Background: Hurricanes and other natural disasters produce public health and economic consequences that last well beyond their immediate aftermath. Resource loss is a core driver of post-traumatic stress disorder (PTSD) after large-scale traumatic events. We examined the effect of restoration of residential and housing-related financial resources on recovery from PTSD in post-disaster contexts. Methods: We built an agent-based model, empiricised with observational and experimental data, to test the effects of differing health service approaches on PTSD recovery, measured by prevalence and persistence. We tested a social services case management (SSCM) approach similar to Psychological First Aid, featuring shelter-based social service provision and linkage to mental health treatment for people who were displaced and had income loss, by comparing the treatment effectiveness of usual care alone, usual care with SSCM, stepped care alone, and stepped care with SSCM. Findings: An SSCM approach to restore housing and provide linkage to mental health services among people who were displaced and had income loss after a large-scale natural disaster resulted in between 1·56 (95% CI 1·55–1·57) and 5·73 (5·04–6·91) times as many remitted PTSD cases as non-SSCM conditions at the end of the first year, and between 1·16 (1·16–1·17) and 2·28 (2·25–2·32) times as many remitted cases at the end of the second year. Interpretation: Restoring economic and housing resources to populations affected by a natural disaster would significantly reduce the mental health burden in populations, particularly those with resource loss, after a disaster. Funding: US Department of Health and Human Services.