DETERMINANTS OF INVESTMENT IN RURAL HOUSEHOLDS IN VIETNAM

DETERMINANTS OF INVESTMENT IN RURAL HOUSEHOLDS IN VIETNAM Tran Thi Thanh Nhan This thesis is presented in partial fulfillment of the requirements for the degree of Master of Natural Resource Management SCHOOL OF AGRICULTURAL AND RESOURCE ECONOMICS Faculty of Natural and Agricultural Sciences November, 2011 1

CERTIFICATION I certify that this thesis has been substantially completed during the course of enrolment in this degree at The University of Western Australia and has not previously been accepted for a degree at this or another institution. All help received in preparing this thesis and all resources used have been acknowledged. The thesis does not contain work that I have published, nor work under review for publication. Tran Thi Thanh Nhan November 2011 i

ACKNOWLEGEMENTS First and foremost I wish to thank Professor Benedict White, Professor Michael Burton and Assistant Professor Chunbo Ma for their roles in supervising my work on this thesis. They are very generous and untiring in their encouragement and direction. They have supported me in terms of knowledge, experiences and worthwhile comments and suggestions for the last two years. I can never thank them enough for what they have done for me. I would like to express my sincere thanks to Australian Centre for International Agricultural Research (ACIAR) for providing me with financial support to study and complete this thesis. Special thanks are also extended to the School of Agricultural and Resource Economics and the International Center of the UWA for their encouragement and administrative supports. My thanks go to Julia Lightfoot for her English editing of the final thesis. Finally, I am thankful to my family and friends, both at home and abroad, for their love, care and encouragement. Without their support, I would not have made it this far. Perth, November 2011 ii

ABSTRACT This study investigates the determinants of rural household investment in Vietnam. Private investment plays a key role in economic development in developing countries in general and in Vietnam in particular. In the study, three components of household investment (in education, farm assets, and non-farm assets) are analyzed. In addition, housing investment is considered as a special type of non-farm asset investment and is estimated in a separated model. The Tobit model is applied using data from the Vietnam Household Living Standard Surveys which was conducted by the General Statistic Office (GSO) in 2004, 2006 and 2008. In analyzing the determinants of household investment, the study focuses on household income, credit access, and land use right as well as other socioeconomic, year and regional dummy variables. The results show that income has a positive effect on all household investments. Regarding credit access, only informal credit sources (from individuals or relatives) are significant in non-farm and housing investment while all credit sources positively affect expenditure on farm assets. Surprisingly, land use right has no significant effect on any of the categories of household investments. The study indicates that there is a close link between the investment and demographic characteristics of households. Also, differences in household investment by regions and time are found in this study. iii

TABLE OF CONTENTS 1 INTRODUCTION... 1 1.1 Background... 1 1.1.1 Rural households in Vietnam... 1 1.1.2 Household investment... 2 1.2 Objective of the Study... 4 1.3 Methodology and data for the study... 5 1.4 Outline of Thesis... 7 2 OVERVIEW OF VIETNAMESE HOUSEHOLD LIVING STANDARDS DATA... 8 2.1 Introduction... 8 2.2 Social characteristics of households... 10 2.2.1 Characteristics of the heads of households... 10 2.2.2 Labour structures... 13 2.3 Economic characteristics of households... 15 2.3.1 Household income... 15 2.3.2 Household expenditure... 17 2.3.3 Land... 19 2.3.4 Household expenditure on education... 20 2.3.5 Household expenditure on farm production assets... 22 2.3.6 Household expenditure on other assets... 23 iv

2.4 Summary... 24 3 LITERATURE REVIEW... 26 3.1 Introduction... 26 3.2 Basic factors affecting household investment... 27 3.2.1 Income and wealth... 27 3.2.2 Land and property... 29 3.3 Theoretical framework... 30 3.3.1 The theory of investment and intertemporal choice.... 30 3.3.2 Utility Maximization Theory... 31 3.4 Empirical models and findings from studies of household investment... 34 3.4.1 Pure theory to empirical investment models... 34 3.4.2 Household education investment... 38 3.4.3 Household farm production asset investment... 41 3.4.4 Household non-farm asset investment... 44 3.5 Household investment analyses in Vietnam: current status and research gap.... 45 3.6 Summary... 47 4 ECONOMETRIC MODEL OF HOUSEHOLD INVESTMENT... 48 4.1 Conceptual framework... 48 4.2 Empirical model of Vietnam household investment... 52 4.3 Data and variables... 54 4.3.1 Data... 54 4.3.2 Dependent variables... 59 v

4.3.3 Explanatory variables... 60 4.4 Summary... 65 5 REGRESSION RESULTS... 67 5.1 Investment in education... 67 5.2 Investment in farm production assets... 72 5.3 Investment in non-farm assets... 75 5.4 Investment in housing... 77 5.5 Summary... 79 6 CONCLUSION AND RECOMMENDATIONS... 82 6.1 Background... 82 6.2 Findings... 83 6.3 Limitations and further study recommendations... 86 7 REFERENCES... 88 vi

LISTS OF TABLES Table 2.1 Number of households by socio-economic region... 9 Table 2.2 Population, area and population density by socio-economic region 2010... 9 Table 2.3 Marital statuses of household heads by region and gender (%)... 10 Table 2.4 Gender, age and ethnicity of household heads by region... 11 Table 2.5 Ethnicity of population by region... 11 Table 2.6 Education level of household heads by region (%)... 12 Table 2.7 Main job of household heads by region (%)... 13 Table 2.8 Household size, number of labourers and average working hours per labourer per week by region... 14 Table 2.9 Household total income, income from wages, income from agriculture by region (thousand dongs)... 15 Table 2.10 Household total income, income from wages, income from agriculture over time (thousand dongs)... 16 Table 2.11 Expenditure, regular expenditure, expenditure per person and regular expenditure per person by region (thousand dongs)... 17 Table 2.12 Expenditure, regular expenditure, expenditure per person and regular expenditure per person by year (thousand dongs)... 18 Table 2.13 Total land area and land area for production by region (m 2 )... 19 Table 2.14 Land area of households in previous years by region (m 2 )... 20 Table 2.15 Number of person attend school, education expenditure, education expenditure per person by region... 21 vii

Table 2.16 Number of person attend school, education expenditure, education expenditure per person by year... 21 Table 2.17 Farm production machines, expenditure on production machines and farm inputs by region and year (thousand dongs)... 22 Table 2.18 Household spending on fixed assets, house, stock, gold and saving accounts by region (thousand dongs)... 23 Table 2.19 Household spending on fixed assets, house, stock, gold and saving accounts by year (thousand dongs)... 24 Table 4.1 Description of variables in the models... 55 Table 4.2 Summary statistics for variables in the models... 57 Table 5.1 Results of Tobit analysis for education investment... 70 Table 5.2 Results of Tobit analysis for farm assets investments... 73 Table 5.3 Results of Tobit analysis for non-farm asset investment... 76 Table 5.4 Results of Tobit analysis for housing investment... 78 viii

LIST OF FIGURES Figure1.1 Map of Vietnam... 6 ix

ABBREVIATION CH FDI LUR MRD NCCC NMM RRD SE USD VHLSS VND Central Highlands Foreign Direct Investment Long-term use right for land Mekong River Delta North Central and Central Coastal Areas Northern Midland and Mountain Areas Red River Delta South East USA Dollar Vietnam Household Living Standards Survey Vietnam Dong x

1 INTRODUCTION 1.1 Background 1.1.1 Rural households in Vietnam The Vietnamese economy has diversified and developed significantly in recent years. However, it remains one of the poorest countries in the world. At USD 1129/per person/year in 2009, Vietnam s Gross Domestic Product (GDP) per capita ranks 149 from a total of 190 countries and territories 1 and 113 on the Human Development Index (HDI) of a total of 169 countries in the world (UNDP) 2. Although the Comprehensive Poverty Reduction and Growth Strategy (CPRGS) was implemented in 2000 the proportion of poor households 3 in the country was still 14.2 per cent in 2009, mainly in rural areas (17.4%) and to a lesser extent in urban areas (6.9%). In 2010, income per capita in urban areas is VND 2130/month, much higher than the VND 1071/month in rural areas (in current market prices). Rural households play an important role in the national economy with more than 70 per cent of the population living in rural areas and 51.5 per cent of labourers working in the agricultural sector (in 2009). According to the results of the 2006 Rural, Agricultural and Fishery Census, Vietnamese rural areas consist of 13.78 million households, of which 10.46 million, or 75.9 per cent, are engaged in agriculture, forestry, and 1 World Bank report, http://data.worldbank.org/indicator/ny.gdp.pcap.cd 2 Human Development Report 2010. http://hdr.undp.org/en/media/hdr_2010_en_table2_reprint.pdf 3 According to the GSO, Poverty rate is calculated by monthly average income per capital of household In 2004, 2006, 2008, it is measured by the Government's poverty line for 2006-2010 period, considering inflation adjustment as follows: 2004: 170 thousand dongs for rural area, 220 thousand dongs for urban area; 2006: 200 thousand dongs for rural area, 260 thousand dongs for urban area; 2008: 290 thousand dongs for rural area, 370 thousand dongs for urban area. In 2010, it is measured by the Government's poverty line for 2011-2015 period which is 400 thousand dongs for rural area and 500 thousand dongs for urban area (Vietnamese). Available from: http://www.gso.gov.vn/default.aspx?tabid=395&idmid=3&itemid=11279 1

aquaculture (GSO 2007, p. 23). However, investment in the agricultural sector 4 has been declining in recent years: as a proportion of total national investment it decreased from 7.4 per cent in 2005 to 6.25 per cent in 2009. Government investment in the agricultural sector fell from 7.14 per cent in 2005 to 5.86 per cent in 2009. A small proportion of foreign direct investment (FDI) was invested in the sector: only 12 FDI projects (about 0.18 per cent of the total FDI value) were approved in 2010. In this context, private investment in this sector will become more important in the future. In addition, high unemployment rates are prevalent in rural regions. In 2009, total labourers in rural areas were 36 million people, or 73.1 per cent of total labourers in the country. Of those, 2.25 per cent were unemployed and 6.51 per cent were underemployed. Every year, there are about one million young people entering the rural labour force, which has tended to increase the rate of rural unemployment. However, rural labourers often lack the skills to apply new technology (GSO 2010b). This requires a production expansion to provide more employment opportunities, or training to improve skills for the rural labour force. 1.1.2 Household investment Household investment plays an important role in economic development in developing countries. Most developing nations do not have a high volume of market capital while they need to implement many development projects that require substantial funds. While the financial resources from the State and foreign investment focus more on projects at the macro level, household investment is considered as an essential source of funds to improve the development and living standards of people at the household level. Household investment has become more important as confidence has been eroded in many financial markets following the 2008 global financial crisis. The global financial crisis has led to a contraction in funding from governments and foreign sources for development projects. 4 In the present study, agricultural sector or agricultural production includes agriculture, forestry, and aquaculture. 2

Economic growth theory confirms that investment is one of the key determinants of economic growth (Coen et al. 2008). However, while many economists emphasize the role of investment in productive capital growth, others also focus on its importance to human capital, people s well-being and in improving living standards. Understanding the determinants of household investment, especially of rural households, is very important for development policy as household investment is an important determinant of economic growth. It also promotes the connection between the service providers such as banks and agricultural services and their customers. Household investment has a particularly important role in the context of Vietnam s rural development. Vietnamese economy, including its rural sector, has achieved remarkable success by adopting reforms, namely Doi moi (renovation) since 1986. It has been changed from a country with food shortages to being one of the largest exporters of agricultural products including rice, coffee and rubber. This achievement is partly attributed to the transformation of the rural household economy and higher household investment. In 2009, the private sector contributed more than 40 per cent of total GDP, of which the proportion of household sector was 29.87 per cent (GSO 2011). This indicates that household funds are a significant source of domestic investment. As an economy under market oriented transformation, Vietnam now faces more opportunities and greater challenges. The Vietnamese legal system promotes domestic investment 5 and encourages all organizations and individuals to invest in different economic activities. In the legal system, favorable conditions are also offered for domestic investment in replanting forestry and in aquaculture as well as for the development in remote and mountainous areas. However, the investment of rural households in production expansion is limited and rural households still face many difficulties in implementing investment activities (CAP 2008). In this context, research on household investment has become more important than previously. 5 The law on promotion of domestic investment was approved in 1994 and amended in 1998 by the National Assembly. 3

This study is expected to contribute to our understanding of rural household investment. Until recently, there has been very little research conducted to investigate this issue in Vietnam and the focus has been on national investment at the macro level which neglects the investment behavior of households (CAP 2008). The most recent and relevant study on this issue was carried out by the Center for Agriculture Policy (CAP) in 2006-2007. However, the main objective of that study was to conduct a survey on household accumulation and investment so only some indicators were analyzed. The study also recommended the need for further research into household investment (CAP 2008). In addition, the present study provides a selective literature review on the relevant economic theory, which is applied to explain decision making at the household level, and the statistical methodology for analyzing limited dependent variable models with a large number of zero values. An empirical model for household investment in developing countries is also developed in the present study. 1.2 Objective of the Study The study aims to evaluate the situation of current investments of households in rural areas, factors determining household investments and to recommend relevant policies for improving household investment in rural areas in Vietnam. Specific objectives are as follows: Evaluate the current household investment in rural areas (in terms of the number of household investors and the average value of investment) from existing data. The difference in investment between regions and its trend over time are also evaluated. Analyze the determinants of household investment. The information on household demographic and socio-economic characteristics is taken into consideration in this analysis. The changes of these determinants by regions and over time are investigated. Assess the effectiveness of some policies relevant to household investment by adding policy relevant variables in the statistical model. 4

Review suitable theoretical models and statistical models for explaining household investment behavior. The study also aims to develop an empirical model for rural household investment in Vietnam. Recommend policy implications for encouraging household investment in rural areas. 1.3 Methodology and data for the study The study uses data from the Vietnam Household Living Standards Survey (VHLSS) in 2004, 2006 and 2008 to analyze the determinants of rural household investment. The surveys were carried out by the General Statistics Office (GSO). This database has a total sample of more than 45 thousand households in each year, of which 9180 households were surveyed for both income and expenditure. Since the study uses data from three different years, only 1281 households are chosen to construct a balanced panel data set. The sample covers almost all provinces in the country and in the study it is divided into six regions based on the socio-economic region division of the GSO (Red River Delta, Northern Midland and Mountain Areas, North Central and Central Coastal Areas, Central Highlands, South East, and Mekong River Delta) (see Figure 1.1). 5

Figure 1.1 Map of Vietnam For the analysis purposes, rural household investment is divided into three types: investment in education, investment in farm assets, and investment in non-farm assets (housing is considered as a special non-farm asset and is represented by a separate 6

model). A Tobit model is estimated for each type of investment using Stata software version 11. 1.4 Outline of Thesis The remainder of the thesis is organized as follows: Chapter 2 Overview of Vietnam Household Living Standards Data summarizes the sample drawn from VHLSS data. In this chapter, information on household demographics, household socio-economic characteristics, and current household investment is presented. Chapter 3 Literature review provides a theoretical framework of utility maximization for the household investment decision. In this model, households maximize their utility under a budget constraint. The households are assumed to face risks when making any investment decision, so the expected utility model has been chosen. The construction process and results of previous empirical household investment studies based on the utility model are also presented in this chapter. Chapter 4 Econometric models of household investment reviews approaches to estimating household investment. Because the data on investment is censored, the Tobit model has been chosen as an appropriate model for household investment in this study. Variable description and the hypothesized relationship between these variables are included. Chapter 5 Regression results presents the econometric estimation of the Tobit model for each type of household investment. Discussion of the marginal effects of the factors determining household investment is provided in this chapter. Chapter 6 Conclusion and Recommendations summarizes the contribution of the thesis. Some conclusions and policy implications for household investment in rural areas of Vietnam are drawn. This chapter also states the limitations of the study and provides suggestions for further research. 7

2 OVERVIEW OF VIETNAMESE HOUSEHOLD LIVING STANDARDS DATA 2.1 Introduction The data used in this study are taken from Vietnam Household Living Standards Survey (VHLSS) in 2004, 2006 and 2008. These surveys are part of the series of VHLSS which were conducted by the General Statistics Office (GSO) from 2002 to 2010. The aim of the survey is to evaluate the living standards of households for policy-making and socio-economic planning and to provide the primary source of information for many other social and economic studies in Vietnam. The survey contains information about many aspects of household living standards, including individual demographics, health, education, employment, household business, expenditures, income, assets, and credit access. Every two years, more than 45,000 households from 13.77 million rural households in Vietnam 6 in 3063 communes across Vietnam are interviewed, of which 9189 households are selected to complete questionnaires on expenditure and income. Of these 9181 households, approximately 6800 households are located in rural areas (see Table 2.1). This sample is rotated from year to year with the principle of 50% of households from the previous survey being re-interviewed in the following survey. In the present study, the panel for three surveys has been formed from 1281 rural households that reported both expenditure and income in all three years 2004, 2006, and 2008 7. In analyzing household investment, the whole country is divided into six regions based on the six socio-economic areas of Vietnam (Red River Delta, Northern Midland and 6 Updated to 31/7/2006 according to AgroCensus 2006 7 Because the way of sampling the survey, only 1281 rural households with both expenditure and income data are available for 3 year 2004, 2006 and 2008. The rotation of surveys also limits the panel structure of survey data to a maximum of 3 years because the sample would be very small if more than 3 years were used. 8

Mountain Areas, North Central and Central Coastal Areas, Central Highlands, South East, and Mekong River Delta). There is a considerable difference in population density in these regions and therefore the number of household samples also varies among regions, as seen in the follow tables. Table 2.1 Number of households by socio-economic region Total of households* Rural household* Sample households in 2004** Sample households in 2006** (Unit: household) Sample households in 2008** Whole country 22,444,322 15,570,642 6,822 6,882 6,837 RRD 5,576,804 3,920,335 1,590 1,593 1,581 NMM 2,770,757 2,235,673 1,320 1,317 1,314 NCCC 4,876,793 3,659,678 1,581 1,566 1,548 CH 1,224,344 844,265 372 417 417 SE 3,693,393 1,607,342 528 516 513 MRD 4,302,231 3,303,349 1,431 1,473 1,464 Source: * Data from the results of the 2009 Vietnam Population and Housing Census (GSO 2010a) ** Calculated from the sample data (rural households that report both income and household expenditure) Table 2.2 Population, area and population density by socio-economic region 2010 Average population (*) Area (**) Population density (Thousand person) (Km 2 ) (Person/km 2 ) Whole country 86927,7 331051,4 263 RRD 19770,0 21063,1 939 NMM 11169,3 95338,8 117 NCCC 18935,5 95885,1 197 CH 5214,2 54640,6 95 SE 14566,5 23605,2 617 MRD 17272,2 40518,5 426 Source: GSO (2011) 9

( * Average population is the average number of population of a certain area in a certain period of time, usually a year (GSO 2011, p.32) ** Area updated to 1/1/2009, according to the Decision No 2097b/QD/BTNMT) 2.2 Social characteristics of households 2.2.1 Characteristics of the heads of households Previous studies identified that there were significant differences in income and investment according to the gender of the head of the household. According to the results from the household investment survey implemented by the Center for Agricultural Policy (CAP) in 2007, income from non-farm activities of households with a male head was higher than those with a female head. On the other hand, income from agricultural activities and wages was higher for female headed households. The survey also showed that the rate of investment in male headed households was higher in both general and specific investment categories (CAP 2008). From the sample of the present study, most of the household heads are male (79%) and the majority of them are married (95%). Proportions of married male heads are higher in the North Central and Central Coastal area (NCCC) and the Central Highlands (CH) regions, and lower in the South East (SE) region. Only a few families in the sample are headed by single or divorced men (Table 2.3). In contrast, in households with female heads, more than half are widowers and almost 10% are single and divorced women. Table 2.3 Marital statuses of household heads by region and gender (%) RRD NMM NCCC CH SE MRD Male head Female head Total N 927 729 936 153 321 777 3037 806 3,843 Single 0.76 1.51 1.92 1.31 3.43 2.06 0.79 5.09 1.69 Married 82.96 90.95 78.1 83.66 78.82 77.35 95.79 29.28 81.84 Widow 14.89 5.35 18.7 11.11 15.58 17.12 2.73 58.19 14.36 Divorce 0.32 1.65 0.75 3.92 0.93 2.06 0.23 4.96 1.22 Separated 1.08 0.55 0.53 0 1.25 1.42 0.46 2.48 0.88 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 The average age of the head of households is 49 years for the total sample. On average, household female heads are older than household male heads with average ages being 10

56 and 48 respectively. Except for the Northern Midland and Mountain (NMM) and CH regions where household heads are younger, the average age of heads is similar in the other regions, as shown in Table 2.4. Table 2.4 Gender, age and ethnicity of household heads by region Gender of household heads (%) Age of household heads (years old) Ethnic of household heads (%) RRD NMM NCCC CH SE MRD Total N 927 729 936 153 321 777 3,843 Male 75.84 88.2 76.6 85.62 74.77 77.61 79.03 Female 24.16 11.8 23.4 14.38 25.23 22.39 20.97 Male 49.91 44.85 47.54 44.18 48.73 49.36 47.83 Female 55.87 48.77 56.19 50.41 53.51 58.68 55.42 Others 2.8 63.1 11.22 42.48 4.98 6.31 18.76 Kinh&Chinese 97.2 36.9 88.78 57.52 95.02 93.69 81.24 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 Ethnicity is an important factor in household economic analysis. There are 54 ethnic groups in Vietnam, and living standards differs between these groups. King is the majority with almost all of its population living in the deltas, cities, and towns. Although not a large ethnic population, people of Chinese origin are usually incorporated into the King group because of the similarities in their culture, living standards, and location. King and Chinese people total more than 80 per cent of the population throughout the whole country (Table 2.5). Table 2.5 Ethnicity of population by region Population of whole country King & Chinese (person) Total population (person) Share (%) Total population (person) Population of rural area King & Chinese (person) Whole country 85,846,997 76,417,498 89 60,410,101 49,958,784 83 RRD 19,584,287 19,290,550 99 13,844,603 13,599,715 98 NMM 11,053,590 5,051,589 46 9,299,879 3,725,856 40 NCCC 18,835,154 17,047,093 91 14,305,583 12,604,031 88 CH 5,115,135 3,333,718 65 3,670,300 2,064,416 56 Share (%) 11

SE 14,067,361 13,705,799 97 6,023,555 5,728,610 95 MRD 17,191,470 15,988,749 93 13,266,181 12,236,156 92 Source: Data from the results of the 2009 Vietnam Population and Housing Census (GSO 2010b) The average income of the King and Chinese groups is double that of the minority ethnic groups (CAP 2008). In addition, according to the results from VHLSS in 2008, the monetary child poverty rate and multi-dimensional child poverty rate in King and Chinese populations are 12.7 per cent and 22.4 per cent respectively, much lower in comparison with 60.7 per cent and 61.5 per cent respectively in other ethnic populations (GSO 2010b). In the present study sample, the Red River Delta (RRD), the South East (SE), and the Mekong River Delta (MRD) are regions where most King and Chinese people live (97.2%, 95.2%, and 93.6%, respectively) while NMM and CH are the locations for the other ethnic people (63.1% and 42.48%). Although the literacy rate of household heads is high (90%), only a small proportion have a higher degree qualification. In the present sample, no region has more than 3 per cent of the household heads who have completed college or university degrees (see Table 2.6). Among household heads, more than half have primary school as their highest level of education and one third only finished lower secondary school. A few household heads have college or university degrees (1.3%). Red River Delta is the region where the highest level of education of household heads is consistently lower or upper secondary school. There are statistically significant differences in the education of household heads by regions but not by years. Table 2.6 Education level of household heads by region (%) RRD NMM NCCC CH SE MRD Total N 927 729 936 153 321 777 3,843 No diploma 0.43 0.82 1.07 1.96 1.87 2.19 1.19 Primary school 33.12 52.95 53.95 69.93 59.5 78.38 54.77 Lower Secondary 50.92 32.24 33.12 20.92 22.43 13.77 31.95 Upper Secondary 14.56 11.93 10.79 5.88 13.71 4.89 10.77 College or higher 0.97 2.06 1.07 1.31 2.49 0.77 1.3 12

Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 All households in the present sample were chosen from the rural areas and, not surprisingly, more than half of household heads state farmer as their main job. This rate varies from region to region. The rate is relatively high in CH region (86.27%), where larger land areas are available for perennial crop farms or in the remote, mountainous NMM region (73.39%), where the infrastructure is not suited to manufacturing or service development. The Red River Delta and the South East are the regions with the lowest rate of household heads working in the agricultural sector. It is not surprising since the RRD and SE are where Hanoi and HoChiMinh City, the two largest cities and economic centers of Vietnam, are located. The rural areas around these cities have more manufacturing and service jobs. The data also shows that the manufacturing sector employs the largest per cent of household heads in the RRD, while it is the service sector attracts in the SE region (see Table 2.7). Table 2.7 Main job of household heads by region (%) RRD NMM NCCC CH SE MRD Total N 927 729 936 153 321 777 3,843 Agriculture 44.01 73.39 57.48 86.27 45.79 59.72 57.87 Manufacture 26.54 7.68 19.55 1.31 16.51 12.61 16.6 Service 9.71 5.08 7.37 3.27 16.51 9.78 8.59 White Collar 6.04 9.47 4.91 3.27 4.98 3.73 5.75 Others/Not Work 13.7 4.39 10.68 5.88 16.2 14.16 11.19 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 Testing for differences in means for the above indicators of household head information shows statistically significant differences among regions. 2.2.2 Labour structures The average number of household members in the total present sample over three years is 4.32 persons per household. Family size is assumed to affect the labour force, income, consumption and investment. The GSO report on the results from VHLSS 2006 indicates that the average number of household members was higher in poor 13

households than in rich households (GSO 2008b, p. 23). In the sample of the present study, the CH and the NMM have a higher average number of household members than in other regions (see Table 2.8). The RRD region has the smallest household size. However, the average number of persons per household of working age in this region is relatively high (2.92 persons). The Central Highlands has the highest average family size (5.10 persons), but has a relatively low number of people of working age (2.81 persons), which is only slightly higher than the average number in the NCCC region. Table 2.8 Household size, number of labourers and average working hours per labourer per week by region Region Household size(person) Labour (person) Average working hour per week Working hours 2004 2006 2008 RRD 4.03 a 2.92 ab 30.57 bc 30.20 30.63 30.90 NMM 4.58 b 3.04 a 30.53 cd 30.64 31.00 29.95 NCCC 4.30 ab 2.76 b 28.76 ad 29.76 x 29.38 xy 27.09 y CH 5.10 b 2.81 ab 30.12 ac 30.88 30.85 28.64 SE 4.37 a 2.99 ab 31.23 c 31.15 30.81 31.74 MRD 4.27 a 2.90 ab 27.04 a 26.35 27.30 27.46 Total 4.32 2.90 29.43 29.49 29.73 29.05 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c,d The different letter superscripts indicate those mean values within a column of the regions that are statistically significantly different from each other at P<0.05 x,y The different letter superscripts indicate those mean values within a row of the years that are statistically significantly different from each other at P<0.05) The average hours worked is 29.43 per week, as shown in Table 2.8. For developing countries like Vietnam, working hour per person is important because it increases income and the difference in working hour was the cause of wider gaps between the rich and the poor (GSO 2008b, p. 26). In the present sample, hours worked per person decreased in 2008 after a slight increase in 2006. On average, the MRD and NCCC have the lowest working hours. The number of working hours has decreased continuously from 29.76 hours per week in 2004 to 27.09 hours in 2008 in the NCCC region. This is also the only region which has a significant difference in the mean of working hours per person over time. 14

2.3 Economic characteristics of households 2.3.1 Household income All the summary monetary data is corrected for inflation by the Consumer Price Index for all goods in Vietnam with the base year 1994. There are differences in income between regions and households in each region. The average total income of households in the MRD region is more than double the total income of the NMM or NCCC households (p<0.05). Although the numbers show that household income in the CH region is much higher than that in the NMM or NCCC regions and is lower than that in the MRD region, the result are not statistically significant. The summary of household incomes is presented in Table 2.9. Table 2.9 Household total income, income from wages, income from agriculture by region (thousand dongs) Region RRD NMM NCCC CH SE MRD Total N 927 729 936 153 321 777 3,843 Total income Income from wage Income from agricultural production 37,830 bc 25,408 a 26,084 a 40,841 abcde 45,695 be 55,346 de 36,931 (96,863) (18,107) (36,256) (39,393) (58,945) (124,548) (79,101) 6,383 b 3,662 a 4,517 ad 3,227 ad 11,423 c 4,982 d 5,424 (8,943) (6,695) (7,075) (5,639) (14,892) (7,461) 8,668 20,149 a 16,895 a 14,885 a 33,776 ab 21,605 a 40,362 b 23,001 (89,977) (14,688) (33,077) (36,104) (56,913) (122,299) 75,442 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c,d,e The different letter superscripts indicate those mean values within a row of the regions that are statistically significantly different from each other at P<0.05 Standard deviations are presented in parentheses) In general, households in the South areas (MRD and SE) have higher incomes in comparison with those in the North area. The MRD has the highest standard deviation (124,548) of the total income of households (more than 2.5 times higher than the mean of incomes in each region). The NMM households have the lowest average income (25,408 thousand dongs/year), however, the income of households in this region is not 15

statistically significantly different from that of the NCCC and CH households. The NMM has the lowest standard deviations in the mean of household income. In the present study sample, income from agriculture accounts for a large proportion of the total household income (62%), while salaries and wages contribute 15%. Consistent with the values for the main job of the household s head, agricultural income as a proportion of total income is highest (83%) in the CH, a region with favorable natural conditions for developing high value perennial crops such as coffee, pepper (Daklak, Gialai, Lam Dong provinces) or fresh vegetables and flowers (Lam Dong province). The Mekong River Delta also has a high proportion of agricultural income proportion in total income as it is relatively well-suited to rice cultivation. The MRD is the only region in which income from agriculture is significantly different from the other regions. The rate of contribution to the total income from agricultural activities is the lowest in the SE region (47%). On the other hand, the contribution from wage income is the highest in this region (25%). The South East is also the only region with income from wages that is significantly different from the other regions. Table 2.10 Household total income, income from wages, income from agriculture over time (thousand dongs) Year Total income Wages income Agricultural income 2004 31,354 a 4,498 a 16,285 a 2006 39,641 b 5,551 b 26,525 b 2008 39,798 b 6,224 b 26,192 b Total 36,931 5,424 23,001 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b The different letter superscripts indicate those mean values within a column of the year 2004,2006,2008 that are statistically significantly different from each other at P<0.05) In the total present study sample, there is a statistically significant difference between the income of households in 2004 and that in 2006, while the income is almost the same in 2006 and 2008. In the period from 2004 to 2006, the total income, wage income, and agricultural income increased by 26 per cent, 23 per cent, and 63 per cent, respectively. According to the GSO (2008b, p. 29), the main reason for the income increase in 2006 16

is the increase in the minimum salary level from October 2005 and that led to an increase in social wages, agricultural and fishery prices like rice, coffee, rubber, cashew, pork (live weight), shrimp, fish increased in comparison with 2004 8 2.3.2 Household expenditure Household expenditure in the VHLSS includes all the household consumption for daily life and other consumption expenditure 9. Following is the summary of household expenditure in the present study sample (see Table 2.11). Table 2.11 Expenditure, regular expenditure, expenditure per person and regular expenditure per person by region (thousand dongs) Criteria RRD NMM NCCC CH SE MRD Total household expenditure 15,547 b 13,561 a 12,880 a 16,795 bd 23,061 c 17,820 d 15,657 (8,686 ) (7,874) (7,516) (10,431) (16,002) (15,048) household regular expenditure 13,814 b 12,304 a 11,745 a 15,414 bd 21,112 c 15,617 d 14,061 (7,545) (6,963) (6,790) (9,563) (15,176) (13,684) expenditure per person 335 b 271 a 263 a 292 ab 479 c 365 d 322 (175) (181) (1530) (191 ) (323) (244) regular expenditure per person 295 bd 242 a 236 a 265 ad 435 c 318 b 286 (143) (141) (121) (164) (294) (206) Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c,d The different letter superscripts indicate those mean values within a row of the regions that are statistically significantly different from each other at P<0.05 The regular expenditure is defined as expenditure on foods and drink and non-food expenditure Standard deviations are presented in parentheses). 8 Here social wages means the wage of hired labour. 9 In the VHLSS, household expenditure includes expenditure on foods and drink (for both normal days and holidays), daily non-food expenditure (gas, cleaning liquid, gas and parking fee for car, bicycle, books, entertainment ), annual expenditure (household items excluding durable goods like clothes, electric wire, shoes, blankets and other annual expenses such as toys, cosmetics, bicycles or automobile tires ) and other household spending (taxes, anniversaries, contribution for funds, cost for overseas study ). 17

On average, expenditure of each household is 15,657 thousand dongs. Although households in the MRD, on average, earn the highest total income, households in the SE region have the largest consumption. The NCCC and the NMM are the regions with the lowest average household expenditure. However, the expenditures in these regions also have the lowest standard deviations. The SE is the region with the highest standard deviation (16,002) on expenditure, as shown in Table 2.11. In the sample, consumption expenditure (regular expenditure) accounts for almost all of total expenditure of households (90%). This proportion is similar for both households and per person expenditures and also is not greatly different between regions, with the highest rate of 92 per cent in the CH and SE regions and the lowest rate of 88 per cent in the MRD region. However, there is a significant difference in the mean expenditures of households or per person among the regions (p<0.05) (see Table 2.12). Table 2.12 Expenditure, regular expenditure, expenditure per person and regular expenditure per person by year (thousand dongs) year Expenditure Regular expenditure (*) Expenditure per person Regular expenditure per person (**) 2004 14,211 a 12,913 a 90.9% 286 a 258 a 90.2% 2006 15,105 a 13,645 a 90.3% 306 a 275 a 89.9% 2008 17,657 b 15,626 b 88.5% 373 b 325 b 87.1% Total 15,657 14,061 89.8% 322 286 88.8% Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: (*) share of household regular expenditure on the total household expenditure (**) share of regular expenditure per person on expenditure per person a,b The different letter superscripts indicate those mean values within a column of the years that are statistically significantly different from each other at p<0.05) The expenditure increased at a higher rate from 2006 to 2008 period than from 2004 to 2006. In 2006, expenditure of the household was 15,105 thousand dongs, an increase of 6 per cent compared with 2004, while the household expenditure in 2008 was 17,657 thousand dongs, an increase of 17 per cent compared with 2006. Similarly, expenditure per person increased 7 per cent in 2004-2006 and a much higher increase of 22 per cent was recorded in the period from 2006 to 2008. Test of mean expenditure over the three 18

surveyed years only shows a significant difference in 2008, compared to the means of the other years. The proportion of consumption for daily life in total household expenditure was very high in the surveyed years. However, this proportion reduced over time from 91 per cent in 2004 to 89 per cent in 2008. 2.3.3 Land Land plays an important role in the economic activities of households, especially of households in rural areas. It is not only a key input for production, but it is also a determinant of household investment decisions. In the present study sample, there is a high degree of variation in the land area among households and among the regions. On average, households in the CH region own the largest land area (17,535 m 2 ) which is five times higher than the average land area of each household in the RRD. The NMM households have the second largest area (15,000 m 2 ). This is not surprising, the CH and the NMM are the central highland and mountain areas where population density is low. Table 2.13 Total land area and land area for production by region (m 2 ) Criteria RRD NMM NCCC CH SE MRD Total land area 3,538 15,006 4,723 17,535 9,674 10,780 (12,287) (34,087) (7,818) (23,099) (19,377) (20,043) Total land with LUR 2,383 10,125 2,369 8,683 7,165 9,103 (11,688) (28,740) (3,300) (22,137) (17,067) (14,407) Land for farm production 3,251 14,345 4,277 16,486 8,320 10,177 (12,253) (34,045) (7,774) (22,855) (19,377) (19,986) LUR Land for farm production 2,150 9,604 2,043 8,260 6,039 8,575 (11,684) (28,696) (3,223) (22,117) (16,984) (14,356) Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: Standard deviations are presented in parentheses) 19

However, as can be seen in Table 2.13, the CH is also the region with the lowest proportion of land area for which households have a long-term use right (LUR). The LUR of land is important factor in Vietnam because it provides some security for longterm investment. In the present study sample, land for production accounts for a large proportion of the total land area. Apart from the SE, all the other regions use more than 90% of the total land for farm production. Besides current land area and value, the household investment activity also relies on the land area that the household owns. Data from the VHLSS includes information on land area households owned in the previous year as shown in Table 2.14. Table 2.14 Land area of households in previous years by region (m 2 ) Year RRD NMM NCCC CH SE MRD Total 2004 2,931 13,820 4,247 13,350 9,420 8,746 7,419 2006 2,985 14,250 4,507 15,658 12,516 10,930 8,273 2008 4,273 16,059 4,003 21,377 7,609 11,201 8,803 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 On average, households in the present study sample increase the area of land owned over the years. As stated earlier, households in the CH and the NMM own the largest land areas. In addition, their land areas have increased gradually. In contrast, the NCCC and the SE are regions where households have smaller land areas and these areas have decreased over time. 2.3.4 Household expenditure on education The average number of household members who go to school is highest in the CH region and lowest in the MRD region (significant different at p<0.05). This could be explained by the fact that the CH has the youngest population with the average number of 2.61 household members under 23 years old, while this number is only 1.26 in the MRD region. However, the average household education expenditure is not as high as in the CH region. Despite the highest rate of school attendance, education expenditure by households in the CH region is only significantly higher than that in the MRD and 20

NMM regions. The SE and the RRD are the regions with a higher education expenditure per person attending school (p<0.05) (see Table 2.15). Table 2.15 Number of person attend school, education expenditure, education expenditure per person by region Criteria RRD NMM NCCC CH SE MRD Number of member under 23 years old 1.35 1.84 1.73 2.61 1.65 1.26 Number of member attend school 1.18 b 1.39 c 1.49 c 1.82 d 1.30 bc 0.87 a Household education expenditures 1,364 bc 852 a 1,272 abc 1,168 b 1,520 b 759 a Education expenditure per schooling person 1,235 b 732 a 869 a 735 a 1,246 b 822 a Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c,d, The different letter superscripts indicate those mean values within a row of the regions that are statistically significantly different from each other at P<0.05) The number of household members under 23 years old and members who attend school decreased gradually over the three surveys, with the greatest decline from 2006 to 2008 period. This is consistent with the decline in birth rate and family size across Vietnam. Table 2.16 Number of person attend school, education expenditure, education expenditure per person by year Criteria 2004 2006/2004 2006 2008/2006 2008 member under 23 years old 1.74 92% 1.61 89% 1.43 (1.40) (1.38) (1.32) member attend school 1.35 95% 1.29 91% 1.17 (1.13) (1.16) (1.09) Household Education 790 a 137% 1,080 b 140% 1,512 c spending (1,310) (1,979) (2,725) Education spend per 613 a 142% 873 b 158% 1,381 c schooling person (861) (1,192) (1,989) Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c, The different letter superscripts indicate those mean values within a row of the years that are statistically significantly different from each other at P<0.05 Standard deviations are presented in parentheses) 21

The expenditure on education increased sharply over the three surveys. In the period from 2004 to 2006, average household education expenditure increased by 37 per cent, with expenditure per person increasing by 42 per cent (see Table 2.16). The expenditure was even higher (p<0.001) in the period from 2006 to 2008 with the rate of 40 per cent and 58 per cent respectively. 2.3.5 Household expenditure on farm production assets Because all households in the sample are located in rural areas, farm production is the main activity of almost all households. More than half of household heads have their main job in the agricultural sector and agriculture contributes to 62 per cent of the average total income of households. For farm production, besides land and labour which have been mentioned above, farm machinery and other inputs such as seed, fertilizer and pesticide are also important. With the exception of the SE region, expenditure on farm machinery increased from 2004 to 2006 period and decreased from 2006 to 2008. Table 2.17 Farm production machines, expenditure on production machines and farm inputs by region and year (thousand dongs) Year RRD NMM NCCC CH SE MRD Total Current value of farm production machines Expenditure on production machines Expenditure on farm production inputs 2004 5,210 4,877 5,723 3,949 10,056 8,821 6,356 2006 4,056 7,951 7,263 5,019 12,221 7,565 7,006 2008 4,696 7,531 6,135 3,652 6,636 8,936 6,562 2004 84 289 133 81 237 70 145 2006 451 872 145 419 85 242 382 2008 356 366 174 273 6 284 267 2004 8,334 a 4,391 a 5,405 a 8,332 ac 13,878 ac 17,050 bc 8,771 2006 11,343 5,227 7,715 12,634 15,995 26,198 12,361 2008 10,684 a 6,259 ac 6,060 ac 14,652 a 8,369 a 21,825 bc 10,936 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c, The different letter superscripts indicate those mean values within a row of the regions that are statistically significantly different from each other at p<0.05) 22

Data from the above table shows that expenditure on farm production inputs, as well as machinery, varies from region to region, and changes from year to year. However, the differences across regions and years are not significant. 2.3.6 Household expenditure on other assets Of the other household expenditure and investment activities (i.e., those other than household consumption, education, medical, or farm production assets), household expenditure on housing accounts for the largest proportion. However, there is no significant difference in this expenditure among the regions. Buying shares and stock is not popular in the rural areas in Vietnam, so the average of household expenditure on this activity over the sample is only 409 thousand dongs. While households in the MRD prefer to keep money in the form of precious metals such as silver and gold, households in the SE region tend to put their money in saving accounts. People in the SE households also spend a lot of money on durable goods and fixed assets. Table 2.18 Household spending on fixed assets, house, stock, gold and saving accounts by region (thousand dongs) Region Fixed asset House Stock Gold Savings RRD 958 b 4,603 423 ab 373 b 1,296 ac NMM 680 ab 2,071 292 ab 125 b 338 ac NCCC 552 a 2,677 219 a 458 b 256 bd CH 997 ab 3,228 199 ab 300 b 3,606 a SE 1,085 ab 4,991 710 ab 1,453 bc 2,835 b MRD 607 ab 2,292 646 b 2,624 ac 161 cd Total 748 3,164 409 889 852 Source: Summary of current study sample, derived from VHLSS 2004, 2006, 2008 (Note: a,b,c,d The different letter superscripts indicate those mean values within a column of the regions that are statistically significantly different from each other at p<0.05) Although there is no significant difference among regions on expenditure for housing, this kind of expenditure increased significantly from 2004 to 2008 (see Table 2.19). 23