Impacts of productive safety net program on the livelihoods of rural households

TABLE OF CONTENTS

STATEMENT OF AUTHOR

AUTHOR

BIOGRAPHICAL SKETCH

ACKNOWLEDGEMENTS

ACRONYMS AND ABBREVIATIONS

LIST OF TABLES

LIST OF FIGURES

LIST OF TABLES IN THE APPENDICE

ABSTRACT

1. INTRODUCTION
1.1. Background of the study
1.2. Statement of the problem
1.3. Research Questions
1.4. Objective of the study
1.5. Significance of the study
1.6. Scope and Limitation of the Study
1.7. Organization of the thesis

2. LITERATURE REVIEW
2.1. Introduction
2.2. Theoretical Literature Review
2.2.1. Definitions and Concepts of PSNP and livelihood
2.2.2. Productive safety net program of Ethiopia
2.2.3. Indicators of sustainable livelihoods
2.2.4. PSNP and Risk Management
2.3. The Empirical Studies of the Impact of PSNP
2.3.1. The Social and Economic Impacts of PSNP in Africa
2.3.2. Social and Economic Impacts of PSNP in Ethiopia
2.4. Impact assessement methods
2.4.1. Experimental evaluation method
2.4.2. Non experimental evaluation method
2.4.3. Quasi Experimental evaluation method
2.4.4. Methodologies to construct counter factual groups
2.4.5. Why PSM method for the study
2.4.6. Steps in the applications of PSM method

3. RESEARCH METHODOLOGY
3.1. Description of the study area
3.1.1. Climate and agroecology
3.1.2. Land use and farming system
3.1.3. Livestock resource
3.1.4. Rural finance
3.1.5. Crop production
3.2. Descriptions of PSNP in the woreda
3.3. Sources and Methods of Data Collection
3.4. Sample Size and Method of sampling design
3.5. Methods of Data Analysis
3.5.1. Descriptive Data Analysis
3.5.2. Econometric Analysis
3.6. Definition and measurement of Variables
3.6.1. Dependent variable
3.6.2. Outcome variables
3.6.3. Independent variables
3.7. Model Diagnostics
3.7.1. Multicollinearity Test
3.7.2. Hetroscedasticity Test

4. RESULT AND DISCUSION
4.1. Descriptive results
4.1.1. Demographic characteristics of sample households
4.1.2. Description of sample households for categorical variables
4.1.3. Descriptive statistics of outcome variables
4.2. Econometric results
4.2.1. Propensity scores estimation
4.2.2. Imposing common support region
4.2.3. Choosing a matching algorithm
4.2.4. Balancing test
4.2.5. Treatment effect on the treated
4.2.6. Sensitivity analysis

5. CONCLUSION AND RECOMMENDATIONS

5.1.Conclusion

5.2.Recomendation

REFERENCES

APPENDICES

APPROVAL SHEET

UNIVERSITY OF GONDAR

COLLEGE OF AGRICULTURE AND RURAL

TRANSFORMATION ,

DEPARTMENT OF AGRICULTURAL ECONOMICS

As thesis research advisors, we hereby certify that we have read and evaluated this thesis prepared under our guidance by Tsegaye Denberie Tesfaye “The Impacts of Productive safety net program on livelihood of rural households: The case of Libokemkem woreda, South Gondar Zone, Ethiopia”. The PSM Estimator model Approach: We recommend that it be submitted as fulfilling the thesis requirement.

illustration not visible in this excerpt

As members of the examining board of the final M.Sc. thesis open defense examination, we certify that we have read and evaluated the thesis prepared by Tsegaye Denberie Tesfaye and examined the candidate. We recommended that the thesis be accepted as fulfilling the thesis requirement for the Degree of Master of Science in Agricultural Economics.

DEDICATION

This thesis work is dedicated to my wife Hiwot Bayu, my sister Sefinesh, my mother Shashitu Getaneh and my father Denberie Tesfaye who lost his life at the beginning of my study , for their unconditional and unbounded love, patience and strength that helped me to complete this work.

STATEMENT OF AUTHOR

I declare that this thesis is my own work and that all sources of materials used for this thesis have been duly acknowledged. This thesis has been submitted in partial fulfillment of the requirements for M.Sc. degree at the University of Gondar and is deposited at the University Library to be made available to borrowers under rules of the library. I solemnly declare that this thesis is not submitted to any other institution anywhere for the award of any academic degree, diploma, or certificate.

Brief quotations from this thesis are allowable without special permission provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the School of Graduate studies when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

Name: Tsegaye Denberie Tesfaye Signature Abbildung in dieser Leseprobe nicht enthalten

Place: University of Gondar, Gondar

Date of Submission:

BIOGRAPHICAL SKETCH

The author was born in May 1990 in South Gondar zone, Debretabour town. He attended his elementary school education at Kimir Dingay town and his secondary school education in Debretabour town Senior Secondary School. Then he joined the Haramay University in 2008 and graduated in July 2010 with Bsc degree in Agribusiness management. Soon after graduation, he has been working for five years in Addis Zemen town in various capacities. Then after, he joined the school of graduate studies at University of Gondar in October 2014 to pursue his M.Sc degree in Agricultural Economics.

ACKNOWLEDGEMENTS

I thank God, for his permission and for helping me to have a good health and courage in accomplishing this research work. I am grateful to many people who have assisted me in doing this research. However, it is nearly impossible to give the full account of all individuals and organizations because of space limitations. The following ones, however, deserve special considerations in the sequential learning process I have passed through.

First, I would like to take this opportunity to thank all my family especially my mother Shashitu Getaneh, my father Denberie Tesfaye, my sisters for their all-round support. I would like to express my sincere appreciation and gratitude to my major advisor Dr. Zemen Ayalew whose inspiring guidance, encouragement, and understanding have vitally contributed a lot to this study and also for his friendly approach towards me that enabled me to express my idea freely. Also, I am most grateful to my co-advisor Getahun Abreham and instructor Abebe Dagnaw for their useful and valuable comments on the main idea of the thesis at an early stage of the work and for their subsequent and unreserved technical support in the whole process of this study. For both of them I could say that no words can express my gratitude for the frequent assistance and close supervision throughout my thesis work.

I would like to say thank you to Development Agents for their help especially in the process of collecting both secondary and primary data. At last, I would like to say thank you very much from the bottom of my heart for my friends Hiwot Bayu, Delelegn Adugna ,Mequanint Tesfaw and Abreham Assefa for their support as a true friend.

LIST OF ACRONYMS AND ABBREVIATIONS

illustration not visible in this excerpt

LIST OF TABLES

1: Land use in the woreda

2: Sample size by kebeles

3: Summary of variables definition and measurement

4: Demographic characteristics of the household heads

5. Description of sample households for categorical variables

6. Description statistics of outcome variables

7: Multicolinearity test for explanatory variables included in the multiple regression

8: Logit results of household program participation

9: Distribution of estimated propensity scores

10: Comparison of the three matching estimators by performance criteria

11: Balancing test results of covariates using kernel band width matching estimator

12: Average treatment effect on treated (ATT)

13: Sensitivity analysis

LIST OF FIGURES

Figure 1: Components of household livelihood security .

Figure 2: PSM-implementation steps..

Figure 3: Map of study the area

Figure 4: region of common support condition

Figure 5: Kernel density of propensity score of all households

Figure 6: Pscore of treated in common support after matching

Figure 7: Pscore of control in common support after matching

LIST OF TABLES IN THE APPENDICES

1: Conversion factor of Tropical Livestock Unit (TLU)

2: Multicolinearity test for continuous Explanatory Variables

3: Heteroskedasticity test

4: Average treatment effect on treated (ATT)

5: Results of sensitivity analysis on ATT results of outcome variables

6: Survey Questionnaire.

IMPACT OF PRODUCTIVE SAFETY NET PROGRAM ON THE LIVELIHOOD OF RURAL HOUSEHOLDS: THE CASE OF LIBO KEMKEM WOREDA OF AMHARA REGIONAL STATE, ETHIOPIA

ABSTRACT

This study evaluated the impact of productive safety net program on the livelihood of rural households of Libo Kemkem woreda. Towards this end, data were collected from 210 randomly selected households of which 119 were program participants and 91 were non-program participant’s selected from four Kebeles of the woreda, where the productive safety net program was implemented.

Data were analyzed using descriptive statistics and econometric analysis. Results from descriptive statistics revealed that among program participants and non participants, the total annual income has increased averagely by 14467.2 birr and 11469.2 birr. The average livestock holding was 3.7230 TLU and 1.4878 TLU for participant and non-participant households, respectively. Thus, the program enables them to through avoidance of forced disposal in response to shock (increase) their livestock holdings.

Applying a propensity score matching technique, it was found that the program has significantly increased participating households’ total income by 59.1%, livestock asset by 14.09% and consumption expenditure by 22.61% compared to non-participating households.

The estimated results also revealed that, households in the program has better access to credit, small land size and better access on agricultural extension, access to aid and less access to irrigation. Finally, physical and biological conservation measures should be widely incorporated, access to extension service for the utilization of new technologies and for policy concern. Generally both households increase their livelihood activities respectively interms of livelihood.

Key words : Productive safety net, impact, livelihood, propensity score matching, Ethiopia.

1. INTRODUCTION

1.1. Background of the Study

Over the past decade, Ethiopia has experienced significant economic growth and progress towards Millennium Development Goals (MDGs). Ethiopia’s annual GDP growth averaged 10.3% between 2004 and 2012, a period that also corresponded to an impressive decrease in poverty rates: in 2004, 38.9% of the population lived below the national poverty line ($0.60 per day); in 2012, this figure was down to 29.6%, (Growth and Transformation Plan, (2010 – 2014).

In the past few years the GDP growth composition by sector shifts from agriculture to other sectors, in 2012 agriculture contributes 46.6%, industry 14.6% and services sector contributes 38.8% of the GDP growth (MoFED, 2012).

However, this impressive growth has been accompanied by inflation pressure, rain fall dependence, natural hazard, which are the main driving force that affects millions of Ethiopian people,(DPRD and MoFED, 2012).Yet much of Ethiopia’s economy depends on agriculture, which is conditioned by adequate and reliable rainfall. Over the year, scanty and erratic rainfall has led to significant drought and subsequent famine. Every year five million people exposed to chronic and transitory (seasonal) food insecurity in particular to rural area (WFP, 2011).

According to DPRD and MoFED (2012), poverty remains widespread in Ethiopia. Using a consumption-based measure of poverty, 38.7 percent of Ethiopians were poor in 2009/2010, implying that 29.2 million people were living below the poverty line. Poverty is slightly higher in rural areas (39.3 percent) than it is in urban areas (35.1 percent).

The government of Ethiopia’s current Growth and Transformation Plan (GTP) aims to enhance productivity and production of smallholder farmers and pastoralists; strengthen marketing systems; improve participation and engagement in livelihood pathways; and reduce the number of chronically food insecure households. Spending on “pro-poor” sectors (health, education, agriculture and natural resources, and rural roads and urban construction) has increased from 52% of general government expenditure in FY 2003 to 70% in FY 2011/12 (MOFED, 2012).

The launch of the PSNP in 2005 represented a pivotal departure from the cycle of annual emergency food aid appeals. Following the drought of 2002/2003, the Government of Ethiopia formed the New Coalition for Food Security to identify key actions to break the cycle of emergency appeals—which saved lives but did little to protect household assets—and comprehensively address food insecurity in Ethiopia. This process resulted in the creation of the Food Security Programme (FSP). Launched in 2003, the FSP was funded by the government of Ethiopia and development Partners and implemented, mostly through government structures, in Amhara, Oromiya, Tigray and Southern Nations, Nationalities and Peoples Region (SNNP), with Harari and Dire Dawa added in 2005. During these first two phases (2005-2009), the FSP comprised three complementary components: the Productive Safety Net Program, the Other Food Security Program, and the Land Access Programme (resettlement).

Amhara Regional State is one of the beneficiary regions in the country. This Service was also initiated with the objective of improving the livelihoods of chronically food insecure households in the PSNP target Woredas through diversifying livelihood options. Therefore, out of 167 rural woredas in the region, the program has been implementing in 64 chronically food insecure weredas (PIM, 2015).

Therefore, the program aimed to help the poor and the vulnerable in coping with the negative impact of the recurrent drought in the study area. This study evaluates the performance of the productive safety net program on the dynamics of household livelihood in the rural kebeles of the woreda beneficiaries.

In this study, improvement in the livelihoods of rural households is assessed by examining the impact of household participation in the program. Household participation is expected to improve household income, consumption and asset holdings, which are used in this study as indicators of the outcomes of the program.

Moreover, the Program is designed to protect household assets and ensure a minimum level of food Consumption. It is also designed to encourage households to increase incomes generated from agricultural activities and to build up assets.

1.2. Statement of the Problem

The interest in developing a safety net program in Ethiopia grows out of the fact that the emergency system in Ethiopia was failing to stabilize livelihoods. Lives were being saved but, livelihoods continued to erode. As a consequence, more and more people were in need, resulting in an overwhelming humanitarian caseload. Each emergency resulted in further asset loss and destitution. As the population lost productive assets and became less able to cope, minor events had the impact of major shocks (SC-UK, 2008). Both government and donors became convinced of the need to this program. For donors, increased interest in budget support mechanisms as well as growing support for social protection also played a part. For the government, concern that the emergency response system was encouraging dependency syndrome and unease about Ethiopia’s image as a ‘basket case’ were strong incentives (SC-UK, 2008).

In Amhara region in general, and in Libo Kemkem woreda in particular, smallholder farmers are characterized by subsistence production and suffered from complex and interrelated socio-economic problems. Shortage of farm land, recurrent drought, and environmental degradation are the most significant problems that challenge the lives of the population (PSNP PIM, 2015). Twenty two kebeles out of the twenty nine kebeles of the woreda are classified as food insecure kebeles. Among others, the regional government has allocated huge amount of resources to protect the rural communities’ asset depletion and diversifications of rural income of households. Although efforts have been made to raise agricultural crop yield, the food insecurity problem is still a major challenge in the woreda, in particular. To increase the productivity of land, the office of agriculture has been promoting adoption and diffusion of improved technologies by farmers. In particular, farmers have been advised to adopt several physical soil conservation measures.

The PSNP is a public program through which food-insecure people are employed in public work for five days a month during the agricultural slack seasons. This is intended to enable households to smooth consumption so that they will not need to sell productive assets in order to overcome food shortages. The public work is also intended to create valuable public goods; moreover, by reducing seasonal liquidity constraints, it is intended to stimulate investments (Anderson et al., 2009).

The program is mainly targeted to help poor farmers who are susceptible to food insecurity about six or more months in a year even though crop failure is not reported. Increasing household asset and preventing asset depletion of the households and community asset building are major points targeted to improve. The proposed study area, Libo Kemkem woreda, is among the 22 woredas identified as chronically food insecure and eligible for the PSNP at the national level.

Despite the fact that the PSNP has been implemented since 2005 in the country to address the problems and shortcomings of the previous practice of assistance that focuses only on saving lives, evaluation of the effects of such programs is not yet given due attention it deserves. But the effectiveness of the program in terms of diversifying livelihood has not been studied in the study area. This study, therefore, attempt to fill this research gap by conducting an empirical study on the impact of the PSNP on farm households’ livelihood in rural Kebeles of the woreda beneficiaries.

1.3. Research Questions

The study attempted to address the following research questions:

- What are the factors influencing rural households’ participation in the PSNP?
- What impacts do the PSNP schemes have on the livelihood (income, livestock holding and consumption) of households in the study area of the beneficiaries?

1.4. Objectives

The general objective of this research is to analyze the impacts of productive safety net program on the livelihood of rural households in Libo Kemkem woreda.

The specific objectives of the research are:

a) To identify factors affecting household’s participation in the productive safety net program.

b) To examine the impact of the PSNP on livelihood of rural beneficiary households.

1.5. Significance of the study

The study contributes to awareness of the impact of PSNP on rural livelihoods and its success in achieving its goals. In other words, it was hoped that this study contributes to the understanding of the impact of PSNP for different stakeholders as well as for anyone who want to use it. In addition, it informs some realities both to the community and policy makers and implementers how to achieve success in livelihoods. Furthermore, the study serve as a bridge for other studies in the future on same and other related issue. The analysis carried out through a comparative assessment of program outcomes of participant households with outcomes of non-participant households. It is also essential for community based organizations working in the study area and other areas with similar socioeconomic settings.

1.6. Scope and Limitations of the Study

Even though the concern of the study, that is productive safety net program is the largest social protection program operating in sub-Saharan Africa, this study is only limited to assessing its impacts on livelihood in four selected rural Kebeles of the woreda. Despite such limited scope, results of the study provide insights into how the program is contributing to its major objectives.

Methodologically, the study uses PSM to assess the impact of PSNP on the rural livelihoods. In doing so, it uses data from non-program participants in order to compare some outcome variables with the result of program participants. However, it can be difficult to find a comparison group (and often an observable) determination and ability that lead the households to join the program. Therefore, the study was undertaken to meet its objectives within the limitations mentioned.

1.7. Organization of the Thesis

The thesis is organized into five chapters. Following this introduction part of the study, the remaining chapters are organized as follows. The second chapter presents review of relevant literature. The third chapter deals with the research methodology. The fourth chapter presents results and discussion. Finally, the fifth chapter presents the conclusion and recommendations of the study.

2. LITERATURE REVIEW

2.1 Introduction

This chapter presents key concepts, theoretical explanations and research findings related to this research. This chapter emphasis on the concepts of productive safety net program, rural livelihood and the existing policies strategies of PSNP, about concepts and approaches of impact evaluation and it also presents empirical studies on the impacts of PSNP on rural livelihood.

2.2 Theoretical Literature Review

Different Economists proposed several theories of investment over different time periods. Therefore, this section reveals definitions and concepts of PSNP and some of the very prominent theoretical literatures on PSNP, rural livelihood and existing policies, strategies and guidelines on PSNP.

2.2.1 Definitions and Concepts of PSNP and livelihood

The Productive safety net program (PSNP) aims to reduce the number of people who rely on annual humanitarian appeals, by providing predictable and timely cash and food (PSNP-PIM, 2015). It aims to shift away from a focus on short-term food needs met through emergency relief to addressing the underlying causes of household food-insecurity.

Households (HHs): CSA defines household as a collection of a persons who normally live together in the same unit or group of housing units and who have common cooking arrangement. The household is the basic unit of analysis in many social, microeconomic and government models. The term refers to all individuals who live in the same dwelling. In economics, a household is a person or a group of people living in the same residence. (CSA, 2012)

Graduations: from the PSNP is defined as a households being able to feed itself for 12 months a year, in the absence of program support, as well as being able to withstand modest shocks (PSNP-PIM, 2010).

Livelihood: The concept of livelihood is widely used in contemporary writings on poverty and rural development, but its meaning can often appear elusive either due to vagueness or to different definitions being encountered in different sources (Ellis and Tengberg, 2000). Moreover, a recent review of livelihoods approaches shows that definitions are far from uniform and prescriptive but are instead constantly evolving and developing. This allows for imaginative adaptations to be made as required, but also renders the concept and use of a livelihoods approach rather difficult to grasp (FAO, 2001). A popular definition is that provided by (Chambers and Conway ,1992) where in a livelihood comprises the capabilities, assets (including both material and social assets) and activities required for a means of living. Briefly, one could describe a livelihood as a combination of the resources used and the activities undertaken in order to live (DFID, 1999)

Household livelihood security: Household livelihood security is defined as adequate and sustainable access to income and resources to meet basic needs (including adequate access to food, potable water, health facilities, educational opportunities, housing, time for community participation and social integration). Livelihoods can be made up of a range of on-farm and off farm activities which together provide a variety of procurement strategies for food and cash. Thus, each household can have several possible sources of entitlement which constitute its livelihood. These entitlements are based on the household's endowments and its position in the legal, political and social fabric of society (Drink water and McEwen, 1992). The risk of livelihood failure determines the level of vulnerability of a household to income, food, health and nutritional insecurity. Therefore, livelihoods are secure when households have secure ownership of, or access to, resources and income earning activities, including reserves and assets, to offset risks, ease shocks and meet contingencies (Chambers, 1989).

A livelihood is sustainable, according to Chambers and Conway (1992), when it "can cope with and recover from the stress and shocks, maintain its capability and assets, and provide sustainable livelihood opportunities for the next generation...” Unfortunately, not all households are equal in their ability to cope with stress and repeated shocks. Poor people balance competing needs for asset preservation, income generation and present and future food supplies in complex ways (Maxwell and Smith, 1992). People may go hungry up to a point to meet another objective. For example, (De Waal 1989) found that during the 1984/85 famine in Darfur, the Sudan, and people chose to go hungry to preserve their assets and future livelihoods. People will tolerate a considerable degree of hunger to preserve seeds for planting, to cultivate their own fields or to avoid selling animals.(Corbett 1988), in exploring the sequential ordering of behavioral responses employed in periods of stress, found that in a number of African and Asian countries preservation of assets takes priority over meeting immediate food needs until the point of destitution. Thus, food and nutritional security are subsets of livelihood security; food needs are not necessarily more important than other basic needs or aspects of subsistence and survival with in households. Food-insecure households juggle among a range of requirements, including immediate consumption and future capacity to produce.

illustration not visible in this excerpt

Figure 1: Components of household livelihood security

Source: CARE USA (2012)

2.2.2. PSNP in Ethiopia

The overall objective of the program is “Food security for those who are able, and food sufficiency for those unable to achieve food security, for male and female members of chronically food insecure households in chronically food insecure areas achieved”. The Productive safety net program classified Food insecurity as chronic or transitory. Some other literatures also include cyclical type of food insecurity as a third kind of food insecurity.

Chronic food insecurity: Chronic (permanent) food insecurity is a continuously inadequate diet resulting from lack of resources to produce or acquire food, or households that are regularly unable to produce or purchase enough food to meet their food needs, even during times of normal rain, are considered chronically food insecure. Structural factors contributing to chronic food insecurity include poverty (as both cause and consequence), the fragile natural resource base, weak institutions and unhelpful or inconsistent government policies. It is argued that chronic food insecurity at the household level is mainly a problem of poor households in most parts of the world (FAO, 2002).

On the other hand , transitory food insecurity ፡ transitory food insecurity is a temporary decline in the household to access enough food (World Bank, 1986; Reutlingen, 1987). When a shock has depleted the food stores and current incomes streams of household to the point that they are unable to meet their immediate food needs, these households are described as transitory food insecure. It results from a temporary decline in household access to food due to crop failure, seasonal scarcities, temporary illness or unemployment, instability in food prices, production, household income or combination of these factors. But, the main triggers of transitory food insecurity in Ethiopia are drought and war. Finally, the cyclical type of food insecurity is caused by seasonality (Osmani 2001, FAO 2006). The PSNP includes measures to protect against transitory food insecurity, and transitory food insecurity is the focus of the emergency relief system.

In general, a household said to be food secure only if it has protection against all kinds of insecurity. The average access to food over the long term should be nutritionally adequate, and a household should be able to cope with short-term vicissitudes (changes) without sacrificing the nutritional needs of any of its members. Finally the concept and definition of food security were developed and clearly explained based on the growing hunger, food insecurity and malnutrition situations in developing countries. From the definitions of food security, slight variations were observed. However, the overall basic principles and definitions of food security, that is, “availability, access and utilization” were stressed in the definitions cited above. Therefore, for the purpose of this study, the definition put forward by Word Food Summit (1996) was taken as a working definition of food security and the household level is considered as the key unit of food security analysis.

2.2.3. Indicators of sustainable livelihoods

There are five key elements that can be recognized, each relating to a wider literature with, established ways of evaluating outcomes. Linking concerns over work and employment with poverty reduction with broader issues of adequacy, security, well-being and capability elements focus on livelihoods. And Livelihood adaptation, vulnerability and resilience, and Natural resource base sustainability elements add the sustainability dimension (Ian Scoones).

A. Creation of working days: This relates to the ability of a particular combination of livelihood strategies to create gainful employment for a certain portion of the year. This may be on or off-farm, part of a wage labour system or subsistence production. Sen (1975: 5) notes three aspects of employment – income (a wage for the employed), production (employment providing a consumable output) and recognition (where employment provides recognition for being engaged in something worthwhile). In terms of the income/production aspects, various target levels have been suggested, but 200 days a year appears to be widely used as a minimum level to create a livelihood (Lipton 1991; 1993). Overall, the number of livelihoods created will be dependent on the proportion of the population available for work.

B. Poverty reduction – The poverty level is a key criterion in the assessment of livelihoods.

Various measures can be used to develop an absolute ‘poverty line’ measure based on income or consumption levels (Ravallion 1992; Baulch 1996). Alternatively, relative poverty and inequality can be assessed using Gini coefficient measures. There are a range of pros and cons for each measure, as well as some major measurement challenges (Greeley 1994). However, such quantitative assessments of poverty can be used in combination with more qualitative indicators of livelihoods (Jodha, 1988; Schaffer 1996).

C. Well-being and capabilities – The notions of ‘well-being’ (cf. Chambers 1995; 1997) and ‘capability’ (Sen 1984; 1987) provide a wider definitional scope for the livelihoods concept. Sen sees capabilities as ‘what people can do or be with their entitlements’, a concept which encompasses far more than the material concerns of food intake or income. Such ideas represent more than the human capital which allows people to do things, but also the intrinsically valued elements of ‘capability’ or ‘well-being’. Chambers (1997) argues that such a well-being approach to poverty and livelihood analysis may allow people themselves to define the criteria which are important. This may result in a range of sustainable livelihood outcome criteria, including diverse factors such as self-esteem, security, happiness, stress, vulnerability, power, exclusion, as well as more conventionally measured material concerns (Chambers 1989).

D. Livelihood adaptation, vulnerability and resilience – The ability of a livelihood to be able to cope with and recover from stresses and shocks is central to the definition of sustainable livelihoods. Such resilience in the face of stresses and shocks is a key to both livelihood adaptation and coping (Davies, 1996). Those who are unable to cope (temporary adjustments in the face of change) or adapt (longer term shifts in livelihood strategies) are inevitably vulnerable and unlikely to achieve sustainable livelihoods. Assessing resilience and the ability to positively adapt or successfully cope requires an analysis of a range of factors, including an evaluation of historical experiences of responses to various shocks and stresses. Different types of shock or stress, in turn, may result in different responses, including avoidance, repartitioning, resistance or tolerance mechanisms (Payne and Lipton 1994: 15).

2.2.4. PSNP and risk management

The PSNP provides cash or food to people who have predictable food needs in a way that enables them to improve their own livelihoods and manage risks today; and therefore become more resilient to the effects of shocks in the future. Independent studies have shown that the PSNP has reversed the pre-2005 trend of decade-on-decade deterioration in livelihoods. The PSNP has shown that providing timely and predictable assistance enables households to manage risk more effectively by preventing costly coping strategies such as sale of vital assets that worsens future food insecurity. The PSNP both protects households from food insecurity and allows them to use their resources more flexibly to smooth out consumption.

However, while the PSNP responds to the chronic food insecurity of households, there are times when a shock results in some households whether within the PSNP or not - facing transitory food insecurity and requiring additional temporary support. In these instances, the PSNP has dedicated Contingency Budgets, designed to meet transitory needs. However, if a shock is too large, the PSNP’s contingency funds can be exhausted before all the transitory needs are met. When the contingency funds are exhausted, the Risk Financing Mechanism (RFM) is designed to address these needs. The RFM is an instrument that allows the PSNP to scale up in times of transitory crisis, in those districts where it is already operational. In particular, the RFM was designed to reduce the ‘typical’ humanitarian timeline for response, so that households would receive assistance before the crisis was felt. In this way, the PNSP can expand and respond as the situation requires. The program can address predictable food needs through usual PSNP operations, can address low-level transitory needs caused by moderate shocks through contingency funds and can address higher levels of transitory needs through the RFM.

According to World Bank, (2013) report, in order for the RFM to function correctly, four conditions need to be fulfilled. These are: Effective early warning systems need to be in place to indicate the need for a response as early as possible (Early Warning); Plans need to be put in place so that when a shock is indicated, key actors know how to respond(Contingency Plans); Resources need to be available to avoid the major time delays associated with the humanitarian appeal process(Contingency Financing) and Institutional arrangements and capacity need to be in place to allow plans to be implemented(Institutions and capacity).

By putting in place effective early warning systems, contingency financing, contingency plans and institutional capacity ahead of the crisis, the ‘typical’ timeline for humanitarian response can be significantly reduced, from 8-9 months to 2 months, as was the case in 2011, when the Horn of Africa was affected by the largest drought in 60 years. In August 2011, Ethiopia used RFM to address the transitory food needs of approx. 9.6 million drought-affected people (World Bank, 2013).

Addressing transitory food insecurity in addition to chronic food insecurity is integral to the transition from relief to development in Ethiopia. With increased vulnerability as a result of climate change, the capacity of communities and Government to manage risks – already being built by the PSNP is becoming increasingly important.

2.3. The Empirical Studies of the Impact of PSNP

There are some empirical studies that have been conducted by different researchers to assess the Impact of PSNP in Ethiopia. Among these studies some of the works tried to assess the impact of the program one year after the onset of the program using cross sectional data - examples include Devereux et al. (2006) and Gilligan et al. (2008). But according to Devereux et al. (2006), since impact might not accrue in the short run, to fully and rigorously evaluate the PSNP, longitudinal Data is needed. Even though some literature did a panel data analysis they did not focus on welfare (poverty), for instance Anderson et al. (2009) and other authors such as Wheelers and Devereux (2010) examined only a change in beneficiary‘s status in time without taking the counterfactual situation.

According to Yibrah (2010) who analyzed the impact of PSNP on rural household’s asset protection and consumption using PSM technique, Productive Safety Net Program intervention enables beneficiary households to retain their assets holdings. The asset values of the PSNP beneficiary households have exceeded that of the non-PSNP beneficiary households. The PSNP beneficiary households, as a result of PSNP intervention, have increased their livestock holdings. Thus, the program enables them to protect (increase) their livestock holdings. The result of this study found that the mean difference of the livestock holdings, in terms of TLU, between the PSNP beneficiary households and the non-PSNP beneficiary households was positive and significant. Therefore, this study will be conducted to evaluate the impact of PSNP on food security, property possession, annual income and consumption expenditure of households’ using propensity score matching technique.

Andersson et al. (2009) analyzed the impact of PSNP on livestock and tree holding of rural household in Ethiopia. The study found that there was no indication of participation in PSNP leads households to disinvest in livestock or tree. In fact, the number of trees increased for households that participated in the program. It could be the case that participation in PSNP (where tree planting and subsequent forest management work on public lands are usual activities) leads to households becoming more skilled in forestry, and that they switch to increased forest planting as a result.

Nonetheless, per the impact evaluation conducted by International Food Policy Research Institute (IFPRI) in 2009 in 68 PSNP Woredas in Tigray, Amhara, Oromia and SNNP regions using a longitudinal (panel) household and community data collected and matching methods, participation in the public works component of the PSNP (defined as receipt of at least 100 birr in payments over the first five months of 2006, 2007, and 2008) has modest effects. It improves food security by 0.40 months and increases growth in livestock holdings by 0.28 Tropical Livestock Units (TLU). Relative to non-beneficiaries, beneficiary households perceive that their welfare has improved (Gilligan et al., 2009).

2.3.1. The Social and Economic Impacts of PSNP in Africa

Different studies have been carried out on the impacts of the social safety net and transfer issues in different countries of Africa. Some of them are Devereux, (2002) assessed the cash transfers intervention in Namibia (social pensions), public works in Zambia, and Mozambique (cash payments to urban destitute). According to this study, the program had identified different poverty and other economic and social outcomes of these income transfers.

Miller et al., 2010, in Malawi, employed both descriptive and econometric techniques of difference-in-differences estimates to analyze the impact of cash transfer on household food security. The results from his study show that intervention households in Malawi allocated 62% of total expenditures to food purchases and the recipients were able to reach what they reported as an acceptable level of food security.

2.3.2 Social and Economic Impacts of PSNP in Ethiopia

In Ethiopia, the PSNP is already having a significant impact and there is clear evidence that several important changes have taken place in terms of nutrition, attitudes, and risk-taking behaviors’, particularly in terms of food consumption, asset protection, asset building, and allowing people to feel secure enough in their income to take productive loans which they previously found too risky (Rachel S., Steve Ashley and Mulugeta T, 2006).

Graduation processes are complex and cannot simply be delivered through a safety net programme alone. Although public work is meant to prevent dependency on the PSNP, findings suggest that it may in fact do the opposite for households with higher numbers of non-workers such as children, people with disabilities and the elderly. The labour requirements of the PSNP draw labour away from households’ own livelihood activities and affect their choice of packages. There is a danger that households become more, not less, dependent on the PSNP because the work requirement reduces their ability to pursue successful alternative livelihood activities.

This suggests that PSNP, especially when transfers are issued as cash, is helping households achieve their wider objectives in terms of investments in human capital (www.wahenga.net lessons from Ethiopia on a scaled-up national safety net programme).

2.4. Impact Assessment Methods

Impact assessment of a designed programme intervention is to show the effect of the programme on participating group and comparator group that did not participate in the programme as a control group, but having similar pre-intervention socio-economic characteristics. Thus, estimating the impact of a programme requires separating its effect from intervening factors which may be correlated with the outcomes, but not caused by the programme (Ravallion, 2005).Impact evaluation of a given intervention programme is intended to determine more broadly whether the programme had desired effects on individual households, organizations, institutions and others as per the programme intervention design. The impact may result in positive or negative effect on beneficiaries (Baker, 1960). Generally there are three impact evaluation methods in estimating treatment group participants and control groups. These are randomization/or experimental design, non-experimental design and quasi-experimental design. Depending on the data availability, ethics to experiment and costs, social science methods deals with randomization/or experimental, non-experimental and quasi-experimental methods (Jalan and Ravallion, 2003).

2.4.1. Experimental evaluation method

Social experiments are intended to analyze policy issues how things react to a type of policy that has never been tried and one which has no available data observed. The concept of social experiment is to assess a group of willing participants, some of whom are randomly assigned to a treatment group and the rest to a control group. The term experimental refers to the group receiving treatments, control refers the group no receiving treatment and random assignment of individuals in to two groups (Colin and Pravin, 2005).

The contribution of the treatment to the outcome difference between the treated and control group can be estimated without confounding bias in the cause where one cannot control for the confounding variables. However, an outcome depends on treatment as well as other observable factors, so controlling for the latter will in general improve the precision of the impact estimate.

A random assignment of households to treatment and non-treatment groups ensures that on average any difference in outcomes of the two groups after intervention can be attributed to the intervention. In randomized experiment the problem of selection bias can be avoided as a best way of assignment in which the participation characteristics is unmeasured or unobserved. In such causes randomization takes place before the program begins (Ezemenari et al., 1999;Smith and Todd, 2005).

2.4.2. Non-experimental method

A non-experimental method is used when the program participant located intentionally. It can be used through the access of cross-sectional survey data after the program is introduced. Accordingly there are two broad categories of non-experimental approach, before and after through cross-sectional estimator. Cross-section estimators use non participants to derive the counterfactual for participants (Bryson et al., 2002).

2.4.3. Quasi-experimental method

A quasi-experimental method is the only alternative utilized where there is no baseline survey or randomization is not a feasible option and not takes place prior the intervention. It involves matching programme participants with a comparable group of individuals, who did not participate in the programme after intervention (Jalan and Ravallion, 2003; Dehejia and Wahba,2002).

2.4.4. Methodologies to construct counterfactual groups

Non experimental methods sometimes are also called statistical methods use statistical techniques to simulate the counterfactual, i.e., the outcome that would have prevailed had there been no intervention. The most frequently used non experimental methods available for evaluating development programs include propensity score matching (PSM), difference indifferences (DD), regression discontinuity design (RDD), and instrumental variables (IV).

a) Propensity Score Matching

The basic idea of the propensity score matching method is to match program participants with non participants typically using individual observable characteristics. Each program participant is paired with a small group of non participants in the comparison group that are most similar in the probability of participating in the program. This probability (called propensity score) is estimated as a function of individual characteristics typically using a statistical model such as logit or probit model. The mean outcomes of these groups of matched non participants form the constructed counterfactual outcome. The mean program impact is estimated by the difference between the observed mean outcome of the project participants and the mean outcome of the constructed counterfactual (Caliendo et al., 2005).

b) Double difference in difference

The difference in difference (or double difference) method entails comparing observed changes in and non participants using a baseline survey before the program. One then repeats this outcome before and after the project for a sample of participants and nonparticipants. Typically, one collects outcome data of both participants survey at some later point(s) after the program is implemented. This repeat survey(s) should be highly comparable with the baseline survey in terms of the questionnaire, the interview, etc. The mean program impact is estimated by comparing the mean difference in outcomes “after” and “before” the intervention between the participant and non participant groups. The underlying assumption of DD method is that project participants would have the same outcomes as individuals in the comparison group in the absence of the project. Since this is highly unlikely in reality, PSM is a natural choice to select a comparison group before calculating the differences in a DD method. For this reason, the PSM and DD methods are often used together in practice (Baker, 2000).

c) Regression discontinuity

The regression discontinuity design method can be used when program participation is determined by an explicitly specified exogenous rule. The method stems from the intuition that individuals around the cut-off point for eligibility are similar and uses individuals just on the other side of the cut-off point as the counterfactual. In other words, RDD compares outcomes of a group of individuals just above the cut-off point for eligibility with a group of individuals just below it. The major technical problem of the RDD method is that it assesses the marginal impact of the program only around the cut-off point for eligibility, and nothing can be said of individuals far away from it. In addition, for the RDD estimate to be valid a threshold has to be applied in practice and individuals should not be able to manipulate the selection score to become eligible (ADB, 2006).

d) Instrument variables

The instrumental variables method works exactly as a standard regression analysis. When the program placement is correlated with participants’ characteristics, then the estimate of program effect using an ordinary least squares regression model is biased. To correct this, one needs to replace the variable characterizing the program placement with another variable(called instrument) such that it mimics the variable being replaced (i.e., correlated with the program placement) but is not directly correlated with the program outcome of interest(Felici, 2008).

2.4.5. Why PSM Method for the Study?

This method is chosen for this study because now a day’s PSM is popular method for program evaluation studies in many applications of interest due to the dimensionality of the observable characteristics is high. This matching method tries to pick an ideal comparison matching based on propensity score in which comparison group is matched with the treatment group on the basis of a set of observed characteristics or by using predicted probability of participation given observed characteristics the closer the propensity score, the better the match(Ravallion,2003).

The PSM method is very useful if there are many potential characteristics to match between a sample of treated individuals and a sample of non-treated individuals. The treatment impact is then the difference in outcomes between the treatment and comparison group (Heckman and Todd, 1997). The PSM method provides a natural weighting scheme that yields unbiased estimates of the treatment. The weights are formed as the inverse of the predicted probability that an individual would make the choice to participate in the treatment. The resulting predicted probabilities are used to create weights that are used in subsequent analyses (Baker,2000). While computing the estimated treatment effect, different matching techniques provide different weights on comparison units. The most frequently estimated parameter for such studies is the average treatment effect on the treated (ATT) which is the difference between expected outcome with and without treatment for those who have actually participated in treatment ( Caliendo and Kopeinig , 2008 ).

PSM neither requires randomization nor pre-intervention data but in practice pre intervention data is used to control for differences in individual characteristics prior to implementation of a given program (This is required if a combination of PSM and DID methods is applied). A second best is to use it in the post-intervention data only (Felici et al., 2008). Unlike econometric regression methods, it does not rely on parametric assumptions to identify the impacts of program and it does not impose a functional form of the outcome thereby avoiding assumptions on functional form and error term distributions ( Rajeeve, et al., 2002). Besides, PSM compares outcome for observation, who share similar observable characteristics using matching methods. This matching method emphasizes the problem of common support thereby avoiding the bias due to extrapolation to non-data region. Results from the matching method are easy to explain to policy makers since the idea of comparison of similar group is quite intuitive. PSM requires large amounts of data both on the universe of variables that could potentially confound the relationship between outcome and intervention, and large numbers of observations to maximize efficiency. Irrespective of its shortcomings, PSM is extensively used in the recent literature (Ravallion, 2005).

2.4.6. Steps in Application of PSM Method

In the estimation of average treatment effect on treated (ATT) using propensity score matching method first the propensity score is estimated using a logit model with maximum likelihood method to estimate the participation probability, a logit model is often preferred due to the consistency of parameter estimation associated with the assumption that error term u in the equation has a logistic distribution (Caliendo and Kopeinig, 2008). Matching estimator is selected based on the data at hand after undertaking matching quality test, overlapping condition or common support condition is identified, the treatment effect is estimated based on the matching estimator selected on the common support region. Finally, sensitivity analysis is undertaken to check the strength of the conditional independence assumption identified. Sensitivity analysis can also be undertaken to check if the influence of an unmeasured variable on the selection process is so strong to undermine the matching procedure.

illustration not visible in this excerpt

Figure 2: PSM-implementation steps

Source: Caliendo and Kopeinig, 2005

3. RESEARCH METHODOLOGY

This chapter describes the overall methodology of the research thesis. This part is divided into four sections. In the first section, the proposed area is described briefly. The second section describes about the productive safety net program in the proposed area. The third section provides information on the sources and methods of data collection while the final section discusses the methods of data analysis.

3.1. Description of the study area

The study has been conducted in Libo Kemkem woreda, South Gondar Zone of Amhara National Regional State, Ethiopia with the distance of 62 km, 80 km, and 85 km from Debre Tabor, Bahir Dar and Gondar respectively. It coordinates at 11057’-12020’N latitude and 37025’-37058’E longitude and it is bordered on the North Belesa woreda, on the south Fogera woreda, on the west Gonder zuria woreda, on the east Ebnat woreda.

It is located at the northern limit of the central highlands of Ethiopia. The landform (altitude) is complex composed of highlands (in the range of 1800 up to 2850 meters above sea level. Topographically, the woreda is characterized by rugged features, plain/flat, mountainous, and undulated which constitute 27%, 35%, 20% and 18% respectively. The land slope of the area is generally undulating to flat land; 50% and 50 % slope (MoWR, 2012).

3.1.1. Climate and agro ecology

Libo kemkem woreda has diversified agro-ecological zones and niches each with distinct soil, geology, vegetation cover and other natural resources. The climate is generally tipped moist mid highland and tipped sub moist mid highland, with the average annual rainfall amount of 900-1400 mm. Most of this rain is received during mid June to September. The rainfall pattern is predominantly uni-modal. Agro-ecologically the climate is in the woyna dega with the largest coverage 78% and dega covers 22%.Its average temperature is 11.1-27.9°C (MoWR., 2009).

illustration not visible in this excerpt

Figure 3. Map of the study area

3.1.2. Land use and Farming system

According to the woreda Environmental protection, land administration and use office annual report, the total area of the woreda is estimated to be 108,157 hectares. The proportion of areas under cultivation, grazing land forest and housing construction is indicated under Table1.Those areas that are covered by bush, shrubs and natural forest are found in the mid-altitude areas and specifically of around the church. ( WEPLAUO annual report, 2012).

Table 1. Land use in Libo kemkem woreda

illustration not visible in this excerpt

Source: WEPLAUO annual report (2012 )

The farming system of the woreda is characterized by crop-livestock mixed farming systems. Average land holding in this area is about one hectare/ 4 Timad / head (WEPLAUO annual report., 2012). Above 85% of people’s income is depending on agricultural production. The woreda is partly labeled as one of the food insecure areas in the south Gondar zone. The major food crops grown in the woreda are Teff, Maize, Wheat, Sorghum, Peas, Beans, Rice, Barley, Potato, seed, Onion, and Tomato which are used both as source of food and income and playing a great role in the Life’s of the people (WADO, 2012).

3.1.3 Livestock resource

Livestock production is an integral part of the production system. Production of cattle (milk, meat), sheep and goat (meat), asses, horse, beekeeping and poultry is a common practice in the woreda. Cattles are exported to the Sudan and used for local market while sheep, goats are mainly used for the local market. Livestock population of the District is cattle 62,609, goats 68,119, sheep 78,161, donkey 10,909, poultry 14,215 and bee hives 5712. (Livestock department, 2016)

3.1.4. Rural finance

The Amhara Credit and Saving Institution (ACSI) is the major provider of credit and saving service for the rural population. The credit repayment schedule varies from one investment type to the other. ACSI has made an agreement with the ANRS DPFSC office called the food security loan distribution agreement since 2011(2003 E.C) to distribute loan for food insecure households to increase their livelihood. Thus, Libo kemkem woreda ACSI sub-branch office in line of its organization also provides loan based on the agreement taken by the ACSI main office at Bahidar town.

Cooperative at kebele and woreda level is one of the rural finance institutions that provide credit and other services for the rural people as well.

3.1.5. Crop Production

The total area of the woreda is 108,157 hectares. A total of 34,812 hectares are used for crop production including 6,519 hectares of irrigable land. The remaining area is for grazing, forest and bushes, roads and other constructions (WoA, 2012).

3.2. Description of the Productive Safety Net Program in the woreda

The Productive safety net program (PSNP) has been implemented since 2005 in the woreda. Since the PSNP has been implemented in 22 food insecure rural kebeles of the 29 total kebeles in the study area, the program has three components; livelihood, direct support and public work component. The livelihood component provides training in the areas of marketing, business and value chain activities and preparation of effective business plan for referral to micro financial institutions to get credit. The remaining two components provide cash and grain to PSNP beneficiaries. The amount of payment was ETB 5 in the starting time of the program and has increased to ETB 41 per day per individual since 2016 in the woreda. The selection criteria of beneficiaries in the woreda as confirmed by food security task forces shows that a community selection based on asset ranking, social status (specially the lowest social status based on their wealth rank). (PSNP-PIM, 2010).

The two components except direct support components, households participate in labor intensive activities such as income generating activities, soil and water conservation activities on communal lands, afforestaion, fencing and construction of schools, construction of feeder roads, and providing local raw materials for construction. The working schedule is from January to June of each year. The participants work for five days per month for at least 6 hours per day and receive 15 kg/person plus 4kg pulse/person. However, the payment is not only for participating individuals in the household rather multiplied by the number of family members. That is, a participating household receives 15Kg of wheat or a cash multiplied by the number of family member considering children and other disabled family members, but those who are able to work should participate in the public work activities.

The PSNP is supplemented by other food security programs (OFSP) in the woreda. PSNP identification card is usually provided to PSNP beneficiary households. Loan is provided for beneficiary households based on their business plan for different livelihoods/investment packages purposes like animal production, fattening, to purchase agricultural inputs, tools and technologies, for off-farm activities.

3.3. Sources and Methods of Data Collection

Both qualitative and quantitative data have been collected from both primary and secondary sources. Households’ demographic and socio-economic characteristics are collected from the sample households by using a semi-structured questionnaire. Trained enumerators fill the questionnaire by interviewing the sample households from users (participants) and non-users (non-participants) of productive safety net program in the proposed area. Concerning households' annual income data, sample households are asked to state their annual income from crop, livestock and off-farm income generating activities. The collected values of annual income items are computed interms of birr of sample households.

Secondary data relevant for this study has been collected from various sources like Bureau of Agriculture and rural development and other relevant private and public institutions like District and Kebeles Administrations in the study area and Woreda food security programe to supplement primary data. In the formal sample survey, semi-structured questionnaire will pre-tested to elicit new information before the formal survey is carried out. Training will be given to enumerators about the questionnaire and follow up has been made to ensure that the process of data collection is smooth. Then the questionnaire has been administered to collect pertinent data.

3.4. Sample Size and Method of sampling design

A three-stage sampling technique is adopted to generate the primary data. Firstly, Libo kemkem woreda out of the five woredas in south Gondar zone, where the program had been operating, was purposively selected. Secondly, out of the twenty two Kebeles four rural Kebeles from dega and weyna dega were randomly selected. Thirdly, households in each of the four Kebeles were grouped into two strata. Stratum one represents PSNP participant and stratum two represents non PSNP participant. Finally the primary data for this study was collected from 210 households from 119 program participants and 91 non-participants in the study areas. Following this procedure, by using a formula provided by Yamane (1967) was used to determine the required sample size at 95% confidence level, 0.5 degree of variability and 9% (0.09) level of precision.

Abbildung in dieser Leseprobe nicht enthalten Where n is the sample size, N is the population size (total household heads size), and e is the level of precision. The above formula provided 118 sample sizes of PSNP participants but equal size of non- participants was selected, however due to different reasons like lack of willingness to response and dislocation of the respondents only 210 (119 participant and 91 non-participant) households were interviewed. (As shown in Table 2) .PSNP has been launched in 22 of 29 Kebeles in the Woreda. The interviews were conducted to the household heads of the sample households.

Table 2: Sample size by kebeles

illustration not visible in this excerpt

Source: own Computation result, 2016

3.5. Methods of Data Analysis

The impact analysis has been used both descriptive statistics and econometric model. Among econometric methods propensity score matching employed to quantify important empirical results. Both descriptive statistics and econometric tools were used to analyze the empirical data. Both qualitative and quantitative data’s are compiled sorted, edited, and represented with appropriate variables for encoding. After the data cleaned, information will coded, arranged into group variables, summarized, and tabulated for interpretation and analysis.

3.5.1 Descriptive Data Analysis

Descriptive statistical tools are very important to have a clear picture of the households included in the sample. Descriptive statistical techniques are employed for the purpose of describing the demographic, socio-economic structure of sample households and the impacts of PSNP on livelihood status in the study area. These analyses are conducted using descriptive statistics such as tabulation, mean, standard deviation, percentage, and to summarize, interpret and conclude the results. Socio- economic data and household attributes have been evaluated using statistical tools. The purpose is to understand the significance and magnitude of households’ livelihood activities taking situation of households program impacts. The study population was categorized using tables, mean difference, and other appropriate statistical tools.

3.5.2 Econometric Analysis

Propensity score matching model was used to address the objectives /to evaluate the impact of PSNP on livelihoods of rural households.

Propensity score matching (PSM) method: According to Khandker et al . (2010) impact evaluation is the act of studying whether the changes in well-being are indeed due to the intervention and not to other factors. The main aim of PSNP was to ensure sustainability of food insecure households in addition to improve their livelihood status. To this effect, there is a need to see whether the intervention of PSNP has significant influence on the participant households or not. However, to compare the before and after intervention difference, baseline survey was not conducted prior to the intervention of the PSNP in the study area. Therefore, this study uses PSM method because PSM is the appropriate method when such kind of problem arises.

Following Caliendo and Kopeinig (2005), there are some steps in implementing PSM. These are: PSM estimation, choosing matching algorithm, checking for overlap (common support), matching quality (effect) estimation and sensitivity analysis.

[...]