COMMUNITY PARTICIPATION FOR DEVELOPING RURAL TRANSPORT INFRASTRUCTURE IN DONG NAI PROVINCE, VIETNAM

Summary. Developing countries are facing sustainability problems in developing rural transport infrastructure; the problems arise from the lack of financial resource and public support. Enhancing community participation promises to resolve these challenges. Demographics has a strong influence on community participation; however, how demographic characteristics interact with community participation has been yet to be focally studied. Dong Nai, a southern province in Vietnam, was selected as the case study owing to its high requirement of rural transport infrastructure. Questionnaires were sent to 438 households. An analysis of data was performed descriptively and with logistic regression. Community participation was recorded as financial, labor, material, and land contribution. Household composition, education, income, and member in household as community leader are among the highest influential factors to community participation. With the importance of community participation, the results of the study benefits both policy makers and local governments in planning, predicting the success of local development programs, and enhancing community participation, which increase the chance of success in rural transport infrastructure.


INTRODUCTION
The current total length of district and commune roads in Vietnam is about 176,863 km, making up to 60.57% of the country's road network. This includes district roads of 45,999 km (15.75%) and commune roads of 130,864 km (44.81%). In addition, there are 2,360 rivers and channels in Vietnam with a total length of 80,000 km, of which 41,900 km with sufficient depth for waterway transportation (Hoang, 2019). The value of the multi-billion-dollar rural transport system, if not properly exploited, managed, and maintained would be a great waste and affect the majority of beneficiaries [5]. Until 2030, Vietnam would need approximately 150-160 billion USD for investment in infrastructure development and improvement (of which 40 billion for electricity, 53 billion for road, three billion for railway, and 25 billion for ports) [2], which would require a capital of approximately 15-16 billion USD every year. The current national capital D.T. Hai, N.X. Quyet, D.A. Nguyen sources including government budget capital, Official Development Assistance (ODA), and bonds can meet only about 50% of that investment demand [13]. Consequently, the government cannot afford to build infrastructure for all levels, and the participation of all other stakeholders is inevitable.
Different from national and urban infrastructure systems, rural infrastructure serves the community in that rural area itself. The participation of the community is therefore essential, as it ensures sustainability [1]. To mobilize community participation in rural infrastructure development, erroneous and incomplete infrastructure should be avoided. Theoretical and practical bases are first necessary to be studied for proposing suitable solutions and policies. Policy makers and local governments in Vietnam are facing conundrums in planning successful investment in rural transport infrastructure and predicting outcomes of current developing programs as these actions request much and in-depth knowledge of the country's current and future resources, local demands, and socio-economic attitudes toward the programs [6,8].
This study aims to explore the current status of community participation in Vietnam through a case study of Dong Nai province, a typical southern delta zone that is in heavy need for rural transport infrastructure. The second purpose of the study is to examine the relationships between demographic characteristics and community participation itself. As demographic and socioeconomic trends and community concerns belong to the most important factors that affect transportation planning, the results of this study can help policy makers and governments in solving the tandem conundrums of developing rural transport infrastructure.

LITERATURE REVIEWS
Present participation of community on developing infrastructure for rural transportation had been studied by many authors in Vietnam and the world (e.g. [15, 10, 3 and 7]). They concluded for developing infrastructure of rural transportation, there are needs of mobilizing social resources with community participation (society, enterprises, community, and local government). From practical view, the World Bank ( [17]) also confirmed a more reliance on Public-Private Partnerships, parts of which are based on community and community network. Public-Private Partnerships belong to a strategy that has been applied for rural transport infrastructure development by many countries in the world.
Owing to the characteristics of rural transport infrastructures, the processes of construction and development happen in the long-term periods, with many community-participated activities, such as define planning demands; establish costs and participated policies; contribute physical resources (money, labour, materials, land); site construction; supervise and check; manage and maintain; and utilize and evaluate. Among these, physical contributions are commonly indicated by several authors (such as [9, 6 and 14]). Therefore, the participation of community on developing infrastructure for rural transportation is affected by 10 factors, including age, sex, background, education, the number of main supporters on households, household classification, household members working as local officer, households' average income, income from agriculture, and living conditions in the area.

RESEARCH METHODOLOGY
From the literature, demographic factors that affect community participation are synthesized. Questionnaires were devised and included in surveys sent to local households in DongNai, a southern province in Vietnam.
Data obtained from the surveys, combined with data from the Department of Statistics of DongNai province [4], were analyzed descriptively to reveal a general snapshot of the demographics of Dong Nai province and the current status of community participation. Then, a logistic regression was performed to explore the influence levels of demographic factors to community participation.
Results were carefully considered to explore potential directions to improve the effectiveness of transportation planning, to predict the outcomes of considered projects, as well as to enhance community participation itself in certain circumstances. . The necessity and urgency of the local people tend to lead to development of rural road infrastructure, alleys, inter-village roads, and inter-village roads. Therefore, on the whole, the need for rural infrastructure development of DongNai is now urgent for the community in general.

Status of Community Participation
a) Analyze the participation by forms of participation -The community surveyed 438 households in the locality. Overall, 94.7% of households provided information for each type of road, and only 5.3% did not participate. Participation depends on the interests of the community itself and the policy of calling for participation, that is, want the community to participate; it is necessary for the community to see the benefits of their participation. At the same time, to improve the effectiveness of participation, it is very important to promote and encourage participation, according to each type of locality.
-The community has contributed professional comments. The survey has 67.8% of households involved in contributing ideas for one or more stages (identification of needs and planning, policy development planning, resource contribution, direct construction, inspection, supervision, operation and maintenance, maintenance, and efficiency evaluation). Contributions are mostly voluntary because participants must have a clear understanding of what they are involved in. As a result, 32.2% of the surveyed households did not comment on any of the stages of the rural infrastructure development process, as they thought they were not well understood, or were not proficient.
-The community has participated in policy formulation: Participation in policy formulation includes participation and calling for investment, calling for participation, and rights and responsibilities of organizations, people, communities, etc. in the development and construction of rural transport infrastructure. Policy formulation has long been understood as the work of local authorities and community organizations.

Analysis of Demographic Factors Affecting Community Participation
The model of Binary Logistic Regression (Equation 1) [11] is often shown as follows: where P(Y1=1) is the odds that the community is involved, P(Y1=0) is the odds of not participating, and β (0, i = 1..n) are the coefficients. The independent varbiables Xi are defined with convention as follows: • X3 -professional training (certificate, 2-year college, 3-year college, university, and postgraduate), • X4 -level of education (primary school, secondary school, and high school), • X5 -number can provide workforce in the household, • X6 -income class (poor, lower income, middle income, and higher income), • X7 -member(s) in the household as community leaderships, • X8 -average income per capita in the household, • X9 -proportion of income from farmwork, and • X10 -terrain of living area (mountainous, hilly, plain, and others). The data were analyzed in IBM SPSS 25. As the first step of any regression procedure, data screening -collinearity check -is performed for the variables (Fig 1). The scatter plot in Fig. 1 shows some correlation between some pairs of variables such as income per capita and education level, as well as income per capita and training level, which are reasonable and predictable. The remaining variables do not appear to form any curvilinear patterns, implying a high robust when applying logistic regression. In the test, P(Y1=1) is the probability that the community financial contribution P(Y1=0) is the probability otherwise. Tab. 3, 4, and 5 show statistics for the test of financial contribution. Tabular statistics of other tests (labor contribution, material contribution, and land contribution) are omitted for the sake of the paper's concision; however, those values are discussed in paragraphs.

Financial Contribution
Regression analysis based on the significance level (Sig.) of the Wald test shows that from X1, to X5 and X7, there is Sig., all > 0.05. Therefore, the correlations were not significant for the Y1 variable, with a 95% confidence interval; The X6 has Sig. < 0.1. Therefore, the correlation coefficients are significant for the Y1 variable, with a reliability of 90%. The variables X8 and X9 have Sig <0.01. Mean, the correlation coefficient was significant for Y1, with a reliability of 99%. Omnibus testing shows Sig. <0.05 (95% confidence). Thus, independent variables are related as shown on Tab. 5.
As a result of the analysis, the average income and income from agriculture and household economy have a direct effect on the contribution of community money to rural infrastructure development. Specifically, up to 42.50% of households decide to donate money for every 100 households and increase their income (from poor to near-poor households / from near-poor to middle-income households / from middle income to good income housholds) to one million VND. The average household income and 1.00% increase in the household's income from agriculture will increase by 3.70% and 2.30%, respectively, of community contribution to rural infrastructure development. Thus, the household economy has the greatest effect on the decision to contribute because if the household income is relatively high, the contribution will be good, and if the economy of households is good, the demand for 52 D.T. Hai, N.X. Quyet, D.A. Nguyen participation will be higher. This is a matter of concern when mobilizing the community to contribute their money.

Labor Contribution
In the test, P(Y2=1) is the probability of the community to contribute labor, and P(Y2=0) is the probability otherwise.
Regression testing based on the significance level (Sig.) of the Wald test shows that X1, X2, X4-X6, and X8-X10 have Sig. all > 0.05. Therefore, the correlations were not significant for the Y2 variable, with a reliability of 95.00%. The X7 has Sig. = 0.012 <0.02. Therefore, the correlation variable is significant Y2, and the reliability is 98%. The variables X3 have Sig <0.01. The correlation coefficients were significant for Y2, with a confidence level of 99%. Validation of the model: with 432 households participating, the model predicted 432 true, so the correct rate is 100%. The remaining 6 households did not participate, but the predicted model was 2, and the true rate was 33.30%. The predicted accuracy rate of the whole model is 99.1%; Omnibus inspections show Sig. < 0.01 (99% confidence). Thus, independent variables have linear relationships with dependent variables in the aggregate. The model selection is appropriate.
In conclusion, having family members working as community leader in households greatly affects the contribution of labor, with the rate of every worker having a cadre of staff at least 7.42 working days for development of Rural Transport Infrastructure, often staffed by staff members, is more likely to be involved in and advocate for "role models". This can be considered as a solution in the promotion of community participation.
The professional qualification of the head of household affects the total number of working days which is contributed with 3.57 unskilled laborers per working day. However, the use of skilled workers should be planned and surveyed for appropriate use in order to effectively engage in, because of the need for manual / non-professional work and vice versa.

Material Contribution
In the test, P(Y3=1) is the probability of community contributing material, and P(Y3=0) is probability otherwise. Validation of the model: with 279 contributing, the model predicts exactly 252 households, so the correct rate is 90.30%. A total of 159 households did not participate, but the predicted model was 28, the true rate was 33.3%. Therefore, the predicted correct rate of the whole model is 63.9%, Omnibus Test Sig. = 0.002 <0.01 (99% confidence). Thus, independent variables have linear relationships with dependent variables in the aggregate. Therefore, this model is selected.
From the results of the analysis, the level of expertise of the community has a high effect on the contribution of materials, with a predicted level of 1 that will contribute more valuable material than 0, VND 51 million per household compared to non-qualified persons. It is argued that those with higher levels of competence are more capable and conscious of contributing; they often suggest different forms of participation when they are not able to participate in forms such as love. They like to work because they are busy because they have no land near the road to donate land.
The site / natural conditions are the conditions that stimulate the community to contribute material, with the level of one household having material conditions in place will tend to contribute material with a value higher than 0.34 million VND / household compared with people without conditions. Thus, the contribution of material depends on the condition, the availability of the community, and/or the ability to mobilize the material.
The contribution of material is not only dependent on the economic condition of the household but also on the contrary to the specific characteristics of the rural economy. On the spot, with the motto "what contributes" and this is considered as a breakthrough solution in enhancing the participation of rural communities for rural infrastructure development. Because this solution encourages many people to participate and limit the difficulties of the People's Council, creating fairness in participation.

Land Contribution
In the test, P(Y4=1) is the probability of community land donation, and P(Y4=0) is probability otherwise.
Regression testing based on the significance level (Sig.) of the Wald test shows that X1, X3-X6, X8, and X10 have Sig. > 0.05. Consequently, the correlations were not significant for the Y4 variable, with a 95% confidence interval. The X2 and X7 have Sig. = 0.086 and 0.030 <0.1. Consequently, the correlations were significant for Y4, with a reliability of 90%. Variable X9 has Sig = 0.002 <0.01. The correlation coefficients were significant with Y4, and the reliability was 99%. Validation of the model: with 307 households involved in land contribution, the model correctly predicted 296 households, and the correct rate is 96.40%. The remaining 131 households did not participate, but the predicted model was 14, and the true rate was 10.70%. Therefore, the predicted correct rate of the whole model is 63.9%; The Omnibus test shows that Sig. = 0.001 < 0.01 (99% confidence). Thus, independent variables have linear relationships with dependent variables in the aggregate. Select the appropriate model.
The percentage of income from agriculture and household with family members working as community leader directly affects land contribution. In which, the number of staff involved in land use and land contribution is better than that of other community members by 0.795m 2 . This is also explained by the "role model". The percentage of income from agriculture also has an impact on the contribution of land, with an impact of 1.80% of the value of the contribution, which is interpreted to mean that rural people have lower incomes than urban land prices. Therefore, the decision on land contribution does not depend much on income, but the land contribution is only related to households with land near the road to be expanded. Beside the relative distance from their houses, contribution propaganda is important to the community of people involved.

DISCUSSION AND RESULT IMPLICATIONS
Results indicate a few factors that have medium to high effect on each category of community participation. A preliminary validation, obtained by common sense arguments, is followed. • Financial contribution: gender (negatively medium), education (positively medium), and household number of workforce (negatively medium). Because of the way gender was coded (0 is female and 1 is male), the result means that households that are represented by male citizens tended to participate more in financial contribution. The reason behind this observation, though interesting, requires another socio-economic study. It is suffice to assume that the more educated citizens are, the more they are aware of the benefits of local transport infrastructure to themselves; furthermore, they can afford to contribute financially. Hence, the effect of education is confirmed. The negative effect of household number of workforces can come from the famous negative correlation between the number of children to the financial well-being of families in Vietnam. • Labor contribution: occupational training level (negatively medium), income class (negatively high), household member as community leader (positively high), and terrain of living area (positively high). If a person is highly trained occupationally, it is not surprised that he/she works fulltime in that occupation, hence less free time to contribute labor to the local projects. Similarly, a high-income household would rather contribute financially than direct labor. The high effect of household member as community leader is interesting. This confirms the motivation of the good social model as verified in literature, as why people volunteer. The positively high effect of terrain of living area was hard to reason, but one potential explanation could be the relationship between types of jobs of the local people and the terrains that surround them. • Material contribution: sex (negatively medium), occupational training level (positively high), education (negatively medium), and terrain of living area (positively medium). Effects of these characteristics are hard to reason about as there should be further studies of which types of materials contributed to the projects. • Land contribution: sex (negatively medium), household member as community leader (positively high), and terrain of living area (positively medium). Land contribution is by far the most interesting observation in community participation, and this type of contribution was one of the most important factors that affect all infrastructure project schedule in Vietnam. Again, household member as community leader plays an important role in determining whether a household will donate their land to community. The medium effect of terrain of living area could be resulted from the land requirements from rural transport projects: projects located in flat surface areas are more common and require more land. The findings of this study can benefit the government in the following areas: a. The planning process for the government in developing rural transport system. "Public [community] participation is a continuous effort of every transportation planning and project development process… transportation planners need to be aware of the expectations of an effective and successful program" [12]. Many variables must be considered in the process of planning transport systems. As the governments are fully aware of each areas' demographics, they can predict demands and resources of these areas, which in turn profiting the planning process.
b. The capability of the government to predict the outcomes of current development programs.
Any model built by the government to predict the outcomes of a development program would use community participation as an input variable. In circumstances where community participation has not been first-order determined (e.g., by surveys), an understanding of demographic characteristics can provide a shorthand for community participation.
c. An opportunity to increase the probability of success of considered programs through increasing community participant.
Revisiting the classification by Shye (2009) [16], causes of community participation come from three major classes: demographic antecedents, motivations, and circumstances. Aside from demographic antecedents, which were determined previously, the latter two notions can be facilitated as well to increase the community participation.
Increasing motivations of local citizens. This is out of the scope of this study but is no doubt an interesting route for future research.
Facilitating circumstances. This route can be started with two activities. Activity 1. Develop flexible participation mechanisms with a well-coordinated process among stakeholders. The purpose of this activity is to organize the community, community, and government together to create regulations that facilitate the participation of stakeholders. There are 4 actors involved: (i) Government: government link with community organization representatives to discuss mechanisms and policies for participatory planning, as well as through organization. This is to implement community consultation. (ii) Community Organization: Representative, Participatory Mechanism / government, or participatory mechanisms, participation mechanisms, and participatory mechanisms; coordinate with mass organizations and / or self-media, consult and consult the community for issues that require broad consultation, such as participation mechanisms, participation, and participation levels. individual of the community; coordinate with specialized units, consultation consultants, consult with professional content related to ICT infrastructure development, to improve the knowledge and communication of community communication, ensure maximum transparency. (iii) Community participation: Participation mechanism is to collaborate with community organizations to engage in discussion, feedback, information provision, etc. on issues related to participation and information. Related news; associate with mass organizations and professional advisory units to capture relevant information in support of participation decision. (iv) Mass organizations: the mechanism of action is to coordinate with the authorities to propagate mechanisms and policies for the participation and information provision of community organizations (coordination in propaganda) and community (extended communication); volunteering to contribute as a pioneer organization of the community organization to involve the community, ...; (v) Specialized units: operational mechanism is to coordinate with the authorities in consultation, discussion, ... issues related to the profession; collaborate with community organizations to explain, provide relevant information, and support communication.
Activity 2. Develop a transparent participation procedure. A transparent procedure will facilitate the readiness of local citizens in volunteering. The procedure is depicted in Fig. 2.
-Step 1: the government or community organization representatives propose / initiate the need for rural infrastructure development and propose participation, with public consultation. -Step 2: the government or community organization representative consult with specialized agencies on issues, content, and participation. -Step 3: the government or community organization representatives consult with mass organizations about communication issues related to ICTD development and participation. -Step 4: the government or community organization representatives agree on issues related to ICTD development and participation. -Step 5: a mass organization and / or technical advisory unit in collaboration with community organization representatives to communicate issues related to ICTD development and participation to the individuals in the community -Step 6: the representative of the community organization consulted the contents to discuss and communicate the information related to the development of rural transport infrastructure and participation. -Step 7: the community entity decides to contribute directly to the process. -Step 8: it is to summarize, evaluate, and draw experience, draw lessons for participation in participatory process, as well as participation of stakeholders.

CONCLUSIONS
In developing countries such as Vietnam, community participation benefits the development of rural's transport infrastructure in many aspects: providing alternatives of financial resources and enhancing public support.
This study examined current status of community participation through a case study and explored the relationship between demographic characteristics and community partipation. Results show that community participation in rural transport programs in Vietnam has appeared as four types: financial contribution, labor contribution, material contribution, and land contribution. Household composition, education, income, member in household as community leader affect community participation more than other factors.
Governments and policy makers can be benefited from this study in the planning process, predicting success outcomes, and increasing the effectiveness of rural transport development. Moreover, policies should make rural transport programs to be clearer to local citizens and should facilitate contribution in various forms. Flexible participation mechanisms and transparent participation procedure are also strongly advised to boost community participation.
Future research might expand to other areas in Vietnam, as well as other developing countries to validate the influence of demographic characteristics.