Do the methods matter? Comparing intra-household versus household level approach to quantifying income in Nepal

Description

Abstract:

Lack of studies comparing household’s income at household and intrahousehold level means that there is scant/no evidence on the differences in income estimates when using household and individual as a unit of analysis across different sources of income. This study tests the null hypothesis that household head cannot provide an accurate estimation of the environmental income of the individual household members while for the wage income interviewing only the household head can provide an accurate estimation of the income of other adult household members. As wage incomes are typically larger, distinctly defined and more regular income sources, they are assumed to be more visible in the household and easier to recall. On the other hand, environmental income is typically irregular and seasonal in nature, short term, and causal and often forgettable. Furthermore, different household members are involved in the collection of various types of forest and non-forest environmental products and hence, the generation of environmental income, predominantly for subsistence use from a different range of habitats. Such variation is the number and use of environmental income as has been reported may result in difficulties for the household head to recall and estimate the environmental income accurately for other household members.

To investigate whether the data yield significantly different estimates of income across different data collection instruments, a total of 328 (78 household heads and 250 individual) respondents from randomly sampled household were interviewed from Chainpur VDC of Chitwan district in Nepal. The result presents that absolute mean environmental income dropped from USD 64.94 to USD 21.53 while wage income increased from 30.82 USD to 188.14 USD across the five years period (2012 and 2017). Increased income and reduced labor availability for forest and non-forest resource collection can be a possible explanation for the decline in environmental income in the study site. Furthermore, the results reveal that in 91.01 percent of household, the household head tends to significantly underestimate and overestimate the household’s total environmental income. In contrast, in 78.2 percent of households, wage income estimates are in line when using the household and the individual as the unit of analysis. Hence, for wage income it is not necessary to interview multiple household members. In contrast, the significant difference in the environmental income estimates may mean that multiple household members needs to be interviewed for a more accurate result. Such significant difference in environmental income when household head accounts for other household member income can have implication on our understanding of the prevalence of poverty and income equalizing effects of the forest and non-forest income.