Constructing HLM to examine multi-level poverty-contributing factors of farmer households: Why and how?

Abstract

Accurately identifying poverty-contributing factors of farmer households in an all-round way is the critical prerequisite and guarantee for taking targeted measures in poverty alleviation. From the combined perspectives of multi-level comprehensive detection and human-nature sustainable development, this study has designed a multi-level index system of householdlevel, village-level, and town-level, and constructed a nested three-level hierarchical linear model to examine the poverty-contributing factors of farmer households, and to reveal the significant ones and their multi-level interaction mechanism. The case test from Fugong County shows that: (1) Poverty-contributing factors are multi-level, showing both individual and background effects. 77.14% of the poverty is caused by household-level factors, 6.24% by village-level ones and 16.62% by town-level factors. (2) Significant poverty-contributing factors at different levels are different, identifying different contribution degrees to poverty gaps of farmer households. Five household-level factors show significant influence on poverty degree and account for 70.95% of the overall poverty gap among poor households, 11.70% for four village-level significant factors and 86.80% for two town-level ones, respectively. (3) Higher-level factors have different degrees of influence on the contribution difference of lower-level ones. The two town-level factors, terrain relief and town per capita annual income have explained 59.38% of the difference of village-level proportion of migrant workers' contribution to poverty degree among towns and 89.89% of the difference of household-level per capita annual income's contribution to poverty degree among towns respectively. (4) Measures such as improving the type of access to roads, developing characteristic planting and breeding, and implementing relocation projects, can help poor households in the study area to reduce poverty. This study provides a new perspective for identifying farmers' poverty-contributing factors and technical reference and decision support for local departments to plan and implement targeted assistance and household-specific development policies.

Examining Multilevel Poverty-Causing Factors in Poor Villages: a Hierarchical Spatial Regression Model

“…Referring to the classic HLM modelling process (Dunifon, 2005;Graves, 2011;Jiang et al, 2020), the HSLM modelling process used in this study is divided into a series of steps. First, the null model (Model I), that is, a hierarchical regression model that does not contain any explanatory variables, is constructed to explore whether there is a background effect on the poverty incidence in poverty-stricken villages.…”

“…Ambitious programmes are required, such as the United Nations 2030 Sustainable Development Goals policy or the comprehensive rural rejuvenation programme proposed by the Chinese government. In recent years, poverty in rural China, especially in contiguous poverty-stricken areas, has changed from a singular economic problem to a complex and diverse problem because of great changes in rural society and rapid growth in economics (Jiang et al, 2020). Academics are required to seriously consider poverty issues and to propose new coping models and methods for improving the relevance and practicability of research.…”

“…These studies have proven that poverty-causing factors are no longer limited to individual characteristics. There may be multilevel factors; that is, the poverty status of the poor may be affected not only by multidimensional factors such as family characteristics and economic development at the individual level but also by multidimensional background factors poverty-stricken people depend on at higher levels, such as economic development, social development, the ecological environment, and so on (Alkire, 2019;Boemi & Papadopoulos, 2019;Ibrahim et al, 2016;Jiang et al, 2020;Liang et al, 2019;Mowat, 2019). This indicates that the occurrence of poverty is driven by both the individual effects (i.e., variations caused by individual characteristics) and background effects (i.e., group effects or pond effects, which refer to variations caused by an individual's environment) of poverty-causing factors at different levels (Jiang et al, 2020;Ma et al, 2018;Park & Nam, 2018;Wang et al, 2019).…”