Moving Out of Poverty
Trends in Community Well-Being and Household Mobility in Nine Cambodian Villages
September 2007

Ingrid FitzGerald and So Sovannarith
with Chan Sophal, Kem Sithen and Tuot Sokphally


CDRI is pleased to present the report of the Moving Out of Poverty Study (MOPS), a major study conducted in partnership with the World Bank. MOPS is the first longitudinal, contextual, mixed-methods study on poverty dynamics conducted in Cambodia.  

The CDRI-World Bank Cambodia study was designed and commissioned initially as part of the World Bank’s global Moving Out of Poverty Study. The scope and methodology were fine tuned to better reflect Cambodian needs and development circumstances. The study provides an innovative methodology and different perspective to past national poverty research and analysis, significantly showing that rural villages and households are not all the same, and that “over-generalisations” about rural poverty in developing countries can be risky and not necessarily useful for policy. It supports the need for careful poverty alleviation targeting, especially of the poorest rural households. Finally, and very importantly, it demonstrates the value of integrating analysis of governance issues and poverty trends, rather than dealing with them separately, and shows the impact of governance, power relations and socio-political forms of inequality on poverty and economic inequality. 

CDRI’s experience and learning from the MOPS study have been rich, underlining not only the value of this type of research for a development research institute aiming to influence government policy, and the capacity building demands and benefits for its researchers, but also the challenges and limitations of undertaking globally designed and brokered studies in a complex environment like Cambodia. It has, however, also provided an opportunity for mutually beneficial learning and the development of close and very beneficial collaborative professional relationships between CDRI and the World Bank’s local poverty team. 

In 2006/07 CDRI completed two major poverty studies: MOPS for the World Bank and the Tonle Sap Participatory Poverty Assessment for the ADB, both challenging and demanding multi-disciplinary studies that have been a significant learning experience for CDRI in research design and methodology and the technical skills and capacity development needs of our researchers. They have taken CDRI’s capacity to conduct quality poverty research, both quantitative and qualitative, a major step forward, while at the same time setting a mid- to long-term poverty research and monitoring agenda for CDRI in support of Cambodia’s National Strategic Development Plan 2006–10.

Through our experience of MOPS and the other recent poverty studies, and interrelated policy research in our major programmes—economy, trade and regional cooperation; governance and public sector reform; natural resources and the environment; and agriculture and rural development—CDRI hopes to continue to build strong local capacity in poverty alleviation research, monitoring and analysis that will serve Cambodia’s future development well.  

Larry Strange
Executive Director, CDRI
September 2007


Executive Summary 

The Moving Out of Poverty Study (MOPS) is a first of its kind in Cambodia, one of 18 studies commissioned by the World Bank to examine poverty dynamics and trends. Conducted in 2004/05, the study revisited nine rural villages in which CDRI had conducted research in 2001, using quantitative and qualitative methods to investigate the extent to which these villages and individual households had been able to move out of poverty and improve prosperity, or had experienced downward mobility and decline.

The study set out to examine: which communities or groups move out of poverty and which remain trapped and why; whether people experience mobility differently in different economic conditions; how and why governance and social networks matter in mobility; what factors explain household and community progress and mobility or decline and stagnation; and the interaction between household and community factors, as well as any variations between villages and types of households.

The study was longitudinal, revisiting households that had previously been included in the 2001 study (using a panel survey) and contextual, exploring local history, geography and trends and their impact on communities and households, and it employed mixed methods, including a household panel survey and in-depth focus group discussions and semi-structured interviews. The nine villages selected for the original 2001 study represent all four of Cambodia’s main rural agro-ecological regions—the Tonle Sap plains, Mekong plains, plateau/mountain region and the coast. The 890 panel households drawn from the nine study villages represent a significant data set, from which statistically valid claims can be made about aggregate and village trends. The panel survey was supplemented by qualitative data from interviews and focus groups with 477 participants from the nine villages, including formal and informal village leaders, mobility groups (with participants who moved into or out of poverty or whose situation remained static) and young men and women.

The framework for analysis involved an examination of the main changes and trends that had affected the study villages, including changes in consumption and incomes and village poverty rates, together with an analysis of the factors contributing to or constraining community development and prosperity and household movement into and out of poverty. The analysis included community factors such as the underlying conditions and development and governance contexts that had shaped the experience of each village, and household factors such as strategies for income generation and the specific vulnerabilities that households experience. Although village and household factors accounting for movement out of poverty or decline and stagnation were explored in detail, individual experience, such as the impact of gender differences on poverty and mobility, was not explored in any depth. As intra-household dynamics are receiving increasing attention in poverty research (Fuwa et al. 2000), this could be a fruitful area for investigation in a further round of the MOPS.

The Study Identifies Trends in Well-Being and Prosperity and Factors Contributing to or Constraining Village Prosperity and Household Mobility

Well-being trends varied significantly between villages and households. In the aggregate, income and consumption rose and poverty fell slightly, representing an overall improvement in well-being. However, income rose in all study villages while consumption fell in three of the study villages and rose in the remaining six: the poverty headcount fell in six villages and rose in the villages experiencing declining consumption. Even in villages where consumption and income rose and poverty fell, poverty remained high and was above the provincial average in all but one village in 2004/05. Four villages had consumption well below the village poverty line in 2004/05. Villages were grouped into three clusters for the purposes of analysis: strongly performing villages, which experienced rising consumption and incomes and falling poverty rates; moderately performing villages, which achieved income and/or consumption growth, or poverty reduction, but not both; and poorly performing villages, which were unable to achieve substantial income growth or consumption growth or poverty reduction.

Among panel households, just over half did not change their status between 2001 and 2004/05, remaining very poor, moderately poor or well off. Of the remaining 48 percent, 26 percent moved up and 22 percent moved down, a net gain of 4 percent in upward mobility over the 3.5 years between surveys. Among non-moving households, 24 percent (of the total sample) remained well off (the comfortably rich), 14 percent remained moderately poor (the static middle) and 14 percent stayed trapped in poverty (the chronically poor)Of households that changed status, 14 percent of the total panel were very poor who became either moderately poor or well off by 2004/05 (escaping poverty), 7 percent were moderately poor who became very poor (deepening poverty) and 12 percent became well off (climbing into wealth). Fifteen percent were well-off households that became moderately or very poor (falling into poverty). The seven mobility groups—comfortably rich, climbing into wealth, escaping poverty, static middle, falling into poverty, deepening poverty and chronically poor—and the three village clusters are the main units of the study’s analysis.

Taking into account measures other than consumption, including income, assets and socio-political measures such as access to networks, resources and decision making, inequality rose between 2001 and 2004/05. While consumption inequality was static, and land-holding inequality fell, income inequality rose, as did the ratio between the value of assets held by the comfortably rich and chronically poor. Poor households are falling behind, with consumption falling further below the poverty line. Rising inequality was a concern in all study villages, focus group participants suggesting that the intersection of economic wealth and socio-political power and influence was responsible for this trend, as resources, opportunities and services are increasingly concentrated in the hands of better off households.

Although the end of armed conflict in the late 1990s was an important milestone in village development, the location, accessibility and geographic endowment of study villages were more significant in determining community well-being. Strongly performing villages were located close to national roads and provincial towns with good roads, productive soil and irrigation, while moderately performing villages were more isolated, with poorer soil, less arable land and no irrigation. Productivity was highest in strongly performing villages and those with natural irrigation, while food insecurity was more common in moderately and poorly performing villages. Strongly performing villages also received more development interventions, including agricultural extension, and were more likely to have clean water and sanitation. Availability of human services (health and education) has improved in all study villages, but more isolated communities continue to experience poorer health outcomes, and the quality of health services is variable, while children in the poorest households, in particular girls, continue to miss out on education.

Awareness and understanding about governance and the role of public institutions were low among ordinary villagers. Participation in social and political processes is largely confined to voting in elections, attendance of community meetings and membership of (usually religious) associations. Examples of poor governance and weak institutional capacity were raised in all study villages, with natural resource-dependent villages the most affected by corruption, intimidation and conflict over natural resource use and management. Trust in and satisfaction with authorities are low, in particular in regard to higher-ups such as forestry and fishery officials and police. Local authorities are regarded as more trustworthy, but are considered unable to protect villagers’ interests or respond effectively to new forms of insecurity such as drug use, youth gangs, domestic violence and other crimes. While peace and security were once critical issues, rural villagers increasingly expect their leaders to deliver improved living standards and express disappointment and frustration about constraints on development and the slowness of poverty reduction.

Households are more reliant on self-employment and wage labour and less reliant on agriculture and common property resources (CPR), including fisheries and forests. Better off households are earning more from agriculture and self-employment, while poorer households are more reliant on wage labour than in the past. Assets for wealth generation, including agricultural land, credit and inputs for agriculture and business, are concentrated in strongly performing villages and among better off households. Rich and upwardly mobile households have larger land-holdings and generate higher yields and profits than poorer households, which are increasingly unable to make a profit from rice farming, resulting in land sales and reliance on wage labour. Not all landless households are poor, however, and losing land does not result in poverty when other income sources, such as wage labour and self-employment, are available. Income from off-farm employment is increasingly important in all villages, and most economically active adults have jobs in addition to farming. Women are typically engaged in wage labour and petty trade, and are more likely to sell their labour locally, while men are concentrated in fishing and forestry and are more likely to migrate for work. Migration, including into Thailand, is an important source of income for better off and upwardly mobile households.

Most panel households are indebted, and most continue to borrow from friends or relatives, or from local moneylenders, rather than from MFIs or village banks. While strongly performing villages and better off households tend to use credit for productive purposes, poor communities and households use it to cope with shocks and crises, and often enter into interlocked credit arrangements. A majority of panel households experienced shocks and crises, illness being the most common, followed by crop damage and death of livestock. Better off households were able to mobilise savings and assets to cope with shocks, while poorer households tended to cut consumption or have family members migrate for work, and were disproportionately affected in income lost. Although shocks and life-cycle events are predictable, poorer households do not plan for or insure against them.

Demographic change, including immigration and population growth, is putting pressure on resources, including land and CPR, in all study villages. Immigrating households are more likely to be poor and landless. Marriage of adult children is expensive due to the costs of traditional wedding ceremonies, and typically leads to land atomisation as households divide their land into smaller parcels in order to give some to their children. Poorer households have higher dependency ratios, while better off households have more earners and fewer dependants. Destructive gender-specific behaviours such as domestic violence, alcohol abuse and young men’s involvement in gangs are impacting on communities and households. Domestic violence appears to be declining in most study villages but is still a serious problem in at least three; it contributes to movement into poverty due to lost income and assets, the costs of illness and injury and divorce and family breakdown. Alcohol abuse and youth gangs were raised as serious concerns in study villages: better off households are more able to absorb the costs of these activities (including paying off authorities when young men are arrested or cause damage).

The MOPS Provides Answers to Key Questions the Research Set Out to Examine

With the exception of individual experiences of poverty, the key questions the research set out to examine were answered in the study findings. Villages that raised living standards were well located and accessible, with year-round roads, productive soil and natural assets such as irrigation from lakes or streams. Households that improved their status were those with opportunities and resources (including economic and socio-political capital) and the capacity to generate new income sources and diversify their earnings. Villages that experienced declining living standards are CPR-reliant and more remote, with limited arable land, less fertile soil and no irrigation. Households experiencing stagnation or downward mobility had smaller land-holdings, fewer earners and more dependants and were reliant on one or two income sources, along with those that experienced shocks and crises or that were affected by risky and destructive behaviours of male household members.

As expected, economic growth enabled some households to move out of poverty, with more upwardly mobile households located in strongly performing villages, and more downwardly mobile and poor households in villages experiencing slower growth or decline. In a context of poor governance and weak institutions, the benefits of growth have not been evenly distributed. Inequality (in particular in income and assets) is increasing significantly in villages that experienced economic growth (where aggregate incomes were higher and rose more rapidly between 2001 and 2004/05). Poor governance and weak institutions are a brake on poverty reduction and contribute to rising inequality. Inequality is potentially destabilising for Cambodia, as the poor are increasingly locked out of opportunities, institutions favour the rich over the poor and trust in public officials is declining. Greater attention to the socio-political dimensions of poverty will therefore be critical in future poverty studies and poverty reduction initiatives.

The concept of khnang (“strong back”) emerged as a central theme in discussions of power and opportunities in focus groups. “Strong back” refers to connections and networks in the “string” (khsae) of patronage relationships, in which those higher in the chain provide favours and protection in return for loyalty, labour and other services from those lower down. While traditional forms of social capital such as labour exchange have largely disappeared, patronage relationships are increasingly important in households’ upward climb because they provide opportunities, assist households in building wealth and help them to secure favourable outcomes in local decision making, conflicts and legal disputes.

The main factors that support households’ climb out of poverty include location in a strongly performing village, many adult earners, multiple income-generation opportunities and the capacity to generate savings and invest in assets that can protect households in the event of shocks or crises and fund investment in human capital (health and education). The main factors driving households into poverty include location in a poorly or moderately performing village, in combination with fewer earners and more dependants, exposure to (multiple) shocks and crises and destructive or risky behaviours by individuals within the household. Importantly, households that fall are often those that have not anticipated crises and shocks, or have not accumulated sufficient assets or savings to weather these events. Chronically poor households have different characteristics. They tend to be located in poor and CPR-reliant villages and have fewer earners and more dependants. According to focus group discussions, they are more likely to have old, disabled or single female household heads. They have limited or no land, and are often CPR-reliant. They do not have the capacity to invest in human capital or the resources to seize new opportunities such as wage labour, and are often locked into indebted labour arrangements. Food insecurity and child labour are common in these households.

Factors that enable households to move out of poverty tend to operate in the village, and include improved access to markets and services, improved agricultural productivity and opportunities for wage labour, and as such are amenable to policy interventions targeting communities (such as road construction and irrigation). However, internal characteristics, such as ambition and risk-taking, and having more adult earners and better health and education services, do enable families to take advantage of opportunities when they arise, as does having connections and “strong back”. Factors leading to households moving down or remaining chronically poor are more often internal to the household and include the balance between earners and dependants, exposure to shocks and illness and destructive behaviours. Incapacity to invest in human capital and lack of connections and networks are also important factors keeping households poor and exposing them to corruption costs in everyday transactions. Poor and downwardly mobile households are of course affected by community factors, in particular the location and geographic endowment of the village, together with the natural disasters that affected many Cambodian villages during 2001–2004/05. Their internal characteristics render them more vulnerable to these external factors than other households.

The interaction between factors varies significantly in different village settings, as one significant change (such as a rise in land speculation or loss of forest access) can alter the fortunes of an entire community, while a country-wide change (such as a growing youth population) can have quite different impacts in different villages due to their individual characteristics (leading to emigration and rising incomes in some communities, and social problems and frustration in others). It is at times difficult to generalise from the study findings, given the specificity of experience in each community.

Although the study was not able to examine individual gender differences in any depth, it does suggest some important trends in gender relations, such as differences between female- and male-headed households. Female-headed households have quite different characteristics and experience poverty and mobility differently than male-headed households. Female-headed households are not always the poorest of the poor; indeed, some of these households were able to move out of poverty or remain wealthy. Often better off female-headed households had unmarried female heads while female heads with spouses who were unable to generate income or were dependent were more likely to be among the very poor. However, upward mobility among the female-headed households in the study is more likely to be unstable because they consumed more and earned less than male-headed households, leaving them with limited savings and fewer assets.

Policy Implications and Recommendations for Future Research

Rural villages and households are not the same, and policy interventions are likely to have quite different impacts on communities and households depending on their history, current situation and status. The study suggests that differences between rural villages and households need to be understood and taken into account in order to develop more effective, targeted interventions. It points to a need for greater public expenditure, together with local devolution of expenditure and revenue raising. The study suggests, however, that poor governance and weak institutional capacity require a national response, because local institutions do not have the administrative or financial capacity or the power to effect changes in Cambodia’s patronage-based political culture.

The study supports greater investment in rural infrastructure, including roads, as well as in agriculture, including in irrigation and extension services. Since not all villages are primarily reliant on agriculture, it also suggests a need for greater investment in rural employment opportunities, accompanied by basic legal protection. In the case of CPR-reliant villages with limited agricultural land, the need for alternatives to agriculture and CPR collection is particularly pressing: even if natural resources are well managed in the future, pressure on CPR is likely to continue as the population grows.

The study supports more specific targeting of rural households, in particular the poor and very poor. It suggests that interventions targeting rural infrastructure and agricultural productivity will be of most benefit to better off households, as will opportunities for trade and skilled employment and affordable, flexible credit. Although some downwardly mobile households may benefit from these interventions, preventive measures, such as greater public investment in health and education, as well as social and legal protection, including health and weather insurance, health equity schemes and basic labour protection, are required to prevent households from moving into poverty. Chronically poor households require interventions that address poverty traps and help them to cope with current circumstances, such as food security programmes, locally available free or heavily subsidised health care and education, more options for local employment and basic labour protection. Stronger and more accountable local government is essential to protect the interests and address the needs of poorer households.

Considerable investment has been made in the MOPS to date, with two rounds conducted in the study villages in 2001 and 2004/05. The value of panel surveys only really becomes evident with three or more rounds, however, and there is a strong case for continued investment in longitudinal research. Future longitudinal, mixed-methods, contextual studies can make a substantial contribution to national poverty monitoring and analysis, including by providing an understanding of the balance between transitory and chronic poverty and a local perspective on national poverty trends. Future rounds of the MOPS could build on lessons from the current study and align current findings and future research more closely with national poverty monitoring and analysis. As there are few panel studies of poverty dynamics internationally, ongoing rounds of the MOPS can potentially make a significant contribution to Cambodia and other countries as well.