Vietnam has made considerable progress in improving the living standards of its people since the reforms implemented under doi moi. This has been reflected in rising per capita expenditure and improvement in overall well-being. These improvements, among others, have been driven predominantly by a diversification of on-farm activities and economic reforms in all sectors of the economy. However, in the future, these successes may not be sustainable, as agricultural diversification is now facing with many constraints. Landlessness and scarcity of land have become major concerns. A high population density, a rapidly growing labour force and an increase in rural - urban migration have created great challenges for development policies. Furthermore, poverty remains a largely rural phenomenon. The rural - urban gap has widened considerably. Underemployment and seasonal unemployment remain a rural and agricultural phenomenon.

Therefore, an effective development strategy should be more focused on promotion of rural non-farm sector in order to create more employment and income opportunities for rural populations, narrow down rural - urban income gap, reduce rural poverty and avoid mass migration flows to cities.

To date, there have been some researches considering the importance and effects of rural non-farm employment and income in Vietnam including Thi (1998), Dien (1997), Bich and Quang (1999), Nam et. al. (1999), Reardon et. al. (1998), MARD (1998), World Bank (1998 and 1999a). However less work has been done on determinants of rural non-farm business income based on data from a nation- wide household level survey. Therefore, a study on rural non-farm business income is worth undertaking.

This research seeks to make three contributions. First, it looks at the importance of the rural non-farm activities in generating employment and income opportunities and the poverty alleviation. Second, it examines rural non-farm businesses and analyzes determinants of their income in Vietnam. Third, it recommends some policy implications to promote rural non-farm business activities in Vietnam in the coming years.

This analysis relies on the hypothesis that rural non-farm activities are operated by both poor households and non-poor households. These activities have enormous potential to play in generating employment and income opportunities, reducing the incidence of poverty and improving household living standards, and have an equalizing effect on rural income distribution. Household wealth and composition, education level, infrastructure, access to credit, and regions are the most important determinants of rural non-farm business income if constraints are addressed and appropriate incentives are provided.

This study focuses on: (i) reviewing theoretical and empirical issues relating to rural non-farm activities and factors affecting rural non-farm business income; (ii) identifying contributions of rural non-farm business activities and determinants of rural non-farm business income in Vietnam (iii) recommending some policies to promote rural non-farm business activities in Vietnam in the coming years.

The scope of this thesis is to examine determinants of income from rural non-farm business activities in Vietnam in 1997/1998. These activities are fully/partly owned or operated by households or household members in rural Vietnam.

The study will focus on addressing the following questions:

1. What are contributions and effects of rural non-farm activities in the economy?
2. What are the determinants of rural non-farm business income in theories and in other developing countries?
3. What are contributions and effects of rural non-farm business activities in the economy of Vietnam?
4. What are the major determinants of rural non-farm business income in Vietnam?
5. What are the recommendations for the government to promote rural non-farm business activities in Vietnam?

This thesis will firstly review the fundamental issues relating to the rural non-farm sector in the economy. Then, it will focus on descriptive analysis of rural non-farm business activities and quantitative analysis of determinants of rural non-farm business income in Vietnam in 1997/1998. Finally, it will suggest some policy implications for promoting rural non-farm activities in Vietnam. To do this, the analysis will employ both qualitative and quantitative techniques of analysis.

The study makes use of both primary and secondary data. Primary data were obtained from Vietnam Living Standards Survey (VLSS) conducted in 1997/1998. Secondary data were collected from various sources, including policy statements, official and unofficial reports, various comments and figures from published studies in the field, newspapers, reports of conferences, and documents and papers on the Internet.

This study would be more interesting if its analysis was developed on the basis of household – level panel data available between VLSS93 and VLSS98 and if the combined rural non-farm self-employment and wage employment were analyzed.

The paper consists of an introduction, four chapters and a bibliography. Chapter I provides an analytical framework and literature review. Chapter II firstly reviews major features of rural reforms and then provides a descriptive analysis of the rural non-farm business activities in Vietnam. Chapter III quantitatively analyzes determinants of rural non-farm business income. Chapter IV provides some conclusions and suggests some policy implications.

The rural non-farm sector is usually defined to include all economic activities in rural areas other than agriculture, livestock, forest, hunting and aquaculture activities. These activities are undertaken by households or household members as independent producers in their homes, the subcontracting of work to farm families by urban-based firms, non-farm wage employment in village and rural town enterprises, and commuting between rural residences and urban non-farm jobs. Income from rural non-farm activities is commonly defined as that derived from wage- paying activities and self-employment in the rural non-farm sector, and remittances. However this study will focus only on non-farm self- employment activities, which are fully/partly owned or operated by households or household members in rural Vietnam. These activities are also referred to as rural non-farm business activities. Rural non-farm business income is calculated as earnings from rural non-farm self- employment activities of households or household members.

Traditionally, the rural non-farm sector is regarded as a low productivity sector that produces low quality goods and is expected to wither away as a country develops and income rises. However, opinions have been moving farther away from this view in recent years largely because it is argued that this sector has an active role in productively absorbing a growing rural labour force, generating employment and income opportunities, reducing income inequality and poverty, dealing with the inherent problems of rural areas, minimizing rural – urban migration and developing intersectoral linkages with the agricultural sector.

The rural non-farm sector is a sizeable sub-sector of the rural economy in terms of income as well as employment. Evidence from nearly 100 studies between the 1970s and 1990s in Africa, Asia and Latin America shows that on average, around 32 percent of rural income in Asia, 40 percent in Latin America and 42 percent in Africa comes from non-farm sources and 44 percent of rural employment in Asia and 25 percent in Latin America engages in the non-farm sector. In many sub-Saharan African and Asian countries, not only do rural non-farm income shares increase over time, but there is also strong evidence of rapid increases in the share of rural non-farm employment. The importance of the rural non-farm sector is thus increasing.

The effect of non-farm activity on rural income distribution can be analyzed by the relationship between non-farm income, on the one hand and total rural income and/or landholdings, on the other. It can also be done by calculating Gini coefficients. Based on these methods however, Reardon et al. (1999) and Lanjouw and Lanjouw (1995) find a mixed picture of the effect of non-farm activity on rural income distribution. In some cases, rural non-farm activity increase overall rural inequality, while in others it affects a decrease. For example, non-agricultural income in rural Ecuador in 1995 had an unequalizing effects on rural income distribution, but non-farm income in rural Pakistan in 1994 had an equalizing effect.

This thesis has demonstrated that rural non-farm income is strongly associated with a lower incidence of poverty and higher household living standards as illustrated in Tanzania, Ecuador and Mexico. Moreover, rural non-farm income likely acts as a impetus that furthers agricultural intensification and expansion. This is drawn from evidence that farmers in Kenya and Rwanda with more rural non-farm income are more willing to grow cash crops or make higher productivity cropping choices. This leads to higher agricultural wages and improved agricultural productivity and thus raises income for farmers

Existing studies show that household wealth and composition, education level, infrastructure, access to credit and regional differences are important determinants of rural non-farm business income. Although there are numerous factors affecting rural non-farm business income, only those mentioned above will be considered in this study. While household wealth and composition largely affect the nature and types of businesses in the rural non-farm sector, education and rural infrastructure are crucial to the increase of rural non-farm business income. Primary education increases the productivity of the workforce and secondary education encourages entrepreneurial skills. Infrastructure widens the size and choice of markets as well as sources of raw materials for rural non-farm activities and encourages sub-contracting arrangements with urban enterprises. Moreover, it facilitates the flow of competing urban goods and introduces urban consumer tastes. Access to credit affects participation in rural non-farm business activities by providing investment and a buffer against risk. Finally, the agroclimatic characteristics of the zone also influence rural non-farm business income. Rural non-farm business income tends to be higher where there are strong urban-rural linkages and where agroclimatic zones are favourable.

This thesis provides empirical evidence on determinants of rural non-farm business income from the cases of Tanzania and Mexico based on econometrics model. It shows that household wealth and composition, education, infrastructure and regions have different effects on rural non-farm business income. While larger household size is usually associated with higher income, the higher dependency ratio reduces ability to earn the income. Gender also has certain influences on the income. Male entrepreneurs tend to earn higher income. There is also a strong relationship between the age of the entrepreneur and rural non-farm business income, which takes the form of an inverted – U age profile. This means that the income rises with age up to a certain level, thereafter it declines. Education is crucial to the increase of rural non-farm business income. Improved infrastructure is associated with higher rural non-farm business income. Regional patterns are also fairly important in explaining the difference of rural non-farm business income among regions. Other factors such as migrant, land owned per capital, cultivating household also affect the rural non-farm business income.

Vietnam’s reform process was initially designed to help the country to overcome economic stagnation and macroeconomic instability in early 1980s. However it failed to deal with more deep-rooted problems. In December 1986, the Sixth Party Congress recognized the shortcomings of old development model and approved a program of major reform of the system of economic management – "doi moi". The major reforms affecting rural economy implemented under doi moi include (i) reversal of the process of collectivization in agriculture; (ii) price liberalization to eliminate the ‘two price’ system, reduction in rationing and subsidies and the removal of barriers to internal trade; (iii) formal acceptance of the private sector and the reaffirmation of equality in various economic sectors; and (iv) partial liberalization of international trade and foreign exchange market. Some other important macroeconomic reforms in banking, foreign investment, the tax system and state owned enterprises have also influenced the rural economy.

The rural economy has responded well to the changes under doi moi. Agricultural output and productivity have risen significantly. A substantial improvement in per capita expenditure and overall wellbeing of the vast majority of people in rural Vietnam has been achieved. Rural hard and soft infrastructure has been continuously developed. Doi moi has liberalized both production capacity and the labour force, and directly created the impetus for rural development.

The vast majority of Vietnam’s population and labour force is located in rural areas. The farm sector absorbs almost three-quarters of the total rural labour force and the rest crucially depends on the non-farm sector for employment. At present, there are over 10 million labourers engaged in the rural non-farm sector, if both wage employment and self-employment are combined. Wage employment only accounts for between 10 and 15 percent of rural non-farm employment. The rest is operated and managed by households or household members.

Rural non-farm business activities in Vietnam play an important role in creating employment and income opportunities. Around one fifth of rural employment and household income was generated from non-farm business activities. Employment and household income in rural non-farm business activities have grown rapidly over 1993 – 1998 period. Employment in these activities has grown by nearly seven percent per annum and rural household income from these activities has also grown by 30 percent in real terms during this period. As limited arable land, over- exploited natural resources and constraints in the diversification of cash crops and expansion of livestock are appearing as major concerns in many locations, rural non-farm business activities will become increasingly important in absorption of labour and the generation of household income in rural areas.

This thesis find that there is a strong positive and nearly linear relationship between rural non-farm household income and total rural household income. As Figure 1 shows, rural non-farm income share rises sharply with the income groups. This suggests an inequalizing effect in rural non-farm business activities on rural household income distribution. Reardon et al. (1998) also found similar findings in rural Vietnam in 1993. Therefore, the pattern of effect of rural non-farm activities on rural income distribution has not changed over 1993 – 1998 period. Although rural non-farm business activities increase the overall rural inequality in Vietnam, these activities have significantly contributed to slowdown the widening of the rural – urban gap and thus reduce social tensions and instability over 1993 – 1998 period.

Participating in rural non-farm business activities allow households to achieve higher standards of living. As Table 1 shows shares of rural non-farm employment in total rural employment rise sharply with living standards, as proxied by consumption expenditure.

By engaging in rural non-farm business activities, rural households can move their families out of poverty. This is drawn from the probit model of probability of poor households in rural Vietnam. The probit model has indicated that poor households in rural Vietnam in 1998 are more likely associated with having female and young household head; lower education level; larger household size; low involvement in non-farm business activities; higher incidence of borrowing; small landholding, and living in communes with less developed infrastructure and few commerce and trading transactions, and in the Northern Uplands (Table 2).

Rural non-farm business activities in Vietnam are very heterogeneous, ranging from farm services, fisheries, food processing, manufacturing, commerce, construction, and transportation to other services. Three- quarters of rural non-farm business activities are concentrated in food processing, manufacturing, commerce and services. These have been mainly developed on the basis of the available local raw materials and human resources. These activities have significantly contributed to increased consumption of agricultural produce, rising agricultural wages and generating economic incentives for agricultural production and extension over the last few years. This suggests that there are production and consumption linkages between the rural non-farm sector and the agricultural sector. Moreover, rural non-farm business activities tend to be most developed where there are strong urban-rural linkages and located in favorable agroclimatic regions, as illustrated in the South East, the Red River Delta and the Mekong Delta.

Most of rural non-farm businesses employ one or two workers, from the household or elsewhere and use small amount of physical capital. While construction, mining and farm service are more likely labour intensive than others, transportation and fishery are more likely capital intensive. There are similarities among regions in terms of the number of workers, but businesses in the South have a higher capital to labour ratio than those in the North.

The diversity and flexibility of rural non-farm business activities are very appropriate for labour mobility and seasonality in rural areas. With low entry and exist cost, rural workers/ entrepreneurs can move easily from less productive to more productive businesses. This has increased the efficiency of labour and resources utilization in rural areas since the beginning of doi moi.

Compared with other forms of enterprise in both private and state sectors, the performance of rural non-farm business activities in 1997-1998 was quite high. On average, each business activity generated a total income of 4.88 million dong over a period of one year. Only nine percent of enterprises are loss- making. Although the rate of tax payers is still low, around one fifth of the enterprises pay taxes, at the average amount of 1.15 million per year. Businesses in the South earn higher income and pay higher taxes than in the North.

This thesis finds that rural non-farm business activities are operated by both the poorest and richest households. This is drawn from positive relationship between the size of businesses as measured by number of workers, production assets and profitability and household living standards as proxied by consumption expenditure. One should not expect, therefore, to see this sector diminish during the development process.

Despite the impressive achievements that have been made over the last few years, poverty remains a largely rural phenomenon, with 90 percent of the poor living in rural areas. Reforms have been associated with indeed widening the rural - urban gap. This is because during 1993 – 1998 period, per capita expenditures in rural areas increased by 5.4 percent per annum, while in urban areas they grew twice as fast. Underemployment and seasonal unemployment were and remain a rural and agricultural phenomenon. Over a quarter of rural labour force was underemployed in 1998.

Rural non-farm business activities are facing with some constraints namely: (i) weak marketing and low competition of products; (ii) poor and obsolete technology; (iii) weak entrepreneurial skills; (iv) insufficient business and market information; and (v) shortage of capital and limited access to credit. Qualification of rural non-farm entrepreneurs is also a major concern. Only one-third of female entrepreneurs and half of the male entrepreneurs have secondary education level or higher. This will cause a lot of difficulties for long term development and expansion of rural non-farm business activities.

Some restrictive regulations on the expansion of household enterprise activities also discourage the development of rural non-farm business activities. The development of non-farm activities has a negative impact on the environment, especially air and water pollution in some traditional and handicraft villages. This has a negative effect on the health of the community and pollutes the ecological environment. The working environment for workers, especially women, in many garment, textile, apparel, and chemical businesses is also very poor. This directly affects their health and reduces labour productivity in the long run.

Rural non-farm business income in Vietnam in 1998 is calculated on the basis of the question that asks how much money the entrepreneur has left after making purchases for his work, combined with the value of any goods or services of the business activity consumed by household members. Thus this measure of the income will firstly include the returns to all factors of production: (i) all paid and unpaid family labour and (ii) family-owned physical capital and inventories. Secondly, the characteristics of the entrepreneur may be important. Age, sex, and level of education may affect the overall performance of the business activity. Thirdly, the income is likely to be affected by the characteristics of the household. Household size, leasing in and leasing out land, and borrowing and lending out money or goods for business may affect the participation of household members in rural non-farm business activities. Fourthly, the income is also influenced by the characteristics of the business activity. Business location, business license and age may be related to the nature and newness of the technology, and the capital embodying that technology. The type of industry will determine profitable opportunities and the degree of competition. Finally, infrastructure factors and regional differences may also affect the income.

This can be demonstrated with the following equation. Let Y_i denote income from the rural non-farm business activity i, X_i is a vector of family- owned inputs, E_i represents a set of the characteristics of the entrepreneur, H_i indicates a set of the characteristics of the household; B_i denotes characteristics of the business activity; I_i is a vector of infrastructure variables, and R_i indicates seven regions of Vietnam. Let u_yi be a random disturbance, capturing unmeasured determinants of and stochastic influences on rural non-farm business income.

Y_i=F(X_i, E_i, H_i, B_i, I_i, R_i, u_yi) (III.1)

The regression model (III.1) will be estimated by single- equation regression techniques. These techniques assume that explanatory variables are exogenous, in other words, that they are not manipulated in response to rural non-farm business income. Yet, according to Vijverberg (1995 and 1998) and Glewwe (1999), within the model (III.1) simultaneously determined inputs (such as capital, inventories and labour) and industry indicators (such as dummy variables for commerce or other services) are likely to be endogenous. Therefore, the inclusion of these variables in the model can cause a simultaneity bias for OLS estimates. To remedy such a bias, these variables are excluded from the model. In the following, the model without inputs and industries is referred to as the "reduced-form specification" and the model with inputs and industries is referred to as the "structural- form specification".

Economic theory offers no guidance as to whether rural non-farm business income should be measured as daily, weekly, monthly or annual values. It is also unknown what the relevant period of production is, and this period may vary among businesses in different industries or even within the same industry. Both the monthly income and daily income from rural non-farm business activities shall be examined.

Economic theory suggests that the impact of inputs is nonlinear due to diminishing returns therefore, a quadratic functional form will be used. Given the nature of cross-sectional data, it is possible that, heteroscedasticity may be present. Based on the methodological exploration of these issues in Vijverberg (1991b), the analysis here will use a log- linear specification, estimated by OLS with White- heteroscedasticity corrected standard errors.

The empirical analysis in this chapter is mainly based on data from Non-farm Self-employment Section of VLSS98. There are 1985 observations available for the quantitative analysis. All observations in the sample were constructed by a multi-stage sampling procedure. Enterprise weights were calculated and attached to each rural non-farm business activity in the dataset to ensure that statistical representativity can be preserved. Information on sample design will be explicitly incorporated in the quantitative analysis.

Table 3 shows definitions of dependent and independent variables used in the regression model. The monthly rural non-farm business income is measured as the ratio of annual rural non-farm business income over the total operating months of the business in the past 12 months. The daily rural non-farm business income is measured as the ratio of the monthly rural non-farm business income over total operating days per month. All nominal values (income, capital and inventory) are converted into real values by adjusting for both monthly consumption price and regional price index.

This thesis analyses quantitatively the determinants of monthly and daily rural non-farm business income in Vietnam in 1998. The results presented in Table 4 and 5 show that returns to physical capital, labour, and human capital are positive and statistically significant.

The parameter estimates in the structural forms indicate that returns to labour and physical capital are considerably high. Although the results show that the returns to physical and human resources exhibit the diminishing returns to scale, there are more opportunities for increasing factors of production (i.e capital and labour) in rural non-farm business activities. This is drawn from the average values of capital and labour presented in Table 3 are much lower than the turning point values of capital and labour. This is very important for rural non-farm development in the coming years.

There is strong evidence of positive rate of return to education of the entrepreneur on rural non-farm business income. The rate of return to education of the non-farm entrepreneurs in rural Vietnam is the same as that estimated by Vijverberg (1995) for Ghana, Lanjouw and Sparrow (1999) for rural Tanzania and Adams (1999) for rural Egypt. The allocative efficiency effect of education of the entrepreneur mainly reflects the better entrepreneurial choices about quantity of inputs, not the industry of economic activity.

There is also a strong relationship between the age of the entrepreneur and rural non-farm business income which takes the form of an inverted- U age profile with a peak at approximately 35 - 39 years of age in the monthly income model or approximately 37 – 40 years of age in the daily income model. Although the participation rate of female entrepreneurs is about half, the average monthly and daily income of female-operated businesses are much lower than that of male-operated businesses. This indicates gender discrimination in rural non-farm business activities. Such is the case in Tanzania (Lanjouw and Sparrow, 1999), Mexico (Lanjouw, 1998a) and Ghana (Glewwe, 1999).

The characteristics of the household also play an important role in the determination of rural non-farm business income. It is found that the larger the household size, the higher the income from rural non-farm business activities. While leasing- in of land was significantly associated with lower monthly income, leasing- out of land was insignificantly associated with higher monthly income. None of the land transaction variables in the daily income model is statistically significant. This suggests that the rural non-farm and agricultural activities present two alternative strategies for raising household income on the monthly basis rather than the daily basis. While households who engage in borrowing transactions of money or goods do not appear to contribute to rural non-farm business income, households lending out money or goods could help to rise the income. The fact that over half of the loans of rural households come from informal financial sources (such as money lenders, relatives, friends) with high interest rate. Many rural household businesses have little or no ability to offer collateral for loans from formal financial institutions. This makes participation in borrowing transactions have little effect on the income.

Features of the business activity, including whether it operates from a fixed location, whether it has a business license and how long it has been operating, have statistically significant effects on rural non-farm business income. The age of the business activity exhibits an inverted U age profile with a peak between 24 and 25 years of operation in the monthly income model or between 29 and 34 years of operation in the daily income model. The monthly and daily income is significantly lower for businesses operating from a fixed location and significantly higher for those having a business license. At the 10 percent significance level, construction and transport in the daily income model and construction, commerce and transportation in the monthly income model do not appear to be statistically significant. The average monthly and daily income in other industries is significantly lower in comparison to the farm services industry (the omitted industry dummy).

Although infrastructure variables have positive effects on monthly rural non-farm business income, none of them are statistically significant at the 10 percent significance level. It is likely that most of rural non-farm business activities are developed on the basis of intersectoral linkages with agricultural activities within the locality. This suggests that there is enormous potential for rural – urban linkages, business linkages and intersectoral linkages between localities to play a more important role in increasing rural non-farm business income.

Regional patterns are also important in explaining the difference of rural non-farm business income among regions. Rural non-farm business income in the Red River Delta and North Central Coast does not statistically differ from that in the Northern Uplands (the omitted regional dummy). The income in all regions in the South tends to be higher than that in the Northern Uplands. The South East could be considered the favorable region for generating rural non-farm business income.