When money and mental health problems pile up: The reciprocal relationship between income and psychological distress

Abstract

Background: Longitudinal studies suggest that socioeconomic status (SES) and mental health have a bidirectional relationship such that SES declines lead to a deterioration of mental health (social causation), while worsening mental health leads to SES declines (social drift). However, the dynamic relationship between income and psychological distress has not been sufficiently studied.

Methods: We use cross-lagged panel models with unit fixed effects (FE-CLPM) and data from a five-wave representative panel (n = 3103) of working-age (18-64) New York City adults. Yearly measures include individual earnings, family income (income-to-needs), and psychological distress. We also examine effects by age, gender, education, and racial/ethnic identification.

Results: We find significant bidirectional effects between earnings and distress. Increases in past-year individual earnings decrease past-month psychological distress (social causation effect [SCE], standardized β= -0.07) and increases in psychological distress reduce next-year individual earnings (social drift effect [SDE], β= -0.03). Family income and distress only have a unidirectional relationship from past-year family income to distress (SCE, β= -.03). Strongest evidence of bidirectional effects between earnings and distress is for prime working-age individuals (SCE, β= -0.1; SDE, β= -0.03), those with less than bachelor's degrees (SCE, β= -0.08; SDE, β= -0.05), and Hispanics (SCE, β= -0.06; SDE, β= -0.08). We also find evidence of reciprocal effects between family income and distress for women (SCE, β= -0.03; SDE, β= -0.05), and Hispanics (SDE, β= -0.04; SDE, β= -0.08).

Conclusions: Individual earnings, which are labor market indicators, may be stronger social determinants of mental health than family income. However, important differences in social causation and social drift effects exist across groups by age, education, gender, and racial/ethnic identities. Future research should examine the types of policies that may buffer the mental health impact of negative income shocks and the declines in income associated with worsening mental health, especially among the most vulnerable.

Fig. 1

This figure depicts the specification of the basic FE-CLPM with unit effects in this study, which controls for individual-specific, time-invariant factors over time. For simplicity, we only show variables associated with cross-lagged and auto-regressive effects, and not covariates. η ^(x) and η ^(y) = unit effect (fixed effects for income and psychiatric distress, respectively); ψ _η^(x) and ψ _η^(y) = unit effect variance; λ (x) and λ (y) = factor loadings (which are allowed to vary over time, not constrained); β_x1^(x) and β_y1^(y) = autoregressive effects; β_x1^(y) and β_y1^(x) = cross-lagged effects; β_x^(y) = contemporaneous effect of x on y.