• Inadequate in 20% households in H’tota, Jaffna, Matale, R’pura 
• Many not receiving support from Samurdhi; 24% households with adequate access to food were Samurdhi recipients 
• Materials used in constructing house walls, roofs, and floors include thatching materials, bricks, mud, wood, cadjan leaves, tiles, sheets, tin, cement
• Toilet types used include septic tanks, open pits, sewerage systems, outside land
• Govt. selection procedure that considers only the income level of the households in providing subsidies is inaccurate
• People who don't have permanent income sources, registered businesses, and income-generating properties are accounted for in low-income households
• If a household can hide its high income level, it can be included as a beneficiary; if the household is unable to reveal its poverty, it is excluded from the beneficiary progs. 

In Sri Lanka, four Districts (Hambantota, Jaffna, Matale, Ratnapura), representing four Provinces (South, North, Central, Sabaragamuwa), have 20 per cent of households experiencing challenges due to inadequate food access with many of them not receiving support from the existing Samurdhi poverty alleviation programme while a substantial number of households with adequate food access were beneficiaries of the same, a recent survey on households food accessibility revealed. This highlights a clear mismatch between programme targeting and actual household needs.

These findings were made in a research article on "Developing an index for measuring household food accessibility in Sri Lanka" which was authored by N. Kuruppuarachchi (attached to the Peradeniya University's Postgraduate Institute of Agriculture, and the Uva Wellassa University's Animal Science and Export Agriculture Faculty), and J. Weerahewa and P. Silva (both attached to the same University's Agriculture Faculty), and published in the Ceylon Journal of Science' 55th Volume's First Issue.

Food security exists when all people, at all times, have physical and economic access to sufficient, safe, and nutritious food that meets their dietary needs and food preferences for an active and healthy life (the World Food Summit 1996, Rome Declaration on World Food Security). Four components are implied in the definition of food security, including availability, accessibility, utilisation, and stability. Availability, accessibility, and utilisation are physical determinants, while stability is a temporal determinant. The inability to reach these four components leads to food insecurity (M. Napoli, P. De Muro and M. Mazziotta's "Towards a food insecurity multidimensional index"). Accordingly, food insecurity is defined as “a situation that exists when people lack secure access to sufficient amounts of safe and nutritious food for normal growth and development and an active and healthy life” (S. McGuire's "Food and Agriculture Organisation, the International Fund for Agricultural Development, and the World Food Programme [WFP]. The state of food insecurity in the world 2015: Meeting the 2015 international hunger targets: Taking stock of uneven progress"). More than 295 million people across 53 countries and territories experienced acute levels of hunger in 2024, an increase of 13.7 million from 2023 (the Global Report on Food Crises' Food Security Information Network and the Global Network Against Food Crises). Sixteen per cent of households in Sri Lanka were moderately food insecure in late 2024, whereas 27% households were consuming inadequate diets as per the Household Food Security Survey conducted by the WFP (Sri Lanka Country Brief February and March of last year [2025]).

Food accessibility is a major aspect of poverty (J. Bhattacharya, J. Currie and S. Haider's "Poverty, food insecurity, and nutritional outcomes in children and adults") and poverty alleviation programmes including the provision of a social safety net, employment opportunities, microfinance, agricultural development, education and skill development, and healthcare access are implemented by Governmental and non-Governmental organisations largely to improve food accessibility (A. Barrientos and D. Hulme's "Social protection for the poor and poorest in developing countries: Reflections on a quiet revolution: Commentary", a Malawian study, B.A. De Aghion and J. Morduch's "The economics of microfinance", and K. Subbarao and C. Del Ninno's "Enhancing social protection in a changing world: A review of the World Bank [WB] experience"). Proper targeting supports the achievement of the proposed outcome of a programme under budget constraints (A. De Janvry and E. Sadoulet's "Economics development: Theory and practice") by avoiding exclusion and inclusion errors. An exclusion error (type I error) occurs when a poor person is categorised as a non-poor person. This will prevent them from getting benefits from the transfer programme. Categorising non-poor persons as poor is an inclusion error (Type II error). It wastes the resources of the programme, and it indirectly increases the exclusion of qualified poor persons.

Materials and methods 

Measurements of food accessibility

Prices of the food, the household income and expenditure, the distribution of the income within the household, low energy availability, the household dietary diversity score (HDDS), the months of household food provisioning, and access to healthcare centres, social protection, safety nets and transfer programmes are considered as measures of food accessibility at the household level (S. Jrad, B. Nahas and H. Baghasa's "Food security models", E.H. Pangaribowo, N. Gerber and M. Torero's "Food and nutrition security indicators: A review", and F. Riely, N. Mock, B. Cogill, L. Bailey and E. Kenefick's "Food security indicators and a framework for use in the monitoring and evaluation of food aid programmes").

Data sources

A dataset developed from the households and home-gardens survey was employed by Kuruppuarachchi et al. A sample of Sri Lanka has been derived, comprising diverse agro-ecological regions (AERs) and various vegetation types. It includes four administrative Districts: Hambantota (the Southern Province), Jaffna (the Northern Province), Matale (the Central Province), and Ratnapura (the Sabaragamuwa Province). In selecting the sample, a transect from forest to urban was identified for sampling within each District, as given in N. Kuruppuarachchi, D.K.N.G. Pushpakumara, G.L.L.P. Silva and L.D.B. Suriyagoda's "A multivariate approach for discriminating home-gardens in different AERs ", and W.A.M. Lowe, G.L.L.P. Silva and D.K.N.G. Pushpakumara's "Home-gardens as a modern carbon storage: Assessment of tree diversity and the above-ground biomass of home-gardens in the Matale District", and running across a maximum number of AERs according to the Latin hypercube sampling technique which facilitates maximising the diversity within the sample while allowing random sampling. About four-five sampling locations identified as circles of five kilometres radius, capturing the maximum number of AERs along each transect, were selected, while eight-12 locations were marked within each circle as the survey points to select representative samples. From each location, three-four households were randomly selected from each point, and thus, 150 households from each District were sampled, and thus the final sample consisted of 600 households. The data were collected through a detailed questionnaire survey. The households and home-gardens survey was conducted in 2019 within three months. The survey was completed before the Covid-19 outbreak began. The household head or the key decision maker was the respondent. A total of 84, 125, 122, and 90 households, respectively, from the Hambantota, Jaffna, Matale, and Ratnapura Districts, respectively, were included in the final data set. Since all the samples were drawn along transects, the final samples in each District represented different AERs. 

The share of food to the total expenditure, the dependency ratio, the inverse value of the HDDS, and the inverse value of the per capita monthly income were the measures of accessibility derived from the data set. The share of food to the total expenditure of the household was calculated by dividing the household’s monthly food expenditure by the household’s monthly total expenditure. The dependency ratio was calculated by dividing the number of dependent members in the family by the number of independent members in the family. The HDDS was calculated based on the indicator given by the WFP. The household per capita monthly income was stated by dividing the household monthly income by the number of family members.

Threshold value for food accessibility

The developed index was employed with the minimum and maximum possible value of each measurement to obtain the maximum possible range of the food accessibility index (FAI). Threshold values for FAI were also derived by utilising the theoretical potential values of the measurements. The theoretical potential values were set as: the HDDS was six as per the HDDS scale (the INDDEX Project's "Data4Diets: Building blocks for diet-related food security analysis"), the per capita monthly income was United States Dollars (US $) 85.79 according to the poverty line (the Department of Census and Statistics' Household Income and Expenditure Survey – 2019 Final Result), the share of food to the total expenditure was 0.65 (“medium food insecure: 50-65”) (L.C. Smith and A. Subandoro's "Measuring food security using household expenditure surveys: Food security in practice"), and the dependency ratio was taken as one by the assumption. According to the threshold value, households were divided into the categories of “adequate access to food” and “inadequate access to food”. 

Comparison of the food accessibility categories with the targeting criteria of Govt. implementations

Samurdhi is the largest poverty alleviation programme implemented by the Government. It was initiated in 1995 and provides cash transfers to the poor. The households that earn less than Rs. 1,500 ($ 5) per month are the cut-off used to identify beneficiaries. 

Results and discussion 

Characteristics of the households

From the studied sample, the majority (87%) of the household heads were males. The mean value of the age of the household head was 53 years plus/minus 13 years, and the mean family size was four. Households had different sources of income. A moderate number of household heads (41%) were involved in farming as their main occupation. The households’ mean monthly income and expenditure were $ 253.73 ± $ 399.61 and $ 165.25 ± $ 108.21, respectively. Further, three households reported a monthly income of zero, stating that they rely on support from their daughters and sons who live separately.

Different types of housing materials have been used by the households in constructing the walls, roofs, and floors of the houses. Thatching materials were used by 5% of the households for the wall and 1% of the households for the roof. Most of the households (73%) applied bricks, and mud and wood were used by 2% and 1% of the households, respectively, in the walls of the houses. The remaining 20% of the households used other types of materials such as cadjan leaves. The majority of the households used tiles (42%) and sheets (53%) as roofing materials. Tin and cadjan leaves were used by the rest of the households for roofs. When building floors, 86% of the households applied cement in their houses, while 13% and 1% of the households applied tiles and mud, respectively. Their sanitary level was visualised by toilet types as 74% of the households used a septic tank, 10%, 12% and 4% of the households used an open pit, a sewerage system, and outside land, respectively. Accordingly, most of the households were occupied with moderate to high levels of housing facilities.

Inadequate food access

A lower number of households have been positioned at the extreme ends of food accessibility. Most households have a moderate level of food accessibility. 

Accordingly, 86 households (20%) were included in the “inadequate access to food” category who were eligible for receiving support. The rest of the 335 households (80%) have adequate access to food and are not eligible for receiving support. The food accessibility level of the two categories was significantly different.

Households classified within the category of having “adequate access to food” exhibit a high HDDS, a higher per capita monthly income, a low share of food to total expenditure, and a low dependency ratio. Most of these households are engaged in farming activities, including the cultivation of rice and vegetables, with some also rear livestock or poultry. Their home-gardens commonly include coconuts, fruits, spices, and vegetables. Households with adequate access to food benefit from a larger proportion of homegrown food in their diets compared to those with inadequate access, which contributes to their relatively higher dietary diversity since these items are obtained free of cost. In contrast, households experiencing inadequate access to food have significantly lower levels of home-garden produce. Additionally, many households with adequate access to food engaged in Governmental or private employment and were earning higher incomes, which also correspond to their lower dependency ratios.

Comparison of the food accessibility categories with the targeting criteria of Govt. implementations

The food accessibility level of the households and the households receiving Samurdhi were not significantly associated. Among the 86 households who have inadequate access to food, 61/71% of the households were Samurdhi non-recipients who should receive the subsidies. The exclusion error (Type I error) in targeting is appearing in this case. Furthermore, 24% of the households with adequate access to food were Samurdhi recipients despite not needing subsidies, highlighting a Type II inclusion error in the targeting process.

Further, the WB's "Sri Lanka development update: Protecting the poor and the vulnerable in a time of crisis" stated that the targeting effectiveness of Samurdhi, which is Sri Lanka’s largest poverty alleviation safety net programme, remains poor, restricting the programme’s potential to enhance equity and promote economic inclusion. The exclusion error is evident as more than 50% of the lowest 20% of the income earners benefited from Samurdhi in 2006, and this had fallen to 38% by 2019. Meanwhile, the inclusion error is shown by the fact that in 2019, 12% of the richest of the 20% of the income earners were receiving Samurdhi benefits.

According to the International Monetary Fund's "Sri Lanka 2021 Article IV consultation — staff report, and statement by the Executive Director for Sri Lanka (Country Report Number 22/91)", almost 31% of the total households have benefited from cash grants of the Samurdhi programme. With this programme, 69% of the households benefited from the poorest 40% of the households in the country (the WB's "The Covid-19 impact on livelihoods and poverty in Sri Lanka. Background note to the Sri Lanka poverty assessment"). Nevertheless, exclusion and inclusion errors have been discovered with this, as 58% of the eligible households have not received the support, and 58% of the ineligible households have benefited (the United Nations Children's Fund's "Tackling the Covid-19 economic crisis in Sri Lanka: Providing universal lifecycle social protection transfers to protect lives and bolster economic recovery"). Even though social safety net reforms have been implemented, both positive and negative progress has been reported. New eligibility criteria have been developed for the Samurdhi programme, which have not yet been implemented. According to an Ethiopian study, household food availability and food aid receipts were not associated, implying the exclusion and inclusion errors. A high level of targeting errors resulted due to the inclusion of both needy and food secure households in food aid programmes, giving priority to female and aged-headed households irrespective of their food needs, focusing only on the historical deficit and specific areas.

Accordingly, most of the households in the category of having “inadequate access to food” do not receive benefits, whereas households that have adequate access to food receive them. Both cases reveal the inaccuracy of the Government selection procedure that considers only the income level of the households in providing subsidies. Much disparity was observed when the households’ food insecurity and the selection criteria by the Government are compared. Therefore, the selection criteria made by the ground-level officers are not accurate. Government officers, including Grama Niladharis, Samurdhi officers, and economic development officers, use mainly the income levels of the households when households are selected for poverty alleviation programmes. However, they are not referring to any income category. Mainly, subsidies are provided to poor people. According to the officers, low-income households are Samurdhi recipients who were identified during their field visits. People who do not have permanent income sources, registered businesses, and income-generating properties are accounted for in the low-income households. According to these scenarios, if a household can hide its high income level, it can be included as a beneficiary. If the household is unable to reveal its poverty, it is excluded from the beneficiary programmes.

Conclusions 

These findings underscore the urgent need for more robust, objective, and multidimensional targeting criteria that better reflect the real food access conditions of households, thereby improving the effectiveness of poverty alleviation interventions. Strengthening targeting mechanisms is essential to enhance the overall impact of social protection programmes aimed at reducing food insecurity and poverty in Sri Lanka.