Lý thuyết về Các nhân tố ảnh hưởng đến thu nhập của người dân, kết quả mô hình về Các nhân tố ảnh hưởng đến thu nhập của người dân, các kiến nghị liên quan đến Các nhân tố ảnh hưởng đến thu nhập của người dân. Tổng quan về Các nhân tố ảnh hưởng đến thu nhập của người dân
Factors Affecting Asian-African Journal to Household of Economics Income andofEconometrics, the Kinh and the Vol.Ethnic 11, No Minority 1, 2011:in135-145 Rural 135 FACTORS AFFECTING TO HOUSEHOLD INCOME OF THE KINH AND THE ETHNIC MINORITY IN RURAL VIETNAM: A CASE STUDY IN THE BUFFER ZONE OF BACH MA NATIONAL PARK Hoang Van Long* and Mitsuyasu Yabe* ABSTRACT This research aims at studying the different effecting factors to household income between the Kinh and the Co Tu farm households in the buffer zone of Bach Ma National Park, Vietnam The linear regression model is employed to explain the effecting factors to household income of these two groups The results show that the household size, age, urban worker, garden, upland and credit effect significantly to household income Research results indicate the followings recommendations for policy implications (1)The Co Tu households should lower their household size and (2) they should access more to urban jobs to increase income for their household In addition, (3) the Co Tu should apply experiences from the Kinh to cultivate their land more efficiency Lastly, (4) the government should give more incentives for the Co Tu to access credit Keywords: Household Income, Agriculture Economics, Ethnic Minority, Co Tu, Vietnam INTRODUCTION Rural economy in Vietnam has improved significantly during the nearly two decades since the introduction of Doi moi 1) ‘economic renovation’ Vietnam has gradually transformed its economy from being agriculture-based to being market-oriented However, the economy is still dominated by agriculture, which occupies a high percentage share in the GDP About 80% of the population lives in rural areas, and two-thirds of the people depend on agriculture for a living Meanwhile, Vietnamese society is ethnically diverse The Kinh (“lowland Vietnamese”) comprise the majority, and account for 84% of the population They co-exist with 53 other smaller ethnic minority groups, Dang, et al [4] Almost all of the ethnic minority groups live in rural mountainous areas and suffer from poverty, Walle and Gunewardena [15] Income inequality or income difference is an economic problem, having social consequences often associated with factors such as economic growth, economic liberalization or location within former centrally planned economies such as Vietnam (See more Adger [1], Walle and Gunewardena [15], Heltberg [11], Epprecht, et al [5]) Understanding the dimensions of income difference and its affecting factors will contribute to the way that rural development and poverty reduction are to be implemented The economic renovation (Doi moi) has resulted in significant * Department of Agriculture and Resource Economics, Kyushu University, Japan, E-mail: hoanglongip@gmail.com, longeeco@agr.kyushu-u.ac.jp 136 Hoang Van Long and Mitsuyasu Yabe achievements, including poverty reduction But there is strong evidence of increasing variability in welfare levels in different regions, between rural and urban households, and along ethnic lines, Epprecht, et al [5] However, the development of inequality in Vietnam is considered to have been modest during the period of economic transformation, Heltberg [11] This study uses regression and decomposition methods to explore the income difference between Kinh and Co Tu2) households in the buffer zone of Bach Ma National Park in Vietnam We study two indicators to reflect household income levels: household characteristics and household income sources In other words, we try to examine (i) whether there is a significant difference in income between Kinh and Co Tu households in the same locality; (ii) what factors affect their incomes; and (iii) the differing sources of household income between the two communities The paper is organized as follows Section is a literature review of income measurement approaches and methods Section describes our methodology and includes our model specialization and data set In Section we present results and discussion, and finally in Section we analyze household income by sources and offer our conclusions and their policy implications LITERATURE REVIEW First, income inequality or income difference refers to differences in income between individuals or families, or between groups, areas, or countries, Black [2] It can be linked with skill, education, opportunities, happiness, health, life expectancy, welfare, assets and social mobility, Heshmati [12] Studies of income inequality have been conducted for over thirty years This section examines the progress of studies of income differentiation using household data, with special attention being given to regression-based approaches Oaxaca [14] and Blinder [3] developed the regression-based method for measuring income inequality They used it to measure wage differences in labor economics They employed many variables known as “individual characteristics” in their regression, and separated their models into two groups to quantify the inequality of wage income Their model tried to explain the reason that the whites earn much higher wages than the blacks and males earn substantially higher wages than females Fields [6] further developed the method by using an income generating equation to “account for” or “decompose” income inequality within a country and its change over time, Fields [6, Fields and Yoo [7] Gunderson [10] identified wage discrimination between male and female In defining the wage gap, he proposed methods such as narrowing defined occupation and regression wage decomposition He applied a regression that decomposed wage differentials using “characteristics” and “structure” differences between male and female Recently, Walle and Gunewardena [15] have applied the Blinder-Oaxaca approach in measuring the sources of inequality between majority and ethnic minority groups in Vietnam The model was adapted to the situation in Vietnam by adding geographic affects to their composition There were two reasons for this First, in the Vietnamese economy, an important characteristic determining household living standards is the location of the household Second, omitting consideration of geographical effects would severely bias estimates of the return to Factors Affecting to Household Income of the Kinh and the Ethnic Minority in Rural 137 non-geographic characteristics Walle and Gunewardera concluded that household income was strongly affected by location, education and land They recommended that the Kinh model of economic earnings should be separated from that of other ethnic minority groups in order to fight poverty Heltberg [11] applied the regression-based composition approach in measuring spatial inequality in Vietnam He employed household characteristics and sub-group indicator variables to regress the logarithm of expenditures against these variables The data were extracted from VLSS in 1992/93 and 1997/98 He concluded that higher education is also important in measuring inequality Moreover, growth of inequality during Vietnam’s reform period appears to have been caused by rising spatial inequality as well as growth of numbers in higher education Policies should strive to create the conditions for more equal returns between regions, Heltberg [11] In this paper, we apply the Walle and Gunewardera [15] approach of income inequality estimation using an econometric analysis of household data This approach explains the factors that affect household income and the levels of importance of these factors However, we modify it to make it appropriate for our data set and the local situation The regression excludes geographical effects because the data are all collected in the same location In addition, we analyze income by sources to show the difference of contribution of these sources to the household income of these two groups METHODOLOGY 3.1 Methodology of Data Analysis The research will use the linear regression model with the log of income per capita (lnW i) of household i The function is written as: ln Wi = α + β1 x1i + β x i + + β k x ki + ε i (1) where x1,…,xk are exogenous household characteristics, ei is an error term orthogonal (not multicollinear) to the explanatory variables Use of the left-hand side variables differs between authors However, it is normal to use demographic variables, land and other assets, education, and occupation In the literature, these variables are treated as exogenous, even if by a strict definition many of them may be endogenous, Heltberg [11] We used the regression model of the log of income per household (Wij) for ith household in a majority or minority group with household characteristics to calculate the factors influencing household income of the Kinh and the Co Tu In addition, to analyze the contribution of income from difference sources, we use the decomposition method to estimate the difference of household from sources between Kinh and Co Tu rural households 3.2 Data Collection and Dataset Data collection was conducted in August of 2004 in Doi village in Thuong Lo commune and in Village in Huong Loc commune, Nam Dong district, Thua Thien Hue province These two villages exhibit the characteristics that are representative of the social economic conditions in 138 Hoang Van Long and Mitsuyasu Yabe the buffer zones of Bach Ma National Park They were selected as our research sites based on several criteria Firstly, they are both in the buffer zone of the park and in rural mountainous areas Next, the ethnicities in these two villages are completely different Finally, their natural condition and settlement location have been undisturbed since the country reunification in 1975 The total sample for interviews was 60 households with 30 households in each village The method to select the sample was stratified sampling The total number of households in these two villages is 129 for Village and 98 for Doi village Based on community wealth ranking, there were three household economic groups: better-off, medium and poor We selected 10 households from each group in each village The households were interviewed by a researcher and four hired enumerators Although the questionnaire was carefully designed and tested before conducting the survey, some mistakes in the data set were unavoidable because of limited time and inexperience of enumerators in household interviewing They involved the accuracy and the representativeness of collected data The household characteristics used in this model were household size, age of household head, number of laborers who work in agriculture, number of laborers who work in urban areas, education level of household head, the political position of household head, area of garden, area of forest, area of upland, area of lowland, area of fishpond, and the total amount of household debt, with responding parameter βj, while εij is a zero mean error term that is assumed to be uncorrelated with Xij Table Definition of Variables Variables Unit Description Income Ethnicity Household size Age Farm workers Urban workers Education Position Forest Garden Upland Lowland Fishpond Credit VND* Log of income per capita Dummy variable to separate the two groups (Kinh = 1; Co Tu = 0) The total number of household member The age the household head Number of labors who work on farm at home Number of labors who work urban areas Number of years household head attended in school Dummy variable for position of household head The area of forest that household has (1,000 m2) The area of garden that household has (1,000 m2) The area of upland that household has (1,000 m2) The area of lowland that household has (1,000 m2) The area of fishpond that household has (1,000 m2) The total loan that the household borrows head year head head years dummy m2 m2 m2 m2 m2 Mil.VND Source: authors Note: *1USD = 16,000 VND in 2004 The dummy variable was employed to estimate the difference between groups The household size and age of household head are employed in the regression as household characteristics Working capacity is one of the most important factors affected by income They include farm and non farm workers However, non-farm workers are expected to have higher income 139 Factors Affecting to Household Income of the Kinh and the Ethnic Minority in Rural than farm workers These two variables were included in our model to reflect direct earning ability of households Education is considered as human capital to access technology in production This can be measured by the level of education of the household head In this research, it was assumed that the household heads of the Kinh and the ethnic minority would have the same quality of education, because of the difficulty in measuring education quality for the two groups Is that the political position of the household head will benefit to income? The dummy variable of political position is used to estimate this All land resources of the household were used in the regression, including forest, garden, upland, lowland (or wetland, or paddy rice field), and fishpond The unit of the land resource variables was recalculated by dividing by 1000 Credit is one of the factors that contribute to increased income The total amount of credit measured in Vietnamese Dong (VND) was used for the model The above variables cover all factors believed by the authors to be significant in evaluating household income Regressions were calculated first for Kinh households and second for Co Tu households using the same dependent variable and independent variables RESULTS AND DISCUSSIONS From the pool of data in the questionnaire, the log of income per capita was selected as the dependent variable, and the household characteristics were taken as independent variables The unit of household income was (VND) Other household characteristics were used as influencing factors to household income Table shows the means and standard deviations for both groups on the household income and its attributes, and production endowments Table Descriptive Statistics Variables Income log Household size Age Farm worker Urban worker Education Position Garden Forest Upland Lowland Fishpond Credit Observations Source: Author’s calculation The Kinh group The Co Tu group Mean Std Dev Mean Std Dev 14.60 7.07 54.53 3.03 1.50 4.60 0.23 1.59 11.74 1.46 0.39 15.38 2.98 30 0.47 2.16 8.14 1.65 1.59 1.48 0.43 0.73 22.87 2.12 0.60 23.31 2.27 14.52 6.10 51.47 2.50 0.07 4.37 0.43 0.59 11.20 1.81 0.86 0.31 6.38 30 0.58 2.31 18.51 1.11 0.37 3.38 0.50 0.47 10.68 2.78 0.98 0.26 6.18 140 Hoang Van Long and Mitsuyasu Yabe Table presents the result of the two regressions between two groups in the buffer zone of BMNP Subtracting the constant term of the regressions, the coefficient for each variable represents its relationship to income Table Coefficients of Regression Models Variables Pool data Coefficient (Constant) Ethnicity Household size Age Farm worker Urban worker Education Position Garden Forest Upland Lowland Fishpond Credit Observations R Square 14.99*** 0.22 -0.15*** -0.006 0.04 0.14** 0.006 -0.01 0.29** -0.01 0.04* -0.02 0.002 0.02* 60 0.521 The Kinh t 37.23 0.95 -4.24 -1.28 0.77 2.52 0.22 -0.07 2.66 -0.10 2.11 -0.34 0.23 2.14 Coefficient 16.34*** -0.11* -0.03** 0.02 0.10* -0.04 -0.39* 0.22 -0.37 -0.36 -0.76** 0.37 -0.02 30 0.648 The Co Tu t 17.15 -1.61 -2.31 26 1.57 -.77 -1.80 74 -1.38 -1.30 -1.86 1.39 -.58 Coefficient 15.34*** -0.16** -0.003 -0.05 0.19 0.007 -0.02 0.11 0.00 0.05* 0.05 0.06 0.03* 30 0.684 t 34.70 -3.19 -.44 -.47 75 18 -.09 49 -.02 1.76 52 15 2.08 ***Significant at 1%; **Significant at 5%; *Significant at 10% Tests for the difference of income per capita between two groups Wilcoxon rank-sum test The z value of this test is 0.429 and the Prob > |z| is 0.6681 The results indicate that the medians are not significantly different at any level smaller than 66.81% Therefore, the different income per capita between the two groups is a level difference and not a structure difference Hence, the difference of income must be explained by the difference of household characteristics and resources Tests for multicollinearity and integrating the regressions Multicollinearity occurs when a single independent variable is correlated with a set of other independent variables According to Gujarati [8], a correlation coefficient greater than 0.80 among explanatory variables is used as a rule of thumb to indicate the existence of serious multicollinearity There is no high correlation among our independent variables This implies that multicollinearity does not exist in the regressions Another test is the Chow test Its purpose is to test whether the intercept and all slope coefficients are the same across the two groups This test follows the F distribution If F > F for significance level á, the hypothesis that the parameters of intercept and slope coefficients are the same for the two sets of observations is accepted, Gujarati [9] Factors Affecting to Household Income of the Kinh and the Ethnic Minority in Rural 141 The F statistic is calculated based on the following formula: Fnk− k = SSR p − ( SSR1 + SSR2 ) n − ( k +1) SSR1 + SSR2 k +1 (2) where k is number of independent variables and n is the total observations SSRp is the sum of squared residuals of pool data, and SSR1, SSR2 are the sum of squared residuals in the separate regression of the two groups Base on the results of regressions of separated data for the two groups and on pool data, we calculate the F statistic as follows: 12 F36 = 7.710 − ( 2.506+3.039) 60 −2(12+1) 2.506+3.039 12+1 = 1.021 (3) The critical F-ratio at 0.05 level for 12, 36 degree of freedom is 2.033 At this level of significance, the computed F (=1.021) is lower than 2.023 As a result, the null hypothesis of the same intercept and slope coefficients for both groups is accepted This result is evidence to accept the hypothesis that pool data be used for both groups in the same regression Even so, it is interesting to add the coefficients of the two separate regressions to understand the factors that affect the two groups differently Returns to demographic characteristics Household size is significantly related to income per capita This is a social issue of a high growth rate of the population However, this is common in rural of areas of Vietnam It needs the effort of government to control population increase Otherwise, economic development with an improvement in living standards is meaningless because economic development cannot catch up the population increase The education level of the household head affects income positively Income per capita is higher in households where the household head has higher education because the household head with higher education can make better decisions regarding income earning activities Political power affects the Kinh negatively and the Co Tu positively However, it relates only slightly to household income For the Kinh, having political power in the community seems to be costly for them It takes much time for community work such as meetings and other services, but salary for these positions is low, so it may not cover the cost On the contrary, an ethnic minority having political power may benefit by getting priority in production such as from supporting services of government or other targeted programs Returns to allocation of labor force Labor force is one of the main factors affecting household income Kinh households have more laborers who work in urban areas so that they can send remittance home The number of Co Tu households with laborers working in urban areas is low (mean = 0.1) For both farm workers and urban workers, the number in Kinh households is higher than in Co Tu households This regression considers only the quantity of labor 142 Hoang Van Long and Mitsuyasu Yabe Returns to land resources Land is one of the most important resources in rural areas Efficiently used land can earn a higher income The households who have greater area of garden and upland tend to have higher income than others In the mountainous areas, most households have a garden Food crops, cash crops, and fruits can be planted in the garden Households may get higher income if they plant crops that can be sold in the market such as high value fruit or cash crops such as pepper and acacia nut These crops take some years to harvest and need high investment with advanced skills Therefore, the Kinh households may be able to afford these costs but the Co Tu households mainly use their gardens for food crops or fallow In addition, the upland is one of the key production resources for households However, shifting cultivation make the land usage less efficient In the past, the Co Tu practiced this cultivation method, but it has become illegal since the government tightened the regulations on forest protection Returns to credit Credit affects income positively to the Co Tu However, this can also be explained the the Kinh not lack credit to invest into their production since they may have more investment sources from savings and remittances than from loans DECOMPOSITION OF INCOME SOURCES It is interesting to analyze the income difference of the two groups by income source The table below shows the average income of the two groups from different sources together with descriptive statistics (mean, standard deviation) Table Income Sources Per household Unit: 1,000 VND Sources The Kinh Group The Co Tu Group Mean Std.Dev Mean Std.Dev Food crops and garden 2,966 3,662 1,568 1,010 Cash crops 4,875 2,744 332 558 Livestock and fishery 4,835 3,734 1,851 2,097 Forest product and forestry 1,581 1,273 4,754 2,511 Salary 232 725 3,610 3,316 Off-farm 730 2,011 480 1,120 Remittance 3,933 6,463 0 Total Income 17,200 9,448 12,500 5,164 Income per capita 19,200 10,300 12,600 5,223 2,425 1,100 2,359 1,342 30 30 Net Income Observations Sources:Authors’ calculation Factors Affecting to Household Income of the Kinh and the Ethnic Minority in Rural 143 Income from agriculture and forestry There is not much difference in income from food crops and gardens (rice, cassava, beans) between the Kinh and the Co Tu households Cassava production is more significant to the Co Tu In fact, this is one of the main crops of the Co Tu farm households Cash crops such as acacia nut, pepper, and sugar contribute significantly to household income for the Kinh Meanwhile the Co Tu farm households not produce these crops Concerning income from livestock and fisheries, livestock is the main activity of the farm households It contributes to their income, but it is not a high profit activity The Kinh households are generally more efficient in raising cattle or pigs In this study, we did not distinguish the breed of pig being raised, but Kinh households are generally known to use more productive pigs than those used by Co Tu households Off-farm income Non-farm activities also contributed to both groups After finishing their farm work, farm workers are involved in other activities such as carpentry, building construction, or selling their labor to provide additional income Co Tu households generally earn more money from these activities Salary Interestingly, the Co Tu households earn higher income from salary Because the Co Tu people were involved in the army during the Vietnam War, many of them have been receiving salary in the form of pensions since the war ended For the Kinh household, salary refers to payment for services as local officers The Kinh people tend to be preferred for appointment to such positions because they generally have higher education levels and work skills Remittance Remittance from household members who work in urban areas and send money back for their family is a significant source for household income and for rural development, but only occurs in Kinh households Rural households use the remittances for many purposes such as health care, education, living expenses and reproduction investments However, the ethnic minority households have only limited access to this source because they have difficulty with language, work skills and cultural barriers To sum up, the Kinh households get a higher income from agriculture production because of farming cash crops The Kinh households also get higher income from remittances On the other hand, the Co Tu households get higher income from forestry and from salary These differences are considered to result from differing group characteristics Economic transition encourages everyone to become involved in the market economy in order to gain a higher income In this respect, the Kinh are more dynamic with income generating activities, market access resulting from wise selection of cash crops, or by taking more jobs in urban areas where the ethnic minorities are considered to be more “backward” 144 Hoang Van Long and Mitsuyasu Yabe CONCLUSIONS AND POLICY IMPLICATIONS By employing both regression and decomposition methods, the empirical results show that the income difference between Kinh and Co Tu households in the same location is modest However, there are considerable differences in the sources of their income The Kinh households had generally higher income and from different sources as compared with Co Tu households The transition of economy blew new winds for household economy in Vietnam as a whole and in the buffer zone of Bach Ma National Park in particular Land tenure and opened markets were the two changes that benefited many farm households In the transition, The Kinh found better access to markets for the production and sale of their agricultural products so they earned more than the Co Tu The income difference between the Kinh and the Co Tu relates to group differences in demographic characteristics such as household size, education, labor forces and political position The Co Tu have practiced their agriculture by shifting cultivation with low productivity for a long time It was considered to be one cause of forest degradation Shifting cultivation is now prohibited by government laws Therefore, the Co Tu have transformed their economy into a more stable and oriented one Although both villages have the same history of settlement and development since the country reunification, the Kinh came from the lowland and brought with them cultivation technology and experiences that helped them develop further in their economy The Co Tu households could apply the techniques and experiences from the Kinh in order to improve productivity and gain greater value from agriculture Interestingly, credit has affected Co Tu households positively While the Co Tu still maintain their life and production by loans, the Kinh invest more for their production by their savings and remittances This is an important distinction relating to income and production between these two groups Investing by savings and remittance yields no return or interest for the Kinh, while on the other hand the Co Tu still get loans with interest from the banks It makes a big difference in production between these two groups and contributes to the difference in their incomes Rural developments in mountainous areas and among ethnic minority groups have long been concerns not only for local authority but also for the government This research looks at only a sample of rural areas in Vietnam We not observe a significant difference of income in structure between the Kinh and the ethnic minority groups The reason is that income from some sources, such as forestry income and salary, is high for the Co Tu but is not sustainable In the future, the forest resource will be depleted and the salary income cannot be continued over the long term Meanwhile, the income sources of the Kinh household are more sustainable For these reasons it is important to find out sustainable ways for the Co Tu households to maintain and improve their income in the future Therefore, our research yields some recommendations to be made in terms of economic development in the rural and mountainous areas such as the buffer zone of Bach Ma National Park: (1) The policies on population control should be maintained in order to improve living standards Factors Affecting to Household Income of the Kinh and the Ethnic Minority in Rural 145 (2) The government should give more incentives for the Co Tu to have better education (3) The Co Tu farm households should cultivate cash crops to increase their income and they should apply experiences from the Kinh in agriculture production (4) Rural credit is an efficient way for an ethnic minority to increase their income Notes Doi moi or economic renovation policy started in 1986 in Viet Nam One representative ethnic minority group lives in the buffer zone of Bach Ma National Park, Vietnam They may have other names such as Ca Tu, Ha, Gao, Phuong, Ca-tang.See more Kiwipedia [13] References Adger, W N (1999), “Exploring Income Inequality in Rural, Coastal Viet Nam,” Journal of Development Studies, 35, No 5, pp 96-119 Black, J (2003), Oxford Dictionary of Economics Second ed Oxford New York: Oxford University Press Blinder, A S (1973), “Wage Discrimination: Reduced Form and Structural Estimates,” Journal of Human Resources, Vol 8, No 4, pp 436-455 Dang, N V., S T Chu, and H V Luu Ethnic Minorities in Vietnam Hanoi: The Gioi Publisher, 2000 Epprecht, M., D Muller, and N Minot (2009), “How Remote are Vietnam’s Ethnic Minorities? An Analysis of Spatial Patterns of Poverty and Inequality,” The Annals of Regional Science Fields, G S (2002), Accounting for Income Inequality and Its Change: A New Method, with Application to the Distribution of Earnings in the United States Cornell University Fields, G S., and G Yoo (2000), “Falling Labor Income Inequality in Korea’s Economic Growth: Patterns and Underlying Causes,” Review of Income and Wealth, Vol 46, No 2, pp 139-159 Gujarati, D N (2003), Basic Econometrics 4th ed: Gary Burke Gujarati, D N (1970), “Use of Dummy Variables in Testing for Equality between Sets of Coefficients in Two Linear Regressions: A Note,” The American Statistician, 24, No 1, pp 50-52 Gunderson, M (1989), “Male-Female Wages Differentials and Policy Responses,” Journal of Economic Literature, Vol 27, No 1, pp 46-72 Heltberg, R (2003), Spatial Inequality in Vietnam: A Regression-based Decomposition World Bank Heshmati, A (2004), A Review of Decomposition of Income Inequality Institute for the Study of Labor Kiwipedia Nguoi Co Tu Oaxaca, R (1973), “Male-Female Wage Differentials in Urban Labor Markets,” International Economic Review, Vol 14, No 3, pp 693-709 Walle, D V D., and D Gunewardena (2001), “Sources of Ethnic Inequality in Viet Nam,” Journal of Development Economics, Vol 65, No 1, pp 177-207