Moderating Effects of Household Type On the Association Between Material Hardship and Depression Among Low-Income Households in Korea

Background and Purpose: With the recent trend of growing poverty and economic polarization in Korea, there has been a tremendous concern over an increasing risk of mental health problems such as depression among the poor. In particular, various social phenomena such as the increase in aging population, delayed marriage, and divorce rate have contributed to producing more single-headed and elderly households that are especially vulnerable to poverty-related stress and mental health problems. However, no research has looked into associations between poverty, depression, and household type in Korea. In light of the urgent need of such studies, this study, using material hardship as a poverty measure to incorporate more concrete and practical understanding of low-income households, aimed 1) to examine the association between material hardship and depression among low-income households; and 2) to investigate whether the association between material hardship and depression differs by household type.

Methods: We used data from the 2006 and 2007 Korean Welfare Panel Study (KWPS). Our final sample consisted of 2,913 low-income households. The dependent variable was depression status at the second year measured by a binary variable (1=depressed and 0=not depressed). The independent variable was material hardship constructed by eight dummy items (skipped meals, unable to pay rent, unable to pay utility bills, utilities shut off, unable to pay children’s tuitions, unable to heat in winter, unable to receive medical services, and had credit problems). Four types of household were constructed by whether or not to be older than 65 (old vs. non-old) and whether or not to have a spouse/partner (single vs. normal): single old, single non-old, normal old, and normal non-old households. To address omitted variable bias, we specified three models: 1) logistic regressions with rich controls; 2) lagged dependent models (the main model); and 3) pooled-sample individual fixed-effect models. In addition, to investigate moderating roles of household type, we added interaction terms of the number of material hardship with household type to the main model.

Results: We found that the number of material hardship was associated with significantly higher probability of being depressed (OR=1.49, p<.001), confirming the robustness across all specification checks. Moreover, we detected varying degrees of association between material hardship and depression across four households specified. Compared to non-old normal households, old single and non-old single households showed significantly higher probability of being depressed as the number of material hardship increases (OR=1.65, p<.001 and OR=1.27, p=.062, respectively), whereas old normal households showed no differences in the association. Three individual items, “unable to pay bills (OR=1.63, p<.001),” “unable to heat in winter (OR=1.94, p<.001),” and “unable to receive medical services (OR=1.68, p<.01),” were associated with significantly higher probability of being depressed.

Conclusion and Implication: Our findings show that experiencing material hardship increases the likelihood of being depressed among low-income households in Korea and that the associations are particularly salient among old single and non-old single households. We suggest that securing material resources for those vulnerable households should be one of the crucial priorities for mental health among low-income households in Korea.