from Andre Briend, who provided extensive advice and support, and Remy Stevens who gave a significant amount of his time pro bono to develop the programme software. The Cost of the Diet project would not have been possible without Arabella Duffield, Claire Chastre, Heather Kindness, Sonya LeJeune, Anna Taylor and the Save the Children teams in Bangladesh, Tanzania, Ethiopia and Myanmar.
This article provides an overview of a novel method developed by Save the Children UK to calculate the cost of an ideal diet, with results from two case studies and a discussion of the limitations to the approach and intended next steps for its development.
Save the Children UK (SC UK) has developed a method known as the 'Cost of the Diet' (CoD) that can calculate the minimum amount of money a family has to spend to meet their macro- and micronutrient requirements using locally available foods. This approach came out of SC UK research, which showed that nutrition education programmes that aim to improve complementary feeding practices in children have had a limited impact because of the economic constraints facing many households in developing countries1. The initial objective for developing this method was to better illustrate the extent of the gap between household income and expenditure, in order to advocate for more appropriate programmes and strategies to reduce child under-nutrition.
Defining the role of cash transfers
Over the last several years, there have been a number of large-scale cash transfer programmes in both Latin America and Africa that have shown positive impacts on feeding practices and nutritional status of children. The donor community has also expressed an interest in wider use of social protection, particularly as a way to achieve Millennium Development Goal (MDG) 12. The recent Lancet series on Maternal and Child Undernutrition emphasised that improvement of complementary feeding among food-insecure populations is best achieved by combining nutrition counselling, food supplements and cash transfers as part of a social protection package3.
A child eating at home
However, there is a distinct lack of tools to help decide what components should go into a social protection package and how large a cash transfer would need to be, to achieve specific nutritional objectives. To make informed decisions when designing a nutrition- focused social protection package, we need to know what foods are available and how much it would actually cost to buy the foods a family needs to meet requirements. The CoD method was developed to do exactly this. As far as SC UK is aware, this is the first tool that can:
- Calculate the minimum cost of a diet for an individual child and the whole family.
- Take into account seasonal variations in price and availability.
- Provide region-specific data on cost and availability.
A number of CoD assessments have been carried out so far and results from two of the pilot studies in Bangladesh and Tanzania are presented here.
The 'Cost of the Diet' programme
The main component of the CoD is a computer programme designed by SC UK that uses linear programming (LP) to calculate lowest cost nutritionally-appropriate diets. The principle of LP is that one can solve a problem (in this case the minimum cost of a diet) whilst fulfilling a range of constraints (e.g. nutrient requirements). The programme builds on work done by the World Health Organisation (WHO), which used LP to put together diets to meet nutritional requirements of children under 2 years of age (http://www.nutrisurvey.de/lp/lp.htm).
The lowest-cost diet is calculated using locally available foods according to the following constraints:
- It must meet (but not exceed) energy requirements of each individual family member.
- It must meet protein, fat and micronutrient requirements for each individual family member4.
- It must not include more than the predetermined allowance for particular food groups.
- For children aged 6-23 months, it must include a fixed amount of breastmilk.
Energy, macro- and micronutrient requirements for individuals are based on WHO recommendations and the nutritional composition of foods is derived from a Food and Agricultural Organisation (FAO) database built into the programme. Maximum allowances for particular food groups were agreed through consultation with experts at University College London, WHO and the University of California, Davis (Table 1). These thresholds have been incorporated to help ensure that the diets are relatively realistic. However, they are not internationally agreed and further work is needed to see whether they can be standardised across countries. The volume of breastmilk that has to be included in the diet for children aged 6-23 months is based on average intakes and is age specific (6-8 months, 674ml; 9-11 months, 616ml; 12-23 months, 549ml)5.
Although the size and composition of a family should be based on what is typical for the region of interest, in this case a standard household was used6.
|Table 1: Maximum percentages of energy requirement for food groups|
|Food group||Maximum percentage of energy requirement that can come from this group|
|Table 2: Country seasons|
|Country||Season 1||Season 2||Season 3||Season 4|
|Mid-November to mid- March||Mid-March to mid-June||Mid-June to mid- September||Mid-September to mid- November|
|Tanzaniab||Pre-harvest (high price)||Post-harvest (low price)||-||-|
|December to April||May to November||-||-|
Data collection done between:
aSeptember and December 2006 (covering the period from April 2005 to March 2006)
bMarch 2006 and July/August 2006 (covering the period from December 2005 to November 2006).
In both Bangladesh and Tanzania, CoD studies were implemented in regions that correspond to livelihood zones identified during Household Economy Approach (HEA) assessments7. Livelihood zones identified using HEA tend to have reasonable commonalities in terms of sources of food, income, expenditure patterns and access to markets.
The number and duration of the seasons in both locations were determined from HEA assessments and discussions with key informants from the community (Table 2). For each season, a comprehensive list of the foods available was compiled together with cost per unit sold. This was done through interviews with community members and local traders. Foods were weighed using Tanita electronic scales. Data were consolidated periodically throughout the process to ensure that information gaps and inconsistencies were identified and filled.
Data entry and analysis
For each location and season, the foods available and price per 100g were entered into the programme. When entering the data, teams were requested to select the equivalent food from the country nearest to the study location from the foods database. The programme was then run to calculate the lowest cost diet that would meet the requirements of the standard household using the foods available. The programme calculates how much it costs to meet daily requirements in a particular season. The average daily cost was calculated as follows:
Average daily cost of the diet = (?(Di x Ci)) /365
where i = season, 1, 2.n,
D = number of days,
C = daily cost of the diet.
For both examples, only the cost of a 'physiological' diet has been calculated; these are diets that meet requirements but that may not be culturally or environmentally feasible. The implications of this are discussed below.
Limitations of the method
The variability in the cost of the diet depended on the foods selected from the composition database, size/structure of the household and the energy requirement limits applied to the food groups. In order to estimate the potential impact of these limitations on the results, the data were re-analysed for the low price season in Tanzania with a number of adjustments, e.g. different household profile, different activity levels, different maximum energy from leafy vegetable sources. Seven diets were costed in this way with various adjustments. Costs ranged from 15% less to 25% more than the calculated average daily cost of the diet. This has been used to calculate the error ranges presented in the results.
To determine the affordability of a nutritionally adequate diet, costs calculated by the programme were compared with household income data from each study location. For Bangladesh, this was obtained from a survey of all households in a village in the livelihood zone8. Income data for Tanzania were taken from HEAs done in 2002/03 and 2004/059,10.
In both cases, income data were derived from a different time period to the food price data and hence income levels were adjusted according to inflation. For Bangladesh, 2004 income data were adjusted to 2006 prices using published inflation rates from the Central Bank of Bangladesh. Official Government inflation rates for Tanzania were used to adjust the total income estimates to 2006 prices. Diet costs were converted from local currency into US dollars (USD) using historical exchange rates (obtained from http://www.oanda.com).
The error range for the HEA income estimates in Tanzania were calculated at 5%. Ranges were not estimated for Bangladesh because individual households were assessed. In the case where the typical household size in a particular wealth group was not 5, income was adjusted to an equivalent 5-person household.
|Table 3: Composition of diets for Bangladesh (A) and Tanzania (B)|
|Bangladesh (A)||Total household of 5 people||Child aged 12 to 23 months|
|Winter (g)||Lean (g)||Summer (g)||Rainy (g)||Winter (g)||Lean (g)||Summer (g)||Rainy (g)|
|Meat/fish and animal products|
|Fresh water fish||0||296||0||287||0||0||0||0|
|White radish (roots, leaves)||89||0||0||0||89||0||0||0|
|Cane sugar molasses||24||0||0||0||0||0||0||0|
|Tanzania (B)||Total household of 5 people||Child aged 12 to 23 months|
|Low price season (g)||High price season (g)||Low price season (g)||High price season (g)|
|Meat/fish and animal products|
Survey team in Tanzania
Range of food items available
The number of foods available varied considerably between the two countries, and to a small extent between seasons. In Bangladesh, a minimum of 60 food items were available in the summer season and a maximum of 65 during the winter season. During both seasons in Tanzania, a total of 33 foods were available. The variation by season in Bangladesh was largely related to availability of fruits and vegetables.
Composition of diets
The diets formulated by the programme to meet nutritional requirements of a total household of 5 people, as well as the child aged 12-23 months, are shown in Table 3. The most immediate observation is that these diets include unrealistic quantities of some foods. Although spinach and amaranth are clearly cheap sources of nutrients in Bangladesh, it is unlikely that a household would consume 1- 1.5kg of these foods on a daily basis. Similarly, it would be surprising if families were able to procure and consume nearly 2 litres of buffalo milk each day. In Tanzania, there are unrealistic quantities of sesame seeds in the diet (which contribute half the household's requirements for zinc, magnesium and calcium, and a significant proportion of fat). The large quantities of leafy vegetables were included in these diets because they were freely available and hence did not affect cost.
Cost of the diet and seasonal variations
The average daily cost of the diets calculated by the programme is shown in Table 4. As expected, the daily cost of the diet varied in both settings according to the season. In Bangladesh, the maximum daily cost was during the lean season (71 Taka) and the minimum was during the winter season (53 Taka). The cost of the diet in the high price season in Tanzania was 1107 Shillings and in the low price season was 708 Shillings.
To estimate affordability we compared annual diet cost with annual income. The annual cost of the diet for a 5 person household in Bangladesh is 22,118 Taka (18,800 - 27,647), which at the time of the assessment was equivalent to 332.5 US dollars per year (282.7 - 415.7). Among the 194 household surveyed, the annual income ranged from less than 1000 to approximately 170,000 Taka. A total of 163 (84%) of these households had an annual income less than the annual cost of the diet; 153 (78.9%) had an income less than the lower limit of this estimate and 173 (89.2%) an income less than the upper limit of this estimate.
The annual cost of a diet for Tanzania was 318,637 Shillings (270,841-398,296), the equivalent of 263.9 US dollars (224.35-329.92). The average income by wealth group according to the HEA is given in Table 5. It is estimated that over half of the very poor and extremely poor (roughly 25%) have an income that is equal to or less than the annual cost of the diet. None of this wealth group and just over half of the 'poor' (roughly 55% in total) are likely to have an income that exceeds the upper range of the annual cost of the diet.
Although a nutritious diet is crucial for good health and development, families clearly need more than this, and it is necessary to take into account the cost of non-food items when interpreting affordability. Research conducted in Tanzania at the same time as this work estimated the annual cost for items such as fuel, medical costs, clothes, school fees and festival costs to be 56,860 Shillings, resulting in an annual food/non-food essential expenditure of just over 375,000 Shillings. This amount would be unaffordable for roughly 55-60% of households in the region.
Comparison with daily labour rate
In both Bangladesh and Tanzania, the poorest households tend to own little or no land and hence are dependent on paid labour. At the time of the assessments, the average labour rate in Bangladesh was between 50 and 60 Taka/day. As shown in Table 4, this is less than the estimated daily cost of a diet, particularly during the lean season. In Tanzania, the labour rate was typically 750 Shillings/day, which is also less than the average daily cost for a nutritionally appropriate diet.
|Table 4: Average daily cost of the diet by country|
|Country||Average daily cost (min/max over the year)|
|Total household of 5 people||Child aged 12-23 months||(% of total household cost)|
|Bangladesh (Taka)||61 (52 - 76)||4 (3 - 5)||(6.6%)|
|(US dollars)||0.91 (0.77 - 1.14)||0.06 (0.05 - 0.07)|
|Tanzania (Shilling)||873 (742 - 1092)||35 (30 - 44)||(4.0%)|
|(US dollars)||0.72 (0.61 - 0.90)||0.03 (0.02 - 0.04)|
|Table 5: Average income by wealth group for Tanzania|
|Wealth group||Percentage of households||Annual food and cash income for a 5 person household (Shillings)|
|Better off||8%||1,657, 000 (1,574,150 - 1,739,850)|
|Middle||23%||951,000 (903,450 - 998,550)|
|Poor||35%||360,000 (342,000 - 378,000)|
|Extremely and very poor||35%||297,000 (282,150 - 311,850)|
The results of the analyses presented here plainly illustrate that the amount of money required to meet nutritional requirements exceeds the income of the poorest households. In Bangladesh and Tanzania, the poorest make up a significant proportion of the population and in both settings, these households are heavily reliant on purchasing food and on seasonal labour. Our research in Tanzania and Bangladesh has shown that diversity of the diets given to children under 2 years is significantly associated with wealth. Clearly, promotion of appropriate dietary practices for young children will have limited impact in these settings because the most vulnerable families will find it difficult to purchase the necessary foods.
The prevalence of chronic malnutrition among young children living in rural areas of Bangladesh and Tanzania is classified as very high, with some regions reporting >40% stunting. These high levels are not only due to inadequate food, but also the monotonous and nutritionally poor diets that result from limited access to an appropriate range of foods. In Bangladesh in particular, the cost of the diet corresponds to the seasonal fluctuations in acute malnutrition, with the peak in rates seen during the lean season when foods are most expensive. Again, the cost of the diet is not the only contributing factor but it certainly plays a key role.
This CoD work is even more relevant given recent concerns about rising food prices. According to the World Bank, global commodity prices rose by 83% over the last 3 years11. It is difficult to determine the extent to which local food prices have been affected by these global trends. However it is widely felt that the impact will be significant, resulting in an even wider gap between income and the cost of the diet, particularly among the urban poor and the landless rural poor.
A woman cooking at home
It is important to note when interpreting the results presented here, that there are a number of limitations to the method. The first relates to the calculation of cost. As mentioned, the diets designed for the two case studies are not entirely realistic but it is not clear to what extent developing more realistic diets will affect cost. We are confident that the results presented are conservative estimates. To create more realistic diets we will need to add additional constraints. These could either increase or maintain cost, but given the principles of LP, they should not lead to a decrease. We are currently undertaking further research to develop appropriate additional constraints that will help in the design of more realistic diets. This is being done in part by investigating how the diets calculated by the programme compare with real dietary practices, but also by looking at dietary recommendations for young children, pregnant/ lactating women and other adults.
One component of the diets that we have not been able to factor in properly is the role of wild foods. Wild foods do form a significant proportion of the diet for population groups in some countries and it is feasible that families could be boosting nutritional intake at no extra cost. Some free foods were included in the analyses and in fact, removing the free green leafy vegetables from the foods available in Tanzania results in a 13% to 32% increase in the cost of the diet. SC UK plan to investigate the role of wild foods in the diets of rural communities but will most likely not systematically include wild foods in CoD analyses because: (i) it is very hard to accurately identify wild plant species and virtually impossible to match these with the limited composition data available for wild foods, (ii) promotion of wild foods as a significant part of diets has implications for sustainability, and (iii) not all families have access or the capacity to collect large quantities of wild foods and hence cost estimates could be misleading. We will, however, continue to include foods that are widely available at no cost but will apply limits to ensure that excessive amounts of these foods are not included in the diets calculated by the programme.
Another major limitation of the method relates to the estimation of affordability. Even if we are sure that the diet costs are realistic, we do not yet have a systematic way to add on the cost of essential non-food items and services. The findings from Tanzania clearly showed that these items make a significant difference to affordability. Further work is needed to develop this aspect of the method so that we can ensure that our estimates of the proportion of households with a gap between income and expenditure are more accurate and hence more useful when designing interventions.
In a similar vein, it is difficult to estimate the affordability of diets because of the type of income data used. First, HEA data do not translate easily into a format that can be used to accurately estimate what percentage of households can or cannot afford a nutritious diet without making potentially incorrect assumption about the distribution of income within the population. This is not a problem when using IHEA (International Health Economics Association) data because income estimates are obtained from individual households. Hence, we are able to plot the distribution of income in the population and so to better estimate the percentage of households that can or cannot afford the diet. Second, the data used came from HEAs implemented several years earlier and the country-specific inflation rates that were applied to adjust these to the time of the CoD may not be particularly representative of what actually happened to income levels. We would like to do more work to track changes in income, alongside changes in costs of foods and non-food items to better understand the relationship between inflation rates and changes in local incomes. Further thought is also needed as to how best to estimate affordability using the income data collected during HEA assessments.
Despite the current limitations, the CoD has the potential to be an extremely valuable tool. The burgeoning work on cash transfers as a means to reduce under-nutrition has prompted a need to better understand the cost and affordability of diets that ensure good nutritional status of children and families. The CoD is the only tool available that can be used to thoroughly and systematically assess this. Once the programme is set up to design more realistic diets, it will also be possible to identify likely requirements for micronutrient supplements and fortified products and to examine the cost impact of providing these types of interventions. Of particular relevance is the scope for using the CoD to monitor and predict the impact of food price rises on the poorest households. SC UK is currently pursuing this line of research.
For further information, including a full copy of the report, contact: Abigail Perry, email: A.Perry@savethechildren.org.uk
Call for input
Although SC UK has a range of options for developing the method further, we are particularly keen to receive input from people working in this sector about how they might want to use the CoD. Our goal is to make this tool freely available in a format that serves the needs of a range of organisations. If you have any ideas or suggestions about uses for the method, or if you would be interested in testing it in the field, contact Abigail Perry, Nutrition Adviser at Save the Children UK, email: A.Perry@savethechildren.org.uk
1Duffield, A. et al. 2003. Thin on the Ground. Save the Children UK.
2MDG 1: Eradicate extreme poverty and hunger.
3 Bhutta, Z.A. et al. 2008. What works? Interventions for maternal and child undernutrition and survival. Lancet, 371(9610):417-40. See summary of Lancet series in Field Exchange 33.
4Requirements for micronutrients to be met include calcium, zinc, magnesium, iron, vitamin A, vitamins B1, B2, B6, B12, niacin, pantothenic acid, folic acid, and vitamin C.
5WHO 1998. Complementary Feeding of Young Children: A Review of Current Scientific Knowledge. WHO, Geneva.
6One child aged 12-23 months, 1 child aged 3-4 years, 1 child aged 7-8 years, 1 male aged 30-59 years (50kg, vigorously active), 1 female aged 30-59 years (45kg, vigorously active, lactating).
7FEG Consulting and Save the Children UK 2008. The Household Economy Approach: a guide for programme planners and policy-makers. Save the Children UK.
8Seaman, J. et al. 2005. A Study of the Relationship between Household Economy and Nutritional Status in a Village in Kurigram, Bangladesh. Save the Children UK, unpublished report.
9Save the Children 2003. Livelihoods of Lindi Rural District: A Household Economy Assessment in Southern Tanzania. Save the Children UK, unpublished report.
10Save the Children 2007. Tackling Extreme Poverty: The Role of Cash Transfers and Complementary Social Protection Measures. Save the Children Tanzania.
11World Bank 2008. Rising food prices: policy options and World Bank response, p.1 http://siteresources.worldbank.org/NEWS/Resources/risingfoodprices_backgroundnote_apr08.pdf
More like this
FEX: Improving minimum dietary diversity for children aged 6-23 months when household affordability is a major constraint in northeast Bangladesh.
By Md Masud Rana, Sheikh Shahed Rahman, Md Al-Amin Shovan, Bazlul Kabir Joarder and Mohammad Raisul Haque Md Masud Rana is a nutrition advisor for Save the Children, with a...
Terms of reference
External Consultant for Cost of Diet (CoD) Analysis for Concern Worldwide’s Irish Aid-funded Liberia WASH Consortium ...
Dr Aliahmad Khan is a nutritionist with the World Food Programme Pakistan Country Office in Islamabad. Muhammad Aslam Shaheen is Chief of Nutrition and SUN Focal Point in...
By Esther Busquet Esther Busquet is a Nutrition Advisor with International Medical Corps (IMC), currently supporting the nutrition programmes in the western, central and...
Summary of Research1 Save the Children. Cost-benefit analysis of the kitchen garden intervention: Women and children/infants improved nutrition in Sindh, Pakistan Report, July...
Summary of review1 A mother prepares a meal in Malawi Prepared by Samuel Hauenstein Swan and Jennifer Stevenson Samuel Hauenstein Swan is Senior Policy and Research Advisor...
FEX: Understanding access to nutritious food by poor urban pregnant women and lactating mothers and their children in Kisumu, Kenya
By Albertien van der Veen, Rik Delnoye and Femke van der Lee Albertien van der Veen is an experienced public health nutritionist/epidemiologist and team-leader of the urban...
Summary of research1 Location: Global What we know: Worldwide 'nutrition transition', including in low-middle income countries is leading to increasing consumption of...
UNICEF, Pakistan with funding support from Department for International Development (DFID) and in collaboration with Ministry of National Health Services Regulation and...
Summary of Published Paper In February 1996, WFP reduced rations for refugees in Uvira, Zaire by 20% when visiting donors noted that food aid, primarily maize and oil, was...
FEX: Linking agriculture with nutrition within SDG2: making a case for a dietary diversity indicator
By Anna Lartey Anna Lartey is Director of FAO's Nutrition Division within the Economic and Social Development Department at FAO in Rome. She was a Professor of Nutrition at...
FEX: Issue 34 Editorial
Issue 34 of Field Exchange has a special focus on infant and young child feeding in emergencies (IFE) and is dedicated to the memory of Tom Marchione. Leading the IFE 'show'...
Poor diet is the number one risk factor driving the world's disease burden. Over 800 million people are undernourished, two billion consume diets that are lacking in one or...
By Kimon Schneider, Pranab K. Roy and Dr. Hasan Kimon Schneider is a development practitioner and Country Delegate with Tdh Foundation. He has a M.A. in Anthropology and a...
Your help We are thinking of conducting a study looking at the impact of our food security interventions (cash transfer, small scale agriculture ) on under five nutrition...
Summary by Kate Sadler and Emily Mitchard Dry riverbed in Shinile Kate Sadler is an Assistant Professor of Nutrition at the Feinstein International Centre, Tufts University...
FEX: The impact of intensive counselling and a mass media campaign on complementary feeding practices and child growth in Bangladesh
Summary of research* Location: Bangladesh. What we know: In Bangladesh there has been little to no progress in improving children's diets and little evidence available on...
Baraka buying millet at Ourafane market with the monthly cash payment from Save the Children Summary of evaluation1 A recent Save the Children survey conducted in southern...
Research Summary 1 Poor nutrition in early life threatens the growth and development of children, which has a knock-on effect on the sustainable development of nations. This...
Summary of research1 Location: Nigeria, Uganda, Ethiopia, Zambia, Mozambique, Tanzania and Nepal What we know: It is generally accepted that agricultural development will...
Reference this page
Abigail Perry (2008). Cost of the Diet – novel approach to estimate affordability of a nutritious diet. Field Exchange 34, October 2008. p20. www.ennonline.net/fex/34/cost