Cambodia Land Titling Rural Baseline Survey Report
December 2007

Executive Summary

The Ministry of Land Management, Urban Planning and Construction (MLMUPC), with support from international donors, is implementing a Land Management and Administration Project (LMAP) to improve land tenure security and strengthen land administration systems. Among other activities, the project has established a systematic land-titling program that will issues one million titles over a five-year period. The project expects that land titles will help: (a) increase farmer access to formal credit; (b) stimulate agricultural and commercial investments in rural and urban areas that will increase productivity and employment; (c) promote more efficient land markets, and (d) promote the use of the official registry to facilitate land transactions and transfers. The LMAP land-titling program is also expected to help achieve the Royal Government of Cambodia’s poverty reduction objectives as outlined in the National Poverty Reduction Strategy, 2006 – 2010 (NPRS).

The Cambodia Development Resource Institute (CDRI) has recently collaborated with MLMUPC to collect baseline data that will be used to assess of the economic and social impact of land titles after three years. The Baseline Survey Project interviewed 1,232 rural households in 40 villages in 10 communes of five provinces during 19 January – 29 February 2004. The four LMAP provinces include Kompong Cham, Kompong Thom, Sihanoukville, and Takeo. The fifth province, Kompong Chhnang, is not in LMAP and serves as the control province for comparison with the four project provinces. Households were randomly selected from village lists according to landholding size and gender. An additional 99 urban households were interviewed in Sihanoukville city (Sangkhat 2) and will be referred to the findings of the urban phase of the baseline survey project in and around Phnom Penh, Siem Reap, and Serei Saophoan (Banteay Meanchey).

The rationale for land titling programs ultimately rests on property rights theories and research that link secure land tenure to investment incentives as well as land values and use. These theories generate a series of hypotheses that can be tested using quasi-experimental methods that compare household and village data from the current BSP (T0) with data that will be collected at a later point (T 0 + x) in both project and non-project areas. Some of the key hypotheses guiding the baseline survey include the following:

  • Access to Credit: People will use land titles as collateral with which to obtain credit from formal lending institutions;

  • Investment: People in rural areas will increase investments in agricultural production and diversification, thus increasing yields and income;

  • Land Markets: As land values increase and transaction costs decrease, land markets will direct land use toward more economically efficient uses;

  • Land Administration: A greater percentage of transactions (e.g., sales; inheritance) will be facilitated through the official registry;

  • Disputes: Secure land titles will reduce the volume and frequency of land disputes by clarifying ownership, parcel boundaries, and transaction procedures.

Distribution of Landholdings

Households with smaller landholdings have fewer agricultural plots that are also smaller in size compared to households with larger landholdings. Indeed, the number and size of plots steadily increases from one landholding interval to another. The most often cited explanation for this pattern begins with the land distribution of 1989 when efforts were made to divide good quality land equally according to the number of working age household members. According to this formula, households with more working members received additional land, and as a result, there was already a degree of structural variation in the 1989 land distribution when one considers landholding size by household.

This land distribution pattern holds for both male- and female-headed households. Male-headed households, however, average 4.4 plots per household and 0.39 hectares per plot, while female-headed households average 3.8 plots and 0.30 hectares per plot. Thirty-four percent of the female-headed households own less than one half hectare of agricultural land, while 18 percent of male-headed households own less than one half hectare. Conversely, 17 percent of the households headed by women own more than 2 hectares of land, while 31 percent of the households headed by males own more than 2 hectares of land.

Households with larger land holdings have a higher percentage of agricultural plot acquisitions through both purchase and clearing than do smaller households, while smaller households have a larger percentage of acquisitions from the Sate and through inheritance. Part of the reason is that larger landholders also have higher average incomes and more potential labour than smaller landholders. Female-headed households have a higher percentage of plot acquisitions from the State (i.e., 1989 land distribution) than do male-headed households. At the same time, the percentage of plots acquired through inheritance is much lower for female-headed households (11.2 percent) than male-headed households (24.6 percent). The percentage of plot acquisitions by purchase and clearing is also lower for female-headed households. The lower percentages for inheritance, purchase, and clearing suggest that female-headed households are less able to acquire additional plots than male-headed households.

Access to Credit

The LMAP survey group reported a total of 743 loans during the six-month period prior to the survey. About 60.0 percent of loans were obtained in the informal sector, which includes relatives and friends (43.7 percent), as well as moneylenders (16.0 percent). The remaining 31.0 percent of loans were obtained in the formal sector, either from Acleda (6.1 percent) or an MFI (24.9 percent). About 9.0 percent of loans were obtained in the “semi-formal” NGO sector (e.g., small savings and loan groups).

Productive investments accounted for 36.0 percent of all credit activity within the survey group, including small businesses (12.0 percent), agricultural production (14.4 percent), and animal raising (9.6 percent). Male-headed households borrowed more for agriculture and business activities, while female-headed households borrowed more for animal raising activities. Healthcare (21.7 percent) and food shortages (17.9 percent), however, accounted for almost 40.0 percent of all loans. A similar percentage of male- and female-headed households borrowed for health care, while a greater percentage of female-headed households borrowed to cover food shortages. The remaining loans (24.5 percent) were for other activities, including social ceremonies, home construction, and transportation.

Land titles are expected to stimulate an increase in the number and average amount of loans for investment in agricultural production and other income generating activities (e.g., animal raising, small business). This assumes that credit markets perform reasonably well in a particular area and that people have the propensity and capacity to borrow. All other factors being equal, we expect to see a larger volume of credit activity in areas where formal credit institutions are more accessible to local farmers. We also expect to see some variation in credit activity according to landholding size and sex of household head in terms of frequency, size, and use of loans.

Agricultural Investments, Productivity, and Land Use

The average amount of agricultural expenditures per household increases along with landholding size. For example, the lowest two smallest landholding intervals have average household expenditures of 13.03 and 21.57 moeun riels respectively, while the upper two intervals expend 31.6 and 51.98 moeun riels respectively. Male-headed households expended about 50 percent more than female-headed households for rice production. Nearly 90 percent of rice production expenditures are financed by “own sources,” followed by loans from relatives and friends (8.3 percent), and credit from “programs,” including semi-formal NGO projects, MFIs, and commercial banks (2.3 percent).

Productivity

The survey data affirms the inverse relationship between farm-size and productivity observed elsewhere in Asia and Cambodia; namely, small farms tend to be more productive in terms of rice yields per hectare than large farms, irrespective of the gender of the household head. One frequently cited reason for this pattern is that small plots are often subdivisions of more fertile land. Another reason is that family labour and other owned inputs are applied more intensively on small farms in the absence of modern farming techniques. According to the survey data, small farmers also appear to apply purchased inputs more intensively than do the larger farms for rice production. For example, the two smallest landholding groups expended 51.3 and 32.4 moeun riels per hectare respectively, while the two largest expended 18.5 and 19.5 moeun riels per hectare respectively.

Although small farms may be more productive in terms of land (i.e., kg per hectare) than larger farms, they are not as productive in terms of investment (i.e., kg per moeun riels). For example, farms with less than 0.5 hectares of land get 39.98 kgs per every moeun riels of expenditure, while farms with 2.0 – 2.99 ha and more than 3.0 ha get 61.89 kgs and 52.1 kgs of rice, respectively, from every moeun riels of expenditure. This suggests that investment efficiency is just as important, if not more so, than the actual level of investment. In terms of land titling impacts, then, increased access to formal credit for agricultural investments needs to be complimented with extension services and infrastructure development that can improve the productivity of capital.

The higher productivity of small farms does not translate into higher amounts of household rice production. Households with less than 0.5 ha of land produced only 640.3 kg of rice, despite their productivity advantage. Meanwhile, the largest farms produced a total of 3.27 MT of rice per household, even though they were only half as productive as the smallest farms. Smaller farms are at a comparative disadvantage, as they must continue to use remaining household financial resources (after farm expenditures) to make up for food shortages rather than invest in other activities. Moreover, if small farmers borrow money to invest in farming that does not produce sufficient rice for home consumption or surplus for sale, they will sink deeper in debt over time. This again highlights the need for extension services and infrastructure development in order to optimize land-titling benefits in specific areas.

Land Use

The survey data affirms the research hypothesis that land use patterns become more diversified as landholding size increases. For example, the percentage of plots allocated for wet-season rice production steadily decreases as landholding size increases, while the percentage for dry-season rice steadily increases along with landholding size. The percentage of plots allocated for chamcar production remains fairly constant across all landholdings, while the percentage of plots that are idle increases along with land size. The percentage of plots allocated for plantation (trees crops) and mixed crops is quite low across all landholdings. As a result, there appears to be considerable scope for crop diversification in many LMAP areas.

It terms of the actual utilization of plots, about 90 percent of all plots are cultivated, though the percentage decreases along with land size. The percentage share of cultivated plots among male- and female-headed is similar across land holding size. Not surprisingly, the percentage of idle plots (7.6 percent) increases with land size, while the percentage of leased plots (1.6 percent) is fairly constant across all land holdings. Female-headed households tend to have a slightly higher percentage of idle land, and they also lease out a higher percentage of their plots than do male-headed households.

We expect that secure land tenure will extend farmer investment horizons. We should observe more diversification of land use involving chamcar production, as well as longer term mixed and plantation crops. We expect that the scope and scale of diversification will increase at a faster rate along with land holding size. We also expect to observe increases in the rate of land utilization across all landholdings sizes as farmers begin to borrow more for agricultural investment. The impact of land titles in this regard may, however, vary according to location and situational factors, including the availability of credit and extension services, infrastructure investments, and market prices.

Land Markets: Values and Transactions

According to the survey data, land values (moeun riels per hectare) decrease as landholding size increases for both male- and female-headed households. However, some variation in land values should also be expected according to plot location (e.g., access to main road, distance from home). Despite the higher value of land per hectare, the average reported value of each plot is less among smaller farms than among larger farms. This is a direct function of the average size of plots on small and large farms. One implication of this pattern concerns access to credit. If the size of a loan depends in part on the amount of collateral that is available, larger farms may be able to obtain larger loans than smaller farms.

Generally speaking, we expect that land values will increase as farm households improve their land utilization rates and diversify in the direction of more economically efficient land uses. Land values will increase at a faster rate along main roads and near administrative and market centres.

Land Transactions

A total of 201 households reported 303 land sales reported since 1989, or 7.8 percent of all the plots in the sample. There is a disproportionate number of sales among the two smallest landholding intervals compared with the three upper intervals. The two lower intervals own 33.7 percents of all the plots in the LMAP survey areas, yet they report selling 50.8 of the total number of plots sold. Meanwhile, the two upper intervals own 38 percent of the plots, but sold only 29 percent of the plots sold.

The most often cited reasons for land sales were health care (24.9 percent), followed by business investments (18.6 percent), and then plot characteristics, including “too small, not profitable,” “poor soil”, or “too far away” (9.7 percent). Another 8.5 percent involved sales to offset food shortages. We expect land sales for these reasons to continue at a similar, if not higher, rate in areas where credit, extension, and affordable health care services are lacking. Other reasons for land sales include loan repayments, funerals, migration costs, and climate-related shocks.

The four areas with the highest average reported land sale prices are all located in areas close to Phnom Penh or provincial towns and/or with infrastructure development projects planned or underway. Land prices are likely to increase in these areas with the advent of more secure land titles. The two communes that have been located off main roads and somewhat away from market and administrative centers have the two lowest reported sale prices. We expect land prices to increase in areas where highway construction currently underway is completed.

Land titles alone, however, will neither slow nor accelerate the rate of land sales among any of the landholding intervals. For example, if lower cost health care services are not available in the survey areas, we can expect to see at least a similar rate of land sales for this reason. If people use land titles in the future to secure loans with which to invest in business or other activities, we may expect to see a decrease in land sales for this reason, unless the investments fail and people sell land in order to repay loans. In either case, land titles may enable people to obtain a better value for their land, though this may be small consolation for those who have no viable alternatives after farming.

Land Management and Administration

About 62.7 percent of the agricultural land plots covered in the LMAP survey areas have never been documented with any kind of paper. Of the documented plots, 61.6 percent have receipts for certificate applications, while another 14.9 percent have land survey investigation papers. Only 8.1 percent have acquired actual land certificates or titles. Another 6.3 percent involve the use of other types of paper. The use of unofficial documentation is consistent across all land holding sizes.

Respondents provided a variety of reasons for not registering their land, including 23 percent who said they did not know the procedures, while another 21 percent said they thought registration was unnecessary. Yet another seven percent said they thought the registration process was too complicated. These represent process-related reasons that suggest a high degree of confusion among people concerning various aspects of the land registration system. This in turn suggests that any increase in the use of the land registry system will depend in large part on the amount and quality of information available to people at the local level, and the degree to which they understand the land registry procedures. It will also depend on the accessibility and efficiency of the system.

Planners expect an increase in the percentage of transactions that are facilitated through the official registry system, particularly in more active land markets where land values are increasing. Such expectations assume that (1) transaction costs associated with official registration will be lower than current costs, (2) people have more confidence in the security of tenure than they do now, and (3) people have sufficient knowledge of the proper procedures and capacity to access the system.

The degree to which people use the official registry to facilitate and record land transfers also depends on the capacity of public administration to govern and enforce property rights effectively. In this sense, people in the baseline survey areas have expressed a great deal of initial faith in the land titles that LMAP is currently providing. The degree to which people use the official system may also vary according to their capacity and willingness to pay related fees and taxes. People will continue to avoid the official registry if they feel tax rates are too high and/or that such measures are not properly or fairly managed.

Land Conflicts

A total of 61 land conflicts were reported in the LMAP survey area since the Commune Council elections. The type of land involved was evenly divided between agricultural and residential land. Boundary conflicts with neighbors accounted for 38.3 percent of the cases, followed by conflicts with other villagers, (21.6 percent), conflicts with relatives (20 percent), and six cases involving encroachment or grabbing on the part of authorities or powerful people. The distribution of land dispute type affirms an earlier observation (So et al., 2002) that most land disputes so far are local in nature, involving boundary conflicts with neighbours, or ownership disputes with relatives. However, other studies have reported higher incidences of land grabbing and encroachment elsewhere in Cambodia, suggesting that the scope and scale of land conflicts may be highly situational.

Households used a variety of methods for resolving disputes, which in several cases required multiple rounds of negotiation. For example, in the first round of resolution, 21 households negotiated directly with the other party, while 23 went to the village chief. Eight households went to the commune chief and three went to the district dispute resolution committee. The clear preference for trying to resolve disputes at the local level can be explained by the fact that the disputes themselves are generally local in nature and that the transaction costs associated with dispute resolution increase once people go outside the village for mediation. Slightly less than half (28) of the first round cases were resolved, while 31 cases were not resolved.

Twenty-seven of 43 households (62.8 percent) indicated they were satisfied with the way their dispute was resolved, while the remaining households were not satisfied. Not surprisingly, 26 of the satisfied households thought the outcome was fair, and the rest thought it was not fair. There is obviously a clear relationship between one’s satisfaction with the outcome and one’s feeling about the fairness of the outcome.

Summary

Landholding size and the gender of the household head tend to be a good predictors of labour, assets, and income, and provide a good indication of a household’s potential capacity to benefit from land titling programs. We therefore predict that households with larger landholdings are in a more favourable position to benefit from LMAP’s systematic land titling projects than smaller landholdings. At the same time, male- headed households also tend to be in a more favourable position to benefit from land titles than female-headed households. High Potential Impact (HPIs) households have more available larger landholdings, more labour, more capital assets, and higher incomes. Low Potential Impact (LPIs) households tend to have less labour, smaller landholdings, fewer capital assets, and lower incomes. The LPIs also include more vulnerable households, such as those headed by single women.

Village location relative to paved roads and local commercial and administrative centres tends to be a good predictor of market access, credit access, and extension services, as well as social services such as health care. We predict that households located near such centres are in a better position to benefit from LMAP’s systematic land titling projects than those located further away. Villages with good soil conditions, access to water resources, diverse land use patterns, and employment opportunities, as well as development inputs are also potential High Capacity Areas (HCAs). Low Capacity Areas (LCAs) include villages that are located some distance from paved roads and/or commercial and administrative centres, have poor soil, lack water resources, and have more homogenous land use patterns and few employment alternatives to farming.

We predict that High Potential Impact (HPI) households will benefit most from land titles in HCA villages, while Low Potential Impact (LPI) households will benefit least in Low Capacity Area (LCA) villages. In between, HPI households in low capacity areas are in a better position to compete for land titling benefits as development occurs. Meanwhile, LPI households will be less able to compete for land titling benefits in high capacity areas unless they are somehow able to access or link up with high capacity factors.

The impact of land titles on social and economic development and poverty reduction in the rural sector can be optimized by targeting land-titling efforts in areas where government agencies, NGOs, and private investors are actively engaged. The benefits for disadvantaged households can also be increased by policies that specifically link land-titling efforts to pro-poor development objectives. In this sense, active consultation and collaboration among all development actors in support of LMAP’s efforts would enhance the benefits from land titles for all landholders.

The research methodology employed in the baseline survey has been quasi-experimental in nature using quantitative data collected in household interviews with a structured, close-ended survey instrument (See Annex A). The follow up survey should incorporate qualitative research approaches and tools into the overall methodology in order to provide more substance and texture to the household survey data, as many of the subtle yet important nuances concerning the economic and social impacts of land titles cannot be effectively captured by a standard household survey instrument.