Sustainable Way to Eradicate Poverty through Social Protection: The Case of Sri Lanka

Abstract

Social protection can be used as an effective policy instrument to achieve zero poverty. A considerable percentage of households in Sri Lanka are still suffering from poverty, reflecting the fact that the existing social protection system does not correctly address the heterogeneity of poverty. This study examines the outreach and impact of social protection on poverty in the context of the spatial heterogeneity of poverty. The reasons and nature of poverty are different according to the different poverty levels and spatial disparities. However, the existing social protection system does not accurately target these disparities. The outreach and impact of social protection are low since the current system does not adjust according to the inflation rate and spatial poverty lines. Hence, it is essential to reformulate the existing social protection system considering the spatial factors and different poverty categories while implementing a rigorous method to select the beneficiaries and benefits of social protection by considering inflation and district-level poverty lines.

1. Introduction

One of the integral parts of sustainable development is eradicating poverty by increasing the social and economic well-being of people (Mustafa et al. 2021). Eradicating poverty is crucial for sustainable development as it ensures equal access to resources, opportunities, and fundamental human rights for all individuals. Poverty alleviation contributes to social stability, reduces inequality, and fosters inclusive economic growth, allowing communities to thrive and reach their full potential. By addressing poverty, a solid foundation for sustainable development can be created. The United Nation’s sustainable development goals emphasize the importance of eradicating poverty through different policy instruments. Social protection is one of the policy instruments suggested by the United Nations. Social protection plays a vital role in eliminating poverty and vulnerability by protecting economic and social stability (ILO 2021). However, only 46.9 percent of people around the world have benefited from at least one social protection program, while 53.1 percent were unprotected from social protection (ILO 2021). Only 30.6 percent of working-age people have legally benefited from a comprehensive social protection system, which includes anything from child and family benefits to old-age coverage (ILO 2021). Conversely, 69.4 percent of working-age people have not been protected or only partially benefited from social protection (ILO 2021). The social protection coverage gap between men and women was 8 percent, reflecting less coverage of women (ILO 2021). The inadequate coverage of social protection for needy people may affect increases in poverty, struggling to achieve zero poverty by 2030.

A considerable number of people still suffer from poverty around the world, including in Sri Lanka, highlighting the need for expanding or restructuring the existing social protection system. Although Sri Lanka has a quite comprehensive social protection system with universal free education and health care, many people still suffer from poverty. According to the updated poverty line based on 2012/2013, the Poverty Head Count Index (PHCI) was 14.3 percent in 2019 (DCS 2022a). Especially, poverty levels are significantly dissimilar in the different parts of Sri Lanka because of geographical, economic, and social factors. The PHCI of the urban, rural, and estate sectors in Sri Lanka were 6, 15, and 33.8 percent, respectively (DCS 2022a). These spatial differences can also be seen in the districts and provincial levels because of disparities in resource endowment, economic opportunities, and infrastructure facilities. The PHCI for the Colombo district, which is the capital of Sri Lanka, was 2.3 percent, while for the Nuwara Eliya district, where most people economically depend on the estate sector, it was 26.3 percent. Additionally, 20.8 percent of people in the Batticaloa district, where a significant number of people depend on the fishing industry, live below the poverty line. At the same time, this percentage is 24.9 percent for the Rathnapura district, where most people depend on the mining industry and the estate sector (DCS 2022a). Currently, Sri Lanka faces a huge economic crisis by increasing poverty and vulnerability. Hence, it is necessary to evaluate the impact of poverty-related social protection programs in the context of the current situation to identify its contribution to achieving zero poverty by 2030.

Poverty is multidimensional. The nature of poverty is different, and it depends on local geographical, economic, and social factors, except for other common factors that cause poverty. Hence, social protection programs should be designed according to the nature of poverty and the reasons for poverty in different geographical areas. A universal social protection system is not always effective in eliminating poverty (Sebastian et al. 2018; Kuss et al. 2022; Beegle et al. 2016). The social protection system should be arranged according to the nature of poverty by targeting the neediest people. This is helpful to prevent leakages of transfer payments for the non-poor and reduce the risk of non-recovering transfer payments for the poor. Although in Sri Lanka, there are many social protection programs to eradicate poverty, those programs do not focus on differences in economic and social opportunities in different geographical areas. Hence, it is essential to identify the reasons and nature of poverty and outreach, and the impact of social protection according to the different geographical areas to target the needy people better so as to achieve zero poverty. Many researchers have examined the impact of social protection on poverty in Sri Lanka (Wickramasinghe 2014; Fact Sheet-UPR 2017; Abeykoon and Elwalagedara 2008; Walsh and Hallegatte 2019; De Silva and Kawasaki 2018; Sebastian et al. 2018; Newhouse et al. 2016). However, it is difficult to find comprehensive research that focuses on all these aspects. Hence, the objectives of this paper are to examine the nature and reasons for poverty and outreach and the impact of social protection programs on different poverty levels by considering spatial disparities. Further, this study attempts to contribute novel findings using a more rigorous method to classify the poverty levels, which especially can be used for poverty-related policy making. Therefore, this study provides valuable insight to policymakers to restructure or implement social protection programs so as to reduce poverty and increase socioeconomic well-being. Additionally, the findings of this study highlight the importance of providing additional facilities (infrastructure, risk management, education, etc.) other than social protection to reduce poverty.

2. Literature Review

Poverty alleviation is crucial for sustainable development as it helps to create a more inclusive and equitable society. When a significant portion of the population is trapped in poverty, their potential contributions to the economy remain untapped, hindering overall growth and progress. Poverty reduction initiatives play a pivotal role in reducing income inequality, which can have far-reaching consequences for social stability and cohesion. Addressing poverty helps to bridge the wealth gap, reducing social tensions and creating a more harmonious society where opportunities are more evenly distributed.

Since poverty alleviation has a high potential for economic growth and social well-being, researchers have analyzed poverty in different aspects. However, poverty cannot be reduced without understanding its nature. Therefore, many researchers have analyzed the spatial heterogeneity of poverty to identify the nature of poverty.

Some researchers have suggested implementing appropriately targeted poverty alleviation programs by considering spatial factors since they found the spatial disparities of poverty. Okwi et al. (2007) examined the relationship between poverty incidence and spatial conditions in rural areas in Kenya using spatial regression. The results indicate that soil type, travel time to public resources, slope, elevation, type of land use, and demographic variables mainly cause the spatial pattern of poverty. They emphasized the importance of implementing targeted anti-poverty programs by considering spatial factors contributing to poverty in different areas.

Ulman and Walega (2014) investigated the levels and composition of poverty in different areas of Poland and the factors that affect the risk of poverty according to the different levels of poverty lines. The results emphasized the spatial disparities and heterogeneity of poverty according to poverty levels. Further, the results indicate that the diversification of poverty is much greater in lower poverty lines.

Mendoza et al. (2016) analyzed the spatial differences of poverty in three provinces in the Philippines using data from the National Household Surveys on Family Income and Labor Market Status. The economic development of the different areas in the Philippines diverges because of geography; ecology; natural resource endowment; and economic, ethnic, and cultural factors. The different levels of economic development have caused poverty in different areas following the factors, such as lack of economic opportunities, low level of skills, education, exposure to shock, poor access to the market, and limited support from policies. Therefore, the authors emphasized the importance of targeted programs for poverty alleviation by considering the reasons for poverty in each area.

Paraguas and Kamil (2005) analyzed the spatial disparities of poverty in Bangladesh using a spatial econometric model. The results confirmed the spatial factor cause for heterogeneity of poverty in Bangladesh. The study recommends introducing targeted anti-poverty programs by assessing spatial factors to reduce poverty.

Agostini et al. (2008) examined the spatial disparities of poverty in Chile by focusing on urban and rural poverty. They found wider variation in poverty in both sectors. They also emphasized the importance of considering the spatial heterogeneity of poverty in formulating anti-poverty programs.

Sri Lanka has three main sectors (urban, rural, and estate), and poverty incidence is diverse because of resource endowment, availability of infrastructure facilities, and economic opportunities. Especially, the poverty rate is higher in the estate sector than in other sectors. Vijayakumar and Olga (2012) analyzed the factors that affect poverty in the estate sector in Sri Lanka. They consider roads, education, access to the market, industrial employment, and agricultural employment as key variables in the model. The results indicate that some factors that cause poverty in the estate sector are associated with spatial factors.

Ranathunga (2010) examined the micro-level determinants of household poverty in Sri Lanka using the Probit and OLS quantile regression models. According to the findings, the significant determinants of poverty are related to the human capital.

As found in the literature, many researchers have pointed out spatial factors caused by the heterogeneity of poverty, and they emphasized the importance of introducing poverty alleviation programs targeting the root causes of poverty in different areas. Social protection is considered an anti-poverty policy that is accepted by policymakers. Many researchers around the world have analyzed the impact of social protection on poverty from different points of view.

Social protection is a vital economic and social policy to reduce poverty and people being vulnerable and poor. Poverty reduction is the main component of economic development and it also boosts economic development in various ways. Mustafa et al. (2021) indicate that social protection expenditure boosted economic growth by reducing poverty in Pakistan, China, India, Bangladesh, and Sri Lanka from 1982 to 2017. They point out that health-related expenditure, especially, has an enormous impact on economic growth by improving labor force efficiency, productivity, and performance. They emphasized the importance of considering social protection expenditure in a dynamic way to achieve economic growth.

Natural disasters are one of the main reasons for poverty and vulnerability. Distortion of property and livelihood, inability to continue education in the usual manner, and resettlement are the most crucial negative aspects of natural disasters. Sri Lanka also faces natural disasters throughout the year in different geographical settings. Floods, landslides, storms, and drought are the main natural disasters that Sri Lanka faces. Walsh and Hallegatte (2019) pointed out that investment in social protection is very effective in reducing well-being losses and resilient people being vulnerable due to climate change, especially considering the flood situation in Sri Lanka. De Silva and Kawasaki (2018) also investigated the relationship between disaster risk, poverty, and associated vulnerabilities in the household sector in Sri Lanka. Especially, they analyzed the impact of flood and drought on local-level communities. The study found that the impact of natural disasters is high for low-income families if their livelihood is based on natural resources and agriculture. However, they did not consider other natural disasters that cause massive damage to some areas in Sir Lanka.

The primary objective of social protection is to uplift people from poverty. Sebastian et al. (2018) found a positive impact of social protection on poverty reduction in Sri Lanka if the programs better target the needy groups and short-term and long-term risks.

Newhouse et al. (2016) found that the impact of social protection on poverty reduction is low in Sri Lanka since the allocated fund for social assistance programs is low and has declined. They found that a large proportion of low-educated poor and near-poor engaged in the agriculture sector faces a high risk of vulnerability, highlighting the need for social protection. Further, they indicate that the ‘Samurdhi’, the main poverty alleviation program, has a decreasing impact on poverty reduction. They concluded these findings without considering the geographical differences.

Risk and uncertainty mainly affect poverty and vulnerabilities. The COVID-19 pandemic is the major health-related risk that has affected poverty around the world recently. The impact of COVID-19 on poverty is vast around the world, highlighting it as an enormous problem in Asia, also. Loss of jobs, lockdown, and health problems mainly affect the economic life of the people, pushing them into vulnerable or poor groups. Kidd et al. (2020) pointed out the importance of social protection to overcome being vulnerable by assessing the economic crisis in Asia. According to them, the social protection system should be reformulated in response to pandemics by reducing the coverage gap to boost the well-being of the people.

Some researchers analyzed the impact of social protection on food security since it is one of the main problems of poor people. Osabohien et al. (2021) analyzed the impact of social protection on food security in the Global South using 15 West African countries. Results show that social protection has a positive impact on food security, and the authors suggested enhancing the social protection system to mitigate the risk faced by poor and vulnerable people. Hidrobo et al. (2018) found that social protection programs cause an increase in food security and assets formulation.

There are various kinds of social protection programs around the world targeting poor people. Cash transfer programs are a more common mechanism. Nirere (2022) found that social protection helps to manage the risk of the poor in Rwanda, even though social protection has not lifted the poor above the national poverty line, by analyzing the cash transfer program. Karakara and Ortsin (2022) found little or no impact of social protection on consumption by analyzing the cash transfer program in Ghana. However, there is a positive impact on non-consumption spending, such as children’s schooling. Katrin et al. (2021) analyzed the outreach and impact of Cash Plus reform on poverty reduction in Zambia by considering specific population sub-groups, changing benefit amounts, and eligibility rules. They found dissimilar impacts on poverty reduction according to the geographical differences and gender of the household head. The results indicate that the reduction in the poverty gap is higher than the reduction in the poverty headcount, emphasizing that the benefits do not always manage individuals out of poverty. Similarly, multiple benefit receivers have a higher impact on moving out of poverty than single benefit receivers. They recommended providing multiple support strategies and a higher level of benefits for a single benefit receiver.

Children also suffer from poverty-related issues. Many countries, especially developing countries, offer special social protection programs targeting children. Cash transfer programs, food programs, and provision of school equipment are some of the social protection programs that aim to reduce poverty-related issues for children. A United Nations Economic and Social Commission for Asia and the Pacific (2021) country team in Mongolia analyzed the impact of Mongolia’s Child Money Program on consumption poverty and inequalities by considering different benefit levels and targeting regimes. According to the findings, the program has a significant but lesser effect on poverty reduction. However, the program has increased the consumption capacity of households, especially the lower deciles groups. Similarly, investment in child benefits has a positive impact on reducing inequalities. The results pointed out that inequality reduces drastically if the benefits are high. However, low benefits have a marginal effect on reducing inequality.

Kuss et al. (2022) found that all beneficiaries did not gain equal benefits from social protection considering the elderly population in Uganda. They emphasize the importance of considering the local economic structure when formulating social protection programs since people in well-developed areas have more opportunities to participate in economic activities rather than in remote and marginalized areas where people have less access to economic opportunities. Similar findings were found by Beegle et al. (2016) in Sub-Saharan Africa, which has a high level of spatial inequalities, especially in transportation, financial services, and markets.

Social protection consists of mainly three branches: social insurance, in-kind transfer, and active labor market programs. Although most researchers have analyzed the impact of in-kind transfer, the impact of social insurance is rarely analyzed. However, social insurance also contributes to poverty alleviation by managing risk and increasing wealth and income. Renuka (2021) found that micro-insurance programs have a significant impact on alleviating poverty, increasing the intellectual investment and growth of marketing strategy, and employment opportunities, increasing small industries and economic development in India using 150 micro-insurance policy-holders of micro insurance, which was implemented by the Insurance Regulatory and Development Authority of India.

A review of the literature shows that many researchers have analyzed the impact of social protection around the world by considering different social protection programs, different risks, and different target groups. Although a significant number of studies have focused on the Sri Lankan situation, it is difficult to find recent studies that comprehensively analyze the poverty situation and the impact of social protection on poverty in different geographical settings. Although Sri Lanka is a small country, there are huge geographical differences with a high risk of natural disasters and inequalities in accessing resources and infrastructure facilities. Those factors widely affect poverty and vulnerability. Additionally, the social protection system should be reformulated according to the heterogeneity of poverty by considering the geographical disparities to eradicate poverty sustainably. However, the research focus on those factors is limited or unavailable in Sri Lanka. Additionally, the inflation rate of Sri Lanka is increasing rapidly, and poverty lines also change significantly. Therefore, the formulation of social protection policies without considering the current situation is not valid to eradicate poverty sustainably. The current study has incorporated recent district-level poverty lines and considers the current trend of the inflation rate. Hence, this study has a significant contribution to filling the existing gaps in poverty alleviation through social protection.

3. Empirical Analysis

3.1. Categorization of Poverty Levels

Primary data collected from five districts (Colombo, Anuradhapura, Rathnapura, Nuwara Eliya, and Batticaloa) in Sri Lanka were used to analyze the objectives of this study. The sample areas were selected by considering the poverty level, the spread of the informal sector, and vulnerability to natural disasters. Purposive sampling techniques were used to select districts, divisional secretariat units, and Grama Seva units (the smallest administrative unit in Sri Lanka), and 1100 household units were selected using simple random sampling techniques.

This paper used a new methodology to categorize households according to the poverty level. The DCS of Sri Lanka used expenditure level to determine the poverty level, but it is not relevant to the aim of this study. The main concern of this paper is to provide rigorous solutions to eradicate poverty through social protection. Hence, households’ minimum monthly income that needs to achieve monthly expenditure (estimated by the DCS) is used to determine the poverty level. This is a novel contribution to the literature since other researchers mostly used expenditure level or nominal monthly income to determine the poverty level (Jayathilaka et al. 2016; Deshappriya 2021; Jayasinghe et al. 2021; Ranathunga 2010). If expenditure is concerned, it is difficult to measure their income poverty since some households’ expenditures consist of debt. Households should have enough income to cover their basic expenditures to have a better life. Hence, social protection programs that target poverty should provide sufficient benefits for the poor according to the factors that affected their poverty so that the poor are able to get rid of poverty in a reasonable period. New categorization will be beneficial to formulate rigorous social protection policies to achieve zero poverty.

The official poverty line indicates the minimum expenditure per person per month to fulfill basic needs. The DCS national level poverty line was LKR 13,777 in December 2022 (DCS 2022b). Six months’ average poverty line was calculated for each selected district so as to remove unrealistic price hikes.

The average household size for every district in the sample is 3.9 members, and the national average family size calculated by the DCS was 3.7 in 2019 for Sri Lanka (CEIC 2021). Hence, the average household size considered for calculation is four members per household. The average monthly income that is needed to achieve the monthly minimum expenditure of the households was calculated by multiplying the average poverty line of each district by the average household size. The average income level is used to identify the poverty levels since households should have income at least to cover their minimum monthly expenditure. The following table (

Table 1. The average monthly household income.

District	Six Months’ Average Poverty Line—LKR	Average Monthly Income (LKR) = Six months’ Average Poverty Line×Average Family Size
Colombo	14,692	58,768
Anuradhapura	13,294	53,176
Batticaloa	13,697	54,788
Nuwara Eliya	14,326	57,304
Rathnapura	13,683	54,732

Source: author calculation based on the DCS data, Available online: http://www.statistics.gov.lk/povertyLine/2022_new, (accessed on 22 March 2023).

) illustrates the average income of a household in each district that needs to fulfill monthly expenditures.

According to the purpose of this study, households are categorized in two different ways.

• Two-way categorization: Poor and Non-poor households
• Four-way categorization: Extremely poor, Poor, Vulnerable to poor, and Non-poor

The threshold level of income for each poverty level was determined by considering the current inflation rate. The year-over-year inflation rate based on the National Consumer Price Index was 53.2 percent in January 2023 (Department of Census and Statistics 2023).

3.2. The Nature of Poverty

Poverty is multidimensional. Hence, the nature of poverty was examined according to the availability of water, sanitation facilities, electricity, and school dropout rate since these factors reflect the different dimensions of poverty.

Households are categorized into four categories: extremely poor, poor, vulnerable to poor, and non-poor, to better capture the nature of poverty in different income groups.

Four-way categorization: Extremely poor, Poor, Vulnerable to poor, and Non-poor

Households are categorized into four different groups using the following criterion.

• Poor Households: income is exactly equal to the average income (average household monthly income = minimum monthly expenditure per person× average household size) or greater than the upper-income level of the extremely poor households;
• Extremely poor households: average income was 50 percent less than poor households;
• Vulnerable households: average income increased by 50 percent relative to the poor households;
• Non-Poor households: average income is above the upper limit of the vulnerable households.

Calculated threshold income levels are appropriate since the inflation rate of January 2023 is 53.2 percent (Department of Census and Statistics 2023). The following table (

Table 2. Four-way categorization: the threshold income level of each income group.

District	Average Monthly Income (Base Category) LKR	Threshold Household Income LKR—Extremely Poor Households	Threshold Household Income LKR—Poor	Threshold Household Income LKR—Vulnerable to Poor	Threshold Household Income LKR—Non-Poor
Colombo	58,768	29,384 or less	29,385–58,768	58,769–88,152	88,153 or above
Anuradhapura	53,176	26,588 or less	26,589–53,176	53,177–79,764	79,765 or above
Batticaloa	54,788	27,394 or less	27,395–54,788	54,789–82,182	82,183 or above
Nuwara Eliya	57,304	28,652 or less	28,653–57,304	57,305–85,956	85,957 or above
Rathnapura	54,732	27,366 or less	27,367–54,732	54,733–82,098	82,099 or above

Source: author calculation based on DCS data.

) illustrates the threshold income level of each income group.

3.3. Determinants of Poverty

Determinants of poverty were examined using two-way categorization (poor and non-poor households)

Two-way categorization: Poor and Non-poor households

The two-way categorization (poor and non-poor households) was determined according to the following criterion.

• Poor households: monthly income is less than or equal to the average monthly income (Table 1);
• Non-poor households: monthly income is greater than average monthly income (Table 1),

The following Table (

Table 3. Two-way categorization: threshold income of the Poor and Non-poor households.

District	Average Monthly Income—LKR (Base Category)	Threshold Income of Poor Households—LKR	Threshold Income of Non-Poor Households—LKR
Colombo	58,768	58,768 or less	58,789 or above
Anuradhapura	53,176	53,176 or less	53,177 or above
Batticaloa	54,788	54,788 or less	54,789 or above
Nuwara Eliya	57,304	57,304 or less	57,305 or above
Rathnapura	54,732	54,732 or less	54,733 or above

Source: author calculation based on DCS data.

) presents the threshold income level of the Poor and Non-poor households.

The Probit model was performed to determine the reasons for poverty. The following table (

Table 4. Variables of the Probit model.

Variables	Measurement
Dependent variable	1—Poor; 0—Otherwise
Independent variables
X1—Main income source	1—Regular; 0—Otherwise (Irregular)
X2—Education of the decision-makers	Number of years of education
X3—Expenditure	1—Expenditure is less than income; 0—Otherwise
X4—Monthly food expenditure	Food expenditure as a percentage of monthly income
X5—Income diversification	1—Income diversified; 0—Otherwise
X6—Number of old-age people above 70 years	Number of people
X7—Number of old-age people above 60 years	Number of people
X8—Number of sick people	Number of people
X9—Reasons for debt	1—Consumption; 0—Otherwise (Investment)
X10—Reasons for debt	1—Education; 0—Otherwise (Investment)
X11—Reasons for debt	1—Health; 0—Otherwise (Investment)
X11—Reasons for debt	1—Housing; 0—Otherwise (Investment)
X12—Availability of own house	1—Available; 0—Otherwise
X13—Households are affected by natural disasters	1—Yes; 0—Otherwise
X14—Households’ income sources are affected by pest diseases	1—Yes; 0—Otherwise
X15—Households are affected by an animal attack	1—Yes; 0—Otherwise
X16—Wealth	1—Available; 0—Otherwise
X15—Social protection	The monetary value of the social protection

Source: prepared by the author.

) illustrates the dependent and independent variables of the Probit model.

3.4. Outreach and Impact of the Social Protection

Outreach and the impact of social protection were analyzed according to the four-way categories of poverty levels (see Table 2). The outreach of social protection was examined according to the percentage of households that received the benefits of at least one social protection program. The impact of the social protection program was examined by calculating the percentage of households that push down other lower categories of income if the social protection benefits are excluded. The monetary value of the poverty-related social protection programs (Samurdhi benefits, old-age benefits, sick and disabled compensation, government pension, and informal sector pension) was considered to analyze the outreach and impact of the social protection since those programs directly or indirectly affect the reduction of poverty.

4. Results and Discussion

4.1. Nature of Poverty

Poverty is multidimensional. The nature of poverty is analyzed according to the availability of water, sanitation, electricity, and school dropout rate since those factors reflect the different dimensions of poverty.

According to the data, the following table (

Table 5. Percentage of households in each poverty level.

District	Percentage of Households
	Extremely Poor	Poor	Vulnerable to Poor	Non-Poor
Colombo	38	34	22	6
Anuradhapura	37	38	15	10
Batticaloa	33	34	21	12
Nuwara Eliya	35	34	22	9
Rathnapura	38	39	12	11

Source: author calculation based on the field survey.

) illustrates the percentage of households that belongs to each poverty level.

According to the results, 67 to 77 percent of households in every district belong to extremely poor or poor categories, highlighting the importance of receiving social protection to get rid of poverty. The main reason for the higher percentage of extremely poor and poor households is the minimum income that determines whether the poverty level is high because of the high inflation rate. As well, sample areas are selected using the purposive sampling method so as to find marginalized areas to target the objectives of this study better. However, it is necessary to examine the nature of poverty to identify the spatial factors that cause poverty.

Poverty may affect the education of children. The following table (

Table 6. Percentage of child drop-out of school education.

Poverty Level	Percentage of Children
	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Extremely poor	12	21	18	19	16
Poor	11	16	16	15	12
Vulnerable to poor	1	4	4	2	3
Non-Poor	0	0	0	0	0

Source: author calculation based on the field survey.

) illustrates the percentage of school-age children that stopped their school education because of financial issues or poverty-related problems.

Sri Lanka has universal free education up to the basic degree in state universities. Apart from that, there are several social protection programs for school children. Under those programs, the government provides textbooks, one uniform per year, a mid-day meal for some schools, insurance for health-related issues, and subsidized transport facilities. According to the current economic crisis in Sri Lanka, inflation is very high, increasing the price of all goods and services rapidly, including school equipment and transport, which affects the education of school children. Even though the government provides these facilities, some households cannot afford other expenditures that their school-aged children need because of financial issues. Apart from the direct impact of financial constraints, financial constraints also affect the education of children in various indirect ways. Some children have to stop their education to look after sick or disabled family members or their younger siblings. If the households have enough money, students do not have to sacrifice their education for those activities. Similarly, some students have to work for money to afford household expenditures, even though it is illegal to work for money. Most financially destitute households have many family problems, especially arising because of poverty. Those problems also negatively affect children’s education. The unavailability of schools close to their residence, lack of transport facilities, and inability to afford the cost of transport are also significant issues that affect education. Natural disasters also have negative consequences on education. Distortion of houses, including their school equipment, re-settlements, and loss of family members, are the negative consequences of natural disasters that damage education. The dropout rate is higher for extremely poor and poor households compared to other income groups. None of the children in non-poor households stopped school education because of poverty-related issues. The school dropout rate is low in the Colombo and Rathnapura districts compared to other districts. The most probable reason is that these two districts have better infrastructure facilities, especially transport facilities and schools, than other districts. One of the main reasons that prevent children from school education in the Nuwara Eliya, Batticaloa, and Rathnapura districts is that children have to work for money to assist their household expenditure since most households depend on the informal sector with seasonal or irregular income. According to the field survey, nearly 30 percent of children in the Nuwara Eliya district work for money on tea estates or vegetable farms without regularly attending school. A significant percentage of children in the Batticaloa district said that they work in the fishing sector to earn money because of financial problems. Most children in the Anuradhapura district, where kidney disease rapidly spreads, have to sacrifice their education to look after sick family members or because of unaffordable health care expenditure. Fourteen children who are aged between 10 and 14 years old said that they have no time and money for school education since their parents suffer from kidney diseases. Accordingly, human–animal conflicts also negatively affect education in the Anuradhapura district. The Colombo and Rathnapura districts are regularly affected by floods, and the costs of these floods are unable to be managed, especially for low-income groups. Landslides, low income, and poor infrastructure negatively affect education in the Nuwara Eliya district. Poverty negatively affects education, and it badly affects the economic and social well-being of the country in the short term and long term. Ranathunga (2010) also pointed out that the main determinant of poverty is related to human capital. Education is one of the most powerful tools that can be used to reduce poverty in future generations. Hence, social protection programs should focus on these spatial factors that negatively affect the education of the children in the poor and extremely poor categories so as to reduce poverty that accumulated because of low education. Especially, education-related social protection coverage should be expanded to capture the issues of poor and extremely poor households since the universal system does not better target their difficulties. Karakara and Ortsin (2022) found that social protection has a positive impact on non-consumption expenditure such as education.

The availability of quality sanitation facilities is also essential for human life. Water, lavatories, and electricity are essential for a quality life. This study considers the availability of clean water sources belonging to their households or vicinity so as to use, without any difficulties throughout the year, household activities such as bathing, washing, cooking, and drinking. The following table (

Table 7. Unavailability of water facilities.

Poverty Levels	Percentage of Households
	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Extremely poor	08	48	32	23	17
Poor	10	44	29	21	15
Vulnerable to poor	9	29	27	4	2
Non -Poor	4	22	28	2	1

Source: author calculation based on field survey.

) illustrates the percentage of households that do not have water facilities according to the different poverty levels.

Availability of water facilities is a big issue for extremely poor and poor households in all the districts rather than the other two categories since poor families do not have sufficient money to receive piped water or have their own wells or tube wells. However, nearly 30 percent of vulnerable and non-poor households in the Anuradhapura and Batticaloa districts do not have sufficient clean water for their daily requirement. This percentage becomes nearly 50 percent for extremely poor and poor households in the Anuradhapura district. Most households in the Anuradhapura district suffer from water problems without their income level because of the long period of drought. Nearly 20 percent of extremely poor and poor women in the Anuradhapura district said that they do not have enough time to work for a salary since they have to spend much time pitching water. Further, they mentioned that sometimes their children also have to spend much time pitching water, sacrificing their education. As well, the quality of water is very low in the Anuradhapura district since water is contaminated by pesticides that are used for agricultural activities. Hence, most households have to buy purified water for drinking. It becomes a big problem for extremely poor and poor households because of financial issues. Most extremely poor and poor households mentioned that they consume contaminated water even though they know the negative consequences of it because of financial problems. The Batticaloa district also has water-related issues, especially in the coastal areas. It is difficult to use water in the drought season since island water is mixed with seawater. Water becomes a big issue only for poor and extremely poor households in all other districts. Lack of water is not just an issue of water. It creates health, social, and economic problems, and its opportunity cost is considerably higher for low-income groups. A lack of clean water causes many health problems, and the main reason for kidney disease in the Anuradhapura district is contaminated water. If the households do not have enough water, they have to spend much time pitching or collecting water. Because of this reason, most women have to scarify their working hours or sometimes they have to refrain from income-earning activities, pushing them to be poorer. Similarly, children face violence when they are left alone at home while the mother is out of the home pitching water or bathing in far-away water sources. Accordingly, children or women travel far for water sometimes, and they also have to face some violence. Setiawati et al. indicated that supplying clean water may increase the labor supply (Setiawati et al. 2023). Hence, it is necessary to solve the water issue since it creates many negative consequences, especially for poor households, and lack of water itself causes poverty.

One of the major parts of sanitation is the availability of lavatories. The unavailability of proper lavatories may create personal and social issues and health problems. The following table (

Table 8. Unavailability of lavatory facilities.

Poverty Levels	Percentage of Households
	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Extremely poor	12	2	16	21	6
Poor	10	3	16	23	4
Vulnerable to poor	0	0	0	0	0
Non-Poor	0	0	0	0	0

Source: author calculation based on the field survey.

) shows the percentage of households in each poverty level that do not have their own lavatory facilities.

The unavailability of the lavatory is an issue for extremely poor and poor households, especially for all the districts except the Anuradhapura and Rathnapura districts. Most households in the Batticaloa district who live in the coastal areas use shared lavatories or public lavatories. Most extremely poor and poor households in the estate sector in the Nuwara Eliya and Rathnapura districts do not have their own lavatories and use public lavatories provided by the estate. However, this is not a big issue in the Rathnapura and Anuradhapura districts. The distortion of the lavatories by natural disasters commonly causes the unavailability of their own lavatories in the Rathnapura and Colombo districts, which are regularly affected by floods.

The type of energy that is used for lighting is also a vital measurement of multidimensional poverty since it measures the quality of life. The following table (

Table 9. Unavailability of electricity.

Poverty Levels	Percentage of Households
	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Extremely poor	10	41	46	39	26
Poor	8	32	31	23	14
Vulnerable to poor	2	15	25	12	7
Non-Poor	0	5	4	0	0

Source: author calculation based on the field survey.

) illustrates the percentage of households that do not have electricity for lighting purposes.

Mainly, there are three major reasons that affect the unavailability of electricity. Households do not have the ability to receive electricity because of financial difficulties even though electricity is available in their area. Similarly, some households cannot connect to electricity facilities because of the unavailability of electricity in their areas. In Sri Lanka, infrastructure facilities are poor in remote areas. Gunawardhana (2004) mentioned that one of the main factors that affecting poverty in the estate sector is poor infrastructure facilities. Another reason is housing conditions. If the houses are built temporarily or with raw materials that cannot use electricity, those households cannot connect to electricity facilities even though electricity is available in those areas. Mostly, extremely poor and poor households have these types of houses. A significant number of extremely poor and poor households do not have electricity facilities compared to the higher income groups because of the abovementioned facts. Lack of electricity facilities is a huge problem in the Anuradhapura and Batticaloa districts because of poor infrastructure facilities and housing conditions. Similarly, a significant percentage of high-income households also do not have electricity facilities in these two districts because of poor infrastructure facilities. A significant percentage of extremely poor and poor households in the Nuwara Eliya district also does not have electricity because of poor housing conditions and infrastructure facilities.

4.2. Determinates of Poverty

According to Table 5, almost the same percentage of households suffer from extreme poverty or poverty in all districts, even though the reasons for their poverty are different because of various reasons. The Probit model was estimated to identify the reasons for poverty in each district. Mainly, two poverty categories (poor and non-poor see Table 3) are used to determine the reasons for poverty since, according to the nature of poverty, extremely poor and poor households have almost the same characteristics. The following table (

Table 10. Results of the Probit model.

	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Variables	Marginal Effect	Marginal Effect	Marginal Effect	Marginal Effect	Marginal Effect
Constant	−0.432 (0.23)	−0.223 (0.32)	0.134 (0.22)	−0.782 (0.19)	−0.954 ** (0.03)
Main source of income	−0.21 *** (0.00)	−0.07 (0.17)	−0.021 ** (0.02)	−0.010 * (0.06)	−0.112 ** (0.02)
Education level of the decision maker	−0.003 ** (0.04)	−0.21 ** (0.05)	−0.74 ** (0.03)	−0.0432 * (0.09)	−0.012 * (0.10)
Expenditure	0.012 (0.03) **	0.132 * (0.07)	0.031 *** (0.00)	0.112 *** (0.00)	0.034 ** (0.04)
Percentage of food expenditure	0.020 *** (0.01)	0.002 (0.13)	0.145 ** (0.03)	0.216 * (0.06)	0.034 ** (0.03)
Income diversification	−0.091 (0.17)	−0.021 * (0.09)	−0.021 * (0.10)	−0.03 ** (0.03)	−0.013 * (0.06)
Number of old-age people (above 70)	0.212 * (0.08)	0.172 * (0.10)	0.082 * (0.09)	0.091 * (0.10)	0.163 * (0.06)
Number of old-age people (years 60–70)	0.003 ** (0.03)	0.023 (0.32)	0.003 (0.13)	0.034 (0.15)	0.04 (0.19)
Number of sick and disabled people	0.231 (0.11)	0.622 ** (0.03)	0.032 ** (0.03)	0.023 * (0.10)	0.121 (0.11)
Reasons for debt Consumption	0.089 *** (0.01)	0.019 (0.28)	0.089 ** (0.03)	0.081 ** (0.03)	0.052 *** (0.01)
Reasons for debt Health	0.230 * (0.10)	0.117 *** (0.00)	0.003 *** (0.01)	0.134 (0.14)	0.002 (0.32)
Reasons for debt Education	0.563 (0.13)	0.139 (0.21)	0.008 (0.13)	0.034 (0.25)	0.432 (0.17)
Reasons for debt Housing	0.023 *** (0.01)	0.032 (0.24)	0.431 (0.36)	0.013 ** (0.02)	0.023 ** (0.03)
Availability of own house	−0.210 * (0.08)	−0.181 (0.12)	−0.041 (0.18)	−0.034 ** (0.04)	−0.021 *** (0.01)
Affect natural disasters	0.080 *** (0.00)	0.002 * (0.10)	0.096 * (0.09)	0.193 *** (0.00)	0.815 *** (0.01)
Human–animal conflicts	0.034 (0.36)	0.061 *** (0.00)	0.004 (0.33)	0.003 (0.23)	0.003 (0.15)
Pet diseases	0.003 (0.23)	0.104 ** (0.03)	0.008 (0.34)	0.012 * (0.09)	0.004 (0.18)
Wealth	−0.10 ** (0.02)	−0.17 ** (0.02)	−0.09 *** (0.01)	−0.031 * (0.06)	−0.121 * (0.09)
Social protection	−0.097 ** (0.02)	−0.02 (0.14)	−0.01 (0.18)	−0.01 (0.11)	−0.01 (0.13)

p-values are presented within brackets. *** Significant at 1%, ** significant at 5%, * significant at 10%. Source: author calculation based on the field survey.

) illustrates the results of the Probit regression model.

The main source of income, education level of the decision maker, expenditure level, income diversification, number of people over 70 years old, number of people over 60 years old, receiving loans for consumption and education, availability of own house, and wealth commonly affected increasing or decreasing trends of poverty in all districts. Other variables have a significant influence on poverty according to spatial factors. According to the results, loans for education and the number of people aged between 60 years and 70 years old did not significantly affect poverty in any district.

Regular monthly income reduced poverty since, most probably, households can manage their expenditure according to their income. However, most poor households have an irregular monthly income that cannot plan their monthly expenditure or their future activities. Hence, it is necessary to assist poor households to have regular income sources so that they can get rid of poverty.

If the decision maker has better education, it significantly affects the reduction of poverty since those households are able to manage their risks by making wise decisions. Similarly, if the decision maker has better education, they can earn more by working a better job.

Nearly 90 percent of poor households’ monthly expenditure is higher than their monthly income. So, they have to borrow money for their essential needs or daily survival. Borrowing for non-income-earning activities greatly affects indebtedness and ultimately increases poverty.

If the households’ food expenditure as a percentage of total expenditure is considerably high, those households have fewer opportunities for savings or investment or less money to spend on other essential needs. Out of the total poor households, 73 percent of households spent more than 80 percent of their income on food, while some households’ food expenditure was higher than their income since food inflation is very high in Sri Lanka. Similarly, the income of poor households is very low. Some women in poor and extremely poor households said that they receive one meal per day since they do not have enough money to spend on food. Except for the Anuradhapura district, the poverty levels of all other districts increase if food expenditure as a percentage of total expenditure is increased. It is essential to provide social protection for food for the districts where food expenditure greatly affects increased poverty. Osabohien et al. (2021) found that the impact of social protection for food security was positive in West African countries. Most households in the Anuradhapura district cultivate all kinds of essential foods, at least for their consumption. So, social protection for food is not essential for this district. However, they need social protection for agricultural activities to enhance their livelihood. Otherwise, they also need social protection facilities for food.

If the households have more than one income source and/or more than one income earner, those households are considered income-diversified households. Income diversification is caused by reducing poverty by improving the ability to cope with risk.

Households have a high probability of being poor or vulnerable if there are sick or disabled family members because of several reasons, such as high expenditure for health care, sick people cannot earn money, and, sometimes, other family members have to sacrifice their occupation to take care of sick people. According to the results, sickness or disability have an increasing impact on poverty, especially in the Anuradhapura, Batticaloa, and Nuwara Eliya districts since most households’ main income sources are based in the informal sector, which does not have decent work conditions. Kidney disease is common in the Anuradhapura district and most households become poor because of this disease. Although Sri Lanka has free health care and the government provides LKR 5000 monthly for the disabled or people who suffer from kidney diseases, those types of assistance do not sufficiently compensate for the risks of those households. Nearly 42 percent of households mention that they did not receive social protection benefits even though there are disabled or sick family members because of various reasons. Therefore, it is essential to provide adequate compensation by targeting all needy households to reduce poverty sustainably. Otherwise, those households push into the poverty cycle, pushing their future generation also into poverty, especially because of low education arising for the financially destitute. Kidd et al. (2020) also emphasized that the coverage gap should be reduced to increase welfare by analyzing the economic crisis in Asia.

Similarly, if there are old-age people in households, those households have to afford extra costs to look after them, and sometimes other family members even have to stay home without working to take care of them. However, according to the findings, people aged between 60 and 70 years do not significantly influence poverty because most of them earn money, especially from the informal sector. However, people aged more than 70 years old significantly influence an increase in poverty since most of them are not able to work and the cost of caring is high. Hence, social protection programs should mainly focus on this vulnerable group.

Debt greatly affects poverty; especially if households receive loans for non-income generating activities, those households have a high chance of being poor or poorer since those loans cause indebtedness. Out of the total sample, a significant number of households have obtained loans for consumption, especially for food or other daily activities. Consumption loans significantly increase poverty in all the districts except the Anuradhapura district, where most households cultivate vegetables and other foods. Borrowing money for health care significantly affects poverty in the Anuradhapura district since most households have kidney patients. Nearly 25 households mentioned that they have to borrow money every month for health care facilities, and because of that, they become indebted. Housing debt also significantly increases poverty in all districts except the Anuradhapura and Batticaloa districts, where the majority of households have their own house even though the condition is low. Construction of a typical house is not a big issue in these two districts because of the low price of land and raw materials. Further, according to the respondents, households in the Rathnapura district have to spend a lot of money to renovate their houses since floods badly damage their houses at least twice a year. Similarly, most households even depend on the loan to pay their house rent. Housing construction and land price is very high in some districts in Sri Lanka. Hence, it is necessary to provide social protection for housing facilities, especially in the districts where housing debt significantly affects poverty.

On the one hand, borrowing money for housing causes poverty; on the other hand, the unattainability of owning a house is another major factor that affects poverty. The unavailability of their own house may lead to economic and social instability. The unavailability of their own houses significantly affects poverty in all districts except the Anuradhapura and Batticaloa districts. The majority of households in the Batticaloa and Anuradhapura districts have their own house even though housing conditions are low.

Floods, landslides, droughts, and storms are the major natural disasters that cause poverty in all districts. Floods and landslides are the main natural disasters that affect the Rathnapura district, while landslides affect poverty in the Nuwara Eliya district. Drought has a significant influence on poverty in the Anuradhapura district, while storms badly affect the Batticaloa district, pushing many households to be poorer. The findings indicate that natural disasters are one of the major reasons for poverty in all districts that are affected in various ways, and the costs of disasters are different. As an example, drought mainly affects agriculture activities in the Anuradhapura district, while floods negatively affect agriculture and the properties in the Colombo and Rathnapura districts. Extremely poor and poor households in flood-affected areas in the Colombo district mentioned that because of floods, they can never get rid of poverty, and it badly affects the education of their children. Storms significantly influence poverty in the Batticaloa district by destroying houses and fishing equipment. Landslides in the Rathnapura and Nuwara Eliya districts had negative consequences on households in numerous ways by damaging their livelihood. Therefore, social protection programs should consider the heterogeneity of the spatial factors that cause poverty to address poverty-related issues correctly. Walsh and Hallegatte (2019) found that social protection has a positive impact on reducing poverty in Sri Lanka.

Human–animal conflicts damage livelihoods, houses, and other properties and are life-threatening for the people in the Anuradhapura district, highlighting the need for special social protection coverage for this consequence. Another major reason for poverty in the Anuradhapura and Nuwara Eliya districts, in which the agriculture sector is widely spread, is pest disease. Hence, it is essential to extend social protection coverage for this issue to reduce poverty in the agricultural community.

If the households inherit substantial wealth, such as gold, financial savings, land, and other properties, those households have significant capacity to reduce poverty since they can manage the risks and other costs better. Hence, it is vital to boost income-earning activities through social protection for poverty reduction.

The main social protection programs, such as Samurdhi, old-age benefits, social protection for diseases and disabilities, and pension, were considered in this analysis to identify the influence of social protection on poverty. According to the findings, social protection has a significant influence on reducing poverty in the Colombo district. At the same time, its impact is not significant for all other districts, even though the sign of this variable shows that social protection contributes to reducing poverty. Katrin et al. (2021) also found a dissimilar impact of social protection because of geographical differences. Most households in the Colombo district receive pensions that are substantially high, and their influence on poverty reduction is considerable. However, the impact of social protection is low for other districts since the current system does not accurately address the specific risks and other economic and social consequences of those areas. As an example, all the districts face different kinds of natural disasters. However, those households did not receive any social protection benefits to recover from those impacts. Accordingly, human–animal conflict, kidney diseases, and pest diseases also significantly affect poverty in the Anuradhapura district, even though those households did not receive social protection to overcome those risks. Inadequacy of the benefits also may affect the insignificance of this variable.

4.3. How Social Protection Alleviates Poverty in Sri Lanka—Outreach and Impact of Social Protection

The outreach of poverty is analyzed using four types of poverty categories. The outreach of social protection is examined by calculating the percentage of poor households who received benefits from at least one social protection program. Pension, old-age benefits, and sick and disability allowances were considered to examine the outreach of the social protection programs. Universal social protection programs (health and education) were not included in this analysis. The following table (

Table 11. Outreach of social protection.

Poverty Levels	Percentage of Households That Benefited from Social Protection
	Colombo	Anuradhapura	Batticaloa	Nuwara Eliya	Rathnapura
Extremely poor	52	74	76	72	68
Poor	13	22	23	24	23
Vulnerable to poor	10	16	3	4	7
Non-Poor	10	11	2	9	10

Source: author calculation based on the field survey.

) illustrates the percentage of households that benefited from at least one social protection program.

According to the information, very few households in other poverty categories, except extremely poor households, receive social protection benefits. Mostly, poor, vulnerable, and non-poor households receive pensions that do not consider income level. However, a few households receive benefits from poverty alleviation programs even though their real income is more than the threshold level of the income. This is a common problem since officers cannot obtain accurate information of income, especially income-earning activities that are not transparent.

The statistics show that nearly half of the extremely poor and poor households only received the benefits of one social protection program. According to the respondents, they were not able to receive social protection facilities because of various reasons such as inability to prove residency, several families living in the same house, complicated procedures, inefficiency, or unfavorable conditions of the officers. A significant percentage of extremely poor households in all districts did not receive the benefits of one social protection program because their threshold income level exceeded the threshold income level that the government uses to select beneficiaries of social protection. The extremely poor households are also even deprived of poverty-related social protection since the threshold income level is not adjusted according to the inflation rate.

The examination of the impact of current social protection programs is essential to boost its impact on reducing poverty and vulnerabilities.

The impact of social protection is estimated by calculating the percentage of households that push down to the next poverty level if social protection benefits are excluded from their monthly income.

To evaluate the impact of social protection on poverty monetary value of Samurdhi benefits, old-age benefits, pension, and social protection for disabled and sick people are considered since the benefits of all these programs are collectively supported to reduce poverty.

The following table (

Table 12. Impact of social protection: percentage of households push to other poverty categories.

Poverty Level	Percentage of Households
	Colombo	Anuradhapura	Nuwara Eliya	Batticaloa	Rathnapura
Extremely poor	9	12	10	11	9
Poor	3	5	7	4	2
Vulnerable to poor	2	3	6	4	2

Source: author calculation based on the field survey.

) illustrates the percentage of households that increased their poverty level after excluding their social protection benefits.

The following table (

Table 13. Percentage of non-poor households that become extremely poor, poor or vulnerable after excluding social protection benefits.

Poverty Level	Percentage of Households
	Colombo	Anuradhapura	Nuwara Eliya	Batticaloa	Rathnapura
Non-poor	8	10	2	7	8

Source: author calculation based on the field survey.

) shows the percentage of non-poor households that push to other poor categories if social protection benefits are removed.

According to the results, a few percentages of households in all the districts were pushed down to a high level of poverty or became vulnerable to being poor if the social protection benefits were excluded. Especially, the households in the lower bound of the poverty level were pushed down when those benefits were excluded. However, a significant percentage of households did not change their poverty level, reflecting the inadequacy of benefits.

A considerable percentage of non-poor households who benefited from the government pension schemes became poor or vulnerable if the pension benefits were removed, reflecting that pension benefits have a huge impact on reducing poverty. The pension benefits are considerably high compared to other social protection benefits. Hence, after removing those benefits, a significant percentage of households became low-income households. These results highlight the importance of providing social protection facilities even for non-poor households.

According to the analysis, the impact of existing social protection programs is very low since only a small percentage of people were pushed down to the other poverty level if social protection benefits were excluded. According to the Probit model, social protection has a positive impact on reducing poverty, and for some districts, it is significant even though the marginal impact is low. Thus, it is evident that social protection has a positive impact on reducing poverty even though the benefit is not adequate to get rid of poverty. Renuka (2021) and Walsh and Hallegatte (2019) also found that social protection contributes to reducing poverty.

Social protection programs should be formulated so that poverty could be eradicated in a reasonable period. Hence, it is necessary to provide adequate benefits according to risks and vulnerabilities. Especially, social protection benefits are not adequate according to the current economic condition in Sri Lanka. The monthly benefits of the Samurdhi program range from LKR 420 to 3500, according to the family size. If a family has fewer than three members, those families receive LKR 1500, while families with four or more members receive LKR 3500 per month. The empowered families receive LKR 420 per month (Department of Samurdhi Development 2018). These benefits are not adequate according to the current inflation rate. Old-age benefits and disability benefits are also very low. The informal sector’s pension schemes also pay very low amounts, and those programs are not continuous or sustainable. Because of these reasons, most informal sector workers are not even enrolled in informal sector pension schemes.

The results show that according to the current situation in Sri Lanka, the existing social protection system is not appropriate to achieve sustainable development goals, especially zero poverty by 2030. Therefore, it is necessary to restructure the existing social protection system by appropriately targeting the geographical heterogeneity of the nature and reasons for poverty.

5. Conclusions and Policy Implications

5.1. Conclusions

The United Nations has declared a goal to achieve zero poverty by 2030. However, according to the current situation in Sri Lanka, poverty is increasing rapidly, and it is questionable whether the existing social protection system has a significant contribution to reducing poverty. Especially, the poverty rate considerably varies according to different geographical settings. Therefore, the aim of this study is to examine the impact of social protection on poverty by considering different income groups and geographical disparities. Primary data were collected from five districts (Colombo, Anuradhapura, Batticaloa, Nuwara Eliya, and Rathnapura) so as to capture the heterogeneity of different geographical settings. Poverty is categorized in two different ways according to the objectives of the study based on the minimum monthly income that needs to achieve the minimum monthly expenditure estimated by the DCS.

• The nature of poverty was analyzed according to the availability of water, electricity, lavatory facilities, and the school dropout rate. The percentage of extremely poor and poor households is higher than other poverty categories since this study considers the recent district poverty lines to categorize poverty levels, and sample areas were selected to represent the majority of marginalized people;
• The percentage of children who stopped school education because of poverty-related issues is high in extremely poor and poor households in all the districts. According to the spatial factors, the school dropout rate is higher in the Anuradhapura, Batticaloa, and Nuwara Eliya districts because of poor infrastructure facilities and the wide spread of the informal sector. Illness and human–animal conflicts also greatly affected the education of children in the Anuradhapura district apart from other factors. Different types of natural disasters also negatively affect the education of children in all the districts;
• Extremely poor and poor households in all the districts have significantly suffered from a lack of water more than other income categories since those households do not have enough money to receive piped water or have their own water sources. However, water also becomes a big issue for other income groups in the Anuradhapura district because of long-term drought, contaminated water, and poor infrastructure facilities. Water also becomes a huge problem in the Batticaloa district because of poor infrastructure facilities and seawater mixed with inland water sources. Therefore, special attention should be paid to the Anuradhapura and Batticaloa districts in reducing poverty related to the water-related issue; meanwhile, in other districts, special attention should be paid to low-income groups;
• The unavailability of lavatories is a big issue for extremely poor and poor households in the Batticaloa and Nuwa Eliya districts rather than other districts, highlighting the need for assistance to build up their own lavatories since the availability of lavatories is a main requirement to have a quality life. Additionally, it is necessary to assist in building up the lavatories that were damaged because of natural disasters. The unavailability of electricity facilities is also a big problem in the Anuradhapura, Batticaloa, and Nuwara Eliya districts because of poor infrastructure facilities;
• A significant percentage of extremely poor and poor households, and even vulnerable households, also do not have electricity since they cannot afford the initial cost and the condition of houses are low;
• Reasons for poverty are analyzed using the Probit regression model. The main source of income, education level of the decision maker, expenditure level, income diversification, number of people over 70 years old, receiving loans for consumption and health, availability of one’s own house, and wealth are common variables that cause an increasing or decreasing trend of poverty in all districts. Apart from those factors, other factors have a significant influence on increasing poverty in different geographical settings. Human–animal conflicts, diseases, and long periods of drought mainly affect poverty in the Anuradhapura district. The unavailability of one’s own house causes poverty in the Colombo, Nuwara Eliya, and Rathnapura districts. Different kinds of natural disasters cause poverty in all the districts. Pest diseases affect poverty by damaging agriculture in the Anuradhapura and Nuwara Eliya districts. Although social protection benefits indicate a decreasing impact on poverty in all the districts, it is significant only for the Colombo district, in which a significant percentage of households benefited from government pension schemes. An inadequacy of social protection benefits may have an insignificant impact on other districts because major risks and needs do not target the existing social protection system. The main reason is the inadequacy of benefits except for government pension benefits that affect reducing poverty;
• The outreach of poverty is examined according to the percentage of households that received benefits from at least one social protection program. The outreach of social protection is low in the Colombo and Rathnapura districts, even for extremely poor households. Nearly 30 percent of extremely poor households in all other districts did not receive Samurdhi benefits, which is the main poverty alleviation program. Nearly 20 percent of poor households receive benefits from social protection in all the districts. In contrast, a percentage of vulnerable and non-poor households receive social protection benefits, especially from the government servant pension scheme. Hence, it is essential to increase the outreach of poverty alleviation programs at least covering all the extremely poor and poor households while providing risk-cooping social protection for vulnerable and non-poor households;
• The impact of social protection was calculated according to the percentage of households that increased their poverty level if they did not receive social protection benefits. Only the households that are on the margin of the poverty line increase their poverty level when social protection benefits are excluded. However, a significant percentage of non-poor households push down low-income categories when pension benefits are removed. The impact of social protection is low because of the inadequacy of the benefits. Additionally, major risks and reasons for poverty are not adequately covered or are not covered by the existing social protection system.

5.2. Policy Implication

The existing social protection system has a marginal impact on reducing poverty since the current system did not accurately target the heterogeneity of spatial factors. The current poverty alleviation programs do not provide adequate benefits so that households can get rid of poverty in a reasonable time. Hence, policymakers should take necessary action to adjust social protection benefits according to the formula that includes the inflation rate. Accordingly, the threshold income level that selects the beneficiaries of social protection should also adjust according to the recent poverty line. Similarly, it is essential to select beneficiaries annually, targeting only needy people. Correspondingly, it is important to consider district-level poverty lines to identify poor households since it reflects the spatial disparities rather than the national-level poverty line.

Reasons for poverty and the nature of poverty are different for different income groups. Hence, it is worthwhile to design different poverty alleviation programs by targeting the heterogeneity of poverty.

Accordingly, the existing social protection system does not cover some risks and reasons for poverty. Therefore, it is important to revise and expand the existing system so as to cover all the reasons and risks that cause poverty. As an example, the cost of natural disasters, human–animal conflicts, and pest diseases are not covered by the existing system. Therefore, new social protection programs should be implemented for uncovered risks by considering the major risks in different geographical areas. These types of formulation also help to increase the sustainability of social protection programs by decreasing social protection expenditure.

Furthermore, spatial factors affecting poverty are closely related to infrastructure facilities, resource endowment, and economic opportunities. Hence, it is essential to enhance these factors to eradicate poverty sustainably while providing social protection facilities to boost economic and social well-being in the short term. Apart from that, macroeconomic stability is also an essential factor that needs to reduce poverty in a sustainable manner. Therefore, policymakers should take necessary actions to reduce the inflation rate and unemployment and implement appropriate monetary and fiscal policies to reduce poverty and vulnerabilities.

5.3. Limitations of the Study and Further Research Options

Although this study examines the nature of poverty, the determinants of poverty, and the impact of social protection on poverty in different areas in Sri Lanka, the findings are subject to some limitations. This study selected only the marginalized areas in the selected districts as the sample area. However, the nature of poverty and the factors that affect poverty may be different in other areas. Accordingly, the impact of social protection on poverty was examined by considering only the main social protection programs provided by the government. However, there are some social protection programs provided by the private sector and non-governmental organizations to reduce poverty. According to the availability of data and other constraints, this study used a straightforward methodology to analyze the impact of social protection programs. Further, this research tested the significance of selected factors affecting poverty. However, some other factors may affect the heterogeneity of poverty.

As explained above, this study contains some limitations. Hence, further research can be conducted to overcome these limitations. Poverty is not only a phenomenon in marginalized areas. It may be a problem in other areas also. Hence, it is necessary to examine the nature of poverty and the impact of social protection in other areas to eradicate poverty sustainably. This study considers only the selected major social protection programs to evaluate the impact of social protection on poverty. However, it is worthwhile to conduct research by considering all the social protection programs to evaluate the impact of those programs on poverty so policymakers can formulate better policies. Further studies can also be conducted using a more comprehensive methodology to analyze the impact of social protection programs. The propensity score-matching techniques are recommended to analyze the impact of social protection programs. Accordingly, the present study did not consider some natural factors such as soil type, slope of land, and climate that can affect household poverty. Therefore, further studies can be conducted, including those factors and other possible factors that may affect poverty.