Published Tier A Psychological Capital

Shared Burden, Separate Scars

Economic Shock, Hardship Landing, and Intra-Household Financial Anxiety Contagion

Balloch, A. (FamNet Research)

Shared Burden, Separate Scars

Abstract

Job loss damages household members in different ways at different times. The household-finance literature has long modelled the economic cost of unemployment as a function of foregone earnings and consumption smoothing — yet the lived experience of families facing income shocks suggests that the dollar value of lost income is only part of the welfare story. The other part — financial-anxiety contagion across the household, from the earner to the partner to the children — has been harder to measure and largely absent from welfare estimates.

This paper traces the full cascade. Using 21 years of HILDA panel data (437,006 observations), we deploy a hybrid competing-risk hit model that combines Cox survival analysis with DeepHit neural networks. The hybrid framework is needed because no single classical method can trace the cascade from a single job-loss event through seven distinct categories of financial hardship to the differential anxiety responses of every household member. The behavioural construct at stake is intra-household psychological-anxiety contagion — the differential, time-staggered way that one member's economic shock propagates into the planning, sleeping, and saving behaviour of others.

We find that 93–96 percent of the post-shock welfare damage is psychological rather than monetary. The father loses his identity; the partner loses sleep; the child loses confidence — and none of these losses are captured by the standard income-substitution measure of welfare cost. The findings sharpen the case for household-level mental-health support around income shocks and demonstrate that the standard welfare measure understates the true burden by an order of magnitude. The paper is a methodological flagship in the broader programme on intra-household financial dynamics and motivates an ongoing project translating each model coefficient into a corresponding lyric in a research-music collaboration — bridging academic finance and public communication.

Data & Methods

Data Source
21 years of HILDA panel data — 437,006 observations
Methods (existing)
Hybrid Competing Risk Hit Model combining Cox survival analysis, DeepHit neural networks, and neural survival analysis. Discovers that 93–96% of hardship damage is psychological.
Published in
FamNet Research (working paper) (2025) (fallback: Journal of Financial Economics, Review of Financial Studies)
SSRN (forthcoming) ResearchGate (forthcoming) Discuss this paper