{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T03:46:50Z","timestamp":1791949610276,"version":"4.3.4"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":0,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>&lt;p&gt;Can complex machine-learning models forecast publicly reported bank risk ratios more reliably than a simple rule that keeps the latest value unchanged? We study 12 institutions in Mozambique using 32 quarterly prudential reports for 2018\u20132025. The audited analytic panel contains 362 bank-quarter records and 349 valid consecutive-quarter forecasting pairs. We compare persistence, observed-variable path regressions, ridge regression, random forest and a regression-plus-random-forest hybrid. Of these pairs, 253 have target dates through 2023; the other 96 cover 2024\u20132025, a period previously examined during exploratory model development. The proposed two-factor latent structural equation model fails measurement checks (CFI = 0.519; RMSEA = 0.248; SRMR = 0.160). For the 96 retrospective evaluation pairs, persistence achieves the lowest mean absolute error for both non-performing loans (0.0339) and liquid assets (0.0556). Random forest achieves liquidity RMSE of 0.0884, nearly equal to persistence (0.0883), but is less accurate for non-performing loans. Rare, large changes and the small number of banks limit the conclusions. The study offers a source-audited and executable assessment of when extra predictive complexity does not add reliable value under public-data constraints.&lt;\/p&gt;<\/jats:p>","DOI":"10.2139\/ssrn.7598999","type":"posted-content","created":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T03:17:40Z","timestamp":1791947860000},"source":"Crossref","is-referenced-by-count":0,"title":["Forecasting Credit and Liquidity Risk in Mozambique (2018\u20132025): A Comparative Evaluation of Machine Learning and Structural Equation Modelling"],"prefix":"10.2139","author":[{"given":"Valter Tito","family":"Manjate","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-1816-0705","authenticated-orcid":true,"given":"Leandro  Tito","family":"Manjate","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T03:17:40Z","timestamp":1791947860000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7598999"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7598999","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}